Pre-emptive iron supplementation prevents myocardial iron deficiency and attenuates adverse remodelling after myocardial infarction

Abstract

Aims

Heart failure (HF) after myocardial infarction (MI) is a major cause of morbidity and mortality. We sought to investigate the functional importance of cardiac iron status after MI and the potential of pre-emptive iron supplementation in preventing cardiac iron deficiency (ID) and attenuating left ventricular (LV) remodelling.

Methods and results

MI was induced in C57BL/6J male mice by left anterior descending coronary artery ligation. Cardiac iron status in the non-infarcted LV myocardium was dynamically regulated after MI: non-haem iron and ferritin increased at 4 weeks but decreased at 24 weeks after MI. Cardiac ID at 24 weeks was associated with reduced expression of iron-dependent electron transport chain (ETC) Complex I compared with sham-operated mice. Hepcidin expression in the non-infarcted LV myocardium was elevated at 4 weeks and suppressed at 24 weeks. Hepcidin suppression at 24 weeks was accompanied by more abundant expression of membrane-localized ferroportin, the iron exporter, in the non-infarcted LV myocardium. Notably, similarly dysregulated iron homeostasis was observed in LV myocardium from failing human hearts, which displayed lower iron content, reduced hepcidin expression, and increased membrane-bound ferroportin. Injecting ferric carboxymaltose (15 µg/g body weight) intravenously at 12, 16, and 20 weeks after MI preserved cardiac iron content and attenuated LV remodelling and dysfunction at 24 weeks compared with saline-injected mice.

Conclusion

We demonstrate, for the first time, that dynamic changes in cardiac iron status after MI are associated with local hepcidin suppression, leading to cardiac ID long term after MI. Pre-emptive iron supplementation maintained cardiac iron content and attenuated adverse remodelling after MI. Our results identify the spontaneous development of cardiac ID as a novel disease mechanism and therapeutic target in post-infarction LV remodelling and HF.

Graphical Abstract

Open in new tabDownload slide

Time of primary review: 21 days

1. Introduction

Developing adverse cardiac remodelling and heart failure (HF) after myocardial infarction (MI) remains a major cause of morbidity and mortality.¹ The remodelling process is driven by increases in wall stress, neurohormonal activation, and inflammation.² Systemic iron deficiency (ID) is a frequent comorbidity in HF that aggravates exercise intolerance and is associated with worse outcomes.^3–5 Iron supplementation improves exercise capacity and prevents hospitalizations in HF patients with systemic ID.^6–12

HF patients may also develop ID in the myocardium.^13–17 Cardiac ID is associated with disease severity and may contribute to abnormal mitochondrial function.^15–17 Indeed, studies in animal models and human embryonic stem cell-derived cardiomyocytes have demonstrated that ID in cardiomyocytes impairs energy metabolism, resulting in cardiomyocyte dysfunction and hypertrophy.^18–22 Importantly, cardiac iron concentration in HF patients is not associated with systemic iron parameters or anaemia,^14–16 suggesting that cardiac and systemic iron homeostasis are independently regulated. However, the mechanisms promoting the development of cardiac ID remain unknown.²³

Systemic and intracellular iron homeostasis is controlled by two interrelated regulatory systems.^24,25 Systemic iron levels depend on intestinal iron absorption and iron recycling from haemoglobin. The liver-derived peptide hormone hepcidin controls duodenal iron uptake as well as iron release from the liver and reticuloendothelial system. Hepcidin binds to ferroportin, the only known cellular iron exporter, inducing its internalization and intracellular degradation,^24,25 thereby regulating iron concentration in the circulation. Intracellular iron availability is secured by iron regulatory proteins (IRPs) promoting transferrin receptor 1 (TFR1) expression and reducing ferroportin expression.^22,25 Transferrin-bound iron is taken up by cells from the circulation via TFR1.²⁵ In addition to regulating systemic iron, hepcidin also controls iron homeostasis in tissues.²⁶ Data from a cardiomyocyte-specific hepcidin knockout mouse model indicated that reduced hepcidin expression leads to cardiac ID due to increased expression of ferroportin.²⁷ Whether the hepcidin-ferroportin axis is involved in the development of cardiac ID in HF remains unknown.

