Approvals
Animal experiments were approved by the Stanford Animal Research Committee (nude rat study at Stanford) and the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES; 33.42502-04-15/1807 and -16/2370; rhesus macaque (Macaca mulatta) studies in Göttingen). The BioVAT-HF-DZHK20 Phase 1/2 clinical trial (ClinicalTrials.gov NCT04396899) was approved by the responsible regulatory agency (Paul-Ehrlich-Institut) and the competent ethics committee (Ethics Committee of the University Medical Center Göttingen under file no. 18/7/20). Patients in BioVAT-HF are under guideline-directed medical therapy for advanced heart failure, which also includes listing for heart transplantation if the patients meet the selection criteria (bridge-to-transplant scenario). Informed consent was obtained for clinical trial participation, clinical trial data use and publication and histopathological analysis of the explanted heart. Echocardiography data from the BioVAT-HF trial were evaluated by a core laboratory (J.-U. V.) at the University of Leuven, Belgium.
Cell lines
We used four rhesus macaque iPS cell lines (Supplementary Table 1) derived from skin fibroblasts using Sendai virus transduction of Oct4, Sox2, Nanog and cMyc (iPS cell 43110-4 and DPZ_iRH25.B1) (refs. 24,38) or episomal plasmid reprogramming (DPZ_iRH34.1 and DPZ_iRH23.1) (ref. 23) and human iPS cell line TC1133, also referred to as LiPSC-GR1.1 (ref. 39), derived by episomal plasmid reprogramming of CD34+ cord blood cells and maintained under Good Manufacturing Practice (GMP) at all times. Cell line authentication was performed by MHC (human) and Mamu (Macaca mulatta MHC) typing. All lines were tested to be free from mycoplasma contamination (in-house testing using Lonza MycoAlert Detection Kit as well as external testing of GMP cell line and clinical implants by Minerva Biolabs, Berlin, Germany).
CM differentiation
CMs were derived from human and rhesus iPS cells according to the previously introduced ABCF+I differentiation protocol8. Briefly, the ABCF+I protocol uses activin A (9 ng ml−1; R&D Systems), bone morphogenic factor 4 (BMP4; 5 ng ml−1; R&D Systems), CHIR99021 (1 μmol l−1; Stemgent) and fibroblast growth factor 2 (FGF2; 5 ng ml−1; PeproTech) in basal medium (RPMI, 2% B27 and 200 μmol l−1 of l-ascorbic acid 2-phosphate sesquimagnesium salt hydrate; Sigma-Aldrich) for 3 days for mesoderm induction (MI) followed by cardiac specification in the presence of IWP4 (5 µmol l−1; Stemgent) in basal medium from culture days 4–10. This was followed by recovery in basal medium and metabolic selection in RPMI without glucose (Thermo Fisher Scientific), sodium lactate (2.2 mmol l−1; Sigma-Aldrich) and β-mercaptoethanol (100 μmol l−1; Sigma-Aldrich). Cultures were performed in laminin (LN521; BioLamina)-coated T flasks (refer to Extended Data Fig. 1a for a summary of the CM differentiation protocols used).
StC differentiation
StCs were derived from human and rhesus iPS cells using the newly developed ABCF-CRAB-VCF protocol. Briefly, the ABCF-CRAB-VCF protocol uses activin A (9 ng ml−1; R&D Systems), BMP4 (5 ng ml−1; R&D Systems), CHIR99021 (1 μmol l−1; Stemgent) and FGF2 (5 ng ml−1; PeproTech) in basal medium (RPMI, 2% B27 and 200 μmol l−1 of l-ascorbic acid 2-phosphate sesquimagnesium salt hydrate; Sigma-Aldrich) for 7 days for MI, followed by epicardial differentiation in the presence of CHIR99021 (1 μmol l−1; Stemgent), retinoic acid (4 µmol l−1; Sigma-Aldrich) and BMP4 (50 ng ml−1; R&D Systems) in basal medium from culture days 8–12. Subsequently, epithelial-to-mesenchymal transition was induced in knockout (KO) DMEM supplemented with 10% KO serum replacement (Thermo Fisher Scientific), vascular endothelial growth factor (VEGF) 165 (25 ng ml−1; PeproTech), CHIR99021 (1 μmol l−1; Stemgent) and FGF2 (50 ng ml−1; PeproTech) from culture days 13–16. Finally, expansion was performed in KO DMEM with 10% KO serum replacement, VEGF (25 ng ml−1) and FGF2 (50 ng ml−1) from culture days 17–29 with repeated passaging of subconfluent cultures. Cultures were performed in laminin-coated (LN521; BioLamina; for MI, epicardial differentiation and epithelial-to-mesenchymal transition) and vitronectin-coated (Thermo Fisher Scientific; for expansion) T flasks (refer to Extended Data Fig. 1b for a summary of the StC differentiation protocol used).
