Non-canonical BIM-regulated energy metabolism determines drug-induced liver necrosis

Paracetamol (acetaminophen, APAP) overdose severely damages mitochondria and triggers several apoptotic processes in hepatocytes, but the final outcome is fulminant necrotic cell death, resulting in acute liver failure and mortality. Here, we studied this switch of cell death modes and demonstrate a non-canonical role of the apoptosis-regulating BCL-2 homolog BIM/Bcl2l11 in promoting necrosis by regulating cellular bioenergetics. BIM deficiency enhanced total ATP production and shifted the bioenergetic profile towards glycolysis, resulting in persistent protection from APAP-induced liver injury. Modulation of glucose levels and deletion of Mitofusins confirmed that severe APAP toxicity occurs only in cells dependent on oxidative phosphorylation. Glycolytic hepatocytes maintained elevated ATP levels and reduced ROS, which enabled lysosomal recycling of damaged mitochondria by mitophagy. The present study highlights how metabolism and bioenergetics affect drug-induced liver toxicity, and identifies BIM as important regulator of glycolysis, mitochondrial respiration, and oxidative stress signaling.


SUPPLEMENTAL FIGURES
A Western Blot of mouse liver (500 mg/kg APAP, 6h), HepG2 and IHH treated for 3 h with APAP or cisplatin and primary human hepatocytes (PHH) treated with 20 mM APAP.
C UMAP plots showing hepatic cell populations from single cell RNA sequencing of mice treated for 6 h with PBS or 500 mg/kg APAP.Cells are colored based on annotation taken from Kolodziejczyk et al. 1 .Hepatocyte-specific genes confirm hepatocyte cluster (right).C-D Western Blot liver lysate from WT and Bim-/-mice treated with 500 mg/kg APAP or 25 µg/kg TNF and 1 g/kg N-galactosamine (GalN) for 6 h with or without injection of 2.5 g/kg glucose (Gluc) 1 h prior to APAP.E Representative immunohistochemistry images of anti-cleaved Caspase-3-stained liver sections from mice treated for 6 h with 500 mg/kg APAP with optional injection of 2.5 g/kg glucose 1 h prior to APAP.As positive control liver from mice treated with 25 µg/kg murine TNF and 1 g/kg Ngalactosamine (GalN) were used.Scale bars 250 µm.F-G Volcano plot and heatmap of top fasting-induced differentially expressed genes (DEGs) from liver bulkRNA sequencing from fasted and non-fasted mice, WTMF = WT male fasted mouse, WTMUF = WT male unfasted mouse.F (PMH) were isolated and after adherence either kept for 1 day at 21% oxygen level or transferred to 10% oxygen chamber and incubated for 1 day to adapt to lower oxygen.PMH were then treated with 20 mM APAP for 16 h to measure cell death measured by MTT assay or for 3 h to measure reactive oxygen species (ROS) levels by ROS-Glo Assay (Promega), n=1.
For experiments, 10-12 weeks old mice were fasted for 12 h before i.p. injection of either 500 mg/kg APAP (SIGMA) in warm PBS for 6 h or, 300 mg/kg APAP for 12 h; or an equal volume of warm PBS (SIGMA (control treatment).Injections were always carried out in the morning.As there is a well-established, substantial difference between sexes in the sensitivity to APAP and it is advisable to analyze sexes independently 2 , we used for the 6 h experiments only male mice and for the 24 h experiments only female mice.For both setups, additional i.p. injection of 2.5 g/kg glucose (SIGMA) took place 1 h prior to APAP/PBS injection.Blood glucose was measured by punctation of warmed tail vein and direct measurement in a GlucoCheck Excellent glucometer (Activemed).Mice were sacrificed by CO 2 and cervical dislocation, and liver lobes were stored in 10% formalin or shock-frozen in liquid nitrogen for further analysis.Serum from heart blood was obtained by coagulation, and centrifugation for 15 min at 1000 xg.Serum Alanine aminotransferase (ALT) levels were measured using the Reflotron® Plus (Roche) with the respective ALT/GPT test stripes.

