Ethics

This study was conducted in accordance with all relevant ethical regulations. The clinical study was approved by the London-Hampstead Research Ethics Committee (07/Q0501/50) and was in accordance with the declaration of Helsinki as reported before29. This study reported multi-lobular necrosis as a predominant histopathological feature being significantly more frequent in non-survivors29. All mouse experiments were carried out in accordance with the UK Home Office regulations (licence 70/8891; protocol 2 or licence PP0604995; protocols 3 and 4 or licence P3F79C606, protocol 2) following ethical approval by the local Animal Welfare and Ethical Review Body at the University of Glasgow.

Patient selection and data collection

The study included 34 consecutive patients with severe acute indeterminate hepatitis who were admitted to a single hospital and underwent transjugular liver biopsy or liver transplantation. The participants were not compensated for participation. Retrospective anonymized data and samples were obtained, following ethical approval and consistent with the UK Human Tissue Act, without informed consent. Of the 34 patients, 14 underwent liver transplantation, and 3 died within 3 months. They were defined as non-survivors. Seventeen patients who recovered spontaneously were defined as survivors. All 34 liver tissue biopsies analysed in this study were obtained at the time of enrolment (baseline biopsies). The following clinical data were collected from the patients: sex, age, date of histopathological examination, history of chronic disease and results of biochemical tests at baseline and at follow-up up to 28 days or the last value before death. These include serum levels of ALT, aspartate transaminase, bilirubin, ALP, international normalized ratio, creatinine, prothrombin time, albumin and hepatic encephalopathy.

Animal studies

Mice were bred and housed in a licensed, pathogen-free facility, under a 12 h light–dark cycle, at stable temperature (19–23 oC) and humidity (55 ± 10%). The mice were bred on a mixed C57Bl6 background, were housed in conventional cages and had ad libitum access to food and water (standard CRM(E) chow; Special Diets Services no. 801730). All experiments were performed on 8–12-week-old male and female mice, according to the guidelines of the Animal Welfare and Ethical Review Body, and are reported in agreement with the ARRIVE guidelines30.

Treatment with AAV was performed as described previously12. Briefly, mice were injected with either AAV8.TBG.PI.Cre.rBG (AAV-Cre) (Addgene, 107787-AAV8) or AAV8.TBG.PI.Null.bGH (AAV-Null) (Addgene, 105536-AAV8) at a dose of 2 × 1011 genomic copies (GC) ml−1unless otherwise stated, in a final volume of 100 μl sterile PBS via a single tail vein injection. Mice in this study weighed 21.9–31.9 g at the time of induction.

For the mouse model of hepatocyte-specific inactivation of Mdm2, male mice homozygous for the Mdm2tm2.1Glo allele (ID: MGI2385439176 (ref. 31)) and the Gt(ROSA)26Sortm14(CAG-tdTomato)Hze allele (ID: MGI3809524177 (ref. 32)) were used (Mdm2E5/E6fl; R26LSL-tdTomato/LSL-tdTomato mice). This model was also crossed to create a p21KO model (allele ID: MGI1888950 (ref. 33)). For the induction of a RAS-induced liver senescence, 8–12-week-old male and female mice heterozygous for the Krastm4Ty allele (ID: MGI:2429948178 (ref. 34)) and the Gt(ROSA)26Sortm14(CAG-tdTomato)Hze allele (ID: MGI3809524177 (ref. 32)) were used (KrasLSL-D12D/WT; R26LSL-tdTomato/LSL-tdTomato mice). Epithelial TGFBR1 deletion experiments used 1 μl g−1of CCl4 diluted 1:4 CCl4 in corn oil administered day 3 in mice with alleles AhCre and TGFBR1fl (IDs: MGI3052655 (ref. 35) and MGI2388050 (ref. 36), respectively). Mice with AhCre+/− TGFBR1fl/fl were used for epithelial induction utilizing β-naphthoflavone (β-NF; Sigma) versus β-NF-treated AhCre−/−Tgfbr1fl/fl controls, while AhCre−/−TGFBR1fl/fl mice were treated with either AAV-TBG-Cre or AAV-Cre-Null as described above at 2 × 1011 GC per mouse. Genetic controls used both AhCre−/−Tgbfr1fl/fl controls given respective controls (β-NF vehicle or AAV-Cre-Null) and AhCre−/−TGFBR1fl/fl mice given no induction agent but time matched CCl4 at day 3. The acute CCl4 model used 1 μl g−1of 1:4 CCl4 in corn oil administration as previously described7. For treatment with TGFβR1i, the mice received either TGFβR1i or vehicle (0.5% hydroxypropyl methylcellulose (HPMC) and 0.1% Tween 80) by oral gavage twice daily19. The dose of TGFβR1i was 50 mg kg−1in 100 ml PBS7. Senolytic and senomorphic agents were administered separately; navitoclax (ABT263), oubaine and rapamycin (or vehicle controls) were administered at 100 mg kg−1(gavage twice weekly), 1 mg kg−1(intraperitoneal injection (i.p.) on days 1 and 3 or twice weekly) and 250 μg (i.p. daily), respectively, were given from days 1 to 7 in the ΔMdm2Hep model and day 7 in the Recovery-Mdm2 model. Rapamycin mice in the ΔMdm2Hep model were collected at day 4. All mice receiving either ABT263 or oubaine either deteriorated clinically and were humanely killed or died shortly (within 6 h) after receiving the senolytics. β-NF was given by 3 × 80 mg i.p. injections for epithelial induction in mice with the AhCre allele as previously described37.

