FIELD OF THE INVENTION

The present invention lies in the field of heart tissue models. More specifically, the invention provides for in vitro heart tissue models named epicardioids, which match the multilayered structure of the ventricular epicardium and myocardium of human embryos. The epicardioids can be used as a heart tissue model. In a further aspect, the invention provides for a method of producing such a heart tissue model in vitro. Moreover, the invention relates to a heart tissue model produced by methods disclosed herein. In addition, the heart tissue model may be used as a cell source. Moreover, the invention relates to heart tissue models which are still developing, named early-epicardioids. Such early-epicardioids may as well be used as a cell source. The invention additionally relates to cells purified from epicardioids or early-epicardioids. Such cells may be for use in therapy and the treatment of heart diseases.

BACKGROUND OF THE INVENTION

Cardiovascular disease (CVD) is the leading cause of death and a major contributor to disability worldwide. Despite this significant public health burden, the development of new cardiovascular therapies has remained very limited in the past decades compared to other clinical areas. Crucially, there is great difficulty in translating preclinical findings into safe and effective therapies: the probability of launch of a cardiovascular drug entering phase I is currently estimated at 4%, one of the lowest among all disease categories. In addition, cardiotoxicity is one of the most common adverse effects across multiple drug classes and an important cause of withdrawal from clinical trials or the market. One root cause is that commonly used preclinical models do not faithfully recapitulate human (patho)physiology. Most studies of cardiac development, function, and disease are based on animal models and particularly mice, but there are many confounding morphological and functional differences between the hearts of humans and rodents. Large animals such as pigs and non-human primates are better suited to translational research, but their use is limited by high costs of handling and ethical considerations. Conversely, classical in vitro cell-based models are amenable to high-throughput testing but suffer from low predictive value.

The epicardium is the mesothelial cell sheet covering the heart's outer surface. Long considered a simple barrier between the pericardial cavity and the myocardium, it is now recognized to hold key functions in cardiac development and repair. The epicardium is largely quiescent in the adult heart; upon injury, it contributes to fibrotic remodeling through inflammatory signaling and modest differentiation into myofibroblasts (Quijada et al. (2020), "The role of the epicardium during heart development and repair", Circulation Research; 126:377-394). By contrast, during embryonic development, it is the source of multiple cell types: a subset of epicardial cells undergoes epithelial-to-mesenchymal transition (EMT) to become epicardial derived cells (EPDCs) that migrate into the myocardium and give rise to the majority of fibroblasts and vascular smooth muscle cells of the heart. Whether EPDCs also differentiate into cardiomyocytes and coronary endothelial cells is still debated, with studies providing conflicting evidence. In addition to these cellular contributions, the epicardium provides signaling factors critical for the development and growth of the myocardium. It has therefore been speculated that these embryonic programs could be exploited for heart repair. This is supported by the fact that the epicardium plays a central role in heart regeneration in species capable of rebuilding adult heart muscle upon injury, such as the zebrafish. Human embryonic stem cell-derived epicardial cells were already shown to enhance the functionality of cardiomyocyte grafts in a rat model of myocardial infarction, which was associated with improved systolic function compared to the transplantation of cardiomyocytes alone (Bargehr et al. (2019), "Epicardial cells derived from human embryonic stem cells augment cardiomyocyte-drived heart regeneration", Nature Biotech., 37:895-906). Although in this setting the epicardial contributions were limited to fibroblast differentiation and trophic support, the study indicates that fetal-like epicardium is a promising target for regenerative medicine. However, the inaccessibility of human embryonic tissue at the early stages of epicardium development, which begins less than four weeks post-conception, has left significant gaps in our understanding of human epicardial development and function. Many questions on the ontogeny of human proepicardial precursors and the functional heterogeneity of epicardial cells are still unresolved, which limits harnessing their full potential for regenerative medicine.

Human pluripotent stem cells (hPSCs) have emerged as a powerful alternative to generate in vitro models of higher physiological relevance. This includes embryonic stem cells (ESCs) and so-called induced pluripotent stem cells (iPSCs) that can be reprogrammed from the somatic cells of any healthy or diseased individual. The present inventors were the first to show that cardiac muscle cells (cardiomyocytes) differentiated from the iPSCs of a patient with long-QT syndrome type I recapitulated functional features of the disease (Moretti et al. (2010), "Patient-specific induced pluripotent stem-cell models for long-QT syndrome", N. Engl. J. Med.; 363:1397-1309). Importantly, pluripotent stem cells have the potential to form organoids, which are defined as self-organized 3D micro-tissues resembling an organ of the body in terms of cell composition, architecture, and function. In contrast to traditional tissue engineering, the generation of organoids relies on the self-patterning of cells upon modulation of signaling pathways that have been found to control the development of the organ in vivo. Organoids of most organs including the intestine, brain, kidney, lung, liver, and retina have already become versatile platforms for disease modeling, drug discovery, and personalized therapy. A notable exception is the heart: although the term organoid has been used in the literature to describe various 3D cardiac structures, no bona fide self-organized cardiac organoid composed of the three layers of epicardium, myocardium, and endocardium has yet been reported. Self-organized cardiac structures containing myocardium and endocardium have been described, but there have been none containing a spontaneously forming and functional epicardial layer.

The present invention provides novel human pluripotent stem cell-derived cardiac micro-tissues showing self-organization of the myocardial and epicardial layers, which were named 'epicardioids'. Epicardioids recapitulate key steps of cardiac development and display the morphological and functional self-patterning typical of the ventricular wall. In addition, the inventors could show that the herein described epicardioids represent an advanced system to model multicellular mechanisms of heart disease.

SUMMARY OF THE INVENTION

In one aspect, the invention relates to a heart tissue model named epicardioid comprising an inner myocardial compartment and an outer epicardial compartment, wherein the outer epicardial compartment comprises from the outside to the inside: a layer of epicardium comprising KRT18 ⁺ mesothelial epicardial cells, and a layer of subepicardium comprising Vim ⁺ epicardial-derived cells (EPDCs); and wherein the inner myocardial compartment comprises from the outside to the inside: a layer of compact-like myocardium comprising densely packed cTnT ⁺ myocardial cells and a layer of trabecular-like myocardium comprising looser packed cTnT ⁺ myocardial cells compared to the compact-like myocardium, wherein the epicardioid has a three- dimensional shape, wherein the myocardial compartment forms the inner core of the epicardioid, wherein the epicardial compartment forms an envelope of the epicardioid, and wherein the epicardioid has been developed from one initial in vitro pluripotent stem cell population.

In an embodiment, the KRT18+ mesothelial epicardial cells may also be positive for the markers CDH1, ZO1, and/or BNC1.

In an embodiment, the initial in vitro cell population is an in vitro pluripotent stem cell population provided in cell culture medium, wherein the pluripotent stem cell population is preferably an induced pluripotent stem cell population, more preferably a human induced pluripotent stem cell population, or wherein the pluripotent stem cell population is preferably an embryonic stem cell population, more preferably a human embryonic stem cell population.

The cells of the outer epicardial compartment may further express epicardial transcription factor TBX18.

The layer of compacted myocardium may comprise a significantly higher proportion of Ki67 ⁺ proliferating cardiomyocytes than the inner trabecular-like layer of myocardium. The inner myocardial compartment may further comprise in both layers, i.e., the compact-like and the trabecular-like myocardium, non-cardiomyocyte cells comprising cardiac fibroblasts, vascular smooth muscle cells, pericytes, and endothelial cells.

In an embodiment, the epicardioid matches with its two compartment structure the multilayered structure of the ventricular epicardium and myocardium of human embryos.

In another aspect, the invention relates to the use of the inventive epicardioid as a heart tissue model. In an embodiment, the invention relates to the use of the inventive epicardioid as an in vitro heart tissue model.

The herein disclosed epicardioid may be used for studying cardiac diseases, preferably left ventricular heart diseases.

The herein disclosed epicardioid may be used for preclinical testing of agents, compounds, drugs or other substances for the treatment of heart diseases. Such heart diseases are preferably hypertrophic and/or fibrotic heart diseases.

The herein disclosed epicardioid may also be used for preclinical testing of drug-induced arrhythmias.

The herein disclosed epicardioid may also be used for preclinical testing of drug-induced cardiotoxicity.

In a further aspect, the herein disclosed epicardioid is provided for use in the treatment of cardiac diseases.

Such treatment may comprise the replacement of cardiomyocytes lost during myocardial infarction with the epicardioid, wherein the replacement is either by re-activating the epicardium's capacity to promote cardiomyocyte proliferation or by triggering de novo differentiation of EPDCs into cardiomyocytes.

In another aspect of the invention, the epicardioids disclosed herein may be used as a cell source. In an embodiment, the epicardioids disclosed herein may be used as a cell source, wherein defined cell populations may be purified from the epicardioid. In an embodiment, the epicardioid of the invention is a cell source for FHF-derived cardiomyocytes, JCF-derived cells comprising mesothelial epicardial cells, EPDCs, fibroblasts, vascular smooth muscle cells, pericytes and cardiomyocytes, or endothelial cells. In an embodiment, the epicardioids disclosed herein may be used as a cell source, wherein defined cell populations may be purified from the epicardioid and wherein these purified cells are for use in the treatment of a heart disease. Cell populations to be isolated from an epicardioid may comprise: FHF-derived cardiomyocytes and JCF-derived cells comprising epicardium cells and cardiomyocytes, epicardium-derived cells comprising EPDCs as well as EPDC-derived cells comprising fibroblasts, vascular smooth muscle cells, pericytes, and cardiomyocytes, or endothelial cells. The defined cell populations may be isolated and purified from the epicardioid at around days 13 to 30.