Histology and Immunohistochemistry
Fixed and paraffin-embedded murine liver tissue was cut in 4 µm sections, rehydrated and stained by hematoxylin and eosin (H&E) with subsequence dehydration and mounting.For the quantification of necrotic area, three images per samples in 5x magnification were taken blinded and scored blinded using ImageJ area measurement based on a grey value threshold of an 8 bit image measuring percentage of pale, necrotic area while leaving out white blood vessels.
For anti-cleaved caspase-3 or anti-phospho-AMPK immunohistochemistry antigens within rehydrated liver sections were retrieved by boiling in 10 mM Sodium citrate buffer (pH 6) for 15 min.Sections were blocked in 3% hydrogen peroxide/PBS for 60 min, followed by incubation in 5% goat serum, 3% BSA in PBS for additional 60 min.Staining with 1:100 of respective primary antibodies (Cell signaling technology) or goat anti-rabbit-IgG (Jackson Immuno) took place over night at 4 °C, was followed by 3 x 10 min wash in TBS-Tween and incubation with 1:100 biotinylated secondary goat anti-rabbit (Jackson Immuno) for 2 h, followed by additional 3 x 10 min wash and incubation in ABC solution of Vectastain ABC Kit (Biologol).After washing, development was achieved by incubation with 1 x DAB substrate solution (Roche) for 5-10 min and washing in running tab water for 5 min.Counterstaining was performed in hematoxylin for 10 sec, followed by dehydration and mounting.The antibodies used are listed in Supplementary Figure 3.
All plates, dishes and slides were pre-coated with 0.3 mg/ml collagen (Sigma-Aldrich) for minimally 2 h prior to seeding.Hepatocytes were kept in culture at 37 °C with 5% CO 2 with daily medium exchange and not longer than 3 days.

Isolation of primary human hepatocytes
Isolation of human hepatocytes was performed at the Insel University Hospital, as described previously 3 .Normal human liver tissue was taken from the periphery of liver specimens from patients undergoing surgical resection for colorectal metastases.Informed consent of the patients was obtained in accordance with institutional guidelines and the local ethics committee.Briefly, cells were isolated using a two-step enzymatic perfusion protocol.The viability of the isolated hepatocytes was determined by trypan blue exclusion, and only preparations of over 90% viability were used.The hepatocytes were seeded onto rat tail collagen-coated tissue culture plastics in DMEM containing 10% FCS, left to attach for 1 to 2 h, and then washed twice with phosphatebuffered saline (PBS) to remove unattached cells.The hepatocytes were cultured in arginine-free Williams E medium (Gibco) supplemented with insulin (0.015 IU/mL), hydrocortisone (5 μmol/L), penicillin (100 IU/mL), streptomycin (100 μg/mL), glutamine (2 mmol/L), and ornithine (0.4 mmol/L) for 24 h before use.
GEV16-mediated inducible overexpression of BCL-2, BCL-x L or MCL-1 was achieved as described in Vince et al. 5 .In brief, two-step cell line generation was performed by first transduction of target cells with HEK293T-dervided virus containing the GEV16 plasmid (pHCA/GAL4(1-93).ER.VP16) and generation of a stable cell line by selection with hygromycin B. The second step included virus production and transduction of generated GEV16-cells with expression plasmids (pF-5xUAS-MCS-SV40-Puro-based) containing either human BCL-2, BCL-x L and MCL-1 genes, each tagged with StrepIII.Cells were further selected with 1 µg/ml puromycin.
Overexpression of target genes was achieved by treatment of cells with 60 nM 4hydroxytamoxifen (4-HT) 24 h prior to experiments.

PI-Hoechst staining
After treatment, cells were stained with 5 µg/ml propidium iodide (PI) (Sigma-Aldrich) and 5 µg/ml Hoechst33342 (BioRad) for 30 min at 37 °C.Fluorescence was measured with the Infinite® 200 PRO plate reader (TECAN) from the top with automatically determined Z-position, in circular 4x4 measurements per well and, optimized gain for red (ex: 535 nm, em: 617 nm) and blue channel (ex: 361 nm, em: 486 nm).The ratio of PI and Hoechst fluorescence per measurement was determined and normalized to untreated control.

