The type I BMP receptor Alk3 is required for the induction of hepatic hepcidin gene expression by interleukin-6

Anemia of chronic disease (ACD), the second most prevalent form of anemia, is associated with chronic infections, autoimmune diseases, and neoplasias.^1-3  ACD is characterized by increased levels of hepcidin, a hepatic hormone that induces degradation of the iron exporter ferroportin-1.^4,5  Ferroportin-1 is responsible for intestinal absorption of dietary iron and for mobilization of iron from intracellular stores in enterocytes, macrophages, and hepatocytes.^6,7  Increased hepcidin levels cause iron to remain inside cells, resulting in decreased serum iron concentrations and impairment of erythropoiesis.

Hepcidin production is transcriptionally regulated by inflammatory cytokines, including IL-6 and bone morphogenetic protein (BMP) signaling. Expression of the gene encoding hepcidin, Hamp-1, is also regulated by a variety of other stimuli, including iron, hypoxia, transforming growth factor-β, tumor necrosis factor-α, hemojuvelin (HJV, a BMP co-receptor), activin B, and the unfolded protein response.^8-12  IL-6 induces hepcidin gene transcription, as well as the transcription of other target genes including the gene encoding heme oxygenase-1 (HO-1), Hmox1, via phosphorylation of the transcription factor STAT3.^13-16  BMP signaling is activated by binding of BMP ligands to a BMP receptor complex, which consists of type II receptors (BMPR2, ActRIIa, and ActRIIb) and type I receptors (Alk1, Alk2, Alk3, and Alk6). Ligand binding induces the type II receptor to phosphorylate and activate the type I receptor. Activated type I receptors phosphorylate BMP-responsive SMAD proteins (SMADs 1, 5, and 8), which translocate together with SMAD4 into the nucleus, where they activate expression of multiple genes including those encoding hepcidin and Id-1.^17,18 

Of the 4 BMP type I receptors, Alk1 is predominantly expressed in endothelial cells.¹⁹  Alk2 and Alk3 are highly expressed in hepatocytes, whereas Alk6 is expressed at much lower levels.²⁰  We previously reported that hepatocyte-specific deletion of Alk2 or Alk3 in mice causes mild or severe iron overload phenotypes, respectively.²¹  In contrast, global deficiency of Alk6 does not alter serum iron concentrations or hepatic tissue iron levels in mice (P. J. Schmidt and M. D. Fleming, unpublished observations).

Inhibition of BMP signaling, using small molecule inhibitors of type I BMP receptor kinase activity (eg, LDN-193189) or recombinant proteins that scavenge BMP ligands (eg, noggin, Alk3-Fc, or HJV-Fc), reduces the ability of IL-6 to induce hepcidin gene expression and decrease serum iron concentrations, thereby, ameliorating ACD in animal models.^18,22-24  In the current study, we sought to identify the BMP type I receptor(s) that contribute(s) to the induction of hepatic hepcidin gene expression by IL-6. We studied mice with hepatocyte-specific deficiency of Alk2 or Alk3 that were exposed to either sustained or transient increases in hepatic IL-6 levels induced by injection with an adenovirus specifying IL-6 (Ad.IL-6) or with recombinant murine IL-6 (mIL-6), respectively. We report that Alk3 is required for the IL-6–mediated induction of hepatic hepcidin gene expression and reduction in serum iron levels.

Human hepatoma cells (HepG2, American Type Culture Collection, Manassas, VA) were cultured in Eagle’s Minimal Essential Medium supplemented with 10% fetal bovine serum, penicillin, streptomycin, and l-glutamine. Cells were transferred to 6-well dishes and starved with Eagle’s Minimal Essential Medium containing no fetal bovine serum for at least 6 hours before exposure to human IL-6.