EHM construction
CMs and StCs were reconstituted as reported previously8 and indicated in Supplementary Table 3 in a mixture of pH neutralized medical grade bovine collagen (EHM, 0.4 mg per 450 µl; Collagen Solutions LLC) and concentrated serum-free medium (2× RPMI and 8% B27 without insulin). Culture was in Iscove’s medium with 4% B27 without insulin (Thermo Fisher Scientific), 1% non-essential amino acids, glutamine (2 mmol l−1), ascorbic acid (300 μmol l−1), IGF-1 (100 ng ml−1), FGF2 (10 ng ml−1), VEGF-165 (5 ng ml−1) and TGF-β1 (5 ng ml−1) (the latter only during culture days 0–3). All growth factors were purchased from PeproTech. Whole-mount immunofluorescence staining for ACTN2 with subsequent confocal analyses (LSM 710, Zen 2.3 SP1; Zeiss), contractility assessments and flow cytometry were performed routinely as reported previously8 for quality control in EHM batches (n = 17 rhesus EHM batches and n = 1 human EHM batch) produced for implantation.
Isometric force measurements
Surrogate EHM loops were suspended under isomeric conditions in thermostatted (37 °C) organ baths filled with Tyrode’s solution (T2397; Sigma-Aldrich). Analysis was under electrical field stimulation at 1.5 Hz with 5-ms square pulses of 150-mA electrical current, as described previously8. EHMs were mechanically stretched at intervals of 125 µm until the maximum twitch force was observed (Lmax) according to the Frank–Starling law. Contractility data were recorded with Bmon and analysed using the Amon software (VitroDat 3.52; Föhr Medical Instruments).
Single-nucleus RNA sequencing
Flash frozen iPS cell-derived CMs, StCs or EHMs were used to prepare cell nuclei for cell composition analyses by snRNA-seq. All steps were performed on ice. Cells and EHM were first homogenized in 500 µl lysis buffer (Sigma) using plastic pestles (45 strokes in a microfuge tube). Then, 1.5 ml of lysis buffer was added and further incubated on ice for 7 min. Following centrifugation at 500g and 4 °C for 5 min, the supernatant was discarded and the crude nuclear pellet was resuspended in 2-ml lysis buffer. After incubation for 7 min and centrifugation, the nuclear pellet was resuspended in 1.8 ml nuclei resuspension buffer (NRB containing 0.5% BSA (Serva), 1:200 RNaseIN Plus Inhibitor (Promega) and 1× protease inhibitor (Roche) in PBS (Invitrogen)) and centrifuged for 5 min at 500g and 4 °C. The resulting nuclear pellet was resuspended in 500-µl NRB and filtered through a 35-µm strainer. To remove the remaining debris, nuclei were further purified by fluorescence-activated cell sorting (BD FACSAria III with an 85-µm nozzle and 45-psi sheath pressure) at 4 °C based on their size and collected into NRB-coated tubes. Following sorting, the nuclei were pelleted by centrifugation for 5 min, resuspended in 100 µl PBS and counted in a Neubauer chamber using 10% trypan blue. snRNA-seq libraries were prepared using Chromium Single Cell 3′-RNAseq v.3 library preparation kit (10X Genomics). Following the nucleus counting step, they were immediately processed to capture 2,000 single nuclei in a 10X chromium controller. Pooled complementary DNA libraries, prepared according to the manufacturer’s guidelines, were then sequenced in an Illumina NextSeq 550 platform. Gene counts were obtained by aligning reads to the hg38 genome (NCBI:GCA_000001405.22; GRCh38.p7) using Cell Ranger software (v.3.0.2; 10X Genomics). The Cell Ranger count pipeline was used to generate a gene-count matrix by mapping reads to the pre-messenger RNA as a reference to account for unspliced nuclear transcripts. The Scanpy package was used for pre-filtering, normalization and clustering40. Nuclei read counts indicating low-quality cells (that is, too many reads suggesting capture of more than one nucleus, too few reads with a read coverage of less than 200 genes or less than 10% housekeeping gene transcript coverage)41 were excluded. Next, counts were scaled by the total library size multiplied by 10,000 and transformed into log space. Principal component analysis was performed on the variable genes, and UMAP was run on the top 50 principal components42. These were used to build a k-nearest-neighbour cell–cell graph with k = 100 neighbours. Subsequently, spectral decomposition over the graph was performed with 50 components, and the Louvain graph clustering algorithm was applied to identify cell clusters. We confirmed that the number of principal components captured almost all of the variance in the data. For each cluster, we assigned a cell-type label by manual evaluation of gene expression for sets of known marker genes (Supplementary Note 1).