Here, we report that mice develop cardiac ID during long-term follow-up after MI and identify cardiac hepcidin suppression and membrane-localized ferroportin up-regulation as a potential mechanism. We show that pre-emptive iron supplementation prevents cardiac ID and attenuates post-MI remodelling. The spontaneous development of cardiac ID therefore emerges as a disease mechanism and therapeutic target after MI.

2. Methods

2.1 Materials

Mouse endothelin-1 (ET1) and interleukin-6 (IL6) were purchased from R&D Systems (Wiesbaden-Nordenstadt, Germany). Ferric carboxymaltose (FCM) was obtained from Vifor Pharma (München, Germany). Unless otherwise stated, all other chemicals were purchased from Sigma-Aldrich (Schnelldorf, Germany).

2.2 Mouse surgery and functional assessment

All surgical procedures were approved by the authorities in Hannover, Germany (Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit; no. 33.12-42502-04-18/2896), and performed according to the European Parliament Directive 2010/63/EU on the protection of animals used for scientific purposes. C57BL6/J mice were purchased from Charles River (Göttingen, Germany) and provided with water and food (TD.80394, ENVIGO, Venray, the Netherlands; iron content 48 ppm) ad libitum. MI was induced in 8-week-old mice by left anterior descending coronary artery (LAD) ligation. In sham-operated control mice, the ligature around the LAD was not tied. Male mice were used, unless otherwise stated. Mice were subcutaneously (s.c.) pre-treated with 2 mg/kg butorphanol (Pfizer, Freiburg, Germany). Anaesthesia was induced with 3–4% isoflurane (Baxter, Unterschleißheim, Germany) inhalation in an induction chamber. After oral intubation, anaesthesia was maintained by 1.5–2% isoflurane. A left thoracotomy was performed, and the LAD was ligated with a Prolene 7-0 (Ethicon, Norderstedt, Germany) suture. After layered wound closure, mice were allowed to recover in a 32°C incubator. High-resolution two-dimensional transthoracic echocardiography (Vevo 3100, VisualSonics, Amsterdam, the Netherlands) was performed with a linear 30 MHz transducer in mice sedated with 1–2% isoflurane. We recorded LV end-diastolic area (LVEDA) and end-systolic area (LVESA) from the long-axis parasternal view and calculated fractional area change as [(LVEDA − LVESA)/LVEDA] × 100 (%). LV pressure–volume loops were recorded with a 1.4 Fr micromanometer-tipped conductance catheter inserted via the right carotid artery (SPR839, Millar Instruments, Houston, TX, USA). Mice were s.c. pre-treated with 2 mg/kg butorphanol and anaesthesia was induced with 3–4% isoflurane. After oral intubation, mice were intraperitoneally injected with 0.8 mg/kg pancuronium (Actavis, Langenfeld, Germany) and anaesthesia was maintained with 2% isoflurane. Steady-state pressure–volume loops were sampled at a rate of 1 kHz and analysed with LabChart 7 Pro software (ADInstruments, Mannheim, Germany). FCM (15 µg/g body weight) or isotonic saline was injected with a 30 G needle in the tail vein at 12, 16, and 20 weeks after MI.

2.3 Tissue collection and analysis

Mice were sacrificed by cervical dislocation under 3–4% isoflurane anaesthesia. Tissues were collected at different time points (4, 10, and 24 weeks) after sham or MI surgery. We used the non-infarcted region of the left ventricle (basal part of the interventricular septum) for all analyses. Corresponding parts of the left ventricle were obtained from sham-operated animals. Tissues were snap-frozen in liquid nitrogen and stored at −80°C. For histology, tissues were either embedded in optimal cutting temperature (OCT, Tissue-Tek, Sakura Finetek Europe, Alphen aan den Rijin, the Netherlands) compound and frozen (for preparing cryosections) or fixed in 4% paraformaldehyde solution (for paraffin sections). Scar sizes were determined in 7 µm LV cryosections stained with Masson’s trichrome. Scar sizes were calculated as the average ratio of scar length to total LV circumference in basal, midventricular, and apical sections. Midventricular cryosections were stained with rhodamine-conjugated wheat germ agglutinin (WGA, RL-1022, Vector Laboratories, Newark, CA, USA) to visualize cardiomyocyte borders. Cell nuclei were detected with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, Thermo Fisher Scientific, Schwerte, Germany).