The mice were killed by CO2 inhalation in a CO2 chamber, cervically dislocated and then weighed. Blood was collected immediately via cardiac puncture for whole blood analysis (EDTA buffer-coated tubes; Sarstedt) and plasma biochemistry (lithium–heparin coated tubes; Sarstedt). The plasma was separated by centrifugation (2,350gfor 10 min at room temperature) within 1–3 h after culling and stored immediately at −80 oC. After weighing the liver, the caudate and left median lobes lobe were snap frozen on dry ice for protein and RNA extraction and for histology studies, respectively. The remaining liver was fixed in 10% neutral buffered formalin (NBF) for 22–24 h before transfer to 70% ethanol for further processing. The left kidney was immediately cut in half, and both halves were snap frozen on dry ice for protein and RNA extraction and for histology. The right kidney, heart, brain and lungs were fixed in 10% NBF for 22–24 h and then changed to 70% ethanol.

Assessment of cognitive functionY-maze test

The mice were individually placed into a testing area measuring 25.66 cm × 17.53 cm onto a Samsung Galaxy Tab 2 10.1 for 5 min. Steps and distance travelled were recorded using MouseTrapp software. The Y-maze arena had three arms of 40 cm identified as A1 A2 A3, each with a differentiating marker at the end of the arm. The mice were assigned different start arms in a rotating allocation and were tested before the start of the experiment and on day 4, with differing arm allocations each time. In the T1 phase, the mice were placed into the maze with only two arms open for 5 min to explore the arena. The mice were then removed to a clean cage for 1 min (fresh cage used per cage of mice). Then, in the T2 phase, the mice were returned to the maze in the starting position with the novel arm opened for 2 min and allowed to explore. A camera was set up above the maze to record movements and the video files analysed via Ethovision XT13 software.

Brain slice electrophysiology

The animals were humanely killed by anaesthetic overdose with inhaled isoflurane and intramuscular injection of ketamine (≥100 mg kg−1) and xylazine (≥10 mg kg−1) as previously described38. The mice were then transcardially perfused with at least 25 ml of sucrose-rich artificial cerebrospinal fluid—composed of 252 mM sucrose, 3.0 mM KCl, 1.25 mM NaH2PO4, 24 mM NaHCO3, 2.0 mM MgSO4, 2.0 mM CaCl2 and 10 mM glucose. The brain was removed and sliced at 450 μm horizontal slices with a Leica VT1000S vibratome in ice-cold sucrose-rich artificial cerebrospinal fluid. The slices were trimmed to the hippocampal region and maintained at 32–34 oC at an air–liquid interface between normal artificial cerebrospinal fluid (sucrose replaced with 126 mM NaCl) and humidified 95% O2/5% CO2. Oscillations were evoked with 10 μM cholinergic agonist carbachol, to activate transmission through acetylcholine receptors. Extracellular recording electrodes were filled with normal artificial cerebrospinal fluid (resistance 2–5 MΩ), and field recordings taken from the border between stratum radiatum and stratum lacunosum moleculare in CA3. The recordings were taken with an Axoclamp-2B amplifier (Axon Instruments) and extracellular data filtered at 0.001–0.4 kHz using Bessel filters. Mains noise was deducted with a Humbug (Digitimer) and data redigitized at 10 kHz using an ITC-16 interface (Digitimer). Axograph 4.6 software (Axon Instruments) was used for data acquisition and analysis.