In another embodiment, epicardioids are used as a source for FHF-derived cardiomyocytes.

In another embodiment, epicardioids are used as a source for JCF-derived epicardial cells.

In one embodiment, epicardioids are used as a source for JCF-derived cardiomyocytes.

In one embodiment, epicardioids are used as a source for epicardium-derived cells.

In one embodiment, epicardioids are used as a source for EPDCs.

In one embodiment, epicardioids are used as a source for EPDC-derived cells.

In one embodiment, epicardioids are used as a source for fibroblasts.

In one embodiment, epicardioids are used as a source for vascular smooth muscle cells.

In one embodiment, epicardioids are used as a source for pericytes.

In one embodiment, epicardioids are used as a source for EPDC-derived cardiomyocytes.

In one embodiment, epicardioids are used as a source for endothelial cells.

The invention further relates to cell populations purified from the disclosed epicardioids.

In an embodiment, the cell population isolated from an epicardioid of the invention is FHF-derived cardiomyocytes.

In an embodiment, the cell population isolated from an epicardioid of the invention is JCF-derived cardiomyocytes.

In an embodiment, the cell population isolated from an epicardioid of the invention is JCF-derived epicardial cells.

In an embodiment, the cell population isolated from an epicardioid of the invention is epicardium- derived cells.

In an embodiment, the cell population isolated from an epicardioid of the invention is EPDCs.

In an embodiment, the cell population isolated from an epicardioid of the invention is EPDC- derived cells.

In an embodiment, the cell population isolated from an epicardioid of the invention is fibroblasts.

In an embodiment, the cell population isolated from an epicardioid of the invention is vascular smooth muscle cells.

In an embodiment, the cell population isolated from an epicardioid of the invention is pericytes. In an embodiment, the cell population isolated from an epicardioid of the invention is EPDC- derived cardiomyocytes.

In an embodiment, the cell population isolated from an epicardioid of the invention is endothelial cells.

In another aspect of the invention, the epicardioids disclosed herein may be used as a cell source for cell-based therapeutic applications.

A further aspect of the invention relates to a cell population purified from the disclosed epicardioids for use in therapy.

In an embodiment, the invention is directed to a cell population isolated from the disclosed epicardioids for use in the treatment of a heart disease.

In an embodiment, the cell population isolated from the disclosed epicardioid for use in therapy may be mesothelial epicardial cells expressing KRT18, KRT19, and/or CDH1, EPDCs expressing VIM, fibroblasts expressing COL1A1 and/or TNC, vascular smooth muscle cells expressing RGS5 and/or MYH11, pericytes expressing MCAM, cardiomyocytes expressing TNNT2, or endothelial cells expressing CD31 and/or CDH5.

In an embodiment, the invention is directed to mesothelial epicardial cells isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the invention is directed to EPDCs isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the invention is directed to fibroblasts isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the invention is directed to vascular smooth muscle cells isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the invention is directed to pericytes isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the invention is directed to cardiomyocytes isolated from a disclosed epicardioid for use in the treatment of a heart disease. In an embodiment the cardiomyocytes are FHF-derived cardiomyocytes. In another embodiment, the cardiomyocytes are JCF-derived cardiomyocytes.

In an embodiment, the invention is directed to endothelial cells isolated from a disclosed epicardioid for use in the treatment of a heart disease.

In an embodiment, the heart disease to be treated comprises myocardial infarction, heart failure, and acquired or inherited forms of cardiomyopathy comprising hypertrophic cardiomyopathy, dilated cardiomyopathy, arrhythmogenic cardiomyopathy, Duchenne muscular dystrophy, and restrictive cardiomyopathy.

In another aspect, the present invention relates to a method of producing a three-dimensional epicardioid comprising an inner myocardial compartment and an outer epicardial compartment, wherein the outer epicardial compartment comprises from the outside to the inside a layer of epicardium comprising KRT18 ⁺ mesothelial epicardial cells, and a layer of subepicardium comprising Vim ⁺ epcardial-derived cells (EPDCs); and wherein the inner myocardial compartment comprises from the outside to the inside a layer of compact-like myocardium comprising densely packed cTnT ⁺ myocardial cells, and a layer of trabecular-like myocardium comprising looser packed cTnT ⁺ myocardial cells compared to the compact-like myocardium, wherein the epicardioid has a three-dimensional shape, wherein the myocardial compartment forms the inner core of the epicardioid, and wherein the epicardial compartment forms the envelope of the epicardioid. The method of producing a three-dimensional epicardioid comprises the steps of: providing pluripotent stem cells (PCSs) and seeding the PSCs in an initial culture medium on day - 1 of the culture; replacing the culture medium with a first basal differentiation medium on day 0 of the culture; replacing the culture medium with a second basal differentiation medium comprising retinoic acid on day 2 of the culture; refreshing the culture medium with the second basal differentiation medium each 24 hours until day 6 of the culture; replacing the culture medium with a third basal differentiation medium on day 6 of the culture; refreshing the culture medium with the third basal differentiation medium on day 7 of the culture; embedding spheroids into a gel comprising collagen type I, covering of the gel with a collagen type I solution, letting the collagen type I solution solidify, and transferring gel sheets to maintenance medium on day 8 of the culture; optionally placing the gel sheets on a rocking shaker on day 10 of the culture; replacing maintenance medium every 2-3 days after day 8 of the culture; isolating epicardioids for further use between days 13 and 100 of the culture, preferably between days 15 and 40.

The three-dimensional epicardioid prepared by this method is a heart tissue model as disclosed herein.

The PSCs are preferably human pluripotent stem cells, preferably human induced pluripotent stem cells (hiPSCs). Alternatively, the PSCs may be embryonic stem cells, preferably human embryonic stem cells.

In a preferred embodiment, the initial number of PSCs on day -1 of the culture is 30,000-40,000 per well of a 96-well plate.

The retinoic acid in the second basal differentiation medium is preferably present in an amount of 0.3-0.75 pM, preferably 0,5 pM. In one embodiment, the initial culture medium comprises ROCK inhibitor thiazovivin. In a preferred embodiment, the initial culture medium is Essential 8 medium comprising 1-5 pM, preferably 2 pM thiazovivin.

In addition, the first basal medium preferably comprises BMP4, Activin A, bFGF, a PI3 kinase inhibitor, preferably LY-29004, and a Wnt activator, preferably 1 CHIR-99021.

The second basal medium preferably further comprises insulin, BMP4, bFGF, and a Wnt- antagonist, preferably IWP2.

The third basal medium preferably comprises insulin, BMP4, bFGF.

The collagen I solution comprises collagen I. In addition, the collagen I solution preferably comprises distilled water, DPBS, NaOH DMEM/F-12 with FBS, non-essential amino acids, and optionally Penicillin-Streptomycin.

In an embodiment, the maintenance medium comprises VEGF.

In a further embodiment, the 96-well plate comprises U-shaped wells which are optionally coated with a low cell adherence agent, preferably with poly-HEMA.

In another aspect, the invention relates to a method for screening or testing a candidate compound for its effects on heart development and/or functionality comprising producing a heart tissue model according to the method disclosed herein, while treating the cells with the candidate compound and comparing development of the heart tissue model with development and/or functionality of a heart tissue model that was not treated with the candidate compound.

In another aspect, the invention relates to a method for screening or testing a candidate compound for its effects on heart development and/or functionality comprising producing an epicardioid according to the method disclosed herein, while treating the cells with the candidate compound and comparing development of the cells with development and/or functionality of the cells that were not treated with the candidate compound.

In an embodiment, the compound to be tested is added to the culture during culture days 1 to 13. In another embodiment, the compound to be tested is added on or after day 13. In a preferred embodiment, the compound to be tested is added on or after day 15 of culture. In an embodiment, the compound is added once, in another embodiment, the compound is added more than once, for instance, each day, each second day or each third day of the culture. The compound to be added may be a drug, a pharmaceutical, a small molecule, an antibody, or a nonpharmaceutical therapeutic agent. The non-pharmaceutical therapeutic agent comprises genetic material (DNA or RNA) and CRISPR/Cas9 components or other gene editing or modulation agents.

In another aspect, the present invention further relates to a heart tissue model prepared according to the herein disclosed method for the production of an epicardioid. In other words, the present invention further relates to a heart tissue model obtainable by the inventive methods for the production of an epicardioid described herein. The heart tissue model is preferably an epicardioid having the structure described herein, and wherein the heart tissue model is isolated from the culture at day 13 or later. Alternatively, the heart tissue model is an early-epicardioid, isolated from the culture at day 4 to 12.

The heart tissue model prepared according to the herein disclosed methods preferably comprises an inner myocardial compartment and an outer epicardial compartment, wherein the outer epicardial compartment comprises from the outside to the inside a layer of epicardium comprising KRT18 ⁺ mesothelial epicardial cells, and a layer of subepicardium comprising Vim ⁺ epcardial-derived cells (EPDCs); and wherein the inner myocardial compartment comprises from the outside to the inside a layer of compact-like myocardium comprising densely packed cTnT ⁺ myocardial cells, and a layer of trabecular-like myocardium comprising looser packed cTnT ⁺ myocardial cells compared to the compact-like myocardium, wherein the epicardioid has a three- dimensional shape, wherein the myocardial compartment forms the inner core of the epicardioid and wherein the epicardial compartment forms the envelope of the epicardioid.

