Human hearts have an intrinsic regenerative potential following myocardial infarction

doi:10.21203/rs.3.rs-3244463/v1

Download PDF

Biological Sciences - Article

Human hearts have an intrinsic regenerative potential following myocardial infarction

https://doi.org/10.21203/rs.3.rs-3244463/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Myocardial infarction (MI) is a leading cause of death worldwide and can eliminate up to a third of the cardiomyocytes (CMs) within the heart. Several non-human mammalian models and a limited number of human heart studies suggest that CMs can proliferate during early development, but most CMs stop regenerating soon after birth. In contrast, rodent MI models show that CMs increase their regenerative capacity in response to ischemia, however this has not been shown in humans. Using immunostaining, multi-omics and a unique pre-mortem human heart collected 5 days post-MI, we show that adult human CMs exhibit an increased intrinsic regenerative potential in response to ischemia. Furthermore, single nucleus ribonucleic acid sequencing (snRNAseq) reveals the unique gene expression of post-MI proliferative human CMs. Finally, we demonstrate a novel clinical method to safely biopsy peri-ischemic post-MI human heart tissue and provide further evidence of ischemic CM regeneration. Developing a means to enhance this intrinsic CM regenerative potential could lead to novel therapeutics for cardiac injury and heart failure.

Biological sciences/Cell biology/Cell division

Health sciences/Diseases/Cardiovascular diseases/Heart failure

Biological sciences/Cell biology/Mechanisms of disease

The adult human heart contains approximately 3 billion cardiomyocytes (CMs). However, significant ischemic events, such as myocardial infarction (MI) caused by coronary artery disease (CAD), can eliminate up to a third of these cells^1,2. Although survival following MI has improved over the last decade due to significant therapeutic advancements, many patients still develop irreversible heart failure³. Thus, identifying alternative options to prevent and reverse heart failure is paramount.

The tenet that the mammalian heart is a terminally differentiated organ in which CMs lack the capacity to undergo proliferation, has been challenged over the last decade^1,4. For instance, several studies have demonstrated that non-human mammalian hearts have an excellent regenerative capacity during normal development². Others have shown neonatal rodents can regenerate the apex of the heart following resection⁵, analogous to non-mammalian models¹. However, after birth, the proliferative potential of CMs decreases, with one study showing that human CMs cease to produce daughter cells after 20 years of age⁶. In contrast to the uninjured adult heart, murine models of induced ischemic MI injury have demonstrated increased CM proliferative capacity in the peri-ischemic border zone^7,8. Whilst these well-established rodent MI models of cardiac regeneration have been used for decades, there is no robust evidence that the observed increase in CM proliferation translates to humans. Indeed, there is an urgent need to regenerate the field of human cardiac regeneration itself⁹.

Herein we demonstrate that human hearts possess an intrinsic CM regenerative potential, which is stimulated by ischemia. We analyzed a unique pre-mortem ischemic human heart using a multi-modal approach by imaging mass cytometry (IMC), multi-omics, immunohistochemistry (IHC) and single nucleus ribonucleic acid sequencing (snRNAseq). Together, these approaches characterized the expression of myocardial transcripts, proteins, and metabolites, in addition to measuring CM proliferation and identifying proliferative CM gene expression. We also developed a novel method to obtain human heart biopsies from patients 7-10 days post-MI, which, following IHC analyses, further confirmed human CMs increase their regenerative potential in response to ischemia.

Characterization of a unique pre-mortem infarcted human heart

We obtained a unique infarcted human heart, whereby a 48yo male patient suffered a catastrophic acute MI, secondary to complete occlusion of the left anterior descending (LAD) coronary artery. This patient was sustained on life support for 5 days post-MI but displayed no neurological activity. The next of kin agreed to organ donation for research (Sydney Heart Bank, University of Sydney Human Research Ethics Committee approval #2016/923) and the heart was pre-mortem cryopreserved within 15 minutes of harvest. Regions of the left ventricle (denoted ‘MI LV’), right ventricle (denoted ‘MI RV’) and right atrium (denoted ‘MI RA’) were collected, providing both positive and negative ischemic internal controls (Fig. 1a). It should be noted that although the MI RV was subtended by a non-diseased right coronary artery, the LAD had the potential to cross the interventricular septum and partially supply the RV. Indeed, we observed ischemia within MI RV, albeit to a lesser degree than the MI LV.

To characterise the infarcted heart, we performed a range of complementary techniques; histology, multiphoton second harmonic generation (SHG) microscopy and immunohistochemistry (IHC). As with post-MI murine models, the ischemic MI LV showed replacement fibrosis, angiogenesis, and immune cell infiltration (Fig. 1b-e). Next, we defined the cellular composition of the post-MI heart using an 18-antibody imaging mass cytometry (IMC) panel. Machine learning multi-cut cell segmentation methodology¹⁰ allowed identification of CM and non-CM cell borders and analysed protein expression within each cell (Fig. 1f-h). Unbiased cluster analysis revealed increased endothelial cells, fibroblasts, M2-macrophages and neutrophils within the MI LV, as compared to remote zone LV (Extended Data Fig. 1, Fig. 1i-k). These data show that this unique human heart exhibits structural and cellular post-MI remodelling similar to the commonly utilised MI mouse models¹¹.

Evidence of increased human cardiomyocyte regenerative potential in the post-MI heart

Increased CM mitosis has been previously reported in post-MI mouse models^7,8. To confirm whether the human heart yielded similar results, we performed IHC analyses. Within MI LV we observed CMs undergoing binucleation and CMs expressing nuclear localisation of mitotic markers phospho-histone H3 (pHH3) and aurora kinase B (AURKB) (Fig. 2a-c). Next, using image analysis, we compared MI LV, MI RV, MI RA, and age/sex matched healthy Donor LV. Analysis of our results demonstrated an increase in mitosis within the highly ischemic MI LV as compared to MI RV, MI RA and Donor LV (Fig. 2d) and also indicated a gradient of mitosis correlating with each region’s degree of ischemia i.e. Donor LV < MI RA < MI RV < MI LV (Fig. 2d). To our knowledge, this proportion of mitotic CMs observed in the MI LV (29.59% pHH3⁺ and 16.78% AURKB⁺) has not been previously reported in humans.

During mitosis, nuclei separate in a process called karyokinesis. Following karyokinesis is cytokinesis, whereby the cytoplasm splits to form two new daughter cells. Although adult CMs are known to complete karyokinesis, during adulthood they often fail to complete cytokinesis, resulting in bi- or multi-nucleated polyploid cells¹. In an injury setting, however, mammalian models have shown mitotic CMs can complete cytokinesis at the infarct border zone^7,8,12. Here, we used IHC analyses to elucidate whether human CMs undergo cytokinesis in response to ischemia. Using this approach, we found instances of CM nuclei splitting into two daughter cells with their cytoplasm partly separated by basement membrane (wheat germ agglutinin⁺, WGA⁺) in MI LV samples (Fig. 3a), suggestive of CMs undergoing cytokinesis. We also detected mitotic kinesin-like protein 1 (MKLP1) accumulation at the cleavage furrow between newly divided CMs (Fig. 3b); providing bona fide evidence of cytokinesis^6,13. Similar to pHH3⁺ CMs and AURKB⁺ CMs, MKLP1⁺ CMs followed a gradient of ischemia: Donor LV (0.01%) < MI RA (0.05%) < MI RV (0.07%) < MI LV (0.34%) (Fig. 3b-d). To our knowledge, only a proportion of 0.005-0.01% MKLP1 positive CMs have previously been detected in non-ischemic human hearts and only in adults below 20yo⁶. Thus, our 48yo MI LV (ischemic) samples exhibited a proportion of MKLP1⁺ CMs significantly higher than previously reported in (non-ischemic) human hearts, in an age range where completed cytokinesis has not previously been observed⁶.

Multi-omics shows the post-MI human heart upregulates transcripts, proteins and metabolites known to increase CM proliferation

To further characterise the infarcted heart, we undertook a multi-omics approach, utilising bulk RNAseq (transcriptomics), proteomics and metabolomics. Using these techniques, we compared MI LV, MI RV and MI RA, to age/sex matched healthy Donor LV. Expectedly, multidimensional scaling (MDS) analysis of RNAseq data showed separation of non-ischemic Donor LV and ischemic MI samples (Fig. 4a). Repeating the comparisons without Donor LV, MDS plots also showed separation of MI LV and MI RV that was not previously apparent (Extended Data Fig. 2a). Differential gene expression (DGE) analysis of MI LV vs Donor LV showed 1830 upregulated transcripts, including genes previously reported to induce adult CM proliferation: YAP1¹⁴, IGF2¹⁵, AGRN¹⁶ and DCN¹⁷ (Fig. 4b). Among the 1036 downregulated transcripts was cell cycle inhibitor CDKN1C (aka p57) (Fig. 4b), which reportedly enhances adult mouse CM cell cycle re-entry when knocked down in genetic mouse models¹⁸.