To generate power spectra Axographs we used Fourier analysis using 60 s per 10 min recording. This was used to calculate peak frequency and area power (area under the peak). The mouse gamma frequency oscillation was measured at frequencies between 15 and 49 Hz. The oscillations were categorized as stable when area power measured within 10% for three consecutive 10 min recording intervals.

HistologyMurine samples

Formalin-fixed, paraffin-embedded sections 4 μm thick were used for simple IHC and for multiplex immunofluorescence. The sections were subjected to heat-induced antigen retrieval, followed by protein blocking to reduce non-specific staining. Incubation with primary antibody overnight at 4 oC or for 1 h at room temperature was followed by secondary antibody incubation and signal detection. The details on the antibodies can be found in Supplementary Table 3.

Photos were taken with a Zeiss Axiovert 200 microscope using a Zeiss Axiocam MRc camera. The stained slides were scanned using a Leica Aperio AT2 slide scanner (Leica Microsystems) at 20× magnification. Automated quantification of positively stained cells or area was performed using the HALO image analysis software (V3.1.1076.363, Indica Labs). Manual quantification of p21+ and BrdU+ kidney cells was performed on 20 random fields at 20× magnification. For p21 IHC quantification on brain sections, total brain area was calculated using the HALO software, and the p21+ cells were manually counted in the whole brain tissue area.

For multiplex immunofluorescence, 4 μm tissue sections underwent heat-induced antigen retrieval by boiling (in a waterbath) in citrate buffer (10 mM Na Citrate (Sigma, W302600), 0.05% Tween 20 (Sigma, P1379), pH 6) for 25 min and were subsequently cooled down for 20 min in the retrieval solution. Peroxidase quenching with 3% H2O2 (Sigma, 95321) was followed by biotin (Invitrogen, 4303) and protein blocking (Abcam, ab64226). The sections were incubated with the primary antibodies either for 1 h at room temperature or overnight at 4 oC, followed by 45 min with the secondary antibodies (conjugated to fluorophor) together with 4,6-diamidino-2-phenylindole (DAPI, 1 mg ml−1). Sudan black B was then used to quench autofluorescence. An aqueous mounting solution (DAKO, S1964) was used for mounting.

The anti-p21 antibody required additional signal amplification, which was achieved by using the tyramide signal amplification system. Briefly, after incubation with the anti-p21 primary antibody, the sections were incubated with a secondary anti-rat biotinylated antibody for 30 min, followed by a 30 min incubation with an avidin–HRP (horse radish peroxidase) complex (Vectastain ABC, Vector, PK-7100). Then, the sections were incubated with tyramide signal amplification fluorescein isothiocyanate (FITC; PerkinElmer, NEL741B001KT) for 6 min (in the dark). After that, the sections that were subjected to a 5 min heat-induced antigen retrieval to remove the anti-p21 antibody complex underwent protein blocking and then were incubated with the other primary antibodies, as described above. The images were taken using a Zeiss 710 upright confocal Z6008 microscope. The Opera Phenix scanner (PerkinElmer) was used to scan the stained sections at 20× magnification. For the analysis of scanned sections, the Columbus software (PerkinElmer) was used to identify hepatocytes and to quantify immunofluorescence staining intensity by hepatocyte in 20 random fields of view.