In an embodiment, the heart tissue model has been isolated from the culture at days 13-30 of the culture.

In an embodiment, the heart tissue model has been isolated from the culture before day 13 of the culture, for instance at day 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 of the culture.

In a further aspect, the invention relates to a method for the isolation and purification of cell populations from the epicardioid disclosed herein, wherein the method comprises the following steps a) generation of an epicardioid according to the methods disclosed herein, b) dissociation of the epicardioid, and c) purification of desired cell populations, wherein dissociation is performed at about day 13 to about day 30 of the epicardioid culture of step a), and wherein purification is performed by flow-based or magnetic separation using surface markers and/or genetically encoded, preferably cell type-specific, fluorescent reporters.

In an embodiment, step a) comprises the steps of: providing pluripotent stem cells (PCSs) and seeding the PSCs in an initial culture medium on day -1 of the culture; replacing the culture medium with a first basal differentiation medium on day 0 of the culture; replacing the culture medium with a second basal differentiation medium comprising retinoic acid on day 2 of the culture; refreshing the culture medium with the second basal differentiation medium each 24 hours until day 6 of the culture; replacing the culture medium with a third basal differentiation medium on day 6 of the culture; refreshingthe culture medium with the third basal differentiation medium on day 7 of the culture; embedding spheroids into a gel comprising collagen type I, covering of the gel with a collagen type I solution, letting the collagen type I solution solidify, and transferring gel sheets to maintenance medium on day 8 of the culture; optionally placing the gel sheets on a rocking shaker on day 10 of the culture; replacing maintenance medium every 2-3 days after day 8 of the culture.

In an embodiment, cardiomyocytes are isolated by use of surface marker SIRPA.

In an embodiment, epicardial cells are isolated by use of surface marker CDH1.

In an embodiment, endothelial cells are isolated by use of surface marker CD31.

In an embodiment, fibroblasts are isolated by use of surface marker CD90.

In an embodiment, mural cells are isolated by use of surface marker PDGFRB.

In an embodiment, dissociation is performed by use of an enzyme-based solution or by mechanical means. Preferably dissociation is performed by use of an enzyme-based solution. In an embodiment, the enzyme of the enzyme-based solution is a protease. In a preferred embodiment, the enzyme is papain. Preferably, dissociation is performed by use of a papainbased enzyme solution. The solution preferably comprises 20 U/mL papain and 1 mM L-cysteine in PBS. In an embodiment, dissociation is performed by use of an enzyme-based solution comprising 20 U/mL papain and 1 mM L-cysteine.

In an embodiment, steps b) and c) are performed immediately after another.

In an embodiment, steps b) and c) are performed at day 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 of the epicardioid culture of step a). Steps b) and c) may be performed at about and at about days 13-30 for isolation of cardiomyocytes, epicardial cells and epicardium- derived cells comprising EPDCs, fibroblasts, mural cells, i.e., vascular smooth muscle cells and pericytes, and cardiomyocytes.

In an embodiment, for the isolation of mesothelial epicardial cells, step b) is performed at around days 13-30 of the epicardioid culture and step c) is performed using markers KRT18, KRT19, and/or CDH1.

In an embodiment, for the isolation of EPDCs, step b) is performed at around days 13-30 of the epicardioid culture and step c) is performed using marker VIM.

In an embodiment, for the isolation of fibroblasts, step b) is performed at around days 13-30 of the epicardioid culture and step c) is performed using markers CD90, COL1A1 and/or TNC. In an embodiment, for the isolation of vascular smooth muscle cells, step b) is performed at around days 13-30 of the epicardioid culture and step c) is performed using markers PDGFRB, RGS5 and/or MYH11.

In an embodiment, for the isolation of pericytes, step b) is performed at around days 13-30 of the epicardioid culture and step c) is performed using markers MCAM and/or KCNJ8.

In an embodiment, for the isolation of cardiomyocytes, step b) is performed at around days 13- 30 of the epicardioid culture and step c) is performed using marker SIRPA or TNNT2.

In an embodiment, for the isolation of endothelial cells, step b) is performed at around days 13- 30 of the epicardioid culture and step c) is performed using markers CD31 and/or CDH5.

In an embodiment, the epicardioids generated in step a) are generated from hiPSCs expressing a fluorescent reporter to track isolated cells in subsequent experiments such as transplantation experiments. The fluorescent reporter may be AAVSl-CAG-eGFP.

In a further aspect, the invention relates to a cell population obtained by a method for the isolation and purification of cells from epicardioids as described herein. Such cell population is mesothelial epicardial cells, EPDCs, fibroblasts, vascular smooth muscle cells, pericytes, cardiomyocytes, or endothelial cells.

In another aspect, the invention related to a cell population obtained by a method for the isolation and purification of cells from epicardioids as described herein for use in the treatment of a heart disease. Such cells may be mesothelial epicardial cells, EPDCs, fibroblasts, vascular smooth muscle cells, pericytes, cardiomyocytes, or endothelial cells. Such heart disease may be myocardial infarction, heart failure, and acquired or inherited forms of cardiomyopathy comprising hypertrophic cardiomyopathy, dilated cardiomyopathy, arrhythmogenic cardiomyopathy, Duchenne muscular dystrophy, and restrictive cardiomyopathy.

In a further aspect, the invention relates to a method for the purification of cell populations from an early-epicardioid, wherein the method comprises the following steps a) providing an epicardioid cell culture according to the methods for the generation of an epicardioid disclosed herein, b) dissociation of the early-epicardioid, and c) purification of desired cell populations from the dissociated early epicardioid obtained by step b), wherein the early-epicardioid is a sphere-shaped non-mature epicardioid at a cell state before day 13 of the culture of step a), wherein dissociation is performed at about day 4 to about day 12 of the epicardioid culture of step a), wherein step c) is performed by flow-based or magnetic separation using surface markers and/or genetically encoded, preferably cell type-specific, fluorescent reporters.

In an embodiment, step a) comprises the steps of: providing pluripotent stem cells (PCSs) and seeding the PSCs in an initial culture medium on day -1 of the culture; replacing the culture medium with a first basal differentiation medium on day 0 of the culture; replacing the culture medium with a second basal differentiation medium comprising retinoic acid on day 2 of the culture; refreshing the culture medium with the second basal differentiation medium each 24 hours until day 6 of the culture; replacing the culture medium with a third basal differentiation medium on day 6 of the culture; refreshingthe culture medium with the third basal differentiation medium on day 7 of the culture; embedding spheroids into a gel comprising collagen type I, covering of the gel with a collagen type I solution, letting the collagen type I solution solidify, and transferring gel sheets to maintenance medium on day 8 of the culture; optionally placing the gel sheets on a rocking shaker on day 10 of the culture; replacing maintenance medium every 2-3 days after day 8 of the culture.

In an embodiment, step b) is performed by use of an enzyme-based solution. Preferably, the enzyme is a protease. In a more preferred embodiment, step b) is performed by use of a papainbased enzyme solution. This solution most preferably comprises 20 U/mL papain and 1 mM L- cysteine in PBS.

In an embodiment, steps b) and c) are performed immediately after another.

In an embodiment, steps b) and c) are performed at day 4, 5, 6, 7, 8, 9, 10, 11 or 12 of the epicardioid culture of step a). Steps b) and c) may be performed at about days 4-5 for isolation of FHF cells and pre-JCF cells, at about days 7-10 for isolation of JCF cells, and at about days 10-12 for mesothelial epicardial cells, EPDCs, cardiomyocytes, and endothelial cells.

In an embodiment, for the isolation of FHF cells, step b) is performed at around days 4-5 of the epicardioid culture and step c) is performed using markers NKX2.5 and/or TBX5. Preferably, for the isolation of FHF cells, step b) is performed at around days 4-5 of the epicardioid culture and step c) is performed using markers NKX2.5 and TBX5.

In an embodiment, for the isolation of pre-JCF cells, step b) is performed at around days 4-5 of the epicardioid culture and step c) is performed using marker ISL1. ISL1 is also expressed in FHF cells, but pre-JCF cells differ from FHF cells by the absence of NKX2.5. Hence, ISL1 and NKX2.5 may be used in combination to distinguish between the two cell populations. In an embodiment, for the isolation of pre-JCF cells, step b) is performed at around days 4-5 of the epicardioid culture and step c) is performed using markers ISL1 and NKX2.5. In an embodiment, for the isolation of JCF cells, step b) is performed at around days 7-10 of the epicardioid culture and step c) is performed using markers HAND1, HOXB6 and/or MAB21L2.

In an embodiment, for the isolation of mesothelial epicardial cells, step b) is performed at around days 10-13 of the epicardioid culture and step c) is performed using markers KRT18, KRT19, and/or CDH1.

In an embodiment, for the isolation of EPDCs, step b) is performed at around days 10-12 of the epicardioid culture and step c) is performed using marker VIM.

In an embodiment, for the isolation of cardiomyocytes, step b) is performed at around days 10- 12 of the epicardioid culture and step c) is performed using marker SIRPA and/or TNNT2.

In an embodiment, for the isolation of endothelial cells, step b) is performed at around days 10- 12 of the epicardioid culture and step c) is performed using markers CD31 and/or CDH5.