Flow cytometry analysis of cell composition
EHMs were washed in PBS and dissociated in Collagenase 1 (2 mg ml−1; Sigma-Aldrich) in PBS with 20% FBS at 37 °C for 1 h followed by Accutase (Millipore), 0.025% trypsin (Thermo Fisher Scientific) and DNase I (20 µg ml−1; Calbiochem) at 20–24 °C for 30 min. After fixation in 70% ice-cold EtOH for more than 10 min, cells were exposed to primary antibodies directed against sarcomeric actinin (ACTN2: 1:4,000; A7811; Sigma) or vimentin (VIM: 1:1,000; ab92547; Abcam) in blocking buffer for 45 min, followed by secondary antibodies in blocking buffer and Hoechst 33342 for 30 min at 4 °C (Supplementary Table 6). Control samples were exposed to undirected IgG1 (MAB002; R&D Systems). Human samples were fixed with 4% formaldehyde and exposed to conjugated antibodies directed against sarcomeric actinin (ACTN2-PE; 1:1,000; 130-106-937; Miltenyi Biotec) and vimentin (VIM-AF647; 1:1,000; 677807; BioLegend) for 15 min at 4 °C. A BD LSRII SORP system (BD Biosciences) or CytoFLEX (Beckman Coulter) was used for the flow cytometry analysis. The gating strategy was as described previously8. BD FACSDiva Software (BD Biosciences) or Kaluza (Beckman) was used for analysis.
I/R injury and EHM implantation in the nude rat study
I/R injury was induced in male 8- to 10-week-old Rowett nude rats (250–350 g; Charles River Laboratories), as described previously24. Briefly, rats were treated with buprenorphine (0.05 mg kg−1 by subcutaneous injection) and cefazolin (50 mg kg−1 by intramuscular injection) followed by anaesthesia under 2% isoflurane. Body temperature monitoring and control were performed using a rectal probe and surgery on a 37 °C heating plate. After intubation, ventilation was with a tidal volume of 0.5 ml kg−1 at a rate of 90 breaths per minute. A left thoracotomy was performed at the fourth intercostal space, and myocardial ischaemia was induced by occlusion of the left anterior descending (LAD) coronary artery. After 60 min of no flow, reperfusion was allowed by releasing the ligation. After closure of the chest wall, carprofen (5 mg kg−1 s.c.) was administered for additional analgesia. Four days after I/R injury, animals were randomly grouped into two groups: (1) implantation of viable rhesus macaque EHM (n = 7) and (2) implantation of non-viable (60 Gy irradiated, as described previously18) rhesus macaque EHM (n = 8). For EHM implantation, the chest was reopened, and EHMs were attached to the left ventricular free wall with eight to 12 stitches using a 7-0 PROLENE suture, as described previously18. EHM implantations were performed by a study-group-blinded microsurgeon.
Echocardiography in the nude rat study
Rats were sedated with 1–2% (v/v) isoflurane in oxygen and placed on a warming pad (37 °C) with ECG respiratory recording. M-mode ultrasound images in the short-axis view were acquired for systolic function using a high-frequency (21 MHz) linear transducer (MS250) on a small animal ultrasound system (Vevo 2100; VisualSonics). Tissue Doppler imaging was performed in a four-chamber view for diastolic function using the same ultrasound probe. Cardiac ultrasound data were analysed using the Vevo LAB software. The inner diameter of the left ventricle was measured from M-mode images at the end of diastole and end of systole. The volumes of the end of diastole and end of systole, fractional shortening and ejection fraction were sequentially calculated. Diastolic relaxation of rat hearts was measured from tissue Doppler images; diastolic dysfunction was defined as E′/A′ < 1.
