Macrophage-HFE controls iron metabolism and immune responses in aged mice

More than one in 200 individuals of northern European origin are affected by HFE-hemochromatosis, a highly prevalent genetic iron overload disorder.¹ Hfe deficiency specifically in hepatocytes causes low expression of the iron regulatory hormone hepcidin, and iron accumulation in organs.² Here we report a previously unrecognized role of HFE in macrophages as a regulator of iron and immune responses, which is independent of the hepatocytic role of HFE in maintaining systemic iron homeostasis.

Our previous work solved a long-standing debate by establishing that HFE acted in hepatocytes to control hepcidin levels, thereby preventing systemic iron overload. ² These studies classified HFE-hereditary hemochromatosis as a liver disease. Low hepcidin levels in HFE-hereditary hemochromatosis increase the stability of the iron exporter ferroportin (FPN), allowing for excess efflux of iron from macrophages and duodenal enterocytes into the circulation.³ As a consequence, macrophages from HFE-patients and Hfe-mutant mice are relatively iron-depleted, despite systemic iro n overload.^4,5

Macrophages play versatile roles in maintaining tissue homeostasis. They are involved in the recycling of aged red blood cells, tissue repair and regeneration, and in immune responses.⁶ Alternative activation of macrophages in response to selected sets of pro- and antiinflammatory stimuli has been linked to changes in intracellular iron levels.^7,8 On the other hand, systemic iron overload increased susceptibility of mice towards infections with Vibrio vulnificus⁹ and Yersinia pseudotuberculosis, ¹⁰ while depletion of iron by deferoxamine improved survival of mice receiving lethal doses of lipopolysaccharides (LPS).¹¹ Interestingly, lack of Hfe in mice was beneficial, protecting against invasive Salmonella enterica serovar typhimurium infection¹² and macrophages derived from Hfe^-/- mice displayed an attenuated inflammatory response to LPS and Salmonella challenge.¹³ One explanation for these findings may be that low iron levels in Hfe-deficient macrophages restrict iron as a nutrient and signal for Salmonella growth. Alternatively, Hfe may exert a previously unrecognized role in immune cells. Given that iron dyshomeostasis is associated with aging and that aging may undermine the immune responses of macrophages, we tested whether selective lack of Hfe in macrophages may affect iron and immune responses in aged mice.

Analyses of iron-related parameters in 45-week old Hfe^LysMCre mutant mice, which carry a specific deletion of Hfe in myeloid cells,² showed significant decreases of non-heme iron levels in the liver, spleen and duodenum (Figure 1A, B). The degree of iron reduction was moderate in comparison to that caused by nutritional iron deficiency (Figure 1A). Consistent with decreased iron levels in the liver and in freshly isolated primary hepatocytes from Hfe^LysMCre mice (Figure 1A, C), serum hepcidin levels were decreased, while circulating iron and erythropoietin levels were unchanged (Figure 1D). Hematologic indices and the numbers of macrophages, granulocytes, T cells and B cells in the spleen and bone marrow were not significantly different between Hfe^LysMCre mutant and control mice (Figure 1E, Online Supplementary Figure S1). In line with low serum hepcidin levels, hepatic expression of hepcidin (Hamp1) and genes known to be co-regulated with hepcidin in a BMP/SMAD-dependent manner (such as Bmp6, Smad7, Id1 and Pai1) were decreased (Figure 1F, G). By contrast, the levels of transferrin receptor 1 (TfR1), which is responsible for transferrin-mediated iron uptake, failed to increase in the liver of Hfe^LysMCre mice (Figure 1G), likely due to a milder decrease in the liver iron stores in Hfe^LysMCre mice than in mice subjected to nutritional iron deficiency (Figure 1A, Online Supplementary Figure S3A).

In the spleen of 45-week old Hfe^LysMCre mice, the decrease in iron levels correlated with increased protein levels of Fpn and TfR1, (Figure 1H), contrasting the effect of nutritional iron deficiency which was associated with low Fpn levels (Online Supplementary Figure S3B). In the duodenum of 45-week old Hfe^LysMCre mice, increased mRNA expression of iron importers such as Dcytb and Dmt1 was observed, whereas the levels of Fpn and TfR1 were not statistically different (Figure 1I).