In a further aspect, the invention relates to a cell population obtained by a method for the isolation and purification of cells from epicardioids as described herein. Such cell population is FHF cells, pre-JCF cells, JCF cells, mesothelial epicardial cells, EPDCs, cardiomyocytes, or endothelial cells.

In another aspect, the invention relates to a cell population obtained by a method for the isolation and purification of cells from epicardioids as described herein for use in the treatment of a heart disease. Such cells may be FHF cells, pre-JCF cells, JCF cells, mesothelial epicardial cells, EPDCs, cardiomyocytes, or endothelial cells. Such heart disease may be myocardial infarction, heart failure, and acquired or inherited forms of cardiomyopathy comprising hypertrophic cardiomyopathy, dilated cardiomyopathy, arrhythmogenic cardiomyopathy, Duchenne muscular dystrophy, and restrictive cardiomyopathy.

ADVANTAGES OF THE INVENTION

In contrast to engineered heart tissues obtained by co-culture of individually differentiated (and/or native) cardiac cells as described in the prior art, the formation of epicardioids is driven by molecular processes that mimic embryonic cardiogenesis. This leads to tissue microarchitecture that more closely resembles in vivo organization, which is of advantage when modeling complex mechanisms of development or disease that depend on the interplay of form and function. Other groups have also generated cardiac microtissues from hPSCs based on developmental principles. Most notably, Hofbauer and colleagues recently generated human pluripotent stem cell-derived 'cardioids' consisting of myocardium forming an inner cavity lined with an endocardium-like layer (subject of patent application EP20164637.9). Drakhlis et al. reported the co-derivation of a cardiac structure - including a myocardial layer and an endocardium-like compartment - with foregut endoderm (Drakhlis et al., (2021), "Human heartforming organoids recapitulate early heart and foregut development"; Nat. Biotehcnol.; 396(39):737-746) . Lewis-Israeli et al. generated cardiac structures containing endothelial cells as well as epicardial cells forming individual clusters (Lewis-Israeli et al., (2021); "Self-assembling human heart organoids for the modeling of cardiac development and congenital heart disease"; Nat. Commun., 12:5142). Others reported in vitro development of early cardiac structures corresponding to the early crescent stage (Rossi et al., (2021); "Capturing Cardiogenesis in Gastruloids"; Cell Stem Cell, 28:230-240. e6; Andersen et al., (2018); Precardiac organoids from two heart fields via Bmp/Wnt signaling"; Nat. Commun., 9:3140)).

However, to our knowledge, herein described epicardioids are the first to show spontaneous organization and maintenance of a continuous epicardial layer as well as an advanced morphological and functional self-patterning of the myocardial layer. The disclosed invention provides for the possibility to established human epicardioids showing retinoic acid-dependent self-organization of ventricular myocardium and epicardium. Epicardioids recapitulate the two major functions of the embryonic epicardium: being the source of progenitors of several cardiac lineages and providing a paracrine milieu driving myocardial compaction and maturation. The latter appears to be key in achieving a high degree of morphological, molecular, and functional self-patterning of the myocardium, which has so far been lacking in cardiac organoid models.

Single-cell transcriptomic analyses of human embryonic and adult heart tissue have provided precious insights into epicardial development. However, isolated tissues represent punctual snapshots that are of limited use for studying dynamic developmental processes, especially those occurring at the earliest stages of cardiogenesis. The epicardioids of the present invention offer a powerful alternative, as they closely mimic the steps of left ventricular development and maturation.

During their in vitro development, the disclosed heart tissue models notably revealed the existence of a human equivalent of the murine juxta-cardiac field (JCF) showing potentiality for myocardial and epicardial fates associated with distinct gene regulatory programs. ISL1 was identified herein as a marker of the human JCF and early embryonic epicardium.

The disclosed heart tissue models, especially the epicardioids of the invention can be used to address open questions related to epicardial fate potential. For instance, transcriptomic and chromatin accessibility profiling combined with lineage tracing supported the still debated myocytic potential of early epicardium. Importantly, for all three epicardial derivatives - fibroblasts, smooth muscle cells, and cardiomyocytes - dynamic changes in fate potential rather than strict pre-determination was observed. This is consistent with recent work in the mouse challenging the long-held notion that there exist distinct epicardial sub-compartments. Finally, the invention demonstrates that mature epicardioids have the unique ability of recapitulating both hypertrophic and fibrotic features of left ventricular hypertrophy. They could therefore be exploited for preclinical testing to identify drugs targeting both aspects of the disease, which are intimately linked during the progression towards heart failure. More broadly, the epicardioids of the invention offer advantages for modeling complex cardiac disorders including congenital heart diseases by allowing dissection of inter- and intra-cellular crosstalk dynamics during disease development.

Insights from epicardioids also lead to new strategies to replace cardiomyocytes lost during myocardial infarction, arguably one of the biggest challenges of modern medicine, either by reactivating the epicardium's capacity to promote cardiomyocyte proliferation or by triggering de novo differentiation of EPDCs into cardiomyocytes.

As such, the disclosed heart tissue models, especially the disclosed epicardioids, offer a unique platform to tackle fundamental questions in developmental biology as well as cardiovascular medicine and drug discovery.

DESCRIPTION OF THE FIGURES

Figure 1: A Schematic drawing of an epicardioid. Shown are the different compartments of the epicardioid, each having characteristic layers of cells. B Confocal microscopy picture showing the epicardioid structure based on different cell stains. KRT18: mesothelial epicardial cells, Vim: subepicardial EPDCs, cTnT: myocardial cells.

Figure 2: Time course single-cell RNA sequencing analysis reveals molecular processes of epicardioid formation. (A,B). Epicardioids were analyzed by single-cell RNA-sequencing (lOx Genomics) at days 2, 3, 4, 5, 7, 10, and 15 of culture. The 31 clusters identified across all samples are broadly annotated (A) and separated by day (B). (C) UMAP feature plots of MESP1, TBX5 and MYL3 in epicardioids across all time points. Heatmaps indicate the respective level of expression of each gene. (D) UMAP feature plot showing cells co-expressing the indicated markers of human ventricular cardiomyocytes, human epicardium, and murine juxta-cardiac field in epicardioids across all time points.

Figure 3: Representative images of cTnT and cytokeratin 18 (KRT18) immunostaining of epicardioids differentiated from four different hPSC lines (3 hiPSC lines and 1 hESC line) at day 15 of culture. The arrowhead indicates an example of a partial epicardial layer.

Figure 4: Generation of human pluripotent stem cell-derived epicardioids showing selforganized myocardium and epicardium. A: Protocol used for 3D cardiac induction of hPSCs,with addition of retinoic acid (RA) or without and representative bright field images at the indicated days. B: BMP4, F: FGF2, A: Activin A, L: LY-29004, C: CHIR-99021, I: IWP2, Ins: Insulin. Scale bars = 500 pm. B,C: Representative images of immunostaining for the cardiomyocyte (CM) marker cardiac troponin T (cTnT) in spheroids differentiated with RA or without (noRA) (B) and immunostaining for cTnT and the epicardial markers TBX18 and TCF21 in spheroids differentiated with RA (C), all at day 15 of culture. Scale bars B - 200 pm, C - 50 pm. D: Representative images of immunostaining for cTnT, vimentin (Vim), and the epithelial markers cytokeratin 18 (KRT18) and tight junction protein 1 (TJP1) in spheroids differentiated with RA, referred to hereafter as epicardioids, at days 15 and 30 of culture. Arrowheads indicate the mesothelial epicardium, arrows indicate subjacent EPDCs. Scale bars - 50 pm. E: (Left) UMAP dimensional reduction plot showing the cell clusters obtained by single-cell RNA sequencing of epicardioids at days 15 and d30, with main cell types annotated (except for unclassified clusters 13 and 14). (Right) Cells from day 15 and day 30 are highlighted in orange and blue, respectively. CM: cardiomyocytes, EPDC: epicardium-derived cells, FB: fibroblasts, MC: mural cells, EC: endothelial cells, prolif.: proliferative, mesoth.: mesothelial, epi: epicardial cells. F: Feature plots showing the expression level of ventricular CM markers (MYH7, MYL2). G: Average expression ratio of the adult to fetal cardiac troponin I isoforms 20 (TNNI3/TNNI1) in the indicated ventricular CM clusters at days 15 and 30 of culture. H: Violin plots showing the expression levels of markers of mesothelial epicardium (CDH1, KRT19, TNNT1, C3), EMT (ZEB2, SNAI2), fibroblasts (FB; COL1A1), and mural cells (MC; RGS5) in the indicated epicardial clusters at days 15 and 30 of culture.

Figure 5: Transcriptional profiles of epicardial cells and their derivatives in epicardioids at days 15 and 30. A: UMAP plot of the subclustering of the epicardial clusters 11, 5, and 10 (shown in Fig. 4E) obtained from scRNA-seq analysis of epicardioids at days 15 and 30, overlaid with the trajectories inferred from RNA velocity; cell types are annotated. EPDCs: epicardium-derived cells, EMT: epithelial-to-mesenchymal transition, CM: cardiomyocyte, diff.: differentiation. B: Cells from day 15 and day 30 are labeled in red and blue, respectively. C: Feature plots showing the level of expression of the canonical epicardial markers WT1, TBX18, BNC1, TCF21, SEMA3D, and ALDH1A2 in the mesothelial subclusters 9 and 5 (presented in A). D: Violin plots showing the expression levels of markers of mesothelial epicardium (CDH1, KRT19), EMT (TWIST1), cardiomyocytes (TNNT2, ACTN2), fibroblasts (FBs; TNC, COL1A1), mural cells (comprising vascular smooth muscle cells and pericytes; RGS5, PDGFRB), and pericytes (MCAM, KCNJ8) in the epicardial subclusters presented in (A).