Proteomics and subsequent pathway analyses of MI LV vs Donor LV showed upregulation of proteins and pathways associated with neutrophils (MPO, ELANE), macrophages (CD14, CD163, CD206), fibroblasts (THY1, CD44, VIM) and angiogenesis (VWF, PECAM1) (Fig. 4c, Extended Data Fig. 2b, Extended Data Fig. 3), providing further evidence this human MI model recapitulates post-infarct remodeling seen in mouse models. In addition, increased post-MI elastogenesis was observed, which we have previously shown in both post-MI rats and humans¹⁹ (Extended Data Fig. 3). Notably, multiple proteins were upregulated in MI LV vs Donor LV that have been previously reported to induce CM proliferation in mouse models: YAP1¹⁴, IGF2¹⁵, POSTN²⁰, TGFBI¹⁶, AGRN¹⁶ and DCN¹⁷ (Fig. 4d,e).

Recently, others have identified multiple metabolic pathways and metabolites that can increase adult mouse CM proliferation post-MI, such as upregulation of the mevalonate pathway²¹, increased malonate²² and increased ketogenesis²³. To assess whether these metabolites were upregulated in the post-MI human heart, we used a combination of hydrophilic interaction (HILIC) and HILIC liquid chromatography with an amide transition paired to tandem mass spectrometry (LC-MS/MS) metabolomics methods to compare MI LV vs Donor LV. This showed MI LV had increased mevalonate, malonate, ketone bodies 3-hydroxybutyrate (3HB), acetoacetate, ketogenic substrate β-Hydroxy β-methylglutaryl-CoA (HMG-CoA), short-chain ketoacid 2-oxobutanoate (2-OB) and central carbon metabolite glyceraldehyde-3-phosphate (G3P, an intermediate in both glycolysis and gluconeogenesis) (Fig. 4f). Of note, the top upregulated metabolite in the MI LV patient (a known smoker) vs Donor LV (non-smokers) was a nicotine metabolite called cotinine²⁴ (Fig. 4f), supporting the validity of our assay. When comparing MI LV vs MI RV, two ketone bodies (acetoacetate and 3HB) again demonstrated statistically significant upregulation (Fig. 4g). Together, these omics datasets show the human ischemic LV upregulates transcripts, proteins, and metabolites that have previously been reported to induce adult CM proliferation in mouse MI models^{14–17,20–23}.

Single nucleus (sn)RNAseq identifies proliferative human cardiomyocyte (ProCM) gene expression following myocardial infarction

To analyse the gene expression of the proliferating CMs we observed in the human heart, we performed single nucleus RNAseq (snRNAseq) on MI LV (n=13,547) and Donor LV (n=5,717). Using gene set enrichment analysis (GSEA) for cell specific markers, we identified 11 cell populations and a distinct separation of MI and Donor cells (Fig. 5a, Extended Data Fig. 4a-c). As expected, due to post-MI CM cell death, MI samples showed a decreased proportion of their CMs:total cells (Extended Data Fig. 4d). Next, using GSEA we defined proliferative cells as having increased S-phase cell cycle gene expression. As expected, likely driven by post-MI stromal cell proliferation, MI samples showed an increased proportion of their proliferative cells:total cells, as compared to Donor LV (Extended Data Fig. 4e). Using these same strategies i.e. enrichment for S-phase markers within the CM population, we identified proliferative CMs (ProCM) (Figure 5b,c) and showed an increased proportion of ProCM:CM in MI samples (Extended Data Fig. 4f).

Differential gene expression analysis (DGE) of MI ProCM identified 235 downregulated and 311 upregulated genes as compared to Donor ProCM. Downregulated genes included CM markers such as MYOM2 and TTN, whilst upregulated genes included EPAS1 (aka HIF2A) POSTN (identified in Fig. 4c proteomics dataset and reported to induce CM proliferation²⁰), CTNNB1 (beta-catenin, reportedly increases CM dedifferentiation and proliferation^25,26), CCND3 (cyclin D3, G1/S-phase cell cycle transition) and RUNX1 (reportedly implicated in CM dedifferentiation and proliferation²⁷). Pathway analysis of MI ProCM showed downregulation of pathways associated with CM structure and upregulation of pathways associated with small GTPases, angiogenesis and cell differentiation, as compared to Donor ProCM (Fig. 5e, top 10 up/downregulated pathways).

To further isolate MI ProCM gene expression, we performed additional comparisons ‘MI ProCM vs MI non-proliferative (nonPro)CM’ and ‘MI nonProCM vs Donor nonProCM’ to remove genes that are differentially expressed solely due to ischemia, rather than uniquely from proliferation (Fig. 5f, Extended Data Fig. 4g-i). This unique MI ProCM gene expression signal (Fig. 5f, red asterisks) showed downregulation of CM proliferation inhibitors CTNNA3²⁸ and SOX6²⁹ and upregulation of TGFBI(paralogue of POSTN, identified in Fig. 4c proteomics dataset and reported to induce CM proliferation¹⁶) and DAB2 (reportedly induces CM dedifferentiation and proliferation²⁷). Proliferation-associated markers POLA1 (DNA polymerase catalytic subunit upregulated in S-phase), NASP (H1-specific histone chaperone) and ATAD2 (reportedly hypoxia-responsive pro-proliferative gene³⁰), were exclusively upregulated in MI ProCMs vs MI nonProCMs (Fig. 5f). Taken together, these data identify the transcriptome of human ProCM following MI.

A safe human biopsy method post-MI shows increased cardiomyocyte mitosis in the peri-infarct border zone

To further demonstrate the regenerative potential of human adult CMs in response to ischemia, we developed a safe method to obtain LV biopsies from patients (<60yo) undergoing coronary artery bypass grafting (CABG) surgery following a recent acute MI (Fig. 6a) (see Supplementary Data for detailed methods and Table S1 clinical characteristics). A time interval of 7-10d post-MI was chosen to capture the window of CM proliferative potential. Using our method, epicardial LV biopsies were taken from the peri-ischemic border zone (‘BZ LV’ hereafter) and an area of healthy myocardium aka remote zone (‘LV RZ’ hereafter) (Fig. 6a). This approach closely approximates the post-MI mouse cardiac regeneration model, whereby BZ LV and RZ LV samples are typically analysed^7,8. As there were no clinical complications throughout the study, we propose this model to be used in future to further investigate early MI remodelling and cardiac regeneration using fresh human myocardium.

To assess whether mitotic CMs could also be identified in LV epicardial biopsies from MI patients, we performed IHC to analyse expression of mitotic markers pHH3 and AURKB. Comparing to Donor LV, both RZ LV and BZ LV regions showed increased pHH3⁺ (0.77% Donor LV, 2.50% RZ LV, 3.50% BZ LV) and AURKB⁺ (0.65% Donor LV, 2.90% RZ LV, 8.71% BZ LV) mitotically active CMs (Fig. 6b-d), albeit lower than observed in earlier MI LV analyses (see Fig. 2). Similar to the MI LV samples, mitosis approximately correlated with each region’s degree of ischemia: Donor LV < RZ LV < BZ LV (figure 6d). These results demonstrate further evidence that adult human CMs increase their proliferative potential in response to ischemia and provides a human post-MI biopsy model for future studies.

Pivotal studies have shown adult human CMs can proliferate at low levels in normal development^6,31,32. However, evidence of the increased CM proliferative potential previously observed following injury in rodent models^7,8 has not been shown in humans. This is relevant given that MI is one of the leading causes of death world-wide. In this study, we demonstrated human adult CMs increase both mitosis and cytokinesis in response to MI. We also characterized the ischemic environment that promoted this intrinsic CM cell division, identifying transcripts, proteins and metabolites previously shown to induce CM proliferation^{14–17,20–23}. Moreover, using snRNAseq analysis, we identified the gene expression of adult human proliferative CMs in response to ischemia. Finally, we developed a novel method to obtain fresh human LV infarct biopsies from MI patients and provided further evidence of increased adult CM proliferation following injury.

