Ex vivo modelling of cardiac injury identifies ferroptosis-related pathways as a potential therapeutic avenue for translational medicine

Abstract

Background: Heart failure (HF) is a burgeoning health problem worldwide. Often arising as a result of cardiac injury, HF has become a major cause of mortality with limited availability of effective treatments. Ferroptotic pathways, triggering an iron-dependent form of cell death, are known to be potential key players in heart disease. This form of cell death does not exhibit typical characteristics of programmed cell death, and is mediated by impaired iron metabolism and lipid peroxidation signalling.

Objectives: The aim of this study is to establish an ex-vivo model of myocardial injury in living myocardial slices (LMS) and to identify novel underlying mechanisms and potential therapeutic druggable target(s).

Methods and results: In this study, we employed LMS as an ex vivo model of cardiac injury to investigate underlying mechanisms and potential therapeutic targets. Cryoinjury was induced in adult rat LMS, resulting in 30 % tissue damage. Cryoinjured LMS demonstrated impaired contractile function, cardiomyocyte hypertrophy, inflammation, and cardiac fibrosis, closely resembling in vivo cardiac injury characteristics. Proteomic analysis revealed an enrichment of factors associated with ferroptosis in the injured LMS, suggesting a potential causative role. To test this hypothesis, we pharmacologically inhibited ferroptotic pathways using ferrostatin (Fer-1) in the cryoinjured rat LMS, resulting in attenuation of structural changes and repression of pro-fibrotic processes. Furthermore, LMS generated from failing human hearts were used as a model of chronic heart failure. In this model, Fer-1 treatment was observed to reduce the expression of ferroptotic genes, enhances contractile function and improves tissue viability. Blocking ferroptosis-associated pathways in human cardiac fibroblasts (HCFs) resulted in a downregulation of fibroblast activation genes, a decrease in fibroblast migration capacity, and a reduction in reactive oxygen species production. RNA sequencing analysis of Fer-1-treated human LMS implicated metallothioneins as a potential underlying mechanism for the inhibition of these pathways. This effect is possibly mediated through the replenishment of glutathione reserves.

Conclusions: Our findings highlight the potential of targeting ferroptosis-related pathways and metallothioneins as a promising strategy for the treatment of heart disease.

Keywords: Cardiac fibroblast; Cardiac injury; Ferrostatin; Living myocardial slice; Metallothionein.

Fig. 1. Cryoinjury in healthy rat LMS induces cardiac dysfunction.

(A) Freshly cut LMS were trimmed based on fiber orientation and attached to plastic rings. Left – intact healthy LMS (rCtrl-LMS). Right – cryioinjured LMS (rAHF-LMS). (B) Representative image of live/dead staining of rAHF-LMS imaged using wide-field microscopy. (C) View of a LMS in biomimetic cultivation chamber (BMCC - InVitroSys). (D) Left – Representation of continuous contractility recording from rCtrl- and rAHF-LMS over 24 h. Middle - Time course of peak amplitude (contractility) during culture in BMCC, shown as fold change to 1 h values (*p < 0.05; two-way ANOVA; n = 6). Right – area under curve (AUC) of contractility over 24 h (*p < 0.05; Student's t-test; n = 6). (E) Force measurement of LMS in the force transducer and quantification of contractile parameters: Maximal contractility, Time to 50 % peak, Time to 90 % relaxation and Tau (τ). (*p < 0.05; Student's t-test; n_rCtrl-LMS = 5, n_rAHF-LMS = 8). Data are displayed as mean ± SEM.

Fig. 2. Cryoinjury of LMS induces remodelling in the region adjacent to the injury.

(A) rCtrl- and rAHF-LMS after 24 h in culture. rAHF were split inpto peri-injury (cryoinjury +1 mm margins) and remote regions. (B) Representative images and quantification of collagen I immunostaining in rCtrl-LMS and the remote regions of rAHF-LMS (**p < 0.01; Student's t-test, n = 4). (C) Cross sections of LMS were stained with wheat germ agglutinin (WGA) to demarcate cell boundaries. Right - quantitative analysis of cardiomyocyte cross-sectional area (*p < 0.05, ***p < 0.001; one-way ANOVA; n_rCtrl = 9, n_peri-injury = 10, n_remote = 12). (D) RT-qPCR quantification of gene expression in LMS. P = peri-injury, R = remote (*p < 0.05, **p < 0.01; one-way ANOVA; n_rCtrl = 6, n_rAHF = 8). (E) Western blot analysis of secreted proteins into the supernatant, normalized to total protein measured via Ponceau S staining (**p < 0.01, ***p < 0.001; Student's t-test; n_rCtrl = 6, n_rAHF = 7). (F) RT-qPCR quantification of miRNAs secreted to the supernatant, normalized to cel-miR-39 (*p < 0.05, **p < 0.01; Student's t-test; n_rCtrl = 5, n_rAHF = 7). Data are displayed as mean ± SEM.

Fig. 3. Mass spectrometry-based proteomics analysis of cryoinjured LMS.