Figure 6: Comparison of the epicardial populations in epicardioids with epicardial cells obtained in 2D differentiation. A: (Top) UMAP plot showing the three clusters obtained by reanalyzing the scRNAseq dataset of Gambardella et al,. (2019), consisting of epicardial cells obtained by directed 2D differentiation of hiPSCs. (Bottom) The distinction between BNCl ^high and TCF21 ^high populations reported by Gambardella and colleagues is conserved, as shown by exclusive expression of BNC1 (red) and TCF21 (green). B: (Top) UMAP plots showing cells co-expressing BNC1, PODXL, WT1, CDH3, NPNT, and CDH1 (BNCl ^high signature reported by Gambardella et al.) in epicardial cells from epicardioids (left) and in 2D epicardial cells from Gambardella et al., 2019 (right). (Bottom) UMAP plots showing cells co-expressing TCF21, THK1, THBS1, FN1, PDGFRA, and TGFB1 (TCF21 ^hl s ^h signature reported by Gambardella et al) in epicardial cells from epicardioids (left) and in 2D epicardial cells (right). The main clusters matching the signature in each dataset are annotated. C: Violin plots showing the expression levels of markers of human epicardium in the 2D epicardium clusters from Gambardella et al. (left) and in the mesothelial subclusters 9 and 5 from epicardioids at days 15 and 30 (shown in Fig. 5A,B). TNNT1, SFRP2, and SFRP5 are expressed in fetal epicardium; LRP2, CALB2, and C3 are shared by fetal and adult epicardium.

Figure 7: Transcriptional profiles of cardiomyocytes in epicardioids at days 15 and 30. A: Violin blots of the expression levels of genes related to cardiomyocyte (CM) Ca2+ handling (ATP2A2, PLN, SLC8A1) in the indicated ventricular CM clusters identified by scRNA-seq. Cells from epicardioids at day 15 and 30 are shown separately. B: Violin plots of the expression levels of markers of pacemaker CMs (SHOX2, HCN4) in the indicated scRNA-seq clusters.

Figure 8: Transcriptional profiles of epicardial cells and their derivatives in epicardioids at days 15 and 30. A: Violin plots of the expression levels of human epicardial genes (S100A10, CCDC80, DCN, TM4SF1, MGP) in the indicated scRNA-seq clusters. B-F: Violin plots of the expression levels of markers of mesothelial epicardium (CALB2, PODXL) (B), epicardium-derived cells (EPDCs; C7, EMILIN1) (C), fibroblasts (FB; THY1, DDR2) (D), smooth muscle cells (SMC; TAGLN, ACTA2) (E), and mitotic cells (TOP2A, PLK1) (F) in the indicated scRNA-seq clusters. Cells from epicardioids at day 15 and 30 are shown separately.

Figure 9: Endothelial cells in epicardioids at days 15 and 30. Violin plots of the expression levels of endothelial cell markers (CDH5, PLVAP) in the indicated scRNA-seq clusters.

Figure 10: Epicardioids recapitulate the cellular crosstalk promoting ventricular patterning. A: Heatmap showing the number of cell-cell interactions between all scRNA-seq clusters in epicardioids at day 15 and 30 of culture. Main cell types are annotated, unclassified clusters (13, 14) were excluded. B: Dot plot showing selected ligand-receptor interactions between mesothelial epicardium (cluster 11) and the indicated cell clusters in epicardioids at day 15 and 30. P-values are indicated by circle size, the mean of the expression level of the two interacting molecules in the respective clusters are indicated by color. C: Violin plots showing the expression levels of markers of human compact (FTH1, FTL, COL3A1) and trabecular (COL2A1, TTN, MALAT1) myocardium in the indicated ventricular cardiomyocyte (CM) clusters. CMs clustering based on high cell cycle activity rather than identity (clusters 6 and 7) were excluded. Dashed lines show the distinction between clusters displaying compact (9, 1), trabecular (2, 0), and intermediate (3, 4) transcriptional profiles.

Figure 11: Cell-cell communication in epicardioids. Dot plot showing ligand-receptor interactions between epicardium-derived cells (cluster 5) and the indicated cell clusters in epicardioids at day

Y1 15 and 30. P-values are indicated by circle size, the mean of the expression level of the two interacting molecules in the respective clusters are indicated by color.

Figure 12: Epicardioids display morphological and functional self-patterning of the myocardium. A,B: Representative images of immunostaining for cardiac troponin T (cTnT) in human ventricular tissue at 6 weeks post conception (wpc) (A) and in noRA spheroids and epicardioids at day 15 (B). The hatched lines mark the outer edge of the myocardium; the dotted lines mark the delimitation between the outer myocardium (OM, defined as the myocardium within 50 pm of the outer edge across all samples) and the inner myocardium (IM). Scale bars in (A) - 500 pm, (B) top - 100 pm, bottom - 50 pm. C: (Left) Density of cardiomyocytes (CMs) in the OM and IM of noRA spheroids and epicardioids at day 15. Dot plots showing all data points; lines connect the values for OM and IM within the same sample. N - 15 per group; 2-way ANOVA matching paired data with Sidak's multiple comparison test. (Right) Corresponding ratio of CM density between OM and IM; box plots indicate the median, 25th and 75th percentile, with whiskers extending to the Sth and 95th percentiles. N - 15 per group; unpaired two-tailed t-test. D: Representative images of immunostaining for cTnT and the cell cycle activity marker Ki67 in noRA spheroids and epicardioids at day 15. Scale bars - 50 pm. E: (Left) Percentage of Ki67+ 15 CMs in the OM and IM of noRA spheroids and epicardioids at day 15. Dot plots showing all data points; lines connect the values for OM and IM within the same sample. N - 12 per group; 2-way ANOVA matching data with Sidak's multiple comparison test. (Right) Corresponding ratio of the percentage of Ki 67+ CMs between OM and IM; box plots indicate the median, 25th and 75 ^th percentile, with whiskers extending to the Sth and 95th percentiles. N - 12 per group; unpaired two tailed t-test. F: Graphical representation of the 3D differentiation of hiPSCs expressing a FRETbased voltage indicator (AAVS1-CAG-VSFP) followed by vibratome cutting of 250 pm-thick slices at day 30 for optical action potential (AP) measurement at day 35. G: Representative map of action potential duration to 50% repolarization (APD50) in a noRA spheroid and epicardioid at day 35. H: Quantification of APD50 (left) and APD90 (right) in the OM and IM of noRA spheroids and epicardioids under 0.5 Hz pacing at day 35. Box plots indicate the median, 25th and 75 ^th percentile, with whiskers extending to the Sth and 95th percentiles. N - 60 APs per layer from 3 noRA spheroids; N - 100 APs per layer from 6 epicardioids; Kruskal-Wallis test with Dunn's multiple comparison test. I: Graphical representation of Ca2+ transient imaging in 250 pm-thick slices cut by vibratome at day 30 and loaded with the fluorescent C 5 a2+ indicator Fluo-4 at day 35. J: Quantification of the time to 50% or 90% peak decay (TD50, left and TD90, right) in the OM and IM of noRA spheroids and epicardioids under 0.5 Hz pacing at day 35. Box plots indicate the median, 25th and 75th percentile, with whiskers extending to the Sth and 95th percentiles. N - 75 transients per layer from 4 noRA spheroids; N - 200 transients per layer from 9 epicardioids; Kruskal-Wallis test with Dunn's multiple comparison test.

Figure 13: The epicardial secretion of IGF2 promotes myocardial proliferation and compaction. A: Representative images of immunostaining for cardiac troponin T (cTnT) and IGF2 (left) or IGF1R (right) in day-15 epicardioids. Scale bar left - 50 pm, right - 100 pm. B: Epicardioids were treated with DMSO or the IGF1R inhibitor Lintisinib (0.25, 0.5 or 1 pM) from days 11 to 15. C: Representative images of immunostaining for cardiac troponin T (cTnT) and the cell cycle activity marker Ki67 in day-15 epicardioids treated with DMSO or Linsitinib. The hatched lines mark the outer edge of the myocardium; the dotted lines separate the outer myocardium (OM) and the inner myocardium (IM). Scale bars - 50 pm. D,E: Percentage of Ki 67 ⁺ cardiomyocytes (CMs) in the OM and IM of day-15 epicardioids treated with DMSO or Linsitinib. Dot plot showing all data points; lines connect the values for OM and IM within the same sample. Two-way ANOVA with Tukey's multiple comparisons test; asterisks indicate the p-values obtained from comparing the OM or IM of Linsitinib-treated samples with the corresponding layer of DMSO controls. +p < 0.05; *p < 0.01; **p < 0.001; ***p < 0.0001. (D) Ratio of CM density between the OM and IM in day-15 epicardioids treated with DMSO or Linsitinib. Mean ± SEM; one-way ANOVA with Tukey's multiple comparisons test. (E) DMSO: N - 7 epicardioids; Linsitinib: N - 6 epicardioids per condition; 3 independent differentiations per group. F: Spheroids differentiated without retinoic acid (noRA), lacking an epicardial layer, were treated with DMSO or recombinant human IGF2 (25, 50 or 100 ng/mL) from days 11 to 15. G: Representative images of immunostaining for cardiac troponin T (cTnT) and Ki67 in DMSO- or IGF2-treated noRA spheroids at day 15. The hatched lines mark the outer edge of the myocardium; the dotted lines separate the outer myocardium (OM, defined as the myocardium within 50 pm of the outer edge) and the inner myocardium (IM). Scale bars - 50 pm. H: Percentage of Ki67 ⁺ cardiomyocytes (CMs) in the OM and IM of DMSO- and IGF2-treated noRA spheroids at day 15. Dot plot showing all data points; lines connect the values for OM and IM within the same sample. Two-way ANOVA with Tukey's multiple comparisons test; *p < 0.05; **p < 0.001; ***p - 0.0001; asterisks indicate the p-values obtained from comparing the OM or IM of IGF2-treated samples with the corresponding layer of DMSO controls. N - 7 spheroids from 3 independent differentiations per group. I: Ratio of CM density (CMs/mm ² ) between the OM and IM in DMSO- and IGF2-treated noRA spheroids at day 15. Mean ± SEM; one-way ANOVA with Tukey's multiple comparisons test. N - 7 spheroids from 3 independent differentiations per group.