All mouse experiments were approved by the Subcommittee on Research Animal Care of the Massachusetts General Hospital, where the experiments were performed. Mice with floxed Alk2 allele (Alk2^+/fl) on a mixed C57BL/6; SV129 background or mice with floxed Alk3 allele (Alk3^+/fl) on a C57BL/6 background^25,26  were bred with B6.Cg-Tg^{(Alb-Cre)21Mgn}/J mice (Jackson Labs). Heterozygous animals were intercrossed to obtain animals homozygous for the Alk2^fl or Alk3^fl alleles with or without the transgene specifying Cre recombinase under the control of the hepatocyte-specific albumin gene promoter (Alb-Cre), as described previously.²¹  Mice with hepatocyte-specific Alk2 deficiency (Alk2^fl/fl; Alb-Cre) were compared with Alk2^fl/fl mice that did not carry the Alb-Cre transgene. Similarly, mice with hepatocyte-specific Alk3 deficiency (Alk3^fl/fl; Alb-Cre) were compared with Alk3^fl/fl mice that did not carry the Alb-Cre transgene. In accordance with the recently published ARRIVE guidelines (http://www.nc3rs.org/ARRIVE), mice were randomized according to their genotypes to either treatment or control group. The investigators, who measured iron concentrations in serum and tissues and who measured messenger RNAs (mRNAs) and protein levels, were blinded as to the experimental groups to which the mice were assigned. The experimental procedure is explained for each method in detail below (eg, adenovirus or recombinant protein injection). Male mice at the age of 8 to 12 weeks were studied, as indicated for each experiment. The average weight of these mice was 20 g. Mice were fed a standard, iron-replete diet (Prolab 5P75 Isopro 3000, 380 ppm iron) and not more than 4 to 5 siblings were housed in each cage.

Recombinant adenovirus specifying GFP was purchased from Vector Biolabs (Ad.GFP, ready-to-use human adenovirus type 5 [dE1/E3], catalog No. 1060, gene name: eGFP, promoter: cytomegalovirus, titer 1 × 10¹⁰ PFU/mL). Recombinant Ad.IL-6 was purchased from Vector Biolabs (Ad.IL-6, ready-to-use human adenovirus type 5 (dE1/E3), catalog No. 1380, gene name: IL6, promoter: cytomegalovirus, titer 1 × 10¹⁰ PFU/mL). Ad.GFP or Ad.IL-6 were propagated in 293 cells and harvested from the cells just before spontaneous lysis. The virus was purified from cell extracts with an Adenovirus Standard Purification Virakit (Virapur). The viral titer was estimated from the optical density at 260 nm.

Mice without and with hepatocyte-specific Alk2 or Alk3 deficiency were injected with 1 × 10¹¹ virus particles (in 100 µL of saline) of either Ad.GFP or Ad.IL-6 via a tail vein. Blood and tissue samples were collected 72 hours after injection.

Mice received a single intraperitoneal injection of recombinant mouse IL-6 (#406-ML-025, 100 ng/g; R&D Systems) or vehicle (0.1% bovine serum antigen [BSA] in phosphate-buffered saline [PBS]). Wild-type mice were anesthetized at 2 and 4 hours after injection, and blood was collected. Mice were euthanized and tissue samples were harvested. Four hours after injection with mIL-6 or vehicle, blood and tissues were harvested from mice with and without hepatocyte-specific Alk2 or Alk3 deficiency.

Blood for measurement of serum iron parameters was obtained by retro-orbital puncture, as described previously.¹⁸  Serum iron concentrations and transferrin saturations were determined using the Iron/UIBC Kit (IRON-SL, #157-30; Thermo Fisher Scientific and SEKISUI Diagnostics) following the manufacturers’ protocols. After mice were euthanized liver and spleen tissues were harvested, and non–heme tissue iron levels were determined, as previously described.¹⁸ 

Total RNA was extracted from liver tissues using Trizol (Invitrogen), as previously described.²¹  Quantitative reverse transcriptase polymerase chain reaction (RT-PCR) was performed on a Mastercycler Realplex2 (Eppendorf) using KAPA SYBR FAST (Kapa Biosystems) for conventional Sybr primers or KAPA PROBE FAST (Kapa Biosystems) for hydrolysis probes (TaqMan Gene Expression Assays, Applied Biosystems, Life Technologies). The relative C_T method was used to normalize the levels of target transcripts to 18S rRNA levels. Primer pairs used for quantitative RT-PCR are listed in supplemental Table 1 (available on the Blood Web site) and are described in our previous publication.²¹ 