In situ hybridization was performed in an autostainer (Leica Bond Rx) using the 2.5 LSx RNAScope kit (Bio-Techne, 322700) according to the manufacturer’s instructions. The probes against Smad7 messenger RNA (Bio-Techne, 429418), TGFβ1 (Bio-Techne, 407758), TGFβ2 (Bio-Techne, 406188) and TGFβR1 (Bio-Techne, 431048) were used for the detection of the respective mRNA, and PPIB (Bio-Techne, 313918) was used as a positive control of gene expression.

Patient samples

Multiplex immunofluorescence staining was performed as previously described29,39,40. Formalin-fixed, paraffin-embedded liver samples were deparaffinized and rehydrated in xylene (Roth) and ethanol (Roth). Antigen retrieval was performed with Tris–EDTA buffer (pH 9) or universal antigen retrieval (Abcam) in a water bath at 98 °C for 30 min, followed by a cooling period of 20–30 minutes. The tissues were blocked with 2% normal goat serum (Thermo Fisher Scientific) to prevent non-specific antibody binding. The slides were incubated overnight at 4 °C with primary antibodies diluted in antibody dilution solution (Life Technology) and stained for 30 min with fluorescently labelled secondary antibodies (Supplementary Table 3) together with DAPI nuclear counterstain (Sigma Aldrich). After scanning the entire slide with a Zeiss Axio Observer7, the images were merged, and the background was subtracted. After each run, the antibodies were stripped by using the 2-mercaptoethanol/SDS method39, and the staining was repeated in multiple cycles over an 3 day period. Subsequently, all scans were aligned, hyperstacked and concatenated using the plugin FIJI HyperStackReg V5.6. For binary images, cell segmentation was performed using Ilastik software (v 1.3.3). Cell identification and counting, as well as fluorescence intensity measurement, were performed using CellProfiler v3.1.9 and plugin FIJI.

SA β-Gal staining

Staining for SA β-Gal was performed as described previously41. Briefly, 10-mm-thick cryosections or cultured cells were fixed in 2% paraformaldehyde or 0.25% glutaraldehyde in PBS for 15 or 5 min, respectively. This was followed by three washes with PBS 1 mM MgCl2 (pH 5.5 or 6 for mouse or human cells and tissues, respectively) and incubation with staining solution (1 mM MgCl2, 0.5 mM K3Fe(CN)6 0.5 mM C6FeK4N6·3H2O and 1 mg ml−1X-Gal in PBS, pH 5.5 or 6 for mouse or human cells and tissues, respectively) overnight (liver sections and cells) or for 2.5 h (kidney sections). After three washes with PBS, the cryosections were counterstained with eosin and mounted, while cells (on coverslips) were mounted immediately. For quantification, SA β-Gal+ and SA β-Gal− cultured cells were counted in 20 random fields of view.

RNA extraction

Whole-tissue RNA was extracted using the Qiagen RNeasy kit (Qiagen, 74104), including the optional DNase step, as described previously12. Briefly, 20–30 mg of snap-frozen tissue were homogenized in 600 ml buffer RLT and 1% β-mercaptoethanol using the Precellys Evolution homogenizer (Bertin Technologies), and the RNA was eluted in 30 μl RNase-free water. The RNA concentration was estimated with the Nanodrop 2000m and only samples with a 260/280 ratio ≥2 were used for further analysis.

Complementary DNA generation and qPCR

cDNA was generated using the QuantiTect Reverse Transcription Kit (Qiagen, 205311) according to the manufacturer’s instructions from 1 μg RNA. A PTC-200 Gradient cycler (MJ Research) was used to perform the genomic DNA wipeout and reverse transcription steps. A sample-free reaction and a reaction without the reverse transcriptase served as the negative controls. A real-time quantitative polymerase chain reaction (qPCR) was performed with the SYBR Green system (Qiagen, 204145) using a QuantStudio 5 real-time polymerase chain reaction system in a 384-well-plate setting (final reaction volume 10 μl per well). All primers used were purchased from Qiagen, as shown in Supplementary Table 4. Each biological replicate (mouse) was run in triplicate, and the 18S ribosomal RNA (Rn18S) was used as a house keeping gene for normalization. Relative expression was calculated using the ΔΔCt method.