Figure 14: Epicardial NRP2 signaling regulates epicardial EMT. A: Representative image of immunostaining for cTnT and NRP2 in day-15 epicardioids, showing signal in the mesothelial epicardium layer (arrowhead) and in the myocardium (asterisk) but not in cells in the subepicardial space (arrow). Scale bar - 50 pm. B: Epicardioids were treated with DMSO or an NRP2 blocking antibody (ab; 200 or 500 pg/mL) from days 11 to 15. C: Representative images of immunostaining for cardiac troponin T (cTnT) in day-15 epicardioids treated with DMSO or an NRP2 blocking antibody. Pink lines indicate the maximum epicardium thickness of each epicardioid. Scale bars - 100 pm. D: Maximum epicardium thickness (pm) in day-15 epicardioids of each condition. Mean ± SEM; one-way ANOVA with Tukey's multiple comparisons test. N - 10 epicardioids from 3 independent differentiations per group. E: Representative images of immunostaining for cTnT, E-cadherin (CDH1) and vimentin (VIM), cTnT and BNC1 or cTnT, CDH1 and TWIST1 in day-15 epicardioids treated with DMSO or 500 pg/mL NRP2 blocking antibody from days 11 to 15. Scale bars = 100 pm. F: Percentage of CDHl /TWISTV, CDH1 ⁺ /TWIST1 ⁺ , CDH1 ⁺ /TWIST1 ^_ , and CDHl /TWIST cells in the epicardial layer (comprising mesothelial epicardium and EPDCs) of day-15 epicardioids treated with DMSO or NRP2 blocking antibody. Mean ± SD; unpaired two-tailed t-test. *p < 0.05; **p < 0.001 N - 9 epicardioids from 3 independent differentiations per group. G: Representative images of immunostaining for cTnT and Ki67 in day-15 epicardioids treated with DMSO or 500 pg/mL NRP2 blocking antibody from days 11 to 15. Scale bars - 100 pm. H: Percentage of Ki67 ⁺ and Ki67“ epicardial cells in day-15 epicardioids treated with DMSO or 500 pg/mL NRP2 blocking antibody from days 11 to 15. Mean ± SD. N - 9 epicardioids from 3 independent differentiations per group. I: The NRP2 blocking antibody was targeted with a secondary antibody alongside immunostaining for cTnT and CDH1 in day-15 epicardioids treated with 500 pg/mL NRP2 blocking antibody from days 11 to 15. Scale bar - 100 pm.

Figure 15: Epicardioids are formed by early segregation of first heart field (FHF) and juxtacardiac field (JCF) progenitors. A: UMAP dimensional reduction plot of the 31 scRNA-seq clusters obtained based on gene expression in (early) epicardioids on days 2, 3, 4, 5, 7, 10, and 15 of differentiation, separated by sample (top) and merged (bottom). Cluster numbers and main cell types are annotated. FHF: first heart field; JCF: juxta-5 cardiac field; EC: endothelial cells; vCM: ventricular cardiomyocytes; Epi: epicardium; FB: fibroblasts; SMC: smooth muscle cells. B: UMAP plot showing cells co-expressing markers of the juxta-cardiac field (JCF; HAND1, MAB21L2, HOXB6, HOXB5, BNC2) in red. C: Computational reconstruction of the developmental trajectories of cells from day 2 to 15 by URD analysis. The indicated clusters corresponding to pre-JCF, JCF, epicardial, and myocytic lineages are highlighted. The asterisk indicates cells segregating based on high mitotic activity. Circles highlight putative unipotent and bipotent JCF populations as well as cardiomyocyte (CMs) appearing downstream of the epicardial lineage. D: Representative images of immunostaining for the cTnT and ISL1 in epicardioids at days 7, 10, and 15. (left) Representative images of immunostaining for cTnT and ISL1 in 15 ventricular tissue from human embryos at 5 and 6 weeks post conception (wpc) (right). Scale bars - 50 pm. E: UMAP visualization of cell clusters identified based on chromatin accessibility via scATAC-seq analysis of epicardioids at days 3, 4, and 5. Main cell types are annotated; uncl.: unclassified. F: UMAP visualization of cells matching the (pre-)JCF signature derived from scRNA-seq analysis (labelled in red) among the scATAC-seq clusters presented in D. Negative 20 cells are shown in grey. G: Heatmap showing average Z-scores of transcription factor motif accessibility across the indicated scATAC-seq clusters. Clusters and motifs relevant to epicardial and myocytic commitment were selected (full heatmap presented in Fig. 20B); defining transcription factors are highlighted. H: UMAP visualization of the gene set activity (AUC) of the MEIS1 and TFAP2A regulons among scRNA-seq clusters at days 3, 4, and 5. Dotted circles highlight FHF and pre-JCF clusters. I: Average coverage plots of KRT7 and NKX2.5 in the scATAC-seq clusters 12, 2, 8, 9, and 3, illustrating potential for epicardial and myocytic fates, respectively.

Figure 16: Transcriptional dynamics during epicardioid development. A: UMAP plot showing the temporal distribution of the 31 clusters obtained by scRNA-seq analysis of epicardioids at days 2, 3, 4, 5, 7, 10, and 15 of differentiation. B: Heatmap showing the average expression of markers of the indicated cell types in all clusters. FHF: first heart field; vCM: ventricular cardiomyocytes; JCF: juxta-cardiac field; EPDCs: epicardium-derived cells; FB: fibroblasts; SMC: smooth muscle cells; EC: endothelial cells. C,D: UMAP plots showing the expression levels of markers of the FHF (TBX5) (C) and anterior SHF (TBX1, FGF8, FGF10) (D) across all time points.

Figure 17: Developmental trajectories of cells during epicardioid development. (Left) Computational reconstruction of the developmental trajectories of cells from day 2 to day 15 by URD analysis, with cells colored by day. (Right) Corresponding plot showing cells colored according to their relative developmental stage in pseudotime.

Figure 18: Progenitors of the JCF and FHF show spatial segregation in early-epicardioids. A: Violin plots showing the expression levels of ISL1 and NKX2.5 in the indicated FHF and pre-JCF scRNA- seq clusters, predominantly present at the indicated days. (B) Representative images of immunostaining for ISL1 and NKX2.5 in epicardioids at days 4 (top) and 5 (bottom). Arrowheads show examples of a layer of ISL1+/NKX2.5- JCF progenitors segregating from ISL1+/NKX2.5+ FHF progenitors. For each day scale bars top - 200 pm, bottom - 50 pm.

Figure 19: ISL1 is a marker of the human JCF and is temporarily maintained in mesothelial epicardium. A: Violin plots showing the expression levels of ISL1 and TNNT2 in the indicated ventricular cardiomyocyte (CM), juxta-cardiac field (JCF) and epicardial (epi) clusters obtained from scRNA-seq analysis of epicardioids from day 2 to day 15; clusters are predominantly present at the indicated days. B: Violin plots of the expression level of ISL1 in the indicated clusters obtained from scRNA-seq analysis of epicardioids at day 15 and 30; cells from day 15 and 30 are shown separately.

Figure 20: Chromatin accessibility profiling by scATAC-seq in early-epicardioids. A: Violin plots showing the predicted expression of markers of cardiac mesoderm (MESP2), cardiomyocytes (TNNT2) and endoderm (FOXA2) in the 13 clusters obtained by unsupervised clustering analysis of scATAC-seq data from epicardioids at days 3, 4, and 5. B: Heatmap showing average Z-scores of transcription factor motif accessibility across the indicated scATAC-seq clusters. Motifs relevant to epicardial and myocytic commitment are framed in blue and presented in Fig. 15G for selected clusters.