Annexin V-PI staining
Cell death in IHH and HepG2 cells was assessed by Annexin V-FITC and PI staining, and flow cytometry analysis.Cells were collected in a V-bottom 96-well plate and pelleted at 500 x g for 10 min with subsequent staining with 100 µl/well self-made Annexin V-FITC (1:1000) in binding buffer (10 mM HEPES, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 ) for 20 min at 4 °C.
After centrifugation, cell pellets were resuspended 100 µl/well binding buffer containing 1 µg/ml PI and directly measured in the BD LSR Fortessa™ flow cytometer.Analysis was performed using FlowJo software by gating for single cells being positive for FITC or PI.

Quantitative RT-PCR
RNA isolation was archived by lysing samples in peqGOLD TriFast™ (VWR) or TRIzol TM Reagent (Invitrogen) according to manufacturer's protocol.Small pieces of frozen murine liver were lysed in 1 ml by 4 Hz shaking in the TissueLyzer II (QIAGEN).Harvested cell pellets were directly lysed in 1 ml peqGOLD TriFast™ (VWR) or 400 µl TRIzol TM Reagent (Invitrogen).After isolation of RNA, chromosomal DNA was digested with DNase I t (NEB), and 2 µg of DNA-free RNA were reverse transcribed into cDNA by High Capacity cDNA Reverse Transcription Kit (Applied Biosystems).
Real-Time-quantitative PCR was performed on a StepOnePlus™ qPCR device (Applied Biosystems) using SYBR® Green (Applied Biosystems) and specific primers.Murine or human beta-actin C T values were used for normalization of respective samples (ΔC T (gene) = C T (gene) -C T (beta-actin)).The used primers are listed in Supplementary Table 2.
As input control, 10% of supernatant was directly boiled in 1 x Laemmli buffer as described.
Remaining supernatant was incubated with bait or isotype-antibodies for 2-3 hours at room temperature under constant rotation.Per sample 30 µl Protein G Sepharose 4 Fast Flow beads (GE Healthcare Life Sciences) were 3x times in CHAPS buffer by centrifugation at 14 000 x g for 1 min each.Beads were added to samples and incubated overnight at 4 °C under constant rotation.On next day, beads were washed 3x to remove unbound proteins and boiled in 2x Laemmli buffer at 95°C for 10 min to release bound proteins.Equal volumes were used for Western Blotting.

Cytosolic and mitochondrial fractionation
To study mitochondrial outer membrane permeabilization, mitochondrial and cytosolic fractions Mitochondrial faction was collected by centrifugation for 10 min at 14000 x g, 4 °C.Samples were boiled in 1 x Laemmli buffer as described before and used in a 4:1 ratio (cytosolic/mitochondria) for Western Blotting.

FADU Assay
Early DNA damage was assessed by automated Fluorimetric Detection of Alkaline DNA Unwinding (FADU) Assay as described by Moreno-Villanueva 6 .In short, cells were seeded and treated in specially prepared 96-well plates and placed in a pipetting robot that conducted cell lysis, alkaline-mediated unwinding of DNA, neutralization step and SYBR Green® administration.Amount of double-stranded DNA was determined in fluorescent plate reader (ex: 485 nm, em: 530 nm) with subsequent calculation of P (sample conditions) to T ratios (100% dsDNA condition) and normalization to untreated control.
Absorption at 405 nm was measured directly every 5 min for 1 h at 37 °C in the Infinite® 200 PRO plate reader (TECAN).Slope of increasing absorbance was normalized to untreated controls.