Whole-tissue (bulk) RNA-seq

For bulk RNA sequencing (RNA-seq), the RNA was extracted as described above. Briefly, the RNA was tested on an Agilent 2200 TapeStation (D1000 ScreenTape) using RNA ScreenTape and only samples with a RIN >7 were further processed for library preparation. A total of 20 ng ml−1of RNA were used to prepare libraries using the TruSeq Stranded mRNA Kit. The Agilent 2200 Tapestation was used to assess library quality and Qubit (Thermo Fisher Scientific) was used to check concentration. The libraries were then run on the Illumina NextSeq 500 using the High Output 75 cycles kit (paired end, 2 × 36 bp cycle, dual index (I5 and I7 Illumina)).

For the bioinformatics analysis, raw data quality checks and trimming were performed using FastQC (versions 0.11.7 and 0.11.9 for Mdm2 and Kras, respectively), FastP (v0.19.3/0.20.1) and FastQ Screen (v0.12.0/0.14.0). The reads were aligned to the mouse genome and annotation (GRCm38.92 version) using HiSat2 version 2.1.0183. Determination and statistical analysis of expression levels was done by a combination of HTSeq version 0.9.1184, the R environment (v3.4.4/4.1.2), utilizing packages from Bioconductor (v3.6/3.15) and differential gene expression analysis based on the negative binomial distribution using the DESeq2 package (v1.18.1186). Pathway analysis was performed using MetaCore (v2018/2022) from Clarivate Analytics (https://portal.genego.com/).

scRNA-seq on kidneys

Three ΔMdm2Hep and three control mice were culled by CO2 inhalation; the blood samples were collected by cardiac puncture, and 30–40 ml cold PBS were used to perfuse the circulatory system via the left heart. The six mice were culled and their kidneys processed on two different days. Two mice (one ΔMdm2Hep and one control) were culled together on one day, and the others (two ΔMdm2Hep and two control) were culled 2 months later. On each occasion, the left kidney was used for dissociation and generation of single-cell suspension, while the right kidney was fixed in 10% NBF. The renal capsule of the left kidney was removed, and the kidney was cut into equally sized pieces and was dissociated using a multi-tissue dissociation kit 1 (Miltenyi, 130-110-201) as per the manufacturer’s instructions. A total of 0.25 g of kidney tissue were placed in a GentleMACS C tube with dissociation buffer (2.35 ml serum-free RPMI (Roswell Park Memorial Institute) culture medium (Gibco), 100 μl enzyme D, 50 μl enzyme R and 12.5 μl enzyme A). Kidney dissociation was performed in a GentleMACS dissociator using the ‘heart_01_01’ programme (15 s). The samples were placed in a waterbath (37 oC) for 30 minutes and then placed back in the dissociator (‘heart_01_02’ programme, 30 s).

After the second round of dissociation, 8 ml of sterile PBS and 10% FBS were added in the C tubes to stop the reaction. The samples were passed through a 40 μm cell strainer into a 50 ml falcon tube. All subsequent steps were performed on wet ice or at 4 oC. The samples were spun at 300g for 5 min and resuspended in 5 ml of cold PBS. They were spun again at 300g for 5 min, resuspended in 1 ml red blood cell lysis buffer (8.29 g NH4Cl, 1 g KHCO3 and 37.2 g Na2EDTA in PBS) and incubated for 30 s on wet ice. The samples were topped with 9 ml of cold PBS and washed twice, as described previously (spin at 300g for 5 min and resuspended in PBS). The samples were resuspended in cold PBS and were subjected to debris removal using the debris removal solution (Miltenyi) according to the manufacturer’s instructions. Finally, the samples were resuspended in 10 ml cold PBS, 10% FBS and 2.5 mM EDTA.

Cell viability and concentration were determined using the Trypan Blue assay and flow cytometry (viability ≥90%). A total of 20,000–40,000 cells were loaded on a 10x Chromium chip (one sample per lane). Cleanup, reverse transcription, cDNA amplification and library preparation were performed using the Chromium Single Cell 3′ Reagent Kits (v3), as per the manufacturer’s instructions.