2.4 Cardiomyocyte isolation and culture

Neonatal rat cardiomyocytes (NRCMs) were isolated from 1- to 3-day-old Sprague–Dawley rats (Charles River).²⁸ To simulate ischaemia, cells were incubated in ischaemia medium under <1% oxygen atmosphere for 3 h, as previously described.²⁸ Adult mouse ventricular cardiomyocytes were isolated by enzymatic digestion using a Langendorff apparatus.²²

2.5 Human tissue samples

We analysed transmural LV basal septum tissue samples from patients undergoing heart transplantation for advanced HF due to ischaemic cardiomyopathy (ICM; n = 11, 10 males, age range 40–70 years) or dilated cardiomyopathy (DCM; n = 10, 7 males, age range 22–59 years). Hearts with septal scarring were not considered. The protocol was approved by the Ethics Committee of Hannover Medical School. All patients provided written informed consent. Control samples were obtained from two sources; we analysed LV tissue samples from individuals who died from non-cardiac causes and had no obvious cardiac disease. These samples were collected during a medico-legal autopsy at the Institute of Forensic Medicine at Hannover Medical School (n = 8, 5 males, age range 44–89 years). The Ethics Committee of Hannover Medical School approved the use of these anonymized tissue samples. Due to the time span from death to sample processing, hepcidin mRNA expression could not be analysed in these samples. Therefore, we also included non-failing human hearts collected within the Wisconsin Donor Network (n = 10, 2 males, age range 21–58 years). The protocol was approved by the Medical College of Wisconsin Institutional Review Board. All samples were flash-frozen in liquid nitrogen and stored at −80°C. All human data were stored in a de-identified manner on a secure database. The investigation was conducted according to the principles of the Declaration of Helsinki.

2.6 Tissue iron measurement

Total iron concentration was measured by inductively coupled plasma-optical emission spectroscopy (ICP-OES) in non-infarcted LV myocardium as previously described.¹⁶ LV myocardial non-haem iron concentration was measured by a modified ferrozine-based colorimetric method.²⁹ Briefly, a homogenized tissue sample was mixed with an equal volume of extraction buffer (25% trichloroacetic acid; 4% sodium pyrophosphate). After heating at 65°C for 2 h, 50 µL of the sample supernatant was mixed with 150 µL of ferrozine solution (50 mmol/L ascorbic acid, 1.7 mmol/L ferrozine, 1.8 mol/L sodium acetate) and the colour density was measured at 562 nm on a microplate reader (Synergy HT, BioTek, Bad Friedrichshall, Germany). To measure LV myocardial haem concentration, a homogenized tissue sample was mixed with 2 mol/L oxalic acid and heated at 95°C for 30 min to generate fluorescent porphyrin from haem. The fluorescence of porphyrin from the sample was detected at 360 nm excitation and 590 nm emission on the microplate reader.³⁰

2.7 Systemic iron status

Plasma iron concentration and unsaturated iron binding capacity (UIBC) were analysed using the Iron Direct Method kit (Ferene) and the Iron UIBC kit (Biolabo, Maizy, France), respectively. Transferrin saturation is the ratio (%) of plasma iron concentration to total iron binding capacity, which is the sum of plasma iron concentration and UIBC. Plasma ferritin and plasma hepcidin concentrations were evaluated by ELISA (ab157713, Abcam, Cambridge, UK and HMC-001, Intrinsic Sciences, La Jolla, CA, USA, respectively). Haematological parameters were analysed with an ABC haematology analyser (Scil Vet, Viernheim, Germany).