HepG2 cells and liver tissues were harvested in radio-immunoprecipitation assay (RIPA) buffer supplemented with protease and phosphatase inhibitors (Sigma-Aldrich). Extracted proteins were separated by electrophoresis using 4% to 12% bis-Tris gels (NuPAGE, Invitrogen) and transferred to polyvinylidene fluoride membranes (Immobilon-FL; Millipore). Membranes were incubated with antibodies directed against total STAT3 (Cat.No. 49045; Cell Signaling Technology), STAT3 phosphorylated at tyrosine 705 (P-STAT3, Cat. No. 9145L; Cell Signaling Technology), total SMAD 1 (Cat. No. LS-C75853; Life Span Bio), phosphorylated SMAD 1/5 (P-SMAD 1/5, Cat. No. 95165; Cell Signaling Technology), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Cat. No. 5174s; Cell Signaling Technology). After washing, membranes were incubated with horseradish peroxidase–linked anti-rabbit IgG or anti-mouse IgG (Cell Signaling Technology). Membranes were incubated with ECL-Plus (GE Healthcare), and chemifluorescence was detected with a Versadoc 4000MP imager (Bio-Rad).

All values were expressed as mean ± standard deviation. Data were analyzed using the Student t test and one-way analysis of variance (ANOVA) Kruskal-Wallis test with Dunn’s correction for multiple comparisons, when applicable. Statistical significance was considered for P values <.05.

We sought to determine which of the 2 BMP type I receptors involved in hepatic iron signaling, Alk2 or Alk3, contributes to the ability of IL-6 to induce hepcidin gene expression and reduce serum iron concentrations. Hepatocyte-specific deletion of a BMP type I receptor was performed by generating mice homozygous for Alk2 or Alk3 alleles flanked by loxP sequences that also carry a transgene in which the albumin promoter directs expression of Cre recombinase. To achieve sustained increases in hepatic IL-6 expression, mice were injected with a recombinant adenovirus (type 5) specifying IL-6 (Ad.IL-6). In pilot studies of wild-type mice injected with Ad.IL-6 and studied 72 hours later, we detected abundant human IL-6 mRNA in the liver, but not in the heart or spleen, associated with increased hepatic hepcidin mRNA levels and reduced serum iron concentrations (data not shown). To ensure that the increase of hepatic hepcidin gene expression was attributable to expression of the transgene rather than a nonspecific effect associated with injection of an adenovirus, we measured hepatic hepcidin and HO-1 mRNA levels in wild-type mice injected with saline, Ad.IL-6, or a control adenovirus, Ad.GFP (supplemental Figure 1A-B). Injection of Ad.IL-6, but not Ad.GFP, induced hepatic hepcidin and HO-1 gene expression.

To test the impact of hepatocyte-specific deficiency of Alk2 or Alk3 on the perturbation of iron homeostasis by IL-6, we compared mice without and with hepatocyte-specific deficiency of Alk2 or Alk3 72 hours after injection with Ad.IL-6 or Ad.GFP. Serum iron levels and transferrin saturations were higher in Ad.GFP-injected mice with hepatocyte-specific Alk2 deficiency (Alk2^fl/fl; Alb-Cre) than in Ad.GFP-injected Alk2^fl/fl mice (Figure 1A-B). These results were consistent with our previous observations that serum iron concentrations are elevated in untreated Alk2^fl/fl; Alb-Cre mice. Exposure to IL-6 for 72 hours decreased serum iron levels and transferrin saturations in both Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (Figure 1A-B). These results suggest that Alk2 is not required for IL-6 to reduce iron levels and transferrin saturations. However, it is of note that the percent reduction in serum iron levels caused by Ad.IL-6 was greater in Alk2^fl/fl mice than in Alk2^fl/fl; Alb-Cre mice.