The samples were sequenced in the Illumina NextSeq 500 using the 2 × 150 bp kit with the following read length parameters: 26 bp Read1, cell barcode and Unique Molecular Identifier (UMI); 8 bp I7 index, sample index; and 98 bp Read2, transcript read. CellRanger v.4.0 (with default parameters) was used to demultiplex Illumina BCL output files and align reads to the ensemble GRCm38.99 reference genome with the addition of the AAV8-TBG-Cre and AAV8-TBG-Null sequences.

All other steps were performed using R v.4.0 and packages from Bioconductor v.3.12. The raw matrices generated by CellRanger (v.4.0) were transformed to SingleCellExperiment objects and they were filtered to exclude empty droplets using the DropletUtils v.1.10 package. SingleCellExperiment objects were merged together to perform further downstream analysis. Following similar parameters used in a previous scRNA-seq analysis murine kidney42, cells with <75 or >3,000 expressed genes or >50% mitochondrial gene expression were filtered out, and only genes expressed in more than ten cells with at least one UMI were kept for further analysis.

The normalization by deconvolution method designed by Lun et al.43 was used to normalize and log transform the counts with functions from Scater package v.1.18. Highly variable genes were computed with default parameters, and the top 10% were used to perform a principal component analysis and UMAP using the top 20 PCs. After examining UMAP plots coloured by batch run, it was determined that batch correction was not required.

Initial clustering was performed with functions from the Seurat package v.4.0 with default parameters and with eps value of 0.5 and resolution sequence of 0.1 to 1 by 0.1. Markers of each cluster were identified by performing a pairwise differential analysis between each pair of clusters with a minimum difference of 20% of cells expressing the gene and log2 fold change threshold of >0.25 and only keeping the differentially expressed markers in all the comparisons. Reclustering of control cells with a new principal component analysis and UMAP was performed, and the main markers were used to manually classify clusters into different cell types. ΔMdm2Hep cells were projected onto the reference UMAP and assigned to the identified clusters. The same methodology was followed for the identified control PTCs, DTCs and the initial cluster ‘mesenchymal cells’ for the projection and assignment of the ΔMdm2Hep cells. Scores for senescence, p21, proliferation, TGFβ, JAK–STAT and regeneration signatures were computed using the AddModuleScore Seurat function. The genes included in each list used to create the signatures can be found in Supplementary Table 5. All transcriptomic data will be made publicly available on the Gene Expression Omnibus (GSE189726) repository at the time of publication. Alternatively, cells expressing senescent features were identified by using an unbiased cell type classifier ‘singleR’, trained on a reference transcriptomic dataset of previously validated murine renal scRNA-seq data14. The hyperparameters used were threshold of 0.4, 150 differentially expressed (DE) genes, quantile of 0.6 and fine-tune parameter set to true.

Differential gene expression (DGE) analysis was performed as described by Giustacchini et al.44. Briefly, the log2 fold change was calculated between groups, and a non-parametric Wilcoxon test was used to compare the expression values. Fisher’s exact test was used to compare expressing cell frequency (percentage of cells per group with at least one UMI count). The Pvalues from both tests were combined using Fisher’s method and adjusted using Benjamini–Hochberg. The genes were considered to be differentially expressed if the P adjusted values were <0.05. The heat maps comparing relative gene expression by cell groups were computed using the ‘DoHeatmap’ function from Seurat R package or ‘plotGroupedHeatmap’ function from the scater R package. In both cases, the scaled value from each group was calculated by substracting the average logcount gene expression from the total mean gene expression and divided by the standard deviation. A two-colour range scale was used to convert scale values into colour intensity. Gene Set Enrichment Analysis was performed in an unsupervised manner using the ‘gseGO’ function from the ClusterProfiler R package; enrichment scores and P values were calculated by the function using an empirical phenotype-based permutation test.

Protein extraction and western blotting

Snap-frozen tissue was homogenized in 300 μl of protein extraction buffer (50 mM HEPES, 100 mM NaF, 150 mM NaCl, 10 mM Na4P207, 10 mM EDTA, 1% Triton X100, 0.1% SDS and 0.5% Na deoxycholate in ddH2O), 1:100 protease inhibitor (Thermo Fisher, 1862209) and 1:100 phosphatase inhibitor (Thermo Fisher, 1862495) in ddH2O using the Precellys Evolution homogenizer (Bertin technologies). The lysate was passed through a 25G needle five to eight times and was then spun twice at 20,800gfor 10 min at 4 oC. The Bradford assay was used to measure protein concentration using the Coomasie Plus reagent (Thermo Fisher, 23236) on a 96-well-plate setting. The absorbance was measured immediately at 596 nm and a standard curve was automatically created using a Spectramax reader.