Histopathology in the nude rat study
The rats were euthanized 4 weeks after EHM implantation. The hearts were excised and fixed in 4% formaldehyde in PBS overnight at 4 °C. After washing in PBS three times for 10 min, the hearts were cryoprotected in 30% sucrose overnight, embedded in optimum cutting temperature compound (Tissue-Tek) on dry ice and stored at −80 °C. Samples were cryosectioned into 10-µm sections and mounted on Superfrost Plus slides (Thermo Fisher Scientific). Permeabilization was performed using 0.1% Triton-X for 15 min at room temperature, followed by blocking with 10% normal goat serum in PBS for 25 min at room temperature. The primary antibodies (Supplementary Table 6) were applied at 4 °C overnight. Appropriate secondary antibodies were applied for 1 h at room temperature followed by 3 × 5 min washes in wash buffer. Slides were sealed with mounting solution containing DAPI for DNA labelling (ProLong Gold Antifade Mountant with DAPI; Thermo Fisher Scientific).
EHM implant preparation in the rhesus macaque study
After transfer to the point of care, EHMs were prepared as 1×, 2× or 5× EHM assemblies by folding (1×) and stacking (2× and 5× EHM) in 3D-printed custom-made holders, followed by suturing (5-0 PROLENE; Ethicon) to a TachoSil membrane (4.5 × 4.5 cm; Takeda) used as (1) a security measure to prevent possible epicardial bleeding, (2) to support targeted surgical administration and (3) to reduce pericardial adhesions43 (Extended Data Fig. 3).
EHM implantation in healthy rhesus macaques
Fourteen (nine male and five female) rhesus macaques (Macaca mulatta) from the breeding colony at the German Primate Center (Deutsches Primatenzentrum) were assigned for the investigation of the feasibility and safety of EHM implantation in healthy animals (cohorts 1 and 2; Supplementary Table 3). The average body weight and age of animals at the time of EHM implantation were 8.4 ± 0.6 kg and 7.5 ± 0.7 years (n = 14). Animals were first anaesthetized by intramuscular injection of ketamine (7 mg kg−1) (100 mg ml−1; Wirtschaftsgenossenschaft deutscher Tierärzte (WDT)) and medetomidine (0.04 mg kg−1) (Domitor, 1 mg ml−1; Vetoquinol GmbH). After endotracheal intubation, total intravenous anaesthesia was used by administration of propofol (Propofol-Lipuro 2%, 10–40 mg kg−1 h−1; B. Braun Melsungen AG) and fentanyl (Fentadon, 10 μg kg−1 h−1; Dechra) via an intravenous line introduced into the vena saphena or vena cephalica using a 22G line (Vasofix Safety, 22 G; B. Braun Melsungen AG) under continuous arterial blood pressure monitoring (via the A. tibialis), ventilation with at least 60% oxygen depending on O2 saturation (Nonin 7500FO) and CO2 concentration in the exhaled air (Siemens Servo Ventilator 900C), and ECG monitoring. Body temperature was monitored using an oesophageal probe. Electrolytes were balanced by continuous Sterofundin infusion (5–10 ml kg−1 h−1; B. Braun Melsungen). Amoxicillin (Duphamox, 15 mg kg−1; Zoetis) and meloxicam (Metacam, 0.2 mg kg−1; Boehringer Ingelheim) or carprofen (Rimadyl, 2–4 mg kg−1; Zoetis) were administered perioperatively. After incision in the fifth intercostal space (according to echocardiographic assessment of the heart position in the operating theatre), the pericardium was opened horizontally, anterior to the phrenic nerve. Epicardial sutures were placed to ensure the proper positioning of the EHM implant at the epicardial target site. After positioning and suturing (six 5-0 PROLENE single-knot sutures) of the EHM onto the epicardial heart wall and closure of the pericardial sac, the chest wall was closed with surgical sutures. The animals were subsequently weaned from anaesthesia and returned to the animal facility under veterinary care. In cohort 1, 1× EHMs (constructed from 40 × 106 cells) were implanted (six allografts and one autograft) in an adaptive study design to first identify optimal immune suppression (n = 2 + 2; tacrolimus only versus tacrolimus plus methylprednisolone) and then extend the study under optimal immune suppression (n = 2) with a parallel autograft without immune suppression as a reference (n = 1); follow-up in cohort 1 was 3 months. Cohort 2 tested the MFD (5× EHM constructed