2.8 Ferroportin membrane localization

Ferroportin membrane localization was visualized in 2 µm LV paraffin sections using a rabbit polyclonal antibody (MTP11-A, Alpha Diagnostics, San Antonio, TX, USA) and an Alexa Fluor 594-labelled secondary antibody (A32740, Invitrogen). Cardiac myocyte borders were visualized with fluorescein-conjugated WGA (FL-1021, Vector Laboratories). Images were acquired with a 40× objective by confocal microscopy (Olympus FV1000, Olympus, Hamburg, Germany) and then recompiled using Image J software. Ferroportin was detected in NRCMs with the MTP11-A antibody and an Alexa Fluor 488-labelled secondary antibody (A32731, Invitrogen). Images were acquired by fluorescence microscopy (Zeiss Axio Observer.Z1, Jena, Germany). Cell nuclei were stained with DAPI.

2.9 Hepcidin silencing

NRCMs were transfected with siRNAs (16 pmol/10⁵ cells) using the Lipofectamine RNAiMAX reagent and pre-designed siRNAs from Thermo Fisher Scientific (hepcidin, Silencer Select pre-designed siRNA, #4390815; scrambled, Silencer Select negative control, #4390771). Cells were cultured for 48 h after transfection and then harvested for RNA preparation or fixed in 3% paraformaldehyde.

2.10 Electrophoretic mobility shift assay

Electrophoretic mobility shift assay (EMSA) was performed using the Lightshift chemiluminescent RNA EMSA kit (Thermo Scientific, Darmstadt, Germany), following the manufacturer’s instructions. Briefly, LV tissue was homogenized in Munroe buffer (10 mmol/L HEPES, pH 7.6; 3 mmol/L MgCl₂; 40 mmol/L KCl; 5% glycerol; 0.2% NP-40) with protease and phosphatase inhibitors (Roche, Mannheim, Germany) and then centrifuged for 10 min at 4°C to remove debris. After determining protein concentration by the Bradford assay (Bio-Rad, Feldkirchen, Germany), 20 µg of protein was incubated with a 3´-biotinylated RNA probe containing ferritin’s iron regulatory element (IRE) region for 10 min at room temperature. RNA–protein complexes were separated on a 6% non-denaturing polyacrylamide gel and transferred to a negatively charged nylon membrane (Amersham, Taufkirchen, Germany). Transferred RNA–protein complexes were cross-linked to the membrane under 120 mJ/cm² UV light (Biometra, Göttingen, Germany) for 1 min. After inducing chemiluminescence from biotin-labelled RNA, the signal was recorded using an ImageQuant LAS 4000 system (GE Healthcare, Uppsala, Sweden).

2.11 Reverse transcription-quantitative polymerase chain reaction

Total RNA was isolated from cells and tissues with the RNeasy RNA isolation kit (Qiagen, Hilden, Germany) followed by reverse transcription into cDNA (SuperScript III reverse transcriptase, Thermo Fisher Scientific). mRNA expression levels were determined by qPCR using SYBR green-based or TaqMan gene expression assays and reagents from Thermo Fisher Scientific or Roche LifeScience (see Supplementary material online, Table S1).

2.12 Isolation of membrane fractions

Non-infarcted LV myocardium or freshly harvested NRCM cell pellets were homogenized in 500 µL homogenizing buffer (20 mmol/L KH₂PO₄, 1 mmol/L EDTA, 135 mmol/L KCl, and protease inhibitor) using a Dounce homogenizer. After centrifugation at 1000g for 5 min, the supernatant was subjected to another centrifugation step at 16 400 g for 1 h. The pellet was resuspended in resuspension buffer (20 mmol/L KH₂PO₄, 10 mmol/L EDTA, 1 mmol/L KCl, and protease inhibitor). All procedures were performed on ice or at 4°C.