Seahorse ATP Rates and Mito Stress Test
To quantify the cellular bioenergetic profile comprising glycolytic and mitochondrial ATP generation, Seahorse XF Real-Time (Induced) ATP Rate Assay Kit (Agilent) was used according to manufacturer's protocol.For the Induced ATP Rate Assay, following adaptions were implemented: In brief, cells were seeded one day prior to experiments into Seahorse XFe24 cell culture plates, medium was carefully replaced to Seahorse XF DMEM medium pH 7.4 (Agilent) supplemented with 18 or 36 mM glucose, 1 mM pyruvate and 2 mM L-glutamine, and cells were incubated at 37°C without CO 2 for 60 min.Directly before measurement, medium was changed again.Seahorse extracellular Flux Analyzer XFe24 (Agilent) was started with pre-hydrated XFe24 sensor cartridge containing, APAP solved in Seahorse XF DMEM Medium pH 7.4 in Port A and B (final concentration together: 15 mM), oligomycin in Port C (final 1.5 µM), a mix of rotenone, antimycin A (both final concentration 0.5 µM) and Hoechst33342 (final 1 µg/ml) in Port D. After Injection of Port A and B, cells were incubated and measured for 1 h before the program continued with oligomycin injection.For the Non-Induced Assay, both APAP ports were excluded, and others moved forward.Afterwards, the cell number was determined based on Hoechst33342-positive nuclei using an automated algorithm in Cellomics ArrayScan TM VTI High-Content Screening (Cellomics TM ).Nuclei were identified based on shape, intensity, and size.Data analysis was performed with Agilent Seahorse Analytics Online Software by normalizing to cell number and to 100% ATP production of untreated wild type control, cultured in normal glucose medium.For calculation of 'maximal ECAR' and basal and maximal respiration in IHH cells, the Seahorse XF Cell Mito Stress Test Kit (Agilent) was used according to manufacturer's protocol with normalization to cell count as described above.The 'maximal ECAR' was calculated as mean of ECAR after oligomycin injection, thus after inhibiting mitochondrial ATP production.Likely, the near-constant ECAR in no glucose is attributed to non-glycolytic acidification, while the increase in ECAR in the normal glucose condition is due to increased glycolysis.

ATP measurement (CellTiter-Glo®)
Intracellular ATP levels were determined with the CellTiter-Glo® 2.0 Assay (Promega) as described in the manufacturers' instructions.Briefly, cells in 96-well plate were lysed with CellTiter-Glo® Reagent and ATP-dependent luminescence was measured after 10 min in Infinite® 200 PRO plate reader (TECAN).

Lactate measurement
Lactate secretion was measured using the LAC 142 Kit (Diaglobal) and the Diaglobal Photometer the according to manufacturers' protocol.For sampling, medium of cells seeded in 6-well plates was changed and supernatant samples were taken at 0, 2, 6 and 24 h.Lactate secretion per sample was calculated as increase normalized to untreated control.

Mitochondrial Morphology by MitoTracker TM confocal microscopy
MitoTracker TM Green (Invitrogen) fluorescent dye was used to monitor mitochondrial morphology.
Cells seeded in 8-well glass bottom IBIDI chambers slides were stained with 100 nM MitoTracker TM Green in phenol red-free DMEM/F-12 medium with supplements for 30 min at 37°C, followed by three times washing with warm medium and further incubation for 30 min at 37°C.Live cell imaging at 37°C, 5% CO 2 was performed with the LSM 880 AxioObserver (Zeiss) at the Bioimaging Center (University of Konstanz) with 488 nm Laser (ex: 488 nm, em: 515 nm) and Plan-Apochromat 40x/1.4Oil DIC M27 objective.Images were acquired with the Zen Black software (Zeiss) and subsequently processed equally with Zen Blue.Mitochondrial morphology was quantified with ImageJ software by calculating the ratio length/width from 30 randomly selected mitochondria per picture (4-9 images per sample, n=3), grouping them into hyperfused (ratio>10), fused (ratio≥5<10) and fragmented (ratio<5) morphology and, finally calculating the percentages per sample.