To the best of our knowledge, there are two human studies (one of which has been placed under expression of concern by The New England Journal of Medicine) that examined changes in regenerative potential of human CMs at the MI BZ^33,34. Specifically, they showed a relative increase in BZ Ki-67⁺ CMs post-MI. Two significant limitations of their approach were the use of Ki-67, a marker present at all stages of the cell cycle (G1, S, G2, and M-phase³⁵), and the use of samples obtained post-mortem, which are subject to degradation in the histological architecture of the myocardium. In contrast, we identified proliferative CMs using pHH3⁺ and AURKB⁺ markers that peak in expression during G2/M phases^35,36, using pre-mortem samples. We therefore provided evidence of CMs in late-stage mitosis, likely undergoing karyokinesis, using fresher samples that were collected within 15 minutes. However, karyokinesis does not imply subsequent cytokinesis, particularly in the human heart, which is notorious for increasing cardiomyocyte ploidy (e.g. to 16n) and bi-nucleation in response to physiologic stresses⁴. Therefore, to show new CM formation following MI, we provided evidence of cytokinesis using MKLP1 localization between newly divided CMs. This strategy was previously employed by Mollova and colleagues⁶, who showed CMs cease to undergo cytokinesis in healthy humans >20yo. Within this context, our findings in the older infarcted 48yo LV, demonstrated MI enhanced CM proliferation in a previously unreported (healthy) age at hitherto undetected levels.

Historically, the failing heart was once regarded as ‘an engine out of fuel’, defined by its decrease in energy metabolism³⁷. Although this paradigm has shifted in recent years³⁸,myocardial energetics remains an essential tool used to deepen our understanding of heart failure. Recent studies have investigated the relationship between metabolites and CM proliferation^21–23. These groups identified that the mevalonate pathway, malonate and ketogenesis, may potentially increase CM proliferation. In these studies, the mevalonate pathway was identified as pro-proliferative via an unbiased screening of small molecule compounds within human embryonic stem cell-derived CMs²¹. In contrast, malonate and ketogenesis were specifically targeted using mouse models^22,23. Within our study, mevalonate, malonate and 5 metabolites involved in ketogenesis, were upregulated. Whilst the potency of each metabolite to induce CM proliferation is unknown, the upregulation of 5 ketogenic metabolites may highlight its importance in the pro-proliferative ischemic setting. However, it should be noted that our metabolomic approach did not include detection of metabolites downstream of mevalonate i.e. mevalonate-5-phosphate, mevalonate pyrophosphate, isopentyl pyrophosphate and dimethylallyl pyrophosphate. Future studies should therefore include additional analysis of these metabolites to determine the full effect of ischemia on the mevalonate pathway.

Mouse models have shown that adult CM dedifferentiation can enable cell cycle initiation, which can be targeted to induce CM proliferation^39–43. Some of these models also demonstrated that proliferating CMs downregulate markers of CM architectural organization, such as sarcomeric structures, and upregulate dedifferentiation markers, such as Runx1 and Dab2^39,43. Our results provide further support for these findings, as we showed ProCMs downregulated CM structural organization pathways such as ‘sarcomere organization’ and ‘muscle organ development’ and upregulated the dedifferentiation markers RUNX1 and DAB2. ProCM also showed upregulated pathways termed ‘cell differentiation’, ‘cell adhesion’, ‘cell migration’ and two small GTPase signaling-related pathways; pathways related to cell morphology and cell-ECM interactions within the surrounding microenvironment. Thus, our data is suggestive that ProCMs undergo transition to a non-typical CM cell state. Interestingly, a recent study similarly described a CM cell state transition within the BZ of human MI⁴⁴ but did not specifically identify or characterize ProCMs.

Recently, Kuppe and colleagues mapped the spatial transcriptomic landscape of myocardial infarction using multi-omics approaches⁴⁴, thereby providing the field an invaluable resource. In agreement with this study, our snRNAseq data showed increased total proliferating cells and decreased CMs in the infarcted sample. Whilst Kuppe and colleagues did not focus on ProCMs, they did identify subpopulations of BZ CMs in a ‘pre-stressed’ stage of CM differentiation, expressing downregulated MYBPC3 expression⁴⁴. This gene reportedly regulates postnatal cardiomyocyte cytokinesis⁴⁵. Although our data did identify downregulation of MYBPC3 in MI ProCM vs Donor ProCM, it was also identified in MI nonProCM vs Donor nonProCM. Thus, in our dataset, MYBPC3 downregulation is considered to be an ischemic, rather than solely proliferative signal, but its role in CM proliferation may still be of importance.

Angiogenesis commonly occurs during hypoxic events. Its function is essential during both heart development and at the post-MI BZ to replace coronary vessels damaged during ischemia⁴⁶. In our study, we showed angiogenesis was upregulated in our infarcted LV using IHC, IMC and proteomics. Further to this, our snRNAseq data showed ProCMs upregulated angiogenic genes (e.g. EPAS1 aka HIF2A, POSTN) and angiogenic pathways. Although CMs are known to stimulate angiogenesis after ischemia⁴⁶, to our knowledge we are the first to show this phenomenon in human ProCMs. Whether this effect is a product of their proximity to an ischemic zone, or an indication ProCMs are signaling for additional blood supply to complete mitosis, is unknown, but this warrants further investigation.

One limitation of our study is the sample size, which is inherent given the relative infrequency of organ donation. Moreover, nearly all healthy donor hearts today are successfully transplanted into heart failure recipients, due to advances in transplant management and accessibility to transplant centers. Pre-mortem human heart tissue is therefore precious, with both MI (non-end-stage heart failure) and donor hearts rarely available for research. For the unique pre-mortem infarcted heart used in our study, the following prerequisites were required: 1) the next of kin consent had been undertaken as part of a national donor program, 2) the patient suffered an out of hospital cardiac arrest secondary to MI but did not display neurological activity compatible with life whilst on life support for 5 days, 4) the next of kin gave consent for organ donation for research (given the heart could not be transplanted), and, 5) the whole heart was not collected post-mortem. As this number of simultaneous events were required to acquire the samples, it is unlikely such a premortem ischemic heart will be available again.

Similar to the infarcted heart, early mammalian cardiac development occurs in a low-oxygen environment⁴⁷, with altered cardiac extracellular matrix (ECM) expression^16,17,20. Our data supports these observations, showing global (bulk RNAseq and proteomics) and localized (snRNAseq) ECM modulation. This included ECM transcripts and proteins known to be upregulated during embryonic development and known to induce CM proliferation inmouse models (e.g., POSTN, AGRN, TGFBI, DCN^16,17,20). Though these findings were the result of comprehensive characterization of a unique ischemic heart, it is unlikely similar tissue will be collected again. As such, we suggest our human LV heart failure biopsy method as a highly suitable candidate model for future studies.

In conclusion, our study shows that the infarcted human heart increases its regenerative potential in response to ischemia. However, although this intrinsic regeneration is not sufficient to replace the CMs lost following MI, studies aiming to therapeutically enhance CM proliferation to reverse heart failure are underway^15,21,22. Developing means to enhance this intrinsic CM regenerative potential could lead to novel therapeutics for cardiac injury and heart failure.

Note: all incubations were at room temperature unless stated otherwise.

Human Heart tissue

Donor hearts deemed suitable for heart transplantation but unable to be transplanted (reasons including transportation logistics, immune incompatibility, and donor-recipient mismatch in size) were procured as previously described⁴⁸. These are not post-mortem samples. The non-diseased donor heart samples were from patients with a non-cardiac cause of death and no significant co-morbidities or cardiac risk factors (hence why the hearts would otherwise have been used for transplant). All hearts (non-diseased donor or ischemic) underwent formal pathological examination by clinical anatomical pathology to confirm either normal histological architecture or ischemic heart disease respectively. LV samples were obtained immediately after harvest and snap frozen in liquid nitrogen (-196 °C). The study was approved by the Human Ethics Committee of The University of Sydney (USYD # 2021/122).

Procurement of a unique infarct human heart

The unique infarcted human heart was acquired from a 48yo male patient, originally on the national donor list, who suffered a catastrophic acute myocardial infarction (MI) due to a blockage of the left anterior descending artery. This patient suffered a cardiac arrest and at the Royal Prince Alfred Hospital, Sydney, whereby a coronary angioplasty and stent insertion could not be achieved. The patient was then kept on life support for 5d post-MI, at which point no neurological activity was detected and the patient was declared braindead. As the heart could not be used for transplant purposes, the next of kin kindly agreed to organ donation for research. The heart was then collected pre-mortem and cryopreserved within 15mins i.e. fragments from the right atrium (MI RA), right ventricle (MI RV) and left ventricle (MI LV) were flash frozen and stored in liquid nitrogen, or fixed in 10% neutral buffered formalin (NBF, Sigma HTS012) overnight, before use.