To determine the impact of hepatocyte-specific Alk3 deficiency on the ability of IL-6 to modulate serum iron concentrations, Alk3^fl/fl and Alk3^fl/fl; Alb-Cre mice were injected with either Ad.IL-6 or Ad.GFP. In Ad.GFP-injected Alk3^fl/fl; Alb-Cre mice, serum iron levels and transferrin saturations were higher than in Ad.GFP-injected Alk3^fl/fl mice (Figure 1C-D). These results were consistent with our previous observations that serum iron concentrations are elevated in untreated mice with hepatocyte-specific Alk3 deficiency. Seventy-two hours after adenovirus injection, serum iron concentrations and transferrin saturations were less in Ad.IL-6–injected Alk3^fl/fl mice than in Ad.GFP-injected Alk3^fl/fl mice. In contrast, serum iron concentrations in Alk3^fl/fl; Alb-Cre mice injected with Ad.IL-6 did not differ from those in Alk3^fl/fl; Alb-Cre mice injected with Ad.GFP. These findings suggest that deficiency of Alk3 abolished the ability of IL-6 to decrease serum iron levels.

To test whether IL-6 requires Alk2 or Alk3 to increase hepcidin gene expression, we measured hepatic hepcidin mRNA levels in Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice, as well as in Alk3^fl/fl and Alk3^fl/fl; Alb-Cre mice, 72 hours after injection with Ad.GFP or Ad.IL-6 (Figure 2). Hepatic hepcidin mRNA levels did not differ in Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice injected with Ad.GFP. However, the ratio of serum iron levels to hepatic hepcidin mRNA levels was markedly higher in the latter, consistent with the previously reported concept that hepatic Alk2 deficiency impairs the responsiveness of hepcidin gene expression to serum iron levels.²¹  Injection with Ad.IL-6 induced hepatic hepcidin gene expression similarly in both Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (2.6- and 2.8-fold, respectively; Figure 2A). The observation that serum iron levels and transferrin saturations remained higher in Alk2^fl/fl; Alb-Cre mice than in Alk2^fl/fl mice even after Ad.IL-6 injection is most likely attributable to the sum of the impact of Alk2 on basal iron metabolism and the Alk2-independent ability of IL-6 to induce hepcidin gene expression.

Hepatic hepcidin mRNA levels were markedly less in Ad.GFP-injected Alk3^fl/fl; Alb-Cre mice than in Ad.GFP-injected Alk3^fl/fl mice, consistent with the previously reported effect of Alk3 deficiency on basal hepatic hepcidin expression. Ad.IL-6 injection induced hepatic hepcidin gene expression in Alk3^fl/fl mice but did not increase hepcidin mRNA levels in the livers of Alk3^fl/fl; Alb-Cre mice (Figure 2B). These results suggest that Alk3, but not Alk2, is required for the induction of hepcidin gene expression by IL-6.

We previously reported that Id-1, a BMP target gene, was also induced by IL-6 in a BMP signaling–dependent manner.^18,24  To examine the importance of Alk2 and Alk3 in the ability of IL-6 to induce Id-1 gene expression, we measured hepatic Id-1 mRNA levels in Alk2^fl/fl and Alk2^fl/fl; Alb-Cre, as well as in Alk3^fl/fl and Alk3^fl/fl; Alb-Cre mice, that were injected with either Ad.GFP or Ad.IL-6. Hepatic Id-1 mRNA levels did not differ in Ad.GFP-injected Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice. Sustained exposure to IL-6 induced hepatic Id-1 gene expression in both Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (Figure 2C). In contrast, hepatic Id-1 mRNA levels were less in Alk3^fl/fl; Alb-Cre mice injected with Ad.GFP than in Alk3^fl/fl mice injected with Ad.GFP. Prolonged exposure to IL-6 induced hepatic Id-1 gene expression in Alk3^fl/fl mice but not in Alk3^fl/fl; Alb-Cre mice (Figure 2D). These data indicate that induction of hepatic Id-1 gene expression by IL-6 is dependent on Alk3.