Mitochondrial ROS by MitoSOX TM confocal microscopy
Mitochondrial reactive oxygen species can be visualized via MitoSOX TM Red (Invitrogen), a fluorogenic dye, which is specifically oxidized by mitochondrial superoxide in living cells.Cells seeded in 8-well glass bottom IBIDI chambers slides were stained with 5 µM MitoSOX TM Red in phenol red-free DMEM/F-12 medium with supplements for 15 min at 37°C and 5% CO 2 , followed by washing with warm medium, and live cell imaging at 37°C, 5% CO 2 was performed immediately with the LSM 880 AxioObserver (Zeiss) at the Bioimaging Center (University of Konstanz) with 514 nm Laser (ex: 514 nm, em: 593 nm) and Plan-Apochromat 40x/1.4Oil DIC M27 objective.
Images were acquired with the Zen Black software (Zeiss) and subsequently processed equally with Zen Blue.The mean fluorescence intensity of each cell in a picture was quantified using the ImageJ software and mean of MFI per technical sample was determined.

MitoSOX TM measurement by flow cytometry
Quantification of mitochondrial ROS in HepG2 and splenocytes was assessed by MitoSOX TM Red (Invitrogen) staining and flow cytometry analysis.Optionally treated cells were harvested using Trypsin/EDTA (Sigma-Aldrich), distributed in V-bottom 96-well plate, and stained with 5 µM MitoSOX TM Red for 10 min at 37 °C.After washing in FACS Buffer (2 mM EDTA, 5% BSA, PBS), cells were resuspended and measured in FACS Buffer containing 1 µg/ml DAPI on a LSR Fortessa™ flow cytometer (BD Biosciences).HepG2 cells were measured in High Throughput Sampler (HTS).Analysis of the Red Median Fluorescence Intensity of single, live cells was done using FlowJo software.

MitoSOX TM measurement with Incucyte®
To study the time course of mitochondrial ROS production, cells were pre-conditioned with glucose for 1.5 h, then stained with MitoSox TM (Invitrogen) in phenol red-free medium with respective glucose concentrations for 15 min, followed by APAP treatment to obtain final concentrations of 5 µM MitoSox TM .Cells were incubated 24 h at 37°C 5% CO 2 in the Incucyte® SX1 system and in intervals of 1 h images with 10x magnification were taken.Images were analyzed using Incucyte® 2021B Software by solely quantifying cytosolic signals (nuclear staining was excluded) and normalizing the measured integrated intensity to the cellular confluence.

Mitochondrial Potential by TMRE staining
For quantification of mitochondrial membrane polarization, cells were harvested by Trypsin/EDTA treatment (Sigma-Aldrich) and stained with 100 nM Mitotracker TM Green and 20 nM TMRE (Sigma-Aldrich) in FACS Buffer (2 mM EDTA, 5% BSA, PBS) for 30 min at 37 °C.After adding DAPI (final concentration 1 µg/ml), samples were directly measured on a LSR Fortessa™ flow cytometer.HepG2 cells were measured in High Throughput Sampler (HTS).Percentage of the single, live, MitoTracker TM Green-positive, and TMRE-positive cells were determined using FlowJo software.

Acridine Orange staining and quantification
To visualize acidic lysosomes, cells were seeded in 8-well glass bottom IBIDI chambers slides, treated, and then stained with 3 µM acridine orange (Sigma-Aldrich) and 5 µg/ml Hoechst33342 for 30 min at 37 °C.Afterwards, cells were carefully washed twice with phenol red-free DMEM/F12 and visualized on an AxioObserver microscope using AxioVision software (Zeiss).Acridine Orange stains acidic components and when applied ad high concentration it can be used to distinguish between green-fluorescent DNA (monomer molecules) and red-fluorescent lysosomes (stacked molecule).Quantification of lysosomes was achieved by taking four images per samples in 20x magnification and blindly determining bright red area using ImageJ area measurement based on a grey value threshold of 8-bit images.To avoid cell density bias, red area (lysosomes) was normalized to blue area (nuclei) per sample.Ratio was normalized to untreated, normal glucose control.

Lysotracker and Cathepsin B staining
To visualize functional lysosomes, cells were seeded in 8-well glass bottom IBIDI chambers slides, treated and then stained with 50 nM LysoTracker TM Green DND-26 (Invitrogen), 1:26 dilution of Cathepsin B Magic Red TM (BIO-RAD), and 5 µg/ml Hoechst33342 (BIO-RAD) for 60 min at 37°C.Before visualization on a AxioObserver microscope, cells were carefully washed twice with phenol red-free DMEM/F12 media.

