Iron chelation as a new therapeutic approach to prevent senescence and liver fibrosis progression

Iron overload and cellular senescence have been implicated in liver fibrosis, but their possible mechanistic connection has not been explored. To address this, we have delved into the role of iron and senescence in an experimental model of chronic liver injury, analyzing whether an iron chelator would prevent liver fibrosis by decreasing hepatocyte senescence. The model of carbon tetrachloride (CCl4) in mice was used as an experimental model of liver fibrosis. Results demonstrated that during the progression of liver fibrosis, accumulation of iron occurs, concomitant with the appearance of fibrotic areas and cells undergoing senescence. Isolated parenchymal hepatocytes from CCl4-treated mice present a gene transcriptomic signature compatible with iron accumulation and senescence, which correlates with induction of Reactive Oxygen Species (ROS)-related genes, activation of the Transforming Growth Factor-beta (TGF-β) pathway and inhibition of oxidative metabolism. Analysis of the iron-related gene signature in a published single-cell RNA-seq dataset from CCl4-treated livers showed iron accumulation correlating with senescence in other non-parenchymal liver cells. Treatment with deferiprone, an iron chelator, attenuated iron accumulation, fibrosis and senescence, concomitant with relevant changes in the senescent-associated secretome (SASP), which switched toward a more anti-inflammatory profile of cytokines. In vitro experiments in human hepatocyte HH4 cells demonstrated that iron accumulates in response to a senescence-inducing reagent, doxorubicin, being deferiprone able to prevent senescence and SASP, attenuating growth arrest and cell death. However, deferiprone did not significantly affect senescence induced by two different agents (doxorubicin and deoxycholic acid) or activation markers in human hepatic stellate LX-2 cells. Transcriptomic data from patients with different etiologies demonstrated the relevance of iron accumulation in the progression of liver chronic damage and fibrosis, correlating with a SASP-related gene signature and pivotal hallmarks of fibrotic changes. Altogether, our study establishes iron accumulation as a clinically exploitable driver to attenuate pathological senescence in hepatocytes.


Isolation of mouse hepatocytes and RNA-seq analysis
After 4 weeks of treatment (with mineral oil or CCl4), livers were perfused with Hank's balanced salt solution supplemented with 10 mM Hepes and 0.2 mM EGTA for 5 min, followed by a 15 min perfusion with William's medium E containing 10 mM Hepes and 0.03% collagenase type 1 (125 U/mg; LS0041, Worthington Biochemical Corp, Lakewood, NJ, USA).Livers were further minced, filtered through a 70 μm cell strainer (BD Biosciences, Franklin Lakes, NJ, USA) and viable hepatocytes were selected by centrifugation in Percoll (17089101, Cytiva, Marlborough, MA, USA) and stored at -80 ºC.
The RNA-seq libraries were prepared with KAPA Stranded mRNA-Seq Illumina® Platforms Kit (Roche, Basel, Switzerland) following the manufacturer's recommendations starting with 500 ng of total RNA as the input material.The library was quality controlled on an Agilent 2100 Bioanalyzer with the DNA 7500 assay.The libraries were sequenced on NovaSeq 6000 (Illumina, San Diego, CA, USA) with a read length of 2x151 bp, following the manufacturer's protocol for dual indexing.Image analysis, base calling and quality scoring of the run were processed using the manufacturer's software Real Time Analysis (RTA v3.4.4).

Analysis of gene expression by RT-qPCR
mRNA expression levels of genes were analyzed by Real Time-quantitative PCR (RT-qPCR).
Total RNA was isolated from the different cells or tissues using EZNA Total RNA Kit II (Omega Bio-tek, Norcross, GA, USA).cDNA was produced using the High-capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA).RT-qPCR was performed in duplicate in a Light Cycler 480 II (Roche).SYBR Green PCR Master Mix was used for PCR reactions (Applied Biosystems).Primers are listed in Suppl.Table 1.RPL32 gene was used as housekeeping.