Having observed that IL-6 was unable to induce hepcidin and Id-1 gene expression in mice with hepatocyte-specific Alk3 deficiency, we investigated the possibility that BMP signaling was required for activation of STAT3 by IL-6 using both cultured cells and genetically-modified mice. HepG2 cells were pretreated with and without LDN-193189, an inhibitor of BMP type I receptor signaling, and then were exposed to IL-6 or BMP6. In a previous study, we reported that LDN-193189 could prevent the induction of hepcidin gene expression in HepG2 cells incubated with IL-6.¹⁸  In the absence of LDN-193189, incubation of HepG2 cells with IL-6 induced phosphorylation of STAT3 (Figure 3). BMP6 induced phosphorylation of SMAD 1/5 but not STAT3 phosphorylation. Pretreatment with LDN-193189 inhibited BMP6-mediated SMAD 1/5 phosphorylation but did not alter the ability of IL-6 to induce STAT3 phosphorylation. Similarly, our in vivo studies revealed that the ability of IL-6 to induce STAT3 phosphorylation did not differ in Alk2^fl/fl; Alb-Cre and Alk2^fl/fl mice or in Alk3^fl/fl; Alb-Cre and Alk3^fl/fl mice (Figure 4A-B).

To further examine the impact of BMP signaling on activation of STAT3 by IL-6 in vivo, we measured HO-1 mRNA levels, a known IL-6/STAT3 target gene,¹⁵  in the livers of Alk2^fl/fl, Alk2^fl/fl; Alb-Cre, Alk3^fl/fl, and Alk3^fl/fl; Alb-Cre mice injected with Ad.IL-6 or Ad.GFP. Hepatic HO-1 mRNA levels did not differ in Ad.GFP-injected Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice. Prolonged exposure to IL-6 increased HO-1 mRNA levels similarly in both Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (Figure 4C). HO-1 mRNA levels did not differ in Ad.GFP-injected Alk3^fl/fl and Alk3^fl/fl; Alb-Cre mice, and stimulation with IL-6 induced HO-1 gene expression similarly in both Alk3^fl/fl and Alk3^fl/fl; Alb-Cre mice (Figure 4D). Taken together, these results indicate that deficiency of Alk2 or Alk3 does not interfere with the ability of IL-6 to induce STAT3 signaling.

To confirm our observations in mice treated with Ad.IL-6, we challenged mice with mIL-6 and measured hepatic hepcidin mRNA levels. A time course study revealed that injection of mIL-6 induced hepatic hepcidin mRNA levels in wild-type mice after 4 hours (supplemental Figure 2A). We did not observe a decrease in serum iron levels in mice challenged with mIL-6 4 hours earlier, likely because there was insufficient time for elevated hepatic hepcidin production to reduce ferroportin levels and sequester iron intracellularly (supplemental Figure 2B).

To determine the impact of the BMP type I receptors Alk2 and Alk3 on the ability of recombinant IL-6 to modulate hepatic hepcidin mRNA levels, mice with and without hepatocyte-specific deficiency of Alk2 or Alk3 were challenged with mIL-6. Four hours after injection, blood was withdrawn for measurement of serum iron levels, and livers were harvested for measurement of hepcidin, Alk2, and Alk3 mRNA levels, as well as liver iron content (LIC). Spleens were also harvested for measurement of splenic iron content (SIC). Challenge with mIL-6 induced hepatic hepcidin mRNA levels similarly in Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (Figure 5A). As anticipated, hepatic Alk2 mRNA levels were markedly reduced in Alk2^fl/fl; Alb-Cre mice (Figure 5B). Serum iron levels were increased in mice with hepatocyte-specific Alk2 deficiency. Injection with mIL-6 did not change serum iron levels at 4 hours in Alk2^fl/fl and Alk2^fl/fl; Alb-Cre mice (Figure 5C). LIC was higher in Alk2^fl/fl; Alb-Cre mice than in Alk2^fl/fl mice. Injection with mIL-6 did not change LIC after 4 hours in Alk2^fl/fl or Alk2^fl/fl; Alb-Cre mice (supplemental Figure 3A). SIC did not differ in Alk2^fl/fl or Alk2^fl/fl; Alb-Cre mice treated with mIL-6 or vehicle (supplemental Figure 3B). Consistent with our findings using Ad.IL-6, these results suggest that the ability of mIL-6 to induce hepatic hepcidin gene expression does not require Alk2.