from 200 × 106 cells, approximately 5 g w/w) as allografts (n = 5) under the cohort 1 identified optimal immune suppression protocol (tacrolimus plus methylprednisolone). An autograft study without concurrent immune suppression (n = 1) served as a reference. Cyclosporin with methylprednisolone was tested in one allografted animal (n = 1) as an alternative to tacrolimus with methylprednisolone. In two allograft animals, immune suppression was stopped 3 months after implantation to investigate the consequences of EHM rejection. Calcineurin inhibitors were started 5 days and methylprednisolone 2 days before EHM implantation to ensure effective immune suppression already at the time of EHM implantation. Therapeutic drug monitoring was applied with dose adjustment, as needed, to ensure target tacrolimus and cyclosporin trough levels (Supplementary Data 1). To establish baseline information, MRI and clinical chemistry data were obtained at two time points: (1) 44 ± 7 days (n = 14) before EHM implantation (baseline 1) and (2) on the day of EHM implantation (baseline 2). Additionally, MRI studies were performed 28, 56 and 84 days (cohort 1) and in addition 168 days (cohort 2) after EHM implantation. Blood draws for clinical chemistry studies were taken at 9, 18, 27, 42, 56, 70 and 84 (±2) days (cohort 1) and in addition 112, 140 and 168 (±2) days (cohort 2) after EHM implantation.
I/R injury in the macaque studies
Twenty rhesus macaques (Macaca mulatta) from the breeding colony at the German Primate Center (Deutsches Primatenzentrum) were assigned to the ‘EHM implantation in chronic heart failure after I/R injury’ cohort (cohort 3; Supplementary Table 3). This study included 14 male and six female macaques. One female macaque (no. 15389) had to be excluded according to the recommendation of the responsible veterinarians before I/R injury induction because of low (6 kg) and no gain in body weight over a time period of more than 2 months after transfer into the experimental animal unit. The average body weight and age of animals at the time of I/R injury were 9.0 ± 0.3 kg and 7.3 ± 0.3 years (n = 19). I/R injury was inflicted under total intravenous anaesthesia, as described above, with repeated control of activated clotting time (ACT; target: 250–350 s; ACT Plus; Medtronic) under heparinization (starting dose of 7,000 U by intravenous injection followed by dose adjustments according to ACT values; Heparin-Natrium-25000-ratiopharm). A 5F Launcher internal mammary artery guiding catheter (Medtronic), bent to fit the aortic root anatomy of the rhesus macaque using heated air, was advanced via the femoral artery into the left main coronary artery under fluoroscopy (Siemens Artis zee multi-purpose). A 0.014 in. coronary guide-wire was advanced into the distal LAD, and a 1.2–2.0 mm over the wire balloon (EMERGE; Boston Scientific) was inflated to 10 bar in the mid-LAD. Occlusion was confirmed via angiography. Balloon pressure (10 bar) was continuously maintained for 3 h and monitored with additional angiographic confirmation of vessel occlusion once per hour (Extended Data Fig. 7). ST-wave elevation myocardial infarction was confirmed by ECG and blood biomarkers (troponin, total creatine kinase/muscle- and brain-type creatine kinase and lactate dehydrogenase; Supplementary Data 3). Five of 19 macaques subjected to I/R injury died because of bleeding complications (n = 3) or sudden cardiac death (n = 2; Supplementary Table 3).
EHM implantation in rhesus macaques with chronic heart failure
Six months (176 ± 6 days; n = 14) after I/R injury and confirmation of chronic heart failure development (Extended Data Fig. 7), rhesus macaques were randomized to be included in control groups (with (n = 4) or without (n = 3) immune suppression) or implanted with either 2× (n = 3) or 5× (n = 4) EHM, as described above. For immune suppression in cohort 3, tacrolimus (0.02 mg kg−1 d−1 PROGRAF i.m. with dose adjustment to reach target trough levels of 15–25 ng ml−1) and methylprednisolone (0.15 mg kg−1 d−1) were administered. Tacrolimus was started 5 days and methylprednisolone 2 days before EHM implantation. The 6-month follow-up (MRI and blood draws) was the same as in cohort 2.