2.13 Immunoblotting

Tissue lysates were prepared in RIPA buffer with protease and phosphatase inhibitors (Roche), separated by SDS–polyacrylamide gel electrophoresis, and transferred to nitrocellulose membranes (Amersham). Membranes were incubated with primary antibodies followed by HRP-linked secondary antibodies for subsequent detection using enhanced chemiluminescence (ECL) solution. Band densities were evaluated with Image J software (version 1.49v, National Institutes of Health). Na⁺K⁺ATPase was used as a loading control when studying membrane fractions. Antibodies were purchased from Abcam [glyceraldehyde-3-phosphate dehydrogenase (GAPDH), clone mAbcam 9484, #ab9482; oxidative phosphorylation (OXPHOS) monoclonal antibody cocktail for the detection of Complex 1–5, #ab110413]; Alpha Diagnostics (ferroportin, polyclonal, #MTP11-A); Cell Signaling Technology, Leiden, the Netherlands [ferritin H (FtH), polyclonal, #3998; Na⁺K⁺ATPase, polyclonal, #3010; phospho-STAT3 (S727), polyclonal, #9134; STAT3, clone 79D7, #4904; phospho-SMAD1/5/9 (S465), clone D5B10, #13820; SMAD1, clone D59D7, #6944; SMAD5, clone D4G2, #12534; phospho-AKT (S473), polyclonal, #9271; total AKT, polyclonal, #9272; phospho-mTOR (S2448), clone D9C2, #5536; total mTOR, clone 7C10, #2983; phospho-ERK1/2 (T202/Y204), clone D13.14.4E, #4370; p44/42 MAPK (ERK1/2), clone 137F5, #4695], and Thermo Fisher (TFR1, clone H68.4, #13-6890).

2.14 Electron transport chain complex I enzyme activity

LV tissue samples were homogenized in ice-cold PBS and Complex I activity was determined in 50 µg/mL tissue lysate using a colorimetric assay from Abcam (ab109721).

2.15 Statistical analysis

Data are presented as median with interquartile range. As some group data were not normally distributed, non-parametric statistical tests were applied to compare experimental groups. The Mann–Whitney U test was used to compare two groups, and the Kruskal–Wallis test with Dunn’s multiple comparisons test was used to compare more than two groups. We considered a two-tailed P-value of <0.05 to indicate statistical significance. All analyses were performed with GraphPad Prism software (version 8).

3. Results

3.1 Mice develop cardiac ID during long-term follow-up after MI

Mice developed LV dilation and progressive LV dysfunction after MI (Figure 1A and B). Compared with sham-operated controls, total LV myocardial iron concentration was unchanged at 4 weeks but reduced at 24 weeks after MI (Figure 1C). LV myocardial non-haem iron concentration was higher 4 weeks after MI compared with sham-operated mice (Figure 1D). Thereafter, non-haem iron declined and was significantly lower (−16%) than in sham-operated mice at 24 weeks (Figure 1D). LV myocardial haem concentration remained unchanged after MI (Figure 1E).

Female mice had similar LV remodelling and comparably reduced LV myocardial non-haem iron concentration 24 weeks after MI (see Supplementary materialonline, Figure S1A–C).

LV myocardial FtH expression was higher at 4 weeks and lower at 24 weeks after MI than in sham-operated mice (Figure 1F). Twenty-four weeks after MI, LV myocardial expression of the iron–sulfur cluster (ISC)-containing Complex I of the mitochondrial electron transport chain (ETC) was significantly reduced compared with sham-operated animals, whereas expression levels of the haem and/or ISC-containing complexes II–IV and complex V were not affected (Figure 1G). These long-term follow-up data indicate, for the first time, that mice spontaneously develop cardiac ID after MI.

3.2 MI dynamically alters LV myocardial iron homeostasis

To explore potential mechanisms leading to cardiac ID, we investigated LV myocardial expression and activity of key iron homeostasis-related proteins in sham-operated and infarcted hearts. LV myocardial IRP1–IRE binding activity was increased at 4 weeks and IRP2–IRE binding activity at 4 and 24 weeks after MI compared with sham-operated mice (see Supplementary material online, Figure S2), which could not explain the observed expression changes in TFR1 and ferroportin after MI.^22,25 Overall and membrane-localized TFR1 expression in the non-infarcted LV myocardium were elevated at 4 weeks but not at 24 weeks after MI (Figure 2A–C). In whole cell lysate, ferroportin expression was higher at 4 and 24 weeks after MI (Figure 2A and D). In membrane fractions, ferroportin abundance was decreased at 4 weeks but increased at 24 weeks after MI compared with sham-operated mice (Figure 2A and E). Likewise, ferroportin expression was augmented, while TFR1 expression was unchanged in ventricular cardiomyocytes isolated from the heart 24 weeks after MI (Figure 2F). Ferroportin immunostaining of the non-infarcted LV myocardium revealed a diffuse intracellular staining pattern at 4 weeks after MI and a shift towards the cell periphery, presumably the plasma membrane, at 24 weeks after MI (Figure 2G). Increased membrane-localized ferroportin abundance (promoting iron efflux) therefore emerged as a potential driver of cardiac ID during post-MI remodelling (Figure 2A, E, and G).