Histology – Picrosirius Red and Verhoeff van Gieson staining

Heart fragments fixed in NBF were washed 3x5min in phosphate buffered saline (PBS), before moving into 70% ethanol. Samples were then embedded in paraffin blocks by infiltrating with paraffin wax overnight in a tissue processor. Paraffin sections (4µm) were cut using a microtome and dried in an oven at 60°C for 2h. Slides were then deparaffinized in xylene (3x5min) and rehydrated using an ethanol:ddH₂O gradient (1x5min: 100%, 100%, 100%, 70% ethanol), followed by ddH₂O.

For picrosirius red staining, slides were rinsed in ddH₂O, incubated Harris haematoxylin (10min), washed in running water, stained with picrosirius solution (1ug/mL Sirius red F3B in saturated aqueous picric acid, PST #C102) for 60min, washed in acidified ddH₂O (1µL/mL HCl), dehydrated 3x2min in 100% ETOH, mounted on glass slides in permanent mounting media (#FNNII065C, Froninecat) and glass cover-slipped. For Verhoeff van Gieson (VvG) staining, sections were stained for 25min using Verhoeff’s solutions A, B, and C (VVGA/B/C/250, POCD) at a 2.5:1:1 ratio, respectively. Sections were rinsed in running water, followed by ferric chloride solution for 5sec (POCD FECL2%/500), rinsed in water, 1min 100% ethanol, 1min de-ionized water and counterstained for 2.5min in Van Gieson’s solution (POCD VANG/1L). Filter paper was used to dry slides, which were then dehydrated in 100% ethanol twice (10 dips each), 3x1min xylene and mounted in permanent mounting media with a glass coverslip. All slides were then imaged on a Nikon NiE widefield microscope at 20x magnification (0.8NA, Nikon Qi2 monochrome camera, NIS-Elements software).

CUBIC clearing/immunohistochemistry (IHC)

Tissue was cleared using a modification of our previous modified CUBIC protocol⁴⁹. Briefly, ~3mm³ fragments were fixed overnight in 10%NBF at 4°C, immersed in CUBIC Reagent 1A at 37°C 3x24h (replacing daily), washed 3x1h in PBS, permeabilized in DeepLabel Solution A (DeepLabel Antibody Staining Kit, Logos, #C33001) for 1h and blocked in 5% normal donkey serum-5% acetylated bovine serum albumin-0.5% TritonX-PBS at 4°C overnight. Tissue was then immunostained with primary antibodies anti-mouse αActinin (1:25, Sigma-Aldrich #A7732), anti-rabbit pHH3 (1:50, Abcam #ab219407), anti-rabbit Aurora Kinase B (1:50, Sigma #SAB5600088) or anti-rabbit MKLP1 (1:50, Abcam #ab174304) diluted in DeepLabel Solution B for 72h on a shaker. Tissue was washed 3x30min with DeepLabel Washing Buffer and incubated in secondary antibodies donkey anti-mouse AlexFluor488 (1:400, Abcam #ab150109) and donkey anti-rabbit AlexaFluor647 (1:400, Abcam #ab150063) in DeepLabel Solution B for 2 days on a shaker protected from light. Tissue was then washed 3x5mins in DeepLabel Washing Buffer, and incubated in DAPI (2µg/mL, Thermo #62248) with wheat germ agglutinin (WGA)-AlexaFluor555 (5µg/mL, ThermoFisher #W32464) in PBS for 1h and washed 3x1h DeepLabel Washing Buffer. Tissue was then optically cleared in X-Clarity (DeepLabel Antibody Staining Kit) at 37°C for 1h, then replaced with fresh X-Clarity before imaging. Tissue immunostained with anti-mouse αActinin and donkey anti-mouse AlexFluor488 was imaged using both confocal and two-photon microscopy (see ‘two-photon second harmonic generation’ below). All other staining combinations were imaged using confocal only (see ‘Confocal Imaging’ section).

Two-photon second harmonic generation (SHG) microscopy

CUBIC cleared and IHC stained tissue (αActinin) was placed in a cover glass-bottomed Petri dish (Ibidi #81158) with 200µL of X-Clarity and imaged on a Leica TCS SP8 STED 3X microscope using 10x or 100x (HC PL APO CS2, 0.4 and 1.4 NA respectively) objectives. αActinin (AlexaFluor488) was imaged using a tuned (490nm^ex/500nm-600nm^em) white light laser z-stack (3.33µm slices, 60µm stack) with a hybrid (photomultiplier (PMT) and avalanche photodiodes) detector (HyD), followed by imaging the same z-stack using a tuned (880nm^ex/420-460nm^em) two-photon laser with an external transmitted light detector to detect SHG and combining the images using FIJI.

Immunohistochemistry (IHC) for paraffin sections

Paraffin embedded sections (see ‘Histology’ methods) were dewaxed in xylene (3x5min), rehydrated (ethanol: 100%, 100%, 95%, 70%, 50%, 30%, ddH₂O; all 1x5mins), antigen retrieved (pH 9.0, tris-EDTA, 95^oC, 30 mins), washed ddH₂O (5mins), washed PBS (5 mins), permeabilized (0.5% TritonX-PBS, 5 mins), blocked in blocking buffer (5% normal donkey serum, 5% acetylated bovine serum albumin, 0.5% TritonX) and stained with primary antibodies anti-mouse α-smooth muscle actin (1:400, Abcam #ab7817) and anti-rabbit CD31 (1:500, Abcam #ab281583), or anti-mouse α-Actinin and anti-rabbit CD45 (1:100, Cell Signaling #13917S), and incubated overnight at 4°C on a shaker. Sections was washed in PBS (3x5min) and stained with secondary antibodies donkey anti-mouse AlexFluor488 (1:400) and donkey anti-rabbit AlexaFluor647 (1:400) for 1h on a shaker protected from light. Sections were washed in PBS (3x5min), nuclei stained with DAPI (1µg/mL), washed in PBS (3x5min), mounted in Prolong Diamond Antifade (ThermoFisher #P36961) and coverslipped. Slides were covered from light and left to cure overnight before imaging on a Nikon NiE widefield microscope at 20x magnification (0.8NA, Nikon Qi2 monochrome camera, NIS-Elements software).

Imaging mass cytometry (IMC) immunostaining

All IMC methods were modified to suit paraffin embedded tissue from our previous frozen section immunostaining protocol¹⁰. Briefly, 7µm sections were cut from paraffin embedded tissue blocks. Sections were then dewaxed, rehydrated, underwent antigen retrieval and permeabilized as described in the above ‘IHC for paraffin sections’ methods section. Tissue was then blocked in Opal Antibody Diluent/Block (Akoya Biosciences, #AKOARD1001EA) for 1h at room temperature, followed by 1h incubation with 1:300 anti-col1α1 (#91144S, Cell Signaling) conjugated to Cy5 using a sulfo-Cyanine5 antibody labelling kit (#3321-10rxn, Lumiprobe) according to the manufacturer’s instructions, hereafter referred to as anti-col1α1-Cy5. Sections were then washed in PBS (3x5mins), followed by overnight incubation in 75µL of IMC metal conjugated antibody cocktail per section (antibodies conjugated using the Maxpar® X8 Multimetal Labeling Kit #201300 Fluidigm kit according to the manufacturer’s instructions). IMC antibody cocktail metals* and clones** consisted of: αSMA-89Y* (1A4**, 1µg/mL), vimentin-113In* (D21H3**, 2µg/mL), CD68-153Eu* (KP1**, 8µg/mL), CD45-154Sm* (D9M8I**, 10µg/mL), CD31-155Gd* (EPR3094**, 6µg/mL), CD14-160Gd* (EPR3653**, 6µg/mL), FXIIIa-161Dy* (polyclonal**, 0.5µg/mL), anti-Cy5-164Dy* (CY5-15**, 4µg/mL), CD66a-167Er* (YTH71.3**, 4µg/mL), CCR1-169Tm* (polyclonal**, 8µg/mL), IRF8-171Yb* (polyclonal**, 4µg/mL) , CD206-172Yb* (685645**, 4µg/mL), HLA-DR-174Yb* (EPR3692**, 1.5µg/mL), αActinin-176Yb* (EP2529Y**, 1.5µg/mL), HH3-209Bi* (D1H2**, 16µg/mL). Sections were washed in PBS (3x5min), followed by incubation in 4% paraformaldehyde (20 mins) and washed in PBS (3x5min). Iridium DNA intercalator diluted in PBS (1:300) was added, incubated for 30 mins, washed in PBS (3x5min), dipped 3x in ultra-pure H₂O 3 times, air dried, then stored in an air tight container before imaging on an IMC with a Hyperion Imaging System (Fluidigm). For data acquisition, the IMC was initially tuned ensuring the resolution (mass 1) >400 and transients cross talk (1) 700. Using a combination of a panoramic bright field image taken on the Hyperion and adjacent picrosirius red stained slides to identify scar regions, region of interest (ROI) were selected in the infarct and remote zones for each section. An energy test was then performed, determining the optimal ablation energy to be 1 dB. ROIs were then ablated and data acquired for downstream analyses. For individual channel visualization, Histocat++ software was used and channels were pseudo-coloured.

