Deletion of BMP6 worsens the phenotype of HJV-deficient mice and attenuates hepcidin levels reached after LPS challenge

Liver sinusoidal endothelial cells produce bone morphogenetic protein 6 (BMP6), which acts on the hepatocyte coreceptor hemojuvelin (HJV) to regulate hepcidin production in a paracrine fashion.^1,2  Both are required to ensure optimal maintenance of iron homeostasis. In humans, mutations in the HJV gene are responsible for juvenile hemochromatosis, an early onset form of hereditary hemochromatosis caused by profound hepcidin insufficiency.³  Whereas loss-of-function mutations in BMP6 have not been described so far, heterozygous mutations in the propeptide have been identified in patients with iron overload and lead to inappropriate hepcidin synthesis.^4,5  In the mouse, deletion of either Bmp6 or Hjv leads to a similar phenotype characterized by hepcidin insufficiency, severe iron loading, and extrahepatic iron accumulation in males,^6-11  which is consistent with the current views that HJV is a coreceptor for BMP6.

Whether BMP6 and HJV may also signal to hepcidin independently of each other is still a topic of discussion.¹¹  To provide direct evidence that BMP6 and HJV can separately stimulate hepcidin, we intercrossed Hjv and Bmp6 knockout mice and looked at whether deletion of both Bmp6 and Hjv in mice of the F2 progeny was aggravating the phenotype of single knockout animals. Whether BMP6 and HJV are both required for the upregulation of hepcidin by inflammatory stimuli is another unresolved issue. We took the opportunity of having genetically comparable single and double knockout animals to challenge them with lipopolysaccharide (LPS) and examine the impact of Bmp6 and/or Hjv deletion on Smad signaling and hepcidin expression after stimulation.

Hjv^−/− mice on a 129S6/SvEvTac background⁹  were bred to Bmp6^tm1Rob mice (Bmp6^−/−) on a CD1 background.¹²  Experiments were done on 12-week-old wild-type, Bmp6^−/−, Hjv^−/−, and Bmp6^−/−Hjv^−/− littermates of the F2 progeny. Experimental protocols were approved by the Midi-Pyrénées Animal Ethics Committee. Animals were given free access to tap water and standard laboratory mouse chow diet (250 mg of iron per kilogram; SAFE, Augy, France). Littermates carrying the different genotypes were also challenged with an intraperitoneal injection of LPS (1 μg/g body weight) and livers were harvested 4 hours later.

Quantitative polymerase chain reactions (PCRs) were run on a LightCycler 480 System (Roche Diagnostics), using primers previously referenced.¹³  Δ Cycle threshold (ΔCt) values were obtained by subtracting the Hprt reference gene Ct to the target gene Ct.

Serum hepcidin levels were quantified using the Intrinsic LifeSciences (La Jolla, CA) Hepcidin-Murine Compete competitive enzyme-linked immunosorbent assay (ELISA).

Transferrin saturation was obtained through the determination of serum iron and latent iron-binding capacity. Quantitative measurement of nonheme iron in the liver, heart, and pancreas was performed according to the method of Torrance and Bothwell.¹⁴  Nontransferrin-bound iron (NTBI) was measured using the FeROS eLPI kit (Aferrix).

Protein extraction and western blot analysis were performed as previously described.¹³ 

Means of quantitative variables (log-transformed for serum hepcidin) were compared with 2-way analysis of variance (ANOVA) followed by Sidak multiple comparison tests of planned contrasts between pairs of means.

We first measured by quantitative PCR the amount of hepcidin messenger RNA (mRNA) in the liver of the genetically comparable F2 littermates (Figure 1A). As previously reported in mice of other genetic backgrounds,¹¹  hepcidin expression in the liver of Bmp6^−/− and Hjv^−/− mice is much lower than in wild-type mice, particularly in males. However, deletion of Bmp6 more dramatically repressed the already reduced hepcidin expression of Hjv^−/− mice, regardless of sex. These observations were confirmed at the protein level (Figure 1B). Gene expression of Id1 (Figure 1C) and Smad7 (Figure 1D), 2 targets of Bmp/Smad signaling, was also further repressed in double knockout mice, compared with single knockouts, providing indirect evidence that deletion of Bmp6 further alters Smad1/5/8 signaling in the liver of Hjv knockout mice. This was confirmed by western blot (Figure 2C lanes 11-13).