3.3 Suppressed myocardial hepcidin expression long term after MI

While increased at 4 weeks, hepcidin (Hamp) mRNA expression in the non-infarcted LV myocardium was unchanged at 10 weeks but strongly down-regulated at 24 weeks after MI, compared with sham-operated controls (Figure 3A). In contrast, hepcidin expression in the liver and plasma hepcidin concentrations remained unchanged during long-term follow-up after MI (Figure 3B and C). Cardiac hepcidin suppression along with increased membrane-localized ferroportin were also observed in female mice 24 weeks after MI (see Supplementary material online, Figure S1D and E).

3.4 Systemic iron status after MI

Peripheral blood counts, haemoglobin concentration, plasma TSAT, and ferritin concentrations did not significantly differ between MI and sham-operated mice at 4 and 24 weeks (see Supplementary material online, Table S2). Plasma TSAT and plasma iron and ferritin concentrations were not correlated with LV non-haem iron at 24 weeks after MI (see Supplementary material online, Figure S3 A–C).

3.5 Exploring the regulation of cardiac hepcidin expression after MI

Next, we explored potential mechanisms regulating cardiac hepcidin expression. LV myocardial expression levels of Il6 mRNA and phosphorylation of STAT3 were increased at 4 weeks after MI, but declined to basal (sham-operated) levels at 24 weeks after MI, which may have contributed to the suppression of hepcidin expression at the later time point (see Supplementary material online, Figure S4A and B).³¹

In cultured cardiomyocytes, IL6 strongly enhanced hepcidin mRNA expression, whereas stimulation with ET1 or simulated ischaemia diminished it (Figure 3D and E). siRNA-mediated hepcidin down-regulation increased membrane-localized ferroportin expression in cultured cardiomyocytes as shown by immunoblotting and immunofluorescence microscopy (Figure 3F–H), suggesting that hepcidin acts as an autocrine regulator of ferroportin in cardiomyocytes. Interestingly, ET1-mediated down-regulation of hepcidin (Figure 3D) was associated with greater membrane-localized ferroportin abundance (Figure 3I).

As simulated ischaemia elevated Hif1α (see Supplementary material online, Figure S4C) and suppressed hepcidin mRNA expression (Figure 3E) in cultured cardiomyocytes, we investigated whether hepcidin down-regulation after MI is associated with LV myocardial hypoxia. However, we observed higher Hif1α expression only at 4, but not 24 weeks after MI, compared with sham-operated mice (see Supplementary material online, Figure S4D), suggesting that hypoxia is not involved in the suppression of hepcidin expression at 24 weeks after MI. Assessing other signalling pathways potentially involved in hepcidin regulation (TGFβ/SMAD4, ERK, PI3K/AKT, and mTOR)^32–34 did not reveal changes that could explain hepcidin suppression in the LV myocardium late after MI (see Supplementary material online, Figure S4E).

3.6 Impaired hepcidin-ferroportin axis in iron-deficient failing human hearts

Consistent with previous reports,^14,15 LV myocardial specimens from patients with advanced HF contained less total and non-haem iron than samples from individuals who died from non-cardiac causes (Figure 4A and B), regardless of HF aetiology (ischaemic or dilated). As observed long term after mouse MI, hepcidin mRNA expression was down-regulated in the failing human myocardium compared with control hearts (Figure 4C). Conversely, expression of ferroportin and its membrane-bound fraction were significantly increased in the failing myocardium (Figure 4D–F), changes that appeared to be independent of age and sex (see Supplementary material online, Figure S5).