Challenge with mIL-6 induced hepatic hepcidin mRNA levels in Alk3^fl/fl mice, but not in Alk3^fl/fl; Alb-Cre mice (Figure 6A). Hepatic Alk3 mRNA levels were markedly reduced in Alk3^fl/fl; Alb-Cre mice (Figure 6B). Serum iron levels were higher in mice with hepatocyte-specific Alk3 deficiency than in Alk3^fl/flmice. Injection with mIL-6 did not change serum iron levels in Alk3^fl/fl or Alk3^fl/fl; Alb-Cre mice (Figure 6C). LIC was greater and SIC was less in Alk3^fl/fl; Alb-Cre mice than in Alk3^fl/fl mice (supplemental Figure 3C-D). Injection of mIL-6 did not change LIC or SIC in either genotype after 4 hours. Consistent with our findings using Ad.IL-6, these observations suggest that the presence of Alk3 is essential for induction of hepatic hepcidin gene expression by IL-6.

Studies in animal models and in patients have revealed an important role for IL-6–induced hepcidin gene expression in the pathogenesis of ACD.^12,27-29  Inhibition of BMP signaling has been shown to reduce hepcidin induction by IL-6^18,23,24  and to prevent the development of ACD.^18,30  However, the BMP type I receptor involved in the induction of hepatic hepcidin gene expression by IL-6 was unknown. We report that hepatocyte-specific Alk2 deficiency does not prevent the ability of IL-6 to decrease serum iron concentrations or increase hepatic hepcidin gene expression. In contrast, hepatocyte-specific Alk3 deficiency markedly impairs the ability of IL-6 to decrease serum iron concentrations and increase hepcidin mRNA levels.

As previously reported,²¹  we observed that hepatocyte-specific deficiency of Alk2 and Alk3 in mice led to mild and severe iron overload, respectively. The presence of Alk3 is the primary determinant of basal hepatic hepcidin mRNA levels, as well as serum and hepatic iron concentrations. In contrast, both Alk2 and Alk3 are required for BMP2 and iron to induce hepatic hepcidin gene expression (in primary hepatocytes and in vivo, respectively).²¹  In the current report, we used adenovirus-mediated gene transfer to direct hepatic expression of human IL-6 for prolonged periods. We observed that Ad.IL-6 induced hepcidin gene expression and reduced serum iron levels in mice with hepatocyte-specific Alk2 deficiency but not in mice with hepatocyte-specific Alk3 deficiency. These results suggest that the Alk3-dependent mechanisms by which IL-6 induces hepcidin gene expression differ from the Alk2- and Alk3-dependent mechanisms required for regulation of hepcidin gene expression by BMP2 and iron.

It is of note that in Alk3-deficient mice, sustained exposure to IL-6 reduced serum transferrin saturations even though serum iron levels did not change. We observed that total iron binding capacity tended to be greater after Ad.IL-6 injection than after Ad.GFP injection in all 4 mouse genotypes. It seems plausible that the impact of IL-6 on serum transferrin saturation represents the combination of an Alk3-dependent effect on iron metabolism and an Alk3-independent effect on transferrin levels.

Iron and BMP ligands are known to activate BMP-responsive SMAD signaling,³¹  whereas IL-6 activates STAT3 signaling.^13,27  The hepcidin gene promoter contains both STAT3-binding and SMAD-binding sites.^27,32  We previously reported that induction of hepcidin gene expression by IL-6 in HepG2 cells required BMP signaling.¹⁸  We considered the possibility that BMP signaling is required for STAT3- signaling in response to IL-6. We observed that the ability of IL-6 to induce STAT3 phosphorylation was not altered in HepG2 cells treated with a small molecule inhibitor of BMP type I receptor kinase activity or in mice with hepatocyte-specific Alk2 or Alk3 deficiency. Moreover, induction of HO-1 gene expression by IL-6/STAT3 signaling, an event reported to occur predominantly in hepatocytes,^33,34  was not altered by deficiency of either Alk2 or Alk3. Taken together, these findings suggest that BMP/Alk3 signaling is required for the ability of IL-6 to induce hepatic hepcidin gene expression, but that reduced BMP signaling does not alter IL-6 signaling via STAT3 phosphorylation.