We conclude that 45-week old Hfe^LysMCre mice show significant alterations in iron metabolism. This is in contrast to the iron state of 12-week old Hfe^LysMCre mice, which showed no changes in any of the iron-related parameters investigated (Figure 1J-M).²

We next monitored systemic responses of 45-week old Hfe^LysMCre mice to iron overload and immune challenges in the absence of macrophage-Hfe. We subjected 45-week old Hfe^LysMCre mutant and control mice to parenteral iron overload. The mice were given a single intra-peritoneal injection of iron-dextran solution. The ability to accumulate iron in tissues and to increase hepcidin was comparable between Hfe^LysMCre mutant and control mice (Figure 2A, B), suggesting that macrophage HFE is dispensable for iron accumulation and hepcidin responses in the setting of the parenteral iron overload used in this study.

By contrast, 45-week old Hfe^LysMCre mice had a better survival in response to LPS-induced endotoxin shock (10 mg/kg; LPS from E. coli 055:B5, L2630) (Figure 1H), compared to 12-week old Hfe^LysMCre mice which succumbed to the endotoxin shock at a similar rate as that in the controls (Figure 2C, D). Likewise, constitutive Hfe^-/- and hepatocyte-specific Hfe-mutant mice (Hfe^AlfpCre), and mice maintained on an iron-deficient diet prior to LPS challenge, were not protected from endotoxin shock (Figure 2E, F). Based on these data we conclude that a specific lack of Hfe in macrophages in 45-week old mice may provide an adaptive mechanism aimed at protecting the host from inflammatory injury.

Given that the release of serum hepcidin, various proand anti-inflammatory cytokines and chemokines is considered a central feature of the inflammatory response, we next measured the expression of a set of serum immune mediators in 45-week old, Hfe^LysMCre mutant mice during endotoxin shock. The overall expression pattern of hepcidin, pro- and anti-inflammatory cytokines and chemokines was similar between mutant Hfe^LysMCre mice and control animals (Online Supplementary Figure S2), implying that the endotoxin shock triggered common pathways. Interestingly, Hfe^LysMCre mutant mice showed a trend towards a higher degree of induction of granulocyte colony-stimulating factor, interferon γ, interleukin (IL)-1b, IL-17α (Online Supplementary Figure S2C) and reduced induction of eotaxin, IL2, IL4, IL5, IL6, IL10, IL13, macrophage inflammatory protein 1α, and tumor necrosis factor α (TNF-α) mediators compared to control mice (Online Supplementary Figure S2D), raising the question of whether differential expression of these immune mediators may explain the remarkably higher tolerance of aged Hfe^LysMCre mice to LPS-induced shock.

To better understand the role of macrophage-Hfe at the cellular level, we next isolated primary macrophages from Hfe^LysMCre, Hfe^AlfpCre (deletion of Hfe in hepatocytes) and constitutive Hfe knock-out (Hfe^-/-) mice and compared iron-related parameters. These mouse models enable the dissociation of the cell-autonomous actions of HFE in macrophages (present in Hfe^LysMCre and Hfe^-/- mice) from those that are a consequence of altered systemic hepcidin expression (present in Hfe^-/- and Hfe^AlfpCre mutant mice).² We showed that lack of Hfe in macrophages derived from aged Hfe^-/- and aged Hfe ^LysMCre mice correlated with reduced intracellular total iron levels whereas macrophages derived from Hfe^AlfpCre mice, which carry a wild-type Hfe allele in all extra-hepatocytic cell types, including macrophages, showed no statistically significant difference in iron levels compared to their respective controls although there was a tendency towards less iron in Hfe^AlfpCre cells (Figure 3A).