IMC multicut cell segmentation

For multicut cell segmentation, individual channels from ROIs within MCD files were converted to a tif file format using the Histocat++ software. Raw tif files were then processed using the RStudio package Spectre and converted into HDF5 files. Ilastik machine learning 12 software used HDF5 files for pixel classification of CM cell borders and nuclei, non-CM cell borders and nuclei and background areas to create probability files. Next boundary-based segmentation with multicut was performed on Ilastik using ROI HDF5 files and their associated probability files using a 0.5 pre-smooth before seeds threshold for water shedding. Training and multicut was then performed on every ROI in the dataset to produce associated multicut segmentation files. Next object classification was performed on Ilastik to identify cell and non-cell objects using ROI and multicut segmentation HDF5 files. Both object predictions and object identities were exported as tif files for each ROI. Ilastik pixel classification was again performed with ROI HDF5 files for region classification of scar and non-scar (myocardium) areas and exported as simple segmentation tif files.

IMC analysis

All cellular analysis was done using the RStudio package Spectre⁵⁰ unless otherwise stated. First, tif files were read into Spectre and spatial objects were created. Next cell mask (object identities), cell type (object predicitons) and region (simple segmentation) .tif files from Ilastik multicut cell segmentation were read into Spectre, followed by polygons and outlines being generated. Initial mask quality control plots were generated to ensure that each mask for each ROI was suitable. Single cell expression plots were then generated for each ROI for each marker, the data extracted for each cell, annotated as ‘cell’ or ‘non-cell’ to identify segmented areas without cells and the area of different regions calculated. Data was then saved as qs, csv and fcs files. For cellular analysis previously generated qs and csv files were read into Spectre on RStudio. An inverse hyperbolic sine function (ArcSinh) was then applied to expression data before rescaling. Non-cells, background and any objects <5µm2 were filtered out of analysis. Then clustering and dimensionality reduction were performed using 20 metaclusters and a perplexity of 200. tSNE plots and an expression heatmap were then plotted using Spectre. Clusters were then annotated according to the heatmap expression levels for each cluster. Data was then saved as csv files and plotted using GraphPad Prism 8.

Confocal imaging - CUBIC cleared IHC stained tissue fragments (mitosis and cytokinesis positive cardiomyocytes)

CUBIC cleared IHC stained tissue was imaged in glass bottomed petri dishes on a Leica TCS SP8 STED 3X microscope in confocal mode. Fluorophores were imaged using an ultraviolet (UV) laser for DAPI and a tunable white light laser for other fluorophores with the following settings: DAPI (UV, 405nm^ex/415nm-485nm^em, PMT detector), AlexaFluor488 (490nm^ex/500nm-550nm^emHyD detector), WGA555 (553nm^ex/563nm-640nm^em HyD detector) and AlexaFluor647 (650nm^ex/660nm-640nm^em HyD detector). For analyses, five 94µm z-stacks (2µm slices) were acquired using a 25x objective (HC Fluotar VISIR, 0.95NA) from each sample. pHH3 and AURKB positive CMs were analyzed using a custom written FIJI script with the BinaryReconstruct plugin, that counted the number of CMs (DAPI⁺αActinin⁺) that were pHH3⁺ or AURKB⁺. For MKLP1 analyses, Huygens Professional v23.04 software object analyzer was used to calculate the number of cells (DAPI objects >20µm size with watershed segmentation). FIJI was then used to manually count the number of CMs undergoing cytokinesis. Manual counts were required to accurately identify MKLP1 localized at the cleavage furrow between two dividing αActinin⁺ CMs. For high resolution images of mitosis/cytokinesis events, sampling rates of less than x=50nm, y=50nm, z=160nm were determined using Nyquist calculations and 100x magnification (HC PL APO CS2, 1.4NA) images with these settings were acquired, followed by deconvolution using Huygens Professional v23.04 software.

Bulk ribonucleic acid sequencing (RNAseq)

Approximately 50mg of tissue was disrupted using a metal bead (Qiagen #69989) in a 2mL tubes with 500uL TRIzol (Ambion #15596018) using the TissueLyser LT (Qiagen) and RNA extraction performed according to TRIzol manufacturer’s instructions. DNA removal was performed using Qiagen RNAse-Free DNase kit (Qiagen #79254) and purified using the RNeasy Mini kit (Qiagen, #74104) according to manufacturer’s instructions. RNA quality was evaluated using a Nanodrop and an RNA Nano Chip (Agilent #5067-1511) on an Agilent Bioanalyzer. Libraries were prepared with Illumina Stranded Total RNA prep Ligation with Ribo Zero Plus kit, according to manufacturer’s instructions. Libraries were sequenced using a paired-end 250bp Illumina NovaSeq 6000 S4 flow cell. Library preparation and sequencing were performed by the Ramaciotti Centre (Sydney, UNSW). FASTQ files were aligned to the human genome (hg38) using Rsubread⁵¹. Mitochondrial genes, rRNA, pseudogenes, unknown genes, and genes without valid symbols were removed. Lowly expressed genes were filtered by the filterByExpr function in edgeR⁵² with default parameters. Trimmed mean of M values (TMM) normalisation⁵³ was then applied. Downstream differential gene expression analyses were performed using limma⁵⁴ and edgeR packages. Voom⁵⁵ was used to fit linear models to compare gene expression across different sampling sites in the MI heart. Multidimensional scaling (MDS) plots of distances were computed by plotMDS function in limma and visualised by ggplot2. P-values are adjusted by the BH method to account for multiple comparisons in the DE analysis. The RStudio package EnhancedVolcano was used to visualize differential expression analyses. Pathway analyses were conducted using Reactome analysis with Reactfoam visualization⁵⁶ and DAVID⁵⁷ analysis with ggplot2 RStudio visualization.

Proteomics

Approximately 10mg of powdered frozen heart tissue was homogenized in 4% sodium deoxycholate (Sigma #D6750) and 100mM Tris-HCl pH 7.5. Samples were denatured at 95^oC, sonicated using QSonica R2 (QSonica) at 70% amplitude, centrifuged (18,000g) and supernatant collected. Next, a BCA assay (ThermoFisher #23225) was used to determine protein concentration and 20ug of protein was digested using trypsin overnight at 37^oC before mass spectrometry preparation as described previously⁵⁸. Peptides were injected onto a 30cm x 70 um C18 (Dr. Maisch, Ammerbuch, Germany, 1.9 µm particle size) fused silica analytical column with a 10 µm pulled tip, coupled online to a nanospray ESI source. To resolve peptides, samples were run over a gradient from 5% - 40% acetonitrile for 120 min (flow rate 300 nL/min). Peptide electrospray ionization was performed at 2.3 kV. Next, MS/MS analysis was performed using a Q-Exactive Fusion Lumos mass spectrometer (ThermoFisher) (27% normalized HCD collision energy for fragmentation). Spectra were attained in a Data-Independent Acquisition (DIA) using 20 variable isolation windows. Statistics analyses on both proteomics and RNA-seq data were performed in RStudio (v4.2.2).

In proteomics data, proteins that were missing in more than 80% samples were filtered out. Protein intensities were normalized by cyclic loess implemented by the normalizeBetweenArrays function in the limma⁵⁴ R package. Differential expression (DE) analysis was performed using lmFit and eBayes functions in limma to compare protein expression levels across different sampling sites in the MI heart. The Benjamini-Hochberg (BH) method was used to calculate adjusted P-values to account for multiple comparisons in the DE analysis. The RStudio package EnhancedVolcano was used to visualize differential expression analyses. Pathway analyses were conducted using Reactome analysis with Reactfoam visualization⁵⁶ and DAVID⁵⁷ analysis with ggplot2 RStudio visualization.