Quantification of IHC results
Slide scans were examined thoroughly for analyses.In Picro-Sirius Red, α-SMA and EPPB preparations, three representative regions from different liver lobes of each animal were selected and the percentage of positive stained area was quantified with ImageJ analysis software v1.44o (National Institutes of Health, Bethesda, MD, USA).For each animal, the arithmetic mean of the stained area from the different selected regions was then calculated.The percentage of p21-positive stained cells was quantified with QuPath software v0.4.4 (ref. 21 in main text) in a whole liver lobe for each animal.To ensure blinding when assessing the outcome in histological preparations, automatization of quantification by QuPath and predetermination of settings in ImageJ was performed.

Western blot analysis
To analyze protein levels by Western blot, liver tissue was lysed in RIPA lysis buffer supplemented with a cocktail of protease inhibitors (11697498001, Roche Diagnostics, Rotkreuz, Switzerland) and orthovanadate (S6508, Sigma) at 4 °C.Lysis was performed in a Tissue Lyser II (QIAGEN, Venlo, The Netherlands) and protein concentration was determined using a Pierce TM BCA protein assay kit (Thermo Fisher Scientific).Proteins were separated with denaturalizing SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a nitrocellulose membrane in wet conditions.For immunoblotting, membranes were incubated in 5% bovine serum albumin (A7906, Sigma) in PBS-Tween 0.05% for 1 h at room temperature.Afterwards, membranes were incubated with the primary antibodies against α-SMA (ab5694, Abcam) and α-Tubulin (T9026, Sigma) both at 1/1000 dilution, overnight at 4 °C.The following day primary antibodies were washed with PBS-Tween 0.05% and binding was developed with secondary antibodies diluted 1/2000 in PBS-Tween 0.05% (NA934V anti-Rabbit and NA931V anti-Mouse for α-SMA and α-Tubulin, respectively), incubated for 1 h at room temperature.Finally, positive hybridization was visualized with a chemiluminescent solution (GE HealthCare, Little Chalfont Amersham, UK) in a ChemiDoc TM Touch Imaging System (Bio-Rad, Munich, Germany) and densitometric analysis was performed using Image Lab TM software (Bio-Rad).

Analysis of the SASP in liver tissue
Liver tissues were lysed in RIPA buffer and centrifuged to separate insoluble debris.
Afterwards, protein concentration was determined using a Pierce TM BCA protein assay kit (Thermo Fisher Scientific) and adjusted so that all samples contained the same amount of protein.Samples were shipped to an external commercial laboratory (Eve Technologies Corp, Calgary, Canada) and a Mouse Cytokine/Chemokine 44-Plex Discovery Assay® Array (MD44) was performed.

In vitro cell culture models
The hepatocyte HH4 cells were plated in MEM with 15% FBS and after 24 h, medium was changed to deplete FBS up to 2% FBS.24 h later, treatments were initiated, and cells were cultured for 6 days more.Treatments were as follows: vehicle; 50 nM doxorubicin (S1208, Selleck Chemicals, Cologne, Germany) to induce senescence; 50 µM deferiprone; and doxorubicin + deferiprone.Culture media and treatments were refreshed after 3 days.
The HSC LX-2 cells were cultured in DMEM with 10% FBS and 24 h later medium was changed to deplete FBS up to 0.5% FBS.The following day, cells were treated with vehicle, 2 ng/mL TGF-β (T7039, Sigma-Aldrich), 50 nM doxorubicin or 80 µM deoxycholic acid (DCA; D6750, Sigma-Aldrich) to induce senescence, or the corresponding senescent inducer combined with TGF-β, both in the absence or presence of 20 µM deferiprone and maintained for 6 days, refreshing the media and treatments after 3 days.Concentration of deferiprone in each case was designed as the amount able to attenuate iron accumulation without serious effects on cell death.LX-2 cells were more sensitive to deleterious effects of deferiprone.