Implantation of telemetry ECG event recorders
Reveal LINQ (Medtronic) telemetry monitors were implanted subcutaneously under general anaesthesia at baseline 1 investigations (44 ± 7 days before EHM implantation in cohorts 1 and 2 (n = 14); 40 ± 7 days before I/R injury (n = 19) and 176 ± 6 days (n = 14) before randomization into the four study groups in cohort 3) according to the supplier’s instructions for heart rate, activity monitoring and arrhythmia event documentation. The optimal position for the ECG recorder was identified by surface ECG assessments on the right parasternal chest. This location was chosen to not interfere with the surgical route of EHM implantation and the cardiac MRI investigations. The event recorder was set to detect tachycardia (greater than or equal to 222 bpm for 16 or more beats), bradycardia (less than or equal to 30 bpm for four or more beats) and asystole (no beats for more than 3 s).
Clinical chemistry
Animals were first anaesthetized by i.m. injection of ketamine (7 mg kg−1) (100 mg ml−1 WDT) and medetomidine (0.04 mg kg−1) (Domitor, 1 mg ml−1; Vetoquinol GmbH). Blood was collected from the femoral vein using the BD Vacutainer system and 22G BD Vacutainer Eclipse blood collection needles. Standard clinical chemistry and haematology analyses were performed at the University Medical Center Central Laboratory.
Magnetic resonance imaging
Animals were sedated with 7 mg kg−1 of ketamine (100 mg ml−1, WDT) and 0.04 mg kg−1 medetomidine (Domitor, 1 mg ml−1; Vetoquinol GmbH) via intramuscular injection. After endotracheal intubation, anaesthesia was maintained via a continuous infusion of Propofol-Lipuro 2% (10–40 mg kg−1 h−1; B. Braun Melsungen AG) via an intravenous line introduced into the vena saphena or vena cephalica using a 22G line (Vasofix Safety, 22G; B. Braun Melsungen AG). Pressure-controlled ventilation (Servo Ventilator 900C; Siemens-Elema AB) with at least 60% oxygen in exhaled air was applied to keep the O2 saturation between 98% and 100% (measured by pulse oximetry; Nonin 7500FO). The CO2 concentration in exhaled air (IntelliVue; Philips) was maintained below 30% to avoid spontaneous breathing. Measurements were performed on a 3T MRI system (MAGNETOM Prisma; Siemens Healthineers) using a 16-channel multipurpose coil (VARIETY; Noras MRI Products) for signal detection, as described previously44. Gadolinium-based contrast agent was applied intravenously as a contrast agent to identify EHM graft perfusion. Data were analysed independently by four blinded investigators as to condition and timing using Segment v.4.0 R12067 (Medviso, segment.heiberg.se) and Medis-Suite v.3.2 software with QMass module v.8.1 (Medis).
Donor-specific antibodies
The iPS cell-derived CMs and StCs, unstimulated and after IFNγ (100 ng ml−1 for 48 h) were exposed to sera obtained before (pre) and at the indicated time points during the study at different dilutions (1:5 to 1:40). A fluorescein isothiocyanate-labelled anti-rhesus immunoglobulin G antibody (4700-02; SouthernBiotech) was used to detect antibodies in the sera bound to the CMs and StCs. The cell mean fluorescence intensity (MFI) and the proportion of stained cells were determined by flow cytometry (LSR II SORP; BD Biosciences). Antibodies that display a selective reactivity to IFNγ-stimulated CMs presumably include DSAs to MHC class I molecules. The pan-human leukocyte antigen antibody W6/32 (BioLegend), which reacts with MHC class I molecules of rhesus macaques, was used to demonstrate the expression of these molecules on CMs and StCs.
Flow cytometry analysis of peripheral immune cells
Whole blood (50 μl) was stained with a mixture of pre-titrated monoclonal antibodies (Supplementary Table 6) for 30 min at room temperature in the dark. Lysis of red blood cells and fixation were performed by incubation with 1-ml red blood cell lysis/fixation solution (BioLegend) for 15 min. Following a washing step with PBS/BSA, cells were analysed using an LSR II cytometer (BD Biosciences) and FlowJo 9.6 software (Tree Star).