3.7 Iron supplementation after MI preserves LV function

To explore whether iron supplementation can prevent cardiac ID and attenuate adverse LV remodelling, we randomized mice to intravenous FCM or saline injections at 12, 16, and 20 weeks after MI (Figure 5A). Consistent with the data shown in Figure 1D, saline-injected MI mice had a lower LV myocardial non-haem iron concentration than saline-injected sham-operated mice at 24 weeks (Figure 5B). Iron supplementation increased non-haem iron to comparable absolute levels in MI and sham-operated mice (Figure 5B). Further, iron supplementation raised plasma ferritin concentrations at 24 weeks in both groups (Table 1). At randomization (12 weeks after MI), both MI groups had similar LV dimensions and systolic dysfunction (Figure 5C and Supplementary material online, Table S3). Subsequently, saline-injected MI mice developed progressive LV dilatation and systolic dysfunction until 24 weeks after MI, whereas FCM injections preserved LV dimensions and function during follow-up (Figure 5C and Supplementary material online, Table S4). At 24 weeks after MI, FCM-treated MI mice had less LV dilatation and more preserved LV ejection fraction (LVEF) and stroke work compared with saline-injected MI mice, as revealed by pressure–volume measurements (Figure 5D–G and Supplementary material online, Table S4). FCM did not affect LV function in sham-operated mice (Figure 5D–G and Supplementary material online, Table S4). Iron supplementation did not alter infarct scar size (see Supplementary material online, Figure S6A) or cardiomyocyte hypertrophy at 24 weeks after MI (see Supplementary material online, Figure S6B). However, iron supplementation rescued the MI-related reductions in ETC Complex I expression and activity (Figure 5H and I).

4. Discussion

Ours is the first study to demonstrate that cardiac ID spontaneously develops long term after MI. Cardiac ID developed independently of systemic ID and was associated with suppressed myocardial hepcidin and increased membrane-localized ferroportin expression in infarcted mouse hearts (which may promote cardiac iron depletion). Notably, these pivotal changes in myocardial iron homeostasis long term after experimental MI were likewise observed in failing human hearts. In mice, pre-emptive iron supplementation prevented cardiac ID and attenuated adverse post-MI remodelling, thus identifying cardiac ID as a therapeutic target.

ID exerts negative effects on cardiac structure and function. In rodent models, ID anaemia results in LV hypertrophy and dysfunction due to inefficient energy metabolism and diminished cardiomyocyte calcium transients.^19,35,36 More specifically, mice genetically engineered to selectively develop cardiac ID exhibit cardiac hypertrophy and dysfunction independent of systemic iron status.^20,22,27 Iron supplementation either prevented or reversed cardiomyopathy in these murine ID models emphasizing the importance of iron for proper cardiac function.^20,22,35,37

Most previous studies investigating cardiac iron homeostasis after MI analysed cardiac iron content only at a single time point.^37–40 Studies reporting increased cardiac iron performed the analysis early after MI,^40,41 while other studies showing no alterations in cardiac iron performed the measurements at somewhat later time points, e.g. at 6–8 weeks.^37–39 In contrast, we serially assessed cardiac iron status at 4, 10, and 24 weeks after MI. This approach enabled us to uncover temporal changes in myocardial iron content and document for the first time the spontaneous development of cardiac ID late after MI.

Cardiac ID may be detrimental also in human HF. Patients with advanced HF often have cardiac ID which is associated with mitochondrial dysfunction.¹⁵ In patients with non-advanced, non-ischaemic HF, low cardiac iron content is associated with more severe disease.¹⁶ Mechanistically, ID in human cardiomyocytes impairs mitochondrial respiration and contractile force owing to reduced expression of the ISC-containing Complexes I, II, and III.^17,21 Along this line, ID is associated with greater cardiac energetic impairment in HF patients.⁴² How cardiac ID develops in chronic HF has remained unknown.

While hepcidin may regulate cardiac iron status in an autocrine manner,²⁷ it is unclear whether this peptide hormone is involved in the development of cardiac ID in HF. Some studies observed elevated cardiac hepcidin expression shortly after MI.^43–45 Another study reported no changes in cardiac hepcidin or iron 8 weeks after MI.³⁷ Our study is in line with these previous reports and found, in addition, that cardiac hepcidin expression decreases long term after MI in mice, and is reduced in failing human hearts. Hepcidin suppression was associated with increased membrane-bound ferroportin and decreased cardiac iron content, suggesting that hepcidin suppression and ferroportin up-regulation may play a causal role in the development of cardiac ID.²⁷ Mechanistically, we found that ET1 reduces hepcidin expression and enhances membrane-localized ferroportin abundance in isolated cardiomyocytes, suggesting that neurohormonal agonist(s) may contribute to cardiac hepcidin suppression after MI.