Single cell RNA sequencing analysis
Annotated single cell RNA sequencing data of APAP-treated mice was downloaded from Kolodziejczyk et al. 1 and analyzed in R environment using R Studio.After initial filtering of the expression matrix as described in Kolodziejczk et al., the R package Seurat (version 4.1.0)was used to normalize, find variable genes, scaling and dimensional reduction (PCA) as described in Butler et al.Clustering was not performed but the provided annotated cell clusters were re-applied and re-checked based on cluster-specific expression patterns.Single cell datasets were visualized using Uniform Manifold Approximation and Projection (UMAP) dimensional reduction using RunUMAP and DimPlot function.Specifically, hepatocyte subset of SPF-housed mice treated with PBS or APAP were analyzed and compared (described in Kolodziejczyk et al., 2020) 1 .
Differentially expressed genes (DEGs) between hepatocytes clusters were calculated using FindMarker function with p-value <0.05.Gene set enrichment and gene ontology/pathway analysis were performed with hepatocyte DEGs using the R packages clusterProfiler (version 4.0.5) and enrichplot (version 1.12.3).

Bulk hepatocyte RNA sequencing analysis
Age and sex-matched C57BL/6J wild type mice were fasted overnight with unlimited access to water.Control animals were not fasted (unfasted) and had permanent access to food and water ad libitum.Following the fasting period, mice were sacrificed, the liver lobes were collected, and total RNA was isolated using the SV Total RNA Isolation System (Promega).Paired-end bulk mRNA sequencing (mRNA-Seq) was performed by the Next Generation Sequencing (NGS) Platform of the University of Bern, Switzerland.Obtained raw paired-end data was processed using the public server usegalaxy.orgwith Trimmomatic, STAR, and featureCounts tools for the generation of gene counts.Quality control was repeatedly checked using FastQC and MultiQC tools.The generated count matrix was downloaded and further analyzed in R environment using R studio.Gene counts were filtered, normalized, log-transformed and DEGs were calculated with an adjusted p-value <0.05 using the R package DESeq2 (version 1.32.0).Visualization was performed using the R package "EnhancedVolcano" (version 1.10.0)and pheatmap (version 1.0.12).Gene set enrichment, gene ontology/pathway and cluster net analysis were performed using the R packages clusterProfiler (version 4.0.5) and enrichplot (version 1.12.3).

Statistics and data display
If not stated otherwise, data obtained from in vivo experiments are median ± standard deviation

DFH
Selected gene ontologies from Top 100 GOs of hepatocyte subset from single cell RNA sequencing data described in (C).E Western Blot of PMH treated with APAP or murine TNF and 30 nM ActD for 12 h.Representative phase contrast microscopy of PMH treated with APAP or 2 ng/ml murine TNF and 30 nM ActD for 12 h.Scale bar shows 50 µm.G MTT assay of PMH (16h) treated in combination with 30 µM z-VAD-fmk or DMSO; UT: untreated.n=3.Quantification of PI-Hoechst stained PMH treated with 20 mM APAP or 2 ng/ml murine TNF and 30 nM ActD for different timepoints.Data show mean +/-SD, n=3

(
SD) and data from ex vivo and in vitro experiments are means ± SD of a minimum of three independent experiments with several technical replicates.GraphPad Prism 8 software was used to perform statistical analysis and data visualization.Statistical significance was calculated by unpaired two-tailed Student's test when comparing two groups, One-Way or Two-Way ANOVA when comparing three or more dependent groups, respectively.Suitable post-hoc tests (Tukey's or Sidak's) were applied and are stated in respective figure legends.Statistical significance is displayed as p-values with ≤ 0.05 being statistically significant (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001).-HA-StreIII-Bcl2 (StrepIII-tagged Bcl-2 overexpression plasmid, cloned from from pF-5xUAS