Figure 1.Lack of macrophage-Hfe in 45-week old mice contributes to moderate systemic iron deficiency in contrast to physiological iron homeostasis present in 12-week old Hfe^LysMCre mice. (A) Non-heme iron levels in the liver, spleen and duodenum from 45-week old Hfe^LysMCre mutant and Hfe^flox control mice, and from mice kept on an iron-deficient diet (IDD). n indicates the number of livers (n=19, 15, 5), spleens (n=19, 15, 5) and duodenums (n=6, 6, 5) isolated from Hfe^LysMCre, Hfe^flox and IDD mice, respectively. (B) Perls staining for iron deposits in the liver, spleen and duodenum of 45-week old Hfe^LysMCre mutant and control mice. Scale bar 50 μm (liver) and 100 μm (spleen and duodenum). Representative stainings of three sections are shown. (C) Intracellular total iron levels measured by total-reflection x-ray fluorescence in primary hepatocytes (HC), Kupffer cells (KC), liver sinusoidal endothelial cells (LSEC), and hepatic stellate cells (HSC) from 45-week old Hfe^LysMCre and Hfe^flox mice (n=3-5). (D) Serum hepcidin, iron and erythropoietin (EPO) levels, and (E) hematologic parameters including red blood cell (RBC) count, hemoglobin (Hgb) concentration, hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and the percentages of white blood cells (WBC) including lymphocytes (Lym), monocytes (Mo), and granulocytes (Gra) in the blood of 45-week old Hfe^LysMCre mutant (n=6) and Hfe^flox control mice (n=6). (F) Relative mRNA expression of iron-related genes in the liver of 45-week old Hfe^LysMCre mutant (n=6) and Hfe^flox control mice (n=9) determined by real-time polymerase chain reaction (PCR). (G, H) Representative immunoblot analysis of pSMAD1, transferrin receptor 1 (TfR1) and ferroportin (FPN), relative to bactin levels (shown in histograms on the right), in the livers and spleens of Hfe^LysMCre mutant (n=8) and control mice (n=8). (I) Relative mRNA expression of iron transporters in the duodenum of 45-week old Hfe^LysMCre mutant (n=4) and Hfe^flox control mice (n=8) measured by real-time PCR. (J-L) Non-heme iron levels (J), serum iron and hepcidin levels (K), and relative mRNA expression of hepcidin (Hamp1) and Bmp6 (L) in the livers of 12-week old Hfe^LysMCre mutant (n=5) and Hfe^flox control mice (n=6). (M) Representative immunoblot analysis of TfR1 and FPN, relative to b-actin levels (shown in histograms on the right), in the spleens of 12-week old Hfe^LysMCre mutant (n=5) and Hfe^flox control mice (n=6). M: Prestained Protein Marker PageRuler. n indicates the number of mice used in the analysis. Data are shown as mean } standard error of mean. Statistically significant differences are indicated as *P<0.05, **P<0.005, ***P<0.0005.

Consistent with our findings in the spleen of Hfe^LysMCre mice (Figure 1G), we showed that FPN protein and mRNA levels were significantly increased in Hfe-deficient macrophages while TfR1 expression was unaltered irrespective of the genotypes (Figure 3B). The mRNA expression of hepcidin and other iron-related genes, pro- and anti-inflammatory cytokines, and toll-like receptors was not affected by the lack of macrophage-Hfe (Figure 3C).

Figure 2.Lack of macrophage-Hfe in 45-week old mice is dispensable during parenteral iron overload but contributes to better survival of Hfe^LysMCre mutant mice during lipopolysaccharide-induced endotoxin shock. (A, B) Non-heme and plasma iron levels (A), and hepcidin mRNA expression (B) in 45-week old Hfe^LysMCre mutant (n=6) and control mice (n=6) following parenteral iron overload. Data are shown as the mean } standard error of mean. Statistically significant differences are indicated as *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001. n indicates the number of mice used in the analysis. (C-F) Survival of 45-week old Hfe^LysMCre mutant and control mice (n=11, 11) (C), 12-week old Hfe^LysMCre mutant and control mice (n=10, 10) (D), 45-week old Hfe^-/- and control mice, and mice maintained on an iron-deficient diet for 4 weeks (IDD) (n=13, 13, 12) (E) and 12-week old Hfe^AlfpCre, Hfe^-/- and control mice (n=9, 9, 9), which were injected with lipopolysaccharide (10 mg/kg) and monitored every 6 h for up to 4 days. The percentage survival rate was expressed by Kaplan-Meier curves. Statistical analyses were performed using the log-rank test and the results were considered statistically significant when *P<0.05. ns: not statistically significant; Fe-ip: intraperitoneal iron injections; LPS: lipopolysaccharide.