Metabolomics

Metabolomic protocols were modified from our previously published protocols^59,60. 50mg of tissue was lysed in a 2mL round bottomed tube with 6µL per 1mg tissue of extraction medium (MeOH:CHCl₃, 2:1, v:v, ThermoFisher #FSBA456-4, Sigma #650498) and a metal bead, using a TissueLyser LT for 4x50cycles cooling on ice between cycles. After removing metal beads, samples were incubated on ice for 30mins. 2µL CHCl₃ per 1mg tissue was added and vortexed for 1 min, followed by 2µL of HPLC H₂O (ThermoFisher, #FSBW6-4) per 1mg tissue with 1min vortexing. Samples were centrifuged at 14000rpm for 20min at 4°C and the aqueous layer transferred into a new microfuge tube. Next 20µL of aqueous layer should be added to 80µL of HILIC IS-SS (10mM L-Phenylalanine-d8 (cambridge Isotope Lab #DLM-372) and 10mM L-Valine-d8 (Sigma #486027) in Acetonitrile:MeOH 75:25 (Fisher Scientific, Optima LC-MS grade, #A955-4 and #A456-4), v:v) for HILIC analysis or 40 aqueous with 60µL of AMIDE IS-SS (10mM L-Phenylalanine-d8, 10mM thymine-d4 (Sigma #487066), 10mM citrate-d4 (Sigma # 485438) in Acetonitrile:MeOH 75:25, v:v) for AMIDE analysis, vortexed briefly then spun at 14000rpm for 20mins at 4°C. The surface 90µL was transferred into glass vials with glass inserts (Waters #186000273, #WAT094179) for LC-MS/MS metabolite analysis.

Metabolite analysis by LC-MS/MS

A Qtrap mass spectrometer (QTRAP 6500⁺ Plus, AB Sciex, CA, U.S.A) connected with a UHPLC system (Nexera 30 series, Shimadzu Corp, Kyoto, Japan) was operated either in positive mode coupled with a HILIC column (Atlantis Silica HILIC Column, 100Å, 3µm, 2.1mm X 150mm) or negative mode with an Amide column (XBridge BEH Amide Column, 130Å, 3.5µm, 2.1mm X 100mm) via electrospray ionization (ESI) to measure in total of 205 metabolites on multiple reaction monitoring (MRM) mode (reference 1&2). The mobile phases for HILIC column comprise A (0.1% formic acid, 10mM ammonium formate in water) and B (0.1% formic acid in acetonitrile). The gradient started at 95% B and decreased to 40% from 4 min to 9.5 min and maintained to 40% till 12.5min, before increased to 95% at 13.5min with flow rate of 0.25mL/min. The column was re-equilibrated for another 10 min at 95% B at flow rate of 0.4 mL/min before next injection. The autosampler temperature was 40°C, and injection volume was 5μL. On the AMIDE column, the mobile phase A is consisted of 20 mM ammonium acetate, 20 mM ammonium hydroxide in 95:5 water: acetonitrile, pH 9.5, and the mobile phase B is acetonitrile. The LC gradient was started with 85% B and decreased to 35% at 8 min and continued to be decreased to 2% B at 9 min; remained at 2% for another min till 10 min before returning to 85% B at 11 min at flow rate of 0.25 mL/min, and then increased the flow rate to 0.5 mL/min at 85% B for re-equilibration of further 4 min before next injection. Pooled plasma samples were extracted and analyzed in every 10 samples for quality control. All the metabolites detected were analyzed using Sciex OS (version 3.0, AB Sciex, CA, U.S.A) and metabolites were confirmed by MRM transitions and retention times match to the original authentic standards. Relative metabolite abundances were log₂ transformed, and normalised using EigenMS⁶¹. The R programming language (V4.2.0), and limma package were used to construct linear regression models for each metabolite, with indicator variables for tissue type included as covariates. An empirical Bayes estimation method was used for the calculation of moderated t-statistics for model coefficients. False discovery rate corrections were applied to all models to control for multiple comparisons using the Benjamini-Hochberg procedure. Differentially abundant metabolites were identified using an alpha of 0.05. Visualization of the data was implemented using the EnhancedVolcano RStudio package.

Single nucleus (sn)RNAseq sample preparation

Tissue samples were dissociated into single nuclei using 10x Chromium Nuclei Isolation Kit (10x Genomics, #PN-1000493) protocol for Single Cell Gene Expression and Chromium Fixed RNA Profiling. Samples were then loaded into a Chromium Next GEM Chip G (10X Genomics, #PN-1000127) for an output of 10,000 nuclei per sample. The library preparation was done according to Chromium Next GEM Single Cell 3’ Reagent Kits v3.1 Dual Index protocol. Quality control was done with Agilent 4200 TapeStation system. The cDNA libraries were sequenced on an Illumina (NovaSeq 6000, S1 100 cycle Flowcell) instrument as a depth of 200M reads/sample (or 20,000 reads/cell). Sequencing was carried out at the Rammaciotti Centre for Genomics, UNSW Sydney.

snRNAseq analysis

Single nuclei data processing: We first checked quality of reads using FASTQC program. Then we mapped reads to NCBI human reference genome (hg 38) and extracted read counts using CellRanger count program (version 7.0.1) with default options. We corrected the read count matrix for potential contamination from ambient RNAs by running CellBender remove-background, with ‘expected-cells’ and ‘total-droplets-included’ options set by cellranger output, –fpr=0.01 and –epoch2=150⁶². We removed cells with (i) the number of sequenced modules less then 2,000 or more than 50,000, (ii) the number of expressed genes less than 2,000, (iii) more than 25% of reads mapped to mitochondrial genes, or (iv) more than 40% of reads mapped to ribosomal genes.Cell clustering:We used Seurat’s FindClusters function (version 4.3.0) with the resolution of 0.8⁶³. Using multiple marker gene sets, we performed gene set enrichment analysis using UCell R package with default options⁶⁴. Subsequently, we visualised and merged cell clusters with similar enrichment patterns, yielding 11 cell clusters including cardiomyocytes. Similarly, we performed enrichment of cell cycling genes to identify proliferating cells as previously described^65,66. We labelled cells in the S phase from this analysis to be proliferative. Proliferative CMs:Proliferative cells in the CM cell cluster were labelled as proliferative CMs. To compare enrichment of proliferative CMs in IHD versus donor control group, we calculated a ratio of proliferative CMs in each treatment group divided by the total number of CMs. For comparison of CMs or proliferative cells between groups, ratios were calculated by dividing the cell count by the total cell count of the group (i.e. donor or MI). Fisher’s exact test was used to test significance of the enrichment (two-tailed).Differential gene expression analysis:We used Seurat’s FindMarkers function with logfc.threshold=0 option to find differentially expressed genes. We performed three comparisons; 1. Proliferative CMs in IHD vs proliferative CMs in donor, 2. Proliferative CMs in IHD vs non-proliferative CMs in IHD, and 3. Non-proliferative CMs in IHD vs non-proliferative CMs in donor. Differentially expressed genes were defined as ones with FDR < 0.05 and average log2-fold change > 0.5 in the first group. For functional annotation of DE genes, we used DAVID gene ontology web-server database with the default options⁶⁵. Biological process GO terms were ranked by statistical significance for the visualisation.

Procuring post-MI LV biopsies from MI patients undergoing coronary artery bypass graft surgery

Consultants from the cardiothoracic team consented appropriate patients for this study. Regional hospital HREC ethics approval was obtained (#X14-039). Broad inclusion criteria were patients (male or female, age > 20 years old) admitted to our hospital with a myocardial infarction (MI), where subsequent coronary angiography demonstrated coronary artery disease that required inpatient CABG. It was at this time, after they were consented for surgery, that recruitment took place. The participants were medically stable and demonstrated full capacity to understand and agree to the consent process. It is not unusual for such patients to wait 3-14 days as an inpatient for their surgery. This reflects their medical stability. Patients, who were hemodynamically unstable, or unable to understand the consent process, were excluded from the recruitment process.