Western blotting was performed on precast gels using the XCell SureLoc Mini-Cell and XCell I Blot Module (Invitrogen, 10572913), and the protein samples were mixed with loading buffer (4× NuPAGE loading buffer (Invitrogen, 11549166) and 5% β-mercaptoethanol) at a concentration of 20 μg μl−1. The samples were heated for 5 min at 95 oC and then were spun for 2 min at 20,800g. A 20 μl sample and 5 μl of protein ladder (Biorad, 1610373) were loaded on a 4–12%, Bis–Tris, 1.0 mm, ten-well precast gel (Invitrogen, NP0321PK2), and the gel was run at 120 V for 1 h and 50 min in a MOPS running buffer (Invitrogen, NP-0001). Transfer onto polyvinylidene difluoride membranes was performed by wet transfer, using the NuPAGE transfer buffer (Invitrogen, NP-0006) for 1 h at 30 V. Transfer efficiency was assessed by Ponceau S staining. The membranes were blocked for 1 h with 5% BSA buffer and then incubated with the primary antibody (diluted in 5% BSA) overnight at 4 °C. This was followed by 45 min incubation with the HRP-conjugated secondary antibody and enhanced chemiluminescence incubation for the appropriate amount of time (2 min for pSmad2, pSmad3, Smad2 and Smad3 and 5 s for the β-actin). Visualization of the bands was performed using the Chemidoc imaging system (Biorad) and quantification and densitometry was done with ImageJ.

ELISA and cytokine arrays

For the detection of cystatin C in murine plasma, the ‘mouse/Rat cystatin C’ immunoassay kit (R&D Technologies, MSCTC0) was used, according to the manufacturer’s instructions. The plate was scanned in a plate reader at 450 nm (wavelength correction was set to 570 nm) within 10 min after assay completion. The standard curve concentration calculations were performed on the ‘Myassays’ website (https://www.myassays.com/).

Cytokine arrays on murine plasma and tissue samples were performed by Eve Technologies, using the TGFβ1, 2, 3 magnetic bead kit for the measurement of the TGFβ ligands and the Milliplex MAP mouse cytokine magnetic bead panel (discovery assay ‘Mouse Cytokine Array/Chemokine Array 31-Plex (MD31)’).

Metabolomics on mouse urine

For the detection of urine amino acids, urine was collected from the mice by free urination 2 and 3 days before AAV-Cre injection on injection day, as well as 3 and 4 days post AAV-Cre injection. The urine collected by free urination after scruffing the mice was diluted 1:50 in cold metabolite extraction buffer (50% methanol, 30% acetonitrile and 20% water) and were vortexed for 30 s. The samples were then centrifuged at 20,800gfor 10 min at 4 °C. Liquid chromatography–mass spectrometry was performed as described previously45. The source data are presented in Supplementary Table 6.

Cell cultureMEFs

WT MEFs were isolated from one E13-E14 WT C57Bl/6 embryo. The MEFs were cultured on 10 cm Petri dishes (Corning) in Dulbecco’s modified Eagle medium supplemented with 10% FBS, 1% penicillin–streptomycin, 1% ʟ-glutamine (cDMEM) at low O2 (3%). The MEFs cultures were confirmed to be free of mycoplasma contamination. For the plasma treatments, passage (P)3–P5 MEFs were trypsinized, and the cell density was determined by using the CASY cell counter (Cambridge Bioscience, 5651808). A total of 30,000 cells (in 1 ml medium) were plated on 24-well-plate wells (with a round coverslip in each well) and were incubated with cDMEM with 1% plasma from either ΔMdm2Hep or control mice for 24 h. The cDMEM with plasma was changed with fresh cDMEM every 2 days, and 6 days after the first medium change, the cells were stained for SA β-Gal as described above. The coverslips were mounted on the slides, and the SA β-Gal+ cells were quantified by manual counting of 20 random fields of view of using a Zeiss Axiovert 200 microscope at 20× magnification.