As a complementary approach that avoided nonspecific effects of adenovirus infection, we measured hepatic hepcidin mRNA levels in mice that received recombinant mIL-6 as a bolus injection. Consistent with our observations using Ad.IL-6,³⁵  we observed that mIL-6 injection induced hepatic hepcidin gene expression in Alk2^fl/fl, Alk2^fl/fl; Alb-Cre, and Alk3^fl/fl mice. In contrast, injection of mIL-6 was insufficient to induce hepatic hepcidin gene expression in Alk3^fl/fl; Alb-Cre mice. These observations confirm that the ability of IL-6 to induce hepatic hepcidin gene expression requires Alk3. The observation that Alk3 is required for the induction of hepcidin gene expression by IL-6 is consistent with the findings of Wang and colleagues, who reported that deficiency of SMAD4 abrogated the ability of IL-6 to induce hepatic hepcidin gene expression.³¹  Our findings are also consistent with those of Verga Falzacappa and colleagues, who reported that binding of phosphorylated STAT3 to the hepcidin gene promoter without activation of the BMP-responsive SMAD binding element is insufficient to induce hepcidin gene transcription, and that mutation of the BMP response element decreases the ability of IL-6 to induce hepcidin gene transcription.³² 

Increased hepcidin levels participate in the development of ACD. Inhibition of BMP signaling by small molecules, such as LDN-193189, decreases hepcidin gene expression and prevents the development of ACD in animal models.^18,30  However, the BMP inhibitors used to treat ACD in animal models can act on multiple BMP type I receptors. It is known that different BMP type I receptors have distinct tissue distribution patterns and different physiological roles. Our current findings suggest that BMP type I receptor–specific inhibitors represent an attractive strategy for the treatment of ACD. The observation that Alk3 plays a critical role in the induction of hepcidin expression by IL-6 suggests that targeting Alk3 selectively might effectively reduce hepcidin gene expression in settings associated with ACD. A therapeutic agent that specifically targets Alk3 might avoid deleterious effects associated with inhibiting Alk1 (eg, reduced endothelial cell survival and angiogenesis)³⁶  or Alk2 (eg, increased bone mineral density caused by decreased osteoclast activity).³⁷ 

In conclusion, we report that induction of hepcidin gene expression by IL-6 requires the presence of the BMP type I receptor, Alk3. Hepatocyte-specific deficiency of Alk3, but not Alk2, markedly impairs the ability of IL-6 to decrease serum iron concentrations. The BMP signaling mechanisms required to induce hepcidin gene expression in response to IL-6 (Alk3-dependent) appear to differ from those required to respond to increased iron levels (dependent on both Alk2 and Alk3). Our findings suggest that Alk3 represents a potential target for the treatment of ACD.

The authors thank Yuji Mishina and Vesa Kaartinen for providing Alk2^+/fl and Alk3^+/fl mice, respectively.

This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG SW 119/3-1), “Innovative Medical Research” of the University of Muenster Medical School (IMF-ST-111206) (A.U.S.), the Deanery of Muenster Medical School (A.U.S.), the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases (R01DK082971) (R.T.P. and K.D.B.), and a grant from the Fondation LeDucq (P.B.Y. and K.D.B.).

Contribution: R.T.P., P.B.Y., D.B.B., K.D.B., and A.U.S. supervised the experimental design; C.M., L.K.L., P.L., S.M.K., A.E.P., S.A.K., R.L.G., E.S., B.Y., D.B.B., and A.U.S. performed experiments; C.M., L.K.L., P.L., S.M.K., A.E.P., R.L.G., D.B.B., K.D.B., and A.U.S. analyzed data; and C.M., R.T.P., D.B.B., K.D.B., and A.U.S. wrote the manuscript.

Conflict-of-interest disclosure: The Massachusetts General Hospital has filed patents related to the use of small molecule inhibitors of BMP signaling to modulate iron metabolism, and P.B.Y., R.T.P., and K.D.B. may be eligible to receive royalties. The remaining authors declare no competing financial interests.

The current affiliation for L.K.L. is Clinic for Hematology, Oncology and Stem Cell Transplantation, University Medical Center Freiburg, Freiburg, Germany.

Correspondence: Andrea U. Steinbicker, Department of Anesthesiology, Intensive Care and Pain Medicine, University Hospital Muenster, Albert-Schweitzer Campus 1, Gebäude A1, 48149 Muenster, Germany; e-mail: andrea.steinbicker@ukmuenster.de, andrea.steinbicker@gmail.com.