Figure 21: Activity of the MEIS1 and TFAP2A regulons in early-epicardioids. A,B: Heatmap showing the level of expression of members of the MEIS1 (A) and TFAP2A (B) regulons in all clusters obtained by scRNAseq analysis of epicardioids from days 2 to 15. Figure 22: Generation and validation of the AAVSl-CAG-FRT-flanked ST0P-mKate2-HA hiPSC reporter line. A: Schematic of the AAVS1 locus targeted with the CAG-FRT-flanked ST0P-mKate2- HA construct. Vertical arrows indicate sgRNA targeting sites. Horizontal arrows represent PCR primers for genotyping AAVS1 locus targeting and homozygosity, respectively. Blue boxes indicate PPP1R12C exons; green boxes indicate the regions of homology. B: Representative PCR genotyping of selected hiPSC clones by amplification of the normal (Pl and P2) and targeted alleles (Pl and P3). C: Normal karyotype confirmed after CRISPR/Cas9 editing in hiPSCs of clone #5 used in further experiments. D: Sanger sequencing of the top 3 predicted off-target sites, confirming no edits at these sites in reporter hiPSCs. E,F: Schematic of the CAG-Puro-2A-FLPo vector used for validation of reporter expression. Puromycin is used to select cells with random integration of the vector; flippase (FLPo) expression mediates the excision of the STOP cassette (FRT-flanked neomycin) in the targeted AAVS1 locus, allowing mKate2 expression. (E) Immunostaining for the HA-tag and detection of endogenous mKate2 after transfection with the CAG-Puro-2A-FLPo vector and 10 days of puromycin treatment confirmed that the FLP-FRT system works correctly in reporter hiPSCs. Scale bar - 25 pm. (F)

Figure 23: Validation of the fate potential of JCF and mesothelial epicardial cells through lineage tracing. A: Schematic representation of the lentiviral vector used for lineage tracing of JCF cells in epicardioids derived from AAVSl-CAG-FRT-flanked STOP-mKate2-HA reporter hiPSCs, consisting of the sequence encoding Tamoxifen-inducible flippase (FLP ^ERT2 ) driven by the MAB21L2 promoter. cPPT: central polypurine tract; WPRE: woodchuck hepatitis virus posttranscriptional regulatory element. B: (Left) Representative images of immunostaining for the HA-tag and ISL1 in epicardioids at day 9, after infection with the MAB21L2-FLP ^ERT2 lentivirus at day 3 and application of hydroxtamoxifen (4-OHT) days 7-8. The arrowhead shows an exemplary HA-tag ⁺ ISL1 ⁺ JCF cell at the outer layer. (Right) Corresponding uninfected negative control. C: (Left) Schematic of the experimental protocol used for lineage tracing of JCF cells. Epicardioids differentiated from AAVSl-FRT-flanked STOP-mKate2-HA reporter hiPSCs were transduced with a lentiviral vector encoding inducible flippase (FLP ^ERT2 ) under the control of the MAB21L2 promoter at day 3 and treated with 4-hydroxytamoxifen (4-OHT) at days 7 and 8 before analysis at day 12. (Middle) Representative images of immunostaining for cTnT, E-cadherin (CDH1) and the HA-tag in infected organoids at day 12. Filled arrowhead: exemplary HA-tag ⁺ mesothelial epicardial cell (epi), empty arrowhead: exemplary HA-tag ⁺ epicardium-derived cell (EPDC), arrow: exemplary HA-tag ⁺ cardiomyocyte (CM). Scale bar - 50 pm. (Right) Percentage of epi, EPDCs and CMs among HA-tag ⁺ cells. Mean ± SEM; N - 14 epicardioids from 3 independent differentiations. D: (Left) Schematic of the experimental protocol used for lineage tracing of early MAB21L2 ⁺ cells. Epicardioids differentiated from AAVSl-FRT-flanked STOP-mKate2-HA reporter hiPSCs were transduced with a lentiviral vector encoding inducible flippase (FLP ^ERT2 ) under the control of the MAB21L2 promoter at day 3 and treated with 4-OHT at days 4 and 5 before analysis at day 12. (c) (Middle) Representative images of immunostaining for cTnT, E-cadherin (CDH1) and the HA-tag in infected organoids at day 12. Filled arrowhead: exemplary HA-tag ⁺ mesothelial epicardial cell (epi), empty arrowhead: exemplary HA-tag ⁺ epicardium-derived cell (EPDC), arrow: exemplary HA-tag ⁺ cardiomyocyte (CM). Scale bar - 50 pm. (Right) Percentage of epi, EPDCs and CMs among HA-tag ⁺ cells. Mean ± SEM; N - 12 epicardioids from 3 independent differentiations. E: Schematic representation of the lentiviral vector used for lineage tracing of mesothelial epicardial cells in epicardioids derived from AAVSl-CAG-FRT-flanked STOP-mKate2-HA reporter hiPSCs, consisting of the sequence encoding flippase (FLP) driven by the CDH1 promoter. F: (Left) Representative images of immunostaining for the HA-tag and CDH1 in epicardioids at day 18, 72 h after infection with the CDH1-FLP lentivirus. The arrowhead shows an exemplary HA-tag ⁺ CDH1 ⁺ epicardial cell at the outer layer. (Right) Corresponding uninfected negative control. G: (Left) Schematic of the experimental protocol used for lineage tracing of mesothelial epicardial cells. Epicardioids differentiated from AAVSl-FRT-flanked STOP-mKate2-HA reporter hiPSCs were transduced with a lentiviral vector encoding flippase (FLP) under the control of the CDH1 promoter at day 15 before analysis at day 24. (Right) Representative images of immunostaining for the HA-tag and the CM marker cardiac troponin T (cTnT) (top) or the HA-tag and the FB marker vimentin (VIM) and the SMC marker calponin (CNN1) (bottom) in infected organoids at day 24. Scale bars - 50 pm. Insets show exemplary marked CM, SMC, and FB at higher magnification; scale bars - 10 pm.

Figure 24: Single-cell whole-transcriptome and chromatin accessibility analyses uncover transcriptional programs guiding fate decisions along the epicardial lineage tree. A: Computational reconstruction of the developmental trajectories of the JCF and epicardial scRNAseq clusters 27, 21, and 23 on days 7, 10, and 15 by URD analysis. B,C: URD plots showing the level of expression of the JCF marker HOXB6, 5 the mesothelial epicardium marker KRT19, ISL1, (C) the myocytic marker MYH7, the epicardium-derived cell (EPDC) marker EMILIN1, the fibroblast marker DDR2, and the smooth muscle marker TAGLN (D) along the trajectories presented in B. Arrows highlight branches with the indicated cell identities. The asterisk indicates cells segregating based on high mitotic activity. D: UMAP visualization of cell clusters identified based on chromatin accessibility after subclustering of the epicardial clusters 5 and 6 obtained via scATAC-seq analysis of epicardioids at days 7, 10, and 15. Main cell types are annotated. Epi: epicardium; EMT: epithelial-to-mesenchymal transition; EPDCs: epicardium-derived cells; JCF: juxta-cardiac field. E: Monocle-generated plot showing the pseudotime ordering and differentiation trajectories of the scATAC-seq clusters presented in (E). Top: cells are colored according to increasing maturation level from dark to light (left) or cluster identity (right). Bottom: plots are colored by cluster identity and split by time point. F: Monocle-generated plots showing the pseudotime-ordered inferred gene activity of markers of the indicated cell states (underlined). Shades of blue indicate reduced gene activity and shades of red indicate increased gene activity. JCF: juxta-cardiac field; epi: epicardium; EMT: epithelial-to-mesenchymal transition; EPDCs: epicardium-derived cells; SMCs: smooth muscle cells. G: Representation of cell states A-F along the pseudotrajectory based on the developmental progression of cells in the epicardial subclusters 0, 1, 2, 3 at days 7, 10, and 15. H: (Left) Percentage of cells showing inferred gene activity for the fibroblast marker DDR2, the smooth muscle cell marker MYH11, and the cardiomyocyte marker TNNT2 and the indicated combinations in each cell state defined in (G). (Right) Percentage of cells showing inferred gene activity for the fibroblast marker DDR2, the smooth muscle cell marker MYH11, the cardiomyocyte marker TNNT2 and the indicated combinations of genes in cell state A at days 7, 10, and 15.

Figure 25: Chromatin accessibility profiling of the epicardial lineage via single-cell ATAC-seq in epicardioids at days 7, 10, and 15. A: UMAP visualization of cell clusters identified based on chromatin accessibility via scATAC-seq analysis of epicardioids at days 7, 10, and 15. Main cell types are annotated. CMs: cardiomyocytes, Epi: epicardium, EPDCs: epicardium-derived cells. B: UMAP visualization of cells matching the epicardial signature derived from scRNA-seq analysis (labelled in red) among the scATACseq clusters presented in (A). Negative cells are shown in grey. (C) Bar chart of gene ontology (GO) categories identified by gene set enrichment analysis (GSEA) of clusters 0, 1, 2, and 3 obtained by subclustering of the epicardial scATAC.seq clusters 5 and 6 (FDR < 0.001). Bars represent the number of genes present in each category.

Figure 26: Chromatin accessibility dynamics infer the fate trajectories of epicardial cells. Monoclegenerated plots showing the pseudotime-ordered inferred gene activity of markers of the indicated cell states (underlined) in the epicardial subclusters 0, 1, 2, 3 at days 7, 10, and 15. Shades of blue indicate reduced gene activity and shades of red indicate increased gene activity. JCF: juxta-cardiac field; epi: epicardium; EMT: epithelial-to-mesenchymal transition; EPDCs: epicardium-derived cells; SMCs: smooth muscle cells.