Epicardial sampling has been described previously as a diagnostic tool for inflammatory or infectious myocardial diseases⁶⁷. Epicardial biopsies were obtained from patients undergoing CABG following admission for myocardial infarction. Two biopsies were taken; one from the peri-infarct border zone subtending the culprit lesion, and the other, from an area of normal (non-ischemic) remote zone left ventricle. Authors SL and PB determined the location for each biopsy prior to and at the time of the operation. Specifically, author SL and an independent cardiologist (not involved in the study) reviewed the clinical history, ECGs, serum chemistry, echocardiograms and coronary angiograms. The culprit lesion was determined by several factors: localizing ST elevation on ECG if present; very tight (>95%) stenosis or subtotal occlusion in the presence of fresh thrombus on coronary angiography; in addition to corresponding regional wall motion abnormalities on echocardiography undertaken during the peri-infarct period. Author SL performed a transesophageal echocardiogram (TEE) at the time of CABG to confirm the presence of the regional wall motion abnormalities and to identify normal (contracting) left ventricle, which was then independently verified by a second study by the duty cardiac anesthetist. Authors SL and PB directly inspected the surgically exposed heart during CABG to confirm both the border zone subtending the culprit lesion and the remote zone left ventricle.

During CABG it is usual practice to incise the epicardium (the outer part of the heart that contains the coronary arteries) in order to expose the native coronary artery and perform the anastomosis of the coronary artery bypass graft. Participants in this study consented to the collection of two thin strips of epicardium (approximately 2-3 mm wide) on either side of the culprit coronary artery, which could then be incorporated into the bypass graft anastomosis. That is, two strips of tissue were excised along an existing incision that was repaired as part of the anastomosis. With respect to the normal left ventricular remote zone epicardial biopsy, a thin strip of epicardium (approximately 2-3 mm wide) was cut in parallel with the myocardial fiber orientation and repaired with a single suture or cauterization. The procedure was performed under direct visualization allowing easy assessment of the suture repair of the biopsy site. As such, the bleeding risk was minimal. The thickness of each biopsy was less than 2-3 mm (the width of the biopsy forceps), far less than the approximate 15 mm thickness of the left ventricular wall. Nonetheless, the epicardial surface was visualized to ensure that there was no injury to the muscle. No complications arose from this study.

IHC of LV biopsies obtained from CABG surgery

CABG biopsies were fixed and subsequently stored in 4% formaldehyde (Sigma-Aldrich, #100496) at 4°C following procurement. Samples underwent sucrose cryo-protection, initially washed in PBS for 5mins before being placed in 15% (w/v) sucrose solution (Sigma-Aldrich #S039, in PBS, filtered at 0.4μm) at 4°C for 2d, followed by 30% (w/v) sucrose solution at 4°C for 5d. After placement in OCT all tissue were cryo-sectioned to 8μm at -20°C and placed on VWR® Superfrost® Plus charged glass slides. Slides were stored at -20°C. Slides were equilibrated to room temperature (RT) for 5mins prior to staining. Sections were outlined with an Immedge™ Hydrophobic Barrier PAP pen. Sections were washed 3x5min in PBS, permeabilized in 0.5% Triton-X 100 in PBS for 15mins, followed by blocking in 20% normal goat serum with 0.2% Tween-20 in PBS for 1h. Sections were then incubated with primary antibodies mouse anti-αActinin (1:200) and rabbit anti-pHH3 (1:200) or rabbit anti-AURKB (1:200) in 5% normal goat serum with 0.2% Tween-20-PBS overnight in a humid chamber at 4°C. Primary antibodies were removed by washing 3x5min in 0.1% NP-40-PBS and incubated with secondary antibodies goat anti-mouse AlexaFluor594 (1:500, Abcam #ab150116) and goat anti-rabbit AlexaFluor488 (1:500, Abcam #ab150077) in 5% normal goat serum with 0.2% Tween-20-PBS in a dark humidity chamber for 1h 45mins. Secondary antibodies were removed by washing 3x5mins in 0.1% NP-40-PBS and incubated in DAPI (1µg/mL) for 10 min. Slides were washed 5mins in dH2O before mounting with 90% glycerol in PBS. A Zeiss Axioimager M2 upright microscope was used to image CABG biopsies. Images were manually analyzed (blinded) to count total cardiomyocyte nuclei and cardiomyocyte nuclei positive for either H3P or AKB. For the normal LV samples, n= 20 sections were counted with random areas on each section selected, likewise, n = 38 sections were counted for the peri-ischemic territory, with the same method. This resulted in > 1000 individual CM nuclei being counted for each group. A second person then randomly selected approximately 20% of the sections for each group and performed counts to validate the percentage differentials between CM positive nuclei.

Data Availability

Upon final submission for publication all data sets will be made available in publicly accessible repositories. Raw proteomics data will be uploaded to EBI PRIDE.

Acknowledgments

We thank the patients and staff of Royal Prince Alfred Hospital and St. Vincent’s Hospital, Sydney, and we thank the Australian Red Cross Blood Service. This work was supported by a grant from the R.T. Hall Trust and the support of the Baird Institute for Applied Heart and Lung Research. N.P acknowledges the following grants supporting this work: Heart Foundation grant 106721 and MRFF grant APP2007625. We acknowledge the CMRI Single Cell Analytics, Advanced Microscopy Centre and the Australian Cancer Research Foundation Telomere Analysis Centre for supporting this work. The authors acknowledge the technical and scientific assistance of Sydney Microscopy & Microanalysis, the University of Sydney node of Microscopy Australia. The authors acknowledge the Sydney Cytometry Core Research Facility, a joint initiative of Centenary Institute and the University of Sydney, for assistance with imaging mass cytometry.

Author contributions

RDH conducted experiments, contributed to experimental design and wrote the manuscript; JW conducted experiments and wrote the manuscript; WJS, ML and ES conducted bioinformatic analysis; CM, DH, XW, LN, LC and ML conducted experiments; JE, FR, AF, JJHC, NP and JOS contributed to experimental design; CH and AG performed analyses; PB contributed to experimental design and conducted experiments; SL conceived the project, designed experiments, conducted experiments and wrote the manuscript.