Pathology
Cardiac arrest was introduced under deep ketamine/medetomidin anaesthesia by injection of pentobarbital (90–120 mg kg−1) in a potassium chloride solution (30 mmol l−1) under online monitoring of heart rate and rhythm via the Reveal LINQ monitor. After confirmed cardiac arrest, the hearts were excised, weighed, photographed (Extended Data Fig. 9) and formaldehyde fixed for histopathological analyses. In addition, the following organs were collected and subjected to macroscopic and microscopic pathology analyses with a particular focus on abnormal cell growth: oesophagus, aorta, lung, trachea, liver, pancreas, spleen, reproductive organs, lymph nodes, thymus, bladder, stomach, small and large intestines, thyroid gland, adrenal gland and parotid gland samples. Histopathological studies were performed on haematoxylin and eosin (H&E)-stained paraffin sections according to standard pathological procedures. The heart explants were sectioned from base to apex in 3- to 5-mm slices before paraffin embedding. Paraffin embedded samples were cut into 2-μm tissue sections and counterstained with Meyer’s haematoxylin (Dako; Agilent Technologies) for 8 min and analysed by light microscopy. Immunohistochemistry reactions were performed after antigen retrieval at 97 °C in citrate buffer (pH 6) or EDTA buffer (pH 9). The following antibodies and dilutions were used: anti-vWF (polyclonal rabbit, EDTA buffer (pH 9), ready-to-use (RTU), 20 min of incubation, Dako; Agilent Technologies), anti-desmin (monoclonal mouse, EDTA buffer (pH 9), RTU, 20 min of incubation, clone D33, Dako; Agilent Technologies) and anti-Ki67 (monoclonal mouse, EDTA buffer (pH 9), RTU, 20 min of incubation, MIB-1, Dako; Agilent Technologies). The sections were incubated with an RTU horseradish peroxidase-labelled secondary antibody at room temperature for 25 min (anti-rabbit/mouse, produced in goat; REAL EnVision Detection System, Dako, Agilent Technologies). The substrate DAB+Chromogen system produces a brown end-product and is applied to visualize the site of the target antigen (REAL DAB+Chromogen; Dako) (refer to an overview of applied antibodies in Supplementary Table 6). Individual CM area was calculated by multiplication of CM length and breadth measured in longitudinally sectioned desmin-positive CMs.
Graft identity assessment
DNA was isolated, as described previously45, from micro-dissected formaldehyde-fixed paraffin embedded (FFPE) slices using the innuPREP FFPE DNA Kit on the InnuPure C16 System (Analytik Jena) according to the manufacturer’s instructions. Samples were obtained from desmin-positive remote and EHM engrafted areas. DNA concentrations were measured on a Qubit 3.0 Fluorometer (Thermo Fisher Scientific). Microsatellite genotyping was performed in macaque samples using a genotyping-by-sequencing approach, as described previously46. Allele calling based on sequence data generated on Illumina’s MiSeq platform (251-bp forward and 51-bp reverse) was done with the CHIIMP pipeline47. In human samples, deep sequencing of a targeted multigene panel (78 genes) was performed on 50 ng genomic DNA. For library preparation, SureSelectXT HS target enrichment kit (Agilent) with enzymatic fragmentation was used following the manufacturer’s protocol. Libraries were sequenced on an Illumina NovaSeq 6000 with 2 × 150-bp read length and with mean coverage of 3,000×. SeqPilot (JSI Medical Systems GmbH) and Varbank 2.0 (Cologne Center for Genomics, University of Cologne) software was used to align sequences to a human reference genome (hg19) and for SNV calling. SNVs were filtered against (1) absence from control area, (2) high coverage and (3) exclusion of sequence artefacts.
Mechanical conditioning study
EHMs (n = 5) were subjected to mechanical stretching in a modified contraction measurement set-up with a programmable actuator48. After suspension and equilibration in EHM medium, EHMs were stretched stepwise from L0 (slack length) to Lmax (preloaded to achieve maximal FOC). Subsequently, preload was adjusted to approximately 80–90% of Lmax, and EHMs were subjected to chronic mechanical stimulation (200-µm extension for 500 ms) at 1 Hz. Contractility was recorded using an automated routine with 2-min data acquisition per hour for the whole study duration of 120 h.
Statistical analyses
Data are reported as mean ± s.e.m. unless otherwise indicated. Microsoft Excel 2019 MSO (16.0.10415.20025) and GraphPad Prism 10.1.2 were used for statistical analyses. Sample numbers (biological replicates) and statistical tests are indicated in the main body of the text, tables and figure legends.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