Figure 27: Endothelin-1 stimulation triggers hypertrophic and fibrotic responses in epicardioids. A: mRNA expression of NPPA, NPPB, ACTA1, and MYH7/MYH6 ratio relative to GAPDH in epicardioids treated with 25 nM or 50 nM endothelin-1 (ET1) compared to untreated controls. Mean ± SEM; N - 3 per group; one-way ANOVA with Sidak's multiple comparison test. B: (Top) Representative images of cardiomyocytes (CMs) re-seeded after dissociation of untreated or ET1- treated epicardioids and stained for cardiac troponin T (cTnT) and the desmosomal marker plakophilin-2 denoting cell membrane (PKP2). Scale bars - 100 pm. (Bottom) Corresponding quantification of CM area, box plots indicate the median, 25th and 75th percentile, with whiskers extending to the Sth and 95th percentiles. N - 260 CMs from 3 differentiations per group; unpaired two-tailed t-test. C: mRNA expression of COL1A2, COL3A1, FN1, and POSTN relative to GAPDH in epicardioids treated with 25 nM or 50 nM ET1 compared to untreated controls. Mean ± SEM; N - 3 per group; one-way ANOVA with Sidak's multiple comparison test. D: Representative images of immunostaining for cTnT and fibronectin in untreated or ETl-treated epicardioids. Scale bars - 200 pm. The square highlights the outer layer of ETl-treated 15 epicardioids further analyzed in (E). E: Representative images of immunostaining for cTnT, a-smooth muscle actin (ct- SMA), fibronectin (FN1), and collagen type III (Coll) in ETl-treated epicardioids. Scale bars - 50 p.m. F: Exemplary traces of Fluo-4 fluorescence in untreated or ETl-treated epicardioids. Blue arrows indicate 0.5 Hz pacing, red arrows indicate examples of arrhythmic events. G: Amplitude of Ca2+ transients in the outer (OM) or inner myocardium (IM) 20 of untreated or ETl-treated epicardioids, box plots indicate the median, 25th and 75th percentile, with whiskers extending to the Sth and 95th percentiles. Untreated: N - 130 transients/layer from 5 epicardioids, ET1: N - 106 transients/layer from 4 epicardioids; Kruskal-Wallis test with Dunn's multiple comparison test. H: Percentage of untreated or ETl-treated epicardioids displaying arrhythmic events in the OM or IM. Mean ± SEM; untreated N - 5 epicardioids, ET1: N - 4 epicardioids; unpaired two-tailed t-test.

Figure 28: Epicardioids recapitulate the myocardial hyperproliferation and fibrosis typical of cardiomyopathy associated with Noonan syndrome. A: Representative images of immunostaining for cTnT, E-cadherin (CDH1) and VIM in day-15 epicardioids generated from the hiPSCs of a patient with Noonan syndrome (NS) carrying a PTPN11 ^N3O8S/+ mutation. Scale bars top - 100 pm, bottom - 50 pm. The arrowhead indicates the mesothelial epicardium layer, the arrow indicates subjacent EPDCs. B: (Left) Representative images of cardiomyocytes (CMs) re-seeded after dissociation of day-30 control or Noonan syndrome (NS) epicardioids and stained for cardiac troponin T (cTnT) and the desmosomal marker plakophili n-2 (PKP2). Scale bars - 100 pm. (Right) Corresponding quantification of CM area, box plots indicate the median, 25 ^th and 75 ^th percentile, with whiskers extending to the 5 ^th and 95 ^th percentiles. Control: N - 379 CMs, NS: N - 360 CMs; 3 differentiations per group; unpaired two-tailed t-test. C: mRNA expression of the hypertrophic markers NPPA, NPPB, ACTA1, and MYH7/MYH6 ratio relative to GAPDH in day-30 control or NS epicardioids. Mean ± SEM; N - 5 epicardioids from 3 independent differentiations per group; unpaired two-tailed t-test. D: (Top) Representative images of immunostaining for cardiac troponin T (cTnT) and Ki67 in control and NS spheroids at day 15. The hatched lines mark the outer edge of the myocardium; the dotted lines separate the outer myocardium (OM, defined as the myocardium within 50 pm of the outer edge) and the inner myocardium (IM). Scale bars - 50 pm. (Bottom) Percentage of Ki67 ⁺ cardiomyocytes (CMs) in the OM and IM of control and NS epicardioids at day 15. Dot plot showing all data points; lines connect the values for OM and IM within the same sample. Control: N - 12 epicardioids; NS: N - 10 epicardioids; 3 independent differentiations per group; two-way ANOVA with Tukey's multiple comparisons test. E: mRNA expression of the extracellular matrix markers COL1A2, COL3A1, FN1, and POSTN relative to GAPDH in day-30 control or NS epicardioids. Mean ± SEM; N - 5 epicardioids from 3 independent differentiations per group; unpaired two-tailed t-test. F: Representative images of immunostaining for cTnT and FN1 in day-15 control and NS organoids. Asterisks indicate areas of fibrotic remodeling. Scale bars - 100 pm. G: Representative images of immunostaining for cTnT, FN1, ot-SMA, and collagen type III in day-15 NS epicardioids. Scale bars - 50 pm. Box plots in (b) and (g) indicate the median, 25 ^th and 75 ^th percentile, with whiskers extending to the 5 ^th and 95 ^th percentiles. DEFINITION OF TERMS

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

The singular forms "a", "an", and "the" include references to the plural unless explicitly stated otherwise in the context.

The term "comprising" means "including" as well as "consisting", e.g., a composition "comprising" X may consist exclusively of X or may include something additional, e.g., X + Y.

The term "epicardioid" is used herein interchangeably with the term "heart tissue model," wherein the epicardioid has been isolated at day 13, preferably at day 15 or later of the culture, and wherein the epicardioid shows the disclosed characteristic structure. Epicardioid is the given name for the herein described mature heart tissue model. The term "epicardioid culture" refers to the cell culture in which an epicardioid is about to develop or has already been developed. In general, organoids are known to be miniaturized and simplified versions of an organ produced in vitro in three dimensions that mimic the key functional, structural and biological complexity. The term epicardioid thus refers to the words organoid and the epicardial compartment of the heart.

The term "epicardium-derived cells" describes any cell derived, directly or indirectly, from the epicardium including EPDCs but also differentiated cells such as fibroblasts. In contrast or addition thereto, the term "EPDCs" is directed to a specific cell population derived from the epicardium excluding further differentiated cells such as fibroblasts, mural cells and the like.

The term "early-epicardioid" is used for epicardioids which are isolated from culture at day 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 of the culture, preferably between day 6 and 10 of the culture, i.e., prior to the day of full epicardioid development at day 13 of the culture. Such early-epicardioids are nevertheless also heart tissue models of the present invention.

The term "pre-JCF cells" describes the cell population that gives rise to the JCF cells in the present epicardioid system. Pre-JCF cells can be interchangeably used with the term "JCF progenitor cells" or "JCF progenitors".

"Myocardial cells" are cardiomyocytes. The terms are used interchangeably.

The term "about" means, in terms of numbers, concentrations, times (minutes, hours, days) and the like, that +/- 10% of the given number, concentration, time (minutes, hours, days) and the like is also comprised. For example, if a certain agent is present in an amount of about 2 pM to about 4 pM in a medium, the agent may be present in an amount of 2 pM (+/- 10%) to 4 pM (+/-10%) in this medium, i.e. in an amount of at least 1,8 pM to at most 4,4 pM. The expression "densely packed" means a high density of cardiomyocytes (CM), on average about between 4000-8500 CM/mm ² , preferably about 6500 CM/mm ² .

The expression "looser packed" means a lower density of cardiomyocytes (CM), on average about between 3000-6200 CM/mm ² , preferably about 4500 CM/mm ² .

The term "compartment" means a defined sub-structure in a higher order structure comprising, for instance, characteristic cell types or a certain pattern or layers of cells, which sub-structure differs from other sub-structures within the higher order structure (here within the epicardioid).

The feature of a "significantly higher" proportion of cells of a specific cell type, e.g., Ki67+ cardiomyocytes, for instance, in a compartment or layer, means that cell numbers of this cell type statistically significantly differ from cell numbers of the same cell type in, e.g., another compartment or layer. In embodiments this means that cell numbers of a certain cell type are at least 40%, 50%, 60%, 70%, 80%, 90%, 100%, 150%, 200%, 250%, 300% or 350% higher compared to cell numbers of the same cell type in, e.g., another compartment or layer.

The term "x days of culture", the expression "at day x", or "at day x of differentiation" can be used interchangeably and mean that the epicardioids were isolated and analyzed after x days of culture.

DETAILED DESCRIPTION OF THE INVENTION

The epicardium is the mesothelial cell sheet covering the heart's outer surface. Long considered a simple barrier between the pericardial cavity and the myocardium, it is now recognized to hold key functions in cardiac development and repair. During embryonic development, a subset of epicardial cells undergoes epithelial-to-mesenchymal transition (EMT) to become epicardial derived cells (EPDCs) that migrate into the myocardium and give rise to the majority of fibroblasts and vascular smooth muscle cells of the heart. Whether EPDCs also differentiate into cardiomyocytes and coronary endothelial cells is still debated, with studies providing conflicting evidence. In addition to these cellular contributions, the epicardium provides signaling factors critical for the development and growth of the myocardium. It also plays a central role in heart regeneration in species capable of rebuilding adult heart muscle upon injury such as the zebrafish, making it a highly promising target for therapy. However, the inaccessibility of human embryonic tissue at the early stages of epicardium development, which begins less than four weeks postconception, has left significant gaps in our understanding of human epicardial development and function. Many questions on the ontogeny of human proepicardial precursors and the functional heterogeneity of epicardial cells are still unresolved, which limits harnessing their full potential for regenerative medicine. Pluripotent stem cell-derived cardiac organoids have emerged as powerful in vitro models of human development and disease, but