Disclosures

Nil

Laflamme, M. A. & Murry, C. E. Heart regeneration. Nature 473, 326–335 (2011).
Senyo, S. E., Lee, R. T. & Kühn, B. Cardiac regeneration based on mechanisms of cardiomyocyte proliferation and differentiation. Stem Cell Res 13, 532–541 (2014).
Ponikowski, P. et al. 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC)Developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J 37, 2129–2200m (2016).
Boström, P. & Frisén, J. New cells in old hearts. N Engl J Med 368, 1358–1360 (2013).
Porrello, E. R. et al. Transient regenerative potential of the neonatal mouse heart. Science 331, 1078–1080 (2011).
Mollova, M. et al. Cardiomyocyte proliferation contributes to heart growth in young humans. Proc Natl Acad Sci U S A 110, 1446–1451 (2013).
Senyo, S. E. et al. Mammalian heart renewal by pre-existing cardiomyocytes. Nature 493, 433–436 (2013).
Wang, W. E. et al. Dedifferentiation, Proliferation, and Redifferentiation of Adult Mammalian Cardiomyocytes After Ischemic Injury. Circulation 136, 834–848 (2017).
Chien, K. R. et al. Regenerating the field of cardiovascular cell therapy. Nat Biotechnol 37, 232–237 (2019).
Hume, R. D. et al. PDGF-AB Reduces Myofibroblast Differentiation Without Increasing Proliferation After Myocardial Infarction. JACC Basic Transl Sci (2023) doi:10.1016/J.JACBTS.2022.11.006/SUPPL_FILE/MMC1.PDF.
Hume, R. D. & Chong, J. J. H. The Cardiac Injury Immune Response as a Target for Regenerative and Cellular Therapies. Clin Ther 42, 1923–1943 (2020).
Nishiyama, C. et al. Prolonged Myocardial Regenerative Capacity in Neonatal Opossum. Circulation 146, 125–139 (2022).
Capalbo, L. et al. The midbody interactome reveals unexpected roles for PP1 phosphatases in cytokinesis. Nat Commun 10, (2019).
Xin, M. et al. Hippo pathway effector Yap promotes cardiac regeneration. Proc Natl Acad Sci U S A 110, 13839–13844 (2013).
Shen, H. et al. Mononuclear diploid cardiomyocytes support neonatal mouse heart regeneration in response to paracrine IGF2 signaling. Elife 9, (2020).
Bassat, E. et al. The extracellular matrix protein agrin promotes heart regeneration in mice. Nature 547, 179 (2017).
Gáspár, R. et al. Decorin Protects Cardiac Myocytes against Simulated Ischemia/Reperfusion Injury. Molecules 25, 3426 (2020).
Di Stefano, V., Giacca, M., Capogrossi, M. C., Crescenzi, M. & Martelli, F. Knockdown of cyclin-dependent kinase inhibitors induces cardiomyocyte re-entry in the cell cycle. J Biol Chem 286, 8644–8654 (2011).
Hume, R. D. et al. Tropoelastin Improves Post-Infarct Cardiac Function. Circ Res 132, 72–86 (2023).
Kühn, B. et al. Periostin induces proliferation of differentiated cardiomyocytes and promotes cardiac repair. Nat Med 13, 962–969 (2007).
Mills, R. J. et al. Drug Screening in Human PSC-Cardiac Organoids Identifies Pro-proliferative Compounds Acting via the Mevalonate Pathway. Cell Stem Cell 24, 895-907.e6 (2019).
Bae, J. et al. Malonate Promotes Adult Cardiomyocyte Proliferation and Heart Regeneration. Circulation 143, 1973–1986 (2021).
Cheng, Y. Y. et al. Metabolic Changes Associated With Cardiomyocyte Dedifferentiation Enable Adult Mammalian Cardiac Regeneration. Circulation 146, 1950–1967 (2022).
Tan, X., Vrana, K. & Ding, Z. M. Cotinine: Pharmacologically Active Metabolite of Nicotine and Neural Mechanisms for Its Actions. Front Behav Neurosci 15, (2021).
Quaife-Ryan, G. A. et al. β-Catenin drives distinct transcriptional networks in proliferative and nonproliferative cardiomyocytes. Development 147, (2020).
Bertozzi, A., Wu, C. C., Hans, S., Brand, M. & Weidinger, G. Wnt/β-catenin signaling acts cell-autonomously to promote cardiomyocyte regeneration in the zebrafish heart. Dev Biol 481, 226–237 (2022).
Zhang, Y. et al. Single-cell imaging and transcriptomic analyses of endogenous cardiomyocyte dedifferentiation and cycling. Cell Discov 5, (2019).
Li, J. et al. Alpha-catenins control cardiomyocyte proliferation by regulating Yap activity. Circ Res 116, 70–79 (2015).
Li, X. et al. MiR-499 regulates cell proliferation and apoptosis during late-stage cardiac differentiation via Sox6 and cyclin D1. PLoS One 8, (2013).
Nayak, A. et al. HIF1α-dependent upregulation of ATAD2 promotes proliferation and migration of stomach cancer cells in response to hypoxia. Biochem Biophys Res Commun 523, 916–923 (2020).
Bergmann, O. et al. Evidence for cardiomyocyte renewal in humans. Science 324, 98–102 (2009).
Bergmann, O. et al. Dynamics of Cell Generation and Turnover in the Human Heart. Cell 161, 1566–1575 (2015).
Meckert, P. C. et al. Endomitosis and polyploidization of myocardial cells in the periphery of human acute myocardial infarction. Cardiovasc Res 67, 116–123 (2005).
AP, B. et al. Evidence that human cardiac myocytes divide after myocardial infarction. N Engl J Med 344, 1248–1249 (2001).
Ramani, P., Taylor, S., Miller, E., Sowa-Avugrah, E. & May, M. T. High Phosphohistone H3 Expression Correlates with Adverse Clinical, Biological, and Pathological Factors in Neuroblastomas. Journal of Histochemistry and Cytochemistry 63, 397 (2015).
Basant, A. et al. Aurora B kinase promotes cytokinesis by inducing centralspindlin oligomers that associate with the plasma membrane. Dev Cell 33, 204–215 (2015).
Neubauer, S. The failing heart--an engine out of fuel. N Engl J Med 356, 1140–1151 (2007).
Ritterhoff, J. & Tian, R. Metabolic mechanisms in physiological and pathological cardiac hypertrophy: new paradigms and challenges. Nat Rev Cardiol (2023) doi:10.1038/S41569-023-00887-X.
Zhang, Y. et al. Single-cell imaging and transcriptomic analyses of endogenous cardiomyocyte dedifferentiation and cycling. Cell Discov 5, (2019).
Cheng, Y. Y. et al. Metabolic Changes Associated With Cardiomyocyte Dedifferentiation Enable Adult Mammalian Cardiac Regeneration. Circulation 146, 1950–1967 (2022).
Zhu, Y., Do, V. D., Richards, A. M. & Foo, R. What we know about cardiomyocyte dedifferentiation. J Mol Cell Cardiol 152, 80–91 (2021).
Bongiovanni, C. et al. Reawakening the Intrinsic Cardiac Regenerative Potential: Molecular Strategies to Boost Dedifferentiation and Proliferation of Endogenous Cardiomyocytes. Front Cardiovasc Med 8, (2021).
Wang, W. E. et al. Dedifferentiation, Proliferation, and Redifferentiation of Adult Mammalian Cardiomyocytes After Ischemic Injury. Circulation 136, 834–848 (2017).
Kuppe, C. et al. Spatial multi-omic map of human myocardial infarction. Nature 608, 766–777 (2022).
Jiang, J. et al. Cardiac myosin binding protein C regulates postnatal myocyte cytokinesis. Proc Natl Acad Sci U S A 112, 9046–9051 (2015).
Wu, X., Reboll, M. R., Korf-Klingebiel, M. & Wollert, K. C. Angiogenesis after acute myocardial infarction. Cardiovasc Res 117, 1257–1273 (2021).
Dunwoodie, S. L. The role of hypoxia in development of the Mammalian embryo. Dev Cell 17, 755–773 (2009).
Lal, S. et al. Best Practice BioBanking of Human Heart Tissue. Biophys Rev 7, 399–406 (2015).
Thavapalachandran, S. et al. Platelet-derived growth factor-AB improves scar mechanics and vascularity after myocardial infarction. Sci Transl Med 12, (2020).
Ashhurst, T. M. et al. Integration, exploration, and analysis of high-dimensional single-cell cytometry data using Spectre. Cytometry A 101, 237–253 (2022).
Liao, Y., Smyth, G. K. & Shi, W. The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Res 47, (2019).
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Robinson, M. D. & Oshlack, A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol 11, (2010).
Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43, e47 (2015).
Law, C. W., Chen, Y., Shi, W. & Smyth, G. K. voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15, (2014).
Gillespie, M. et al. The reactome pathway knowledgebase 2022. Nucleic Acids Res 50, D687–D692 (2022).
Huang, D. W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44–57 (2009).
Harney, D. J. et al. Small-protein Enrichment Assay Enables the Rapid, Unbiased Analysis of Over 100 Low Abundance Factors from Human Plasma. Mol Cell Proteomics 18, 1899–1915 (2019).
Koay, Y. C. et al. Effect of chronic exercise in healthy young male adults: a metabolomic analysis. Cardiovasc Res 117, 613–622 (2021).
O’Sullivan, J. F. et al. Dimethylguanidino valeric acid is a marker of liver fat and predicts diabetes. J Clin Invest 127, 4394–4402 (2017).
Karpievitch, Y. V., Nikolic, S. B., Wilson, R., Sharman, J. E. & Edwards, L. M. Metabolomics data normalization with EigenMS. PLoS One 9, (2014).
Fleming, S. J. et al. Unsupervised removal of systematic background noise from droplet-based single-cell experiments using CellBender. bioRXIV Preprint (2022).
Stuart, T. et al. Comprehensive Integration of Single-Cell Data. Cell 177, 1888-1902.e21 (2019).
Andreatta, M. & Carmona, S. J. UCell: Robust and scalable single-cell gene signature scoring. Comput Struct Biotechnol J 19, 3796–3798 (2021).
Nestorowa, S. et al. A single-cell resolution map of mouse hematopoietic stem and progenitor cell differentiation. Blood 128, e20–e31 (2016).
Tirosh, I. et al. Dissecting the multicellular ecosystem of metastatic melanoma by single-cell RNA-seq. Science 352, 189–196 (2016).
Lehtonen, J. Y. A., Jokinen, J. J., Holmström, M. & Kupari, M. Open chest core needle biopsy of left ventricle in the evaluation of suspected focal myocardial inflammation. J Thorac Cardiovasc Surg 149, e99–e102 (2015).

There is NO Competing Interest.

Download PDF

Version 1

posted

You are reading this latest preprint version

Human hearts have an intrinsic regenerative potential following myocardial infarction

Status:

Version 1

Abstract

Figures

Main

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1