Figure 3.Hfe in macrophages controls cellular iron homeostasis via ferroportin. (A) Intracellular total iron levels in bone marrow-derived macrophages (BMDM) from 45-week old Hfe^-/-, Hfe^LysMCre and Hfe^AlfpCre mutant mice (n=3, 4). (B) Representative immunoblot analysis of transferrin receptor 1 (TfR1) and ferroportin (FPN) relative to b-actin levels (shown in histograms below) in 45-week old Hfe^-/-, Hfe^LysMCre and Hfe^AlfpCre mutant mice. (C) Relative mRNA quantification of iron-related genes, inflammatory cytokines and toll-like receptors genes by real-time polymerase chain reaction in BMDM derived from 45-week old Hfe^LysMCre mutant and control mice (n=6, 6). (D, E) Intracellular total iron levels (D) and immunoblot analysis (E) of TfR1 and FPN, relative to b-actin levels (shown in histograms on the right), in BMDM from 12-week old Hfe^LysMCre mutant and Hfe^flox control mice (n=5, 4). M: Prestained Protein Marker PageRuler (Thermo Scientific). (F) Schematic illustration of the Hfe cDNA construct cloned into the pcDNA6.2 vector under the control of the cytomegalovirus (CMV) promoter, tagged with V5 epitope at the C-terminal end and a poly-A tail. (G) Relative mRNA expression of Hfe and Fpn, and (H) intracellular total iron levels in transiently transfected BMDM with vector carrying the Hfe cDNA construct (indicated as ‘o/e Hfe’) or empty vector (‘mock’) (n=4, 5). (I) Representative immunoblot analysis, from two independent experiments, of FPN, TfR1, ferritin and relative quantification to β-actin levels (shown in histograms on the right) in BMDM transiently transfected with a vector carrying Hfe cDNA or an empty vector (‘mock’) (n=4, 6). (J) Phosphorylation status of STAT3(Ser727) in the BMDM from Hfe^LysMCre mutant and Hfe^flox control mice (n=7, 7) and in BMDM transiently transfected with a vector carrying Hfe cDNA or an empty vector (‘mock’) (n=9, 9) measured by Bio-Plex Pro Cell Signaling MAPK Panel. AU: arbitrary units. Data are shown as mean } standard error of mean. Statistically significant differences are indicated as *P<0.05, **P<0.005. n indicates the number of mice used in the analysis.

Importantly, intracellular total iron levels, FPN and TfR1 protein expression, were not changed in macrophages isolated from 12-week old Hfe^LysMCre mutant mice (Figure 3D, E), supporting our previous findings.²

Conversely, overexpression of Hfe in primary wild-type macrophages counteracted FPN induction and intracellular iron deficiency (Figure 3F-I). We observed an interesting correlation between HFE expression in macrophages and the STAT3 signaling pathway, whereby the latter’s activity was significantly decreased in Hfe-deficient macrophages and 3-fold increased upon Hfe overexpression (Figure 3J). Additional phospho-proteins studied, including pp38MAPK, pERK1/2, pMEK1, pJNK, pATF-2, pp90RSK, pHSP27, and p53, were not significantly different between Hfe-deficient and Hfe-overexpressing macrophages (data not shown). The contribution of STAT3 signaling to FPN expression in Hfe-deficient macrophages deserves further investigations.

Collectively, the results from primary macrophages show that macrophage-Hfe expression controls iron accumulation by regulating expression of the iron exporter FPN. However, the LysM-Cre does not target all macrophage types with the same efficacy and its expression varies among tissues. This may explain the lack of an iron-poor phenotype in Küpffer cells in contrast to bone marrow-derived macropahges. Our findings are in line with an earlier study by Drakesmith et al., who showed that HFE expression in THP1 macrophages inversely correlated with FPN levels and iron release, whereas iron uptake by TfR1 was not affected.¹⁴ Other studies using mouse or human Hfe/HFE-deficient blood monocyte/macrophage cultures^4,5 also demonstrated increased FPN expression in isolated cells. Similarly, reciprocal bone marrow¹⁵ and liver transplantation⁵ studies between wild-type and Hfe^-/- mice suggested a diminished capacity of macrophages to store iron, being consistent with a function of HFE in inhibiting iron release.¹⁴ However, under all these conditions, the observed effects could not be dissociated from the systemic effects of hepcidin.

In summary, our findings provide compelling new data on the long-standing question of the role of HFE in macrophages.

Footnotes

Correspondence

MAJA VUJIĆ SPASIĆ - maja.vujic@uni-ulm.de

Disclosures: no conflicts of interests to disclose.

Contributions: conceptualization: MVS ; methodology: NKT, DY, KW, AG, SC and KL; investigation and formal analysis: NKT, KW, AG, SC, KL, MUM and MVS; original draft: MVS; review & editing: all authors contributed.

Funding

This work was supported by the German Research Foundation (DFG, VU75/2-1) and by Ulm University (support for MVS). MUM acknowledges funding from the DFG (SFB1036) and support from the late Dr. Andre Bachmann, who strongly believed in a beneficial role of the C282Y genotype.

Acknowledgments

We thank the staff of the Animal Facility at Ulm University and Prof. Gröne (DKFZ, Heidelberg, Germany) for their contribution to this work and Dr. S. Altamura for providing the samples from FpnKI mice.