In patients with HF and systemic ID, intravenous iron supplementation reduces HF hospitalizations and improves symptoms, exercise capacity, and health-related quality of life.^6–12 Only a few studies investigated whether iron supplementation affects cardiac function and structure. These studies indicate that iron supplementation in patients with established HF and systemic ID improves LV systolic function and augments the beneficial structural effects of cardiac resynchronization therapy.^46,47 We establish that pre-emptive iron supplementation, starting before cardiac ID has developed, can attenuate/prevent further remodelling after MI.

Changes in, and the effects of, cardiac iron status might be different in the setting of an acute coronary syndrome (ACS). Elevated cardiac iron may promote oxidative stress and ferroptosis (iron-dependent cell death) during myocardial ischaemia–reperfusion injury.⁴⁸ Whether systemic ID is harmful or beneficial may depend on the time point after ACS.^49–53 For example, patients with ST-segment elevation MI and systemic ID were found to develop less myocardial reperfusion injury and to have a better in-hospital course.⁴⁹ However, studies with long-term follow-up after ACS found that systemic ID is associated with increased HF severity and mortality.^54,55 Our data suggest that cardiac ID spontaneously develops long-term after MI, contributes to adverse LV remodelling, and can be targeted by pre-emptive iron supplementation.

Our study has limitations that warrant further consideration. It was technically not feasible to quantify murine or human hepcidin protein expression in the myocardium by ELISA or liquid chromatography–mass spectrometry. We therefore determined hepcidin mRNA by RT-qPCR, as previously done by others.^43,56 While we found that a neurohormonal agonist (ET1) suppressed hepcidin expression in cardiomyocytes, the exact regulatory pathways controlling cardiac hepcidin expression in the remodelled and/or failing myocardium remain to be defined. We acknowledge that the proportion of women was higher in the non-failing group than in the advanced HF group. In addition, the age of heart donors was lower than the age of patients with ICM. However, our observations in the human samples appeared to be independent of age and sex. Experiments were conducted in male mice. However, key findings were confirmed in female mice. Iron supplementation increased plasma ferritin and moderately overcorrected the cardiac iron deficit. We used a rather high FCM dose and did not analyse iron re-distribution to the liver and other organs. Yet, in iron-supplemented sham-operated mice, a neutral effect on cardiac function was noted.

In conclusion, we identify cardiac ID as a disease mechanism promoting adverse LV remodelling long term after MI. Independent of systemic iron status, cardiac ID was associated with impaired hepcidin expression and increased membrane-bound ferroportin in murine hearts. The same pattern of dysregulated cardiac iron homeostasis and cardiac ID was found in failing human hearts. Pre-emptive iron supplementation attenuated adverse LV remodelling after MI in mice. Our findings highlight the significance of cardiac iron in preventing HF after MI.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Authors’ contributions

B.C., M.T., Y.W., F.R., V.G.H., Z.M., A.R., and C.W. performed experiments. M.K., A.B., C.B., J.D.S., and D.J. provided human tissues. B.C., M.T., J.B., K.C.W., and T.K. contributed to data interpretation and analysis. B.C., M.T., K.C.W., and T.K. devised the experimental plan and wrote the manuscript. B.C. and T.K. supervised all aspects of this study, including design, execution, and data interpretation. All authors participated in critical reviewing and revising the manuscript.

Acknowledgements

We thank Anja Guba-Quint, Ivonne Marquard, Silke Pretzer, and Luisa von Wolffersdorf for technical assistance. We are grateful to the Core Unit Laser Microscopy at Hannover Medical School for providing support with confocal microscopy.

Funding

This work was supported by the German Research Foundation (Clinical Research Unit [KFO311] to J.D.S., D.J., J.B., K.C.W., and T.K.).

Data availability

The data underlying this article will be shared upon reasonable request to the corresponding author.

References

Author notes