In line with the repression of hepcidin, targeted disruption of the Bmp6 and/or the Hjv gene leads to a strong increase in transferrin saturation (supplemental Figure 1A, available on the Blood Web site) and liver iron accumulation (supplemental Figure 1B) in mice of both sexes. However, whereas Bmp6^−/− or Hjv^−/− males have iron deposits in acinar cells of the exocrine pancreas (Figure 1E; supplemental Figure 2A) and in the heart (Figure 1F; supplemental Figure 2B), females do not accumulate iron in these extrahepatic tissues, which reflects the fact that, in the absence of testosterone, their hepcidin is less strongly repressed¹⁰  and their NTBI is lower (Figure 1G) than in males. Nevertheless, the concomitant loss of Bmp6 suppresses this hepcidin advantage over males. As a consequence, and in contrast to single Bmp6^−/− or Hjv^−/− females, double knockout females have substantial NTBI amounts and massive iron loading in all extrahepatic tissues examined (supplemental Figure 2A-B).

Interestingly, although LPS significantly induces hepcidin mRNA (Figure 2A) and protein (Figure 2B) not only in wild-type males and females but also in single and double knockout animals, the level reached after stimulation depends on basal expression of hepcidin. It is highest in wild-type mice, intermediate in Bmp6^−/− and in Hjv^−/− mice, and lowest in double knockout animals. This is compatible with the proposed synergy between interleukin 6/STAT3 and BMP/SMAD signaling in regulating hepcidin during inflammation.¹⁵  Smad5 signaling was previously shown to be activated by activin B as a consequence of LPS stimulation¹⁶  but to have no impact on hepcidin induction.¹³  Here, although activation of Smad5 by LPS was similar in wild-type and in Bmp6^−/− mice (Figure 2C-D lanes 3-6), the level of hepcidin reached after stimulation was much lower in Bmp6^−/− mice. In contrast, Hjv^−/− mice present with reduced Smad5 activation compared with Bmp6^−/− mice (Figure 2C-D lanes 5-8) but have similar induction of hepcidin. These data confirm recent in vitro observations showing that HJV augments Smad5 signaling by activin B.¹⁷  They also definitely indicate the lack of relationship between activation of Smad1/5/8 signaling by inflammatory stimuli, which is facilitated by HJV, and elevation of hepcidin expression.

In conclusion, analysis of this Bmp6^−/− × Hjv^−/− intercross clearly shows that deletion of both Bmp6 and Hjv further represses hepcidin and aggravates the phenotype of single knockout animals. This indicates that, when 1 actor of the major hepcidin signaling pathway is lacking, alternative pathways, although less efficient to activate Smad1/5/8, succeed in maintaining hepcidin to a level avoiding extrahepatic iron accumulation in females. These can be initiated by interaction of HJV with liver sinusoidal endothelial cell–produced BMP2 whose deletion in mice also leads to iron overload,^18,19  or by direct binding of BMP6 to preformed BMP type I/type II receptor complexes that exist at the membrane in the absence of ligands and coreceptors.^20-22  The suppression of these alternative pathways, as here in Bmp6/Hjv double knockout animals, leads to greater repression of hepcidin in mice of both sexes and substantial exacerbation of the extrahepatic iron overload phenotype in females. Our data also show that induction of hepcidin by LPS in vivo is linked to BMP/Smad signaling before but not after stimulation. Notably, the less severely affected Bmp6^−/− or Hjv^−/− females produce, when challenged with LPS, more hepcidin than unchallenged wild-type mice. Thus, in females, treatments targeting BMP type I receptors, previously shown to attenuate induction of hepcidin gene expression by various inflammatory stimuli,^23-25  will be more effective against anemia of inflammation than treatments that would target only HJV, as the latter would not prevent BMP6 to signal to hepcidin independently of HJV.

The authors thank Rachel Balouzat, Florence Capilla, and Yara Barreira (US006 ANEXPLO, Toulouse) for their technical assistance and help in the mouse breeding.

This work was supported by grants from Fondation pour la Recherche Médicale (DEQ2000326528), Agence Nationale de la Recherche (ANR-13-BSV3-0015-01), and the Programme des Investissements d'Avenir Aninfimip (ANR-11-EQPX-0003).

Contribution: C.L. genotyped the mice; C.L. and O.G. killed the mice and sampled the different tissues; C.L. performed reverse transcription–PCR experiments and hepcidin ELISAs; C.B.-F. did quantitative iron measurements and western blots; O.G. performed Perls Prussian blue staining of deparaffinized tissue sections; D.M. helped with data analysis and interpretation; and M.-P.R. and H.C. led and supervised the project through all stages, helped in data analyses, and wrote the manuscript with suggestions and comments from all authors.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Hélène Coppin, Institut de Recherche en Santé Digestive, INSERM U1220 Bat B, CHU Purpan, Place du Docteur Baylac, CS 60039, F-31024 Toulouse Cedex 3, France; e-mail: helene.coppin@inserm.fr; and Marie-Paule Roth, Institut de Recherche en Santé Digestive, INSERM U1220 Bat B, CHU Purpan, Place du Docteur Baylac, CS 60039, F-31024 Toulouse Cedex 3, France; e-mail: marie-paule.roth@inserm.fr.