(A) Diagram of proteomics analysis. 3 groups were analyzed: rCtrl-LMS, rAHF-remote (rAHF-R) and rAHF-peri-injury (rAHF-PI) (p-ajd < 0.05; |Log₂FC| > 0.5; n = 3). (B) Heatmap illustrating protein levels. Color scaling shows the levels for given proteins, red - highest, blue – lowest. (C) Venn diagram showing the overlap between DAPs in our proteomics dataset and DEGs found in a previously published RNA seq dataset from ICM and DCM patients (GSE116250). (D + E) Functional enrichment analysis of proteomics data. Selected clusters from different databases are included. X-axis: Normalized enrichment score rAHF-R/rAHF-P vs. rCtrl. Bubble size illustrates the ratio between the proteins connected to a process and the overall query size. Color of the bubble correlates with the database source. FDR < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Cardioprotective effects via inhibition of ferroptotic pathways in acute heart failure model ex vivo.

(A) Experimental design of ferroptosis inhibition (via Fer-1) in acute heart failure model in adult rat LMS. (B) Representative images and quantification of collagen I immunostaining in rCtrl and rAHF treated with either DMSO (1:1000) or Fer-1 [10 μM] (*p < 0.05, **p < 0.01; one-way ANOVA, n = 7). (C) RT-qPCR quantification of gene expression in rCtrl- and rAHF-LMS (*p < 0.05, **p < 0.01, ****p < 0.0001; two-way ANOVA; n = 12 for all conditions, except for ctrl + Fer-1 with n = 2). (D) Left - time course of peak amplitude (contractility) during culture in BMCC, shown as fold change to 1 h values (n = 4). Middle – AUC of contractility over 24 h (*p < 0.05; one-way ANOVA; n = 4). Right – time to peak analysis (*p < 0.05; one-way ANOVA; n = 4). (E) Representative images and quantification of TUNEL staining in rCtrl and rAHF treated with either DMSO (1:1000) or Fer-1 [10 μM] (****p < 0.0001; one-way ANOVA, n = 3). (F) Experimental design of ferroptosis inhibition (via Fer-1) in a non-injury based fibrotic model in adult rat LMS stimulated with 50 ng/μL TGF-β1. (G) RT-qPCR quantification of gene expression in rCtrl- and + TGF-β LMS (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two-way ANOVA; n = 6). (H) LDH release measured in supernatants collected from rCtrl- and + TGF-β LMS (***p < 0.001; one-way ANOVA; n = 6).

Fig. 5. Cardioprotective effects via Fer-1 mediated inhibition of ferroptotic pathways in chronic heart failure ex vivo model.

(A) Experimental design of ferroptosis inhibition (via Fer-1) in chronic heart failure model in human LMS. (B) RT-qPCR quantification of ferroptosis-related gene expression in hCHF (*p < 0.05, **p < 0.01; Student's t-test; n_DMSO = 11, n_Fer-1 = 12). (C) LDH release detected in supernatants collected from cultured hCHF LMS (**p < 0.01; one-way ANOVA; n = 4). (D) Left - time course of peak amplitude (contractility) during culture in BMCC, shown as fold change to 1 h values (***p < 0.001; two-way ANOVA; n_DMSO = 5, n_Fer-1 = 6). Right – AUC of contractility over 72 h (*p < 0.05; Student's t-test; n = 6). Data are displayed as mean ± SEM.

Fig. 6. Inhibition of ferroptosis-related pathways in vitro exhibits anti-fibrotic and anti-inflammatory effects.

(A) ROS production assay in human cardiac fibroblasts (HCF) over 6 h, with and without H₂O₂ stimulation. Fer-1 [10 μM] and AFC [100 μM] were used to inhibit or boost ferroptosis, respectively. Luminiscence was measured via Cytation 1 (BioTek) (*p < 0.05, ****p < 0.0001; two-way ANOVA; n = 4). (B) RT-qPCR quantification of gene expression (*p < 0.05; Student's t-test; n = 5). (C) Migration (scratch) assay to measure HCF migration index. Left - Representative images of 0 h and 24 h scratches under Fer-1 [10 μM] or DMSO (1:1000) administration. Right – Migration index over 26 h – measurement were done in 2 h intervals (****p < 0.0001; two-way ANOVA; n = 3). Data are displayed as mean ± SEM.

Fig. 7. Metallothionein overexpression is associated with beneficial effects in human LMS and cardiac fibroblasts.

(A) Diagram of RNA seq transcriptomic analysis. 2 groups were analyzed: hCHF-LMS treated with Fer-1 [10 μM] and hCHF-LMS trated with DMSO (1:1000) (p-ajd < 0.05; |Log₂FC| > 0.5; n = 4). (B) Heatmap illustrating the expression of differentially expressed genes. Color scaling shows the levels for given proteins, blue - highest, red – lowest. (C) Functional enrichment analysis of RNA seq data. Selected clusters from different databases are included. X-axis - Normalized enrichment score hCHF+Fer-1 vs. hCHF+DMSO. FDR < 0.05. (D) Volcano plot of RNA seq dataset (p-ajd < 0.05; |Log₂FC| > 0.7; n = 4). Differentially expressed Metallothionein 1 subtypes are highlighted. (E) RT-qPCR quantification of ferroptosis-related gene expression (*p < 0.05; Student's t-test; n = 5). Data are displayed as mean ± SEM. (F) Ferroptosis sensitization assays in HCF. Cell death was measured using CellTox™ Green or WST-1 following a 6-h treatment with increasing doses of RSL-3. (*p < 0.05; two-way ANOVA; n = 3). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)