Detection of intracellular iron
Labile iron was measured using the FerroOrange probe (36104, Cell Signaling Technology, Danvers, MA, USA) following manufacturer's instructions.A 24 h treatment with an aqueous solution of 330 µM iron sulfate (F8633, Sigma-Aldrich) and 330 µM iron nitrate (F8508, Sigma-Aldrich) was used as positive control.Cells were examined through a Leica DM IRB Inverted Fluorescence Microscope (Leica Microsystems, Wetzlar, Germany).For analysis by flow cytometry, cells were dissociated and processed in a MoFlo Astrios Cell Sorter (Beckman Coulter, Brea, CA, USA) at the Biology-Bellvitge Unit from Scientific and Technological Centers (CCiTUB), Universitat de Barcelona.

SA-β-GAL staining
Cells were stained for senescence-associated β-galactosidase activity with a commercial

Analysis of ferroptosis by flow cytometry
Lipid peroxidation as a hallmark of ferroptosis was determined using the C11-BODIPY probe (D3861, Invitrogen, Carlsbad, CA, USA).Briefly, cells were incubated with 5 µM C11-BODIPY for 40 min.Then, both floating (dead) and attached (alive) cells were collected for each condition and run on a Gallios TM Cytometer (Beckman Coulter) at the Biology-Bellvitge Unit from Scientific and Technological Centers (CCiTUB), Universitat de Barcelona.The ferroptosis inducer RSL3 (S8155, Selleck Chemicals) was used at 2 µM as positive control and incubated overnight.To determine viability, cells were stained with 4′,6-diamidino-2phenylindole (DAPI).

Gene signatures obtained from public data bases
ROS and TGF-β signaling gene signatures were obtained from Hallmark collection.
Oxidative phosphorylation signature (In-house) was generated selecting genes that were in at least 2 out of the 3 oxidative phosphorylation signatures from MSigDB v2023.1 (Hallmark, WikiPathway, and Gene Ontology Biological Process collections).

Species
kit (#9860, Cell Signaling Technology), following manufacturer's instructions.Staining of mouse liver tissue was done as stated in Maus et al. (ref.19 in main text).
Single-cell RNA-sequencing data from Yang et al. cohort (ref.30 in main text) was accessed through GEO accession number GSE171904 and the processed Seurat object was used to quantify gene expression signatures in non-parenchymal liver cell types with the AddModuleScore function.Comparisons between conditions were assessed using Mann-Whitney U-test and adjusted for multiple comparisons.Gene expression from acute CCl4 damage in mouse hepatocytes was assessed using Godoy et al. cohort (accessed through ArrayExpress E-MTAB-4444) (ref.29 in main text).Raw data was downloaded and normalized using robust multi-array average (RMA) (1).Additionally, mouse liver gene expression data from chronic CCl4-bulk sequencing (Hammad et al. cohort, ref. 31 in main text) was also downloaded and normalized using RMA through Gene Expression Omnibus (GEO) accession number GSE222576.To analyze human gene expression, Fujiwara et al. cohort of liver biopsies from HCC-naïve MASLD patients (ref.34 in main text) was accessed through Gene Expression Omnibus (GEO) accession number GSE193066.Relative log-expression normalized data was directly downloaded from GEO. Trepo et al. cohort of liver biopsies from HCC-naïve ASH and alcoholic cirrhosis patients (ref.35 in main text) was accessed through GEO accession GSE103580 and raw data was downloaded and normalized using RMA.For each cohort, a collection of gene signatures was obtained from MSigDB v2023 (2) and gene set variation analysis (GSVA) (3) was used to assess the relative activation of the signatures in the samples.Heatmaps showing relative activation for each gene signature were plotted using ComplexHeatmap package for R. Correlation in human samples between two gene signatures was assessed using Pearson correlation.Kendall's τ was used to assess the association between the gene signatures and fibrosis stage in Fujiwara et al. cohort.All p-values were adjusted for multiple testing with Bonferroni test correction.All analyses were performed using R v4.0.4 (4).