NS cell-derived neuronal cells

Neuronal cells were derived from human foetal neural crest progenitors as previously described46. The cells were passaged using trypsin-EDTA solution and were seeded into a geltrex coated 48-well cell culture plate at a density of 5,000 cells per well with 250 μl of proliferation medium. They were allowed to proliferate for 2 days before the proliferation medium was withdrawn and replaced with differentiation medium. The growth medium was replaced with differentiation medium for 16 days before beginning the experimental procedure. The composition of the proliferation and differentiation media has been described previously47.

For the plasma treatment experiments, plasma samples were thawed on ice, vortexed and heat inactivated for 30 min at 60 oC. They were diluted in media to the desired concentration (1:100) and added to the plate for 24 h. For the additional treatment with TGFβR1i, either TGFβR1i (0.2 mM) or vehicle (dimethylsulfoxide, DMSO) were mixed with the same plasma-containing medium and stayed on the cells for 24 h. Two days later, the cells were stained for SA β-Gal, and the positive cells were quantified as described in ‘SA β-Gal staining’ section.

PCLS and PCKSEthics

Human kidney tissue will be collected from donor kidneys declined for transplantation and accepted for research, through the Newcastle Transplant Tissue Biobank under project IOT054.

Kidney slice culture

Eight millimetre human kidney and murine kidney (WT) or liver (Mdm2flfl or ΔMdm2Hep) tissue cores were generated using an 8 mm Stiefel biopsy punch, placed in a metal mould and 3% low gelling temperature agarose and allowed to set. The slices were then generated and cultured according to methods previously described48,49. Human kidney slices were cultured with 10 ng TGFβ1 to induce fibrosis and treated with ±10 μM TGFβR1i SB-525334 (Sigma Aldrich, S8822, batch number: 000014491).

Media transfer experiment

Murine PCLS were generated from liver (Mdm2flfl or ΔMdm2Hep), and the medium was collected and pooled after 24 and 48 h and concentrated using Amicon Ultra-15 Centrifugal Filter Units (Merck, UFC9003). WT murine kidney slices were cultured in PCKS medium containing 0.5% v/v of concentrated cultured PCLS-conditioned medium (equivalent to 25% PCLS condition medium at final concentration) from the livers of control Mdm2flfl or ΔMdm2Hep mice, with or without 200 nM AZ12601011.

Statistical analyses and graphs

The Prism 9 Software (GraphPad Software) was used for statistical analyses. The Shapiro–Wilk test was used to assess data normality. For normally distributed data, the one-way analysis of variance (ANOVA), two-way ANOVA, the Brown–Forsythe and Welch ANOVA test, paired and unpaired t-tests or the Welch’s t-test, were used to test for statistical significance between data groups. The Kruskal–Wallis test or the Mann–Whitney test were performed for non-parametric data. All statistical tests comparing two groups were two-tailed. All figures were created using the Scribus Software (v1.4.7, G.N.U. general public licence). Unless otherwise stated, all data points on the line or bar graphs represent the mean ± standard error of the mean (s.e.m.), and each dot represents a single mouse.

Statistics and reproducibility

No statistical methods were used to pre-determine sample sizes, but our sample sizes are similar to those reported in previous publications3,7,11,12. For animal experiments, the biological replicate sizes were chosen taking into account the variability observed in pilot and prior studies using AAV-TBG-Cre in the Mdm2 model. For all experiments, the animal/sample assignment was matched for sex and age-matched controls and were assigned based upon randomly assigned mouse identification markings. Batched staining and analyses alongside controls were used throughout. The investigators were not blinded for the in vivo experiments. No animals or data points were excluded from analyses. The technical staff administering the therapy were blinded to the mouse genotypes. All subsequent tissue handling and analysis were blinded and/or performed using standardized automated analyses where possible. Western blot studies were performed without blinding, given that samples were run by condition for visualization. In the figure legends, n represents the number of mice, unless explicitly stated. The data distribution for normality and testing of equal variances were assessed using Prism 9 Software. No animals or data points were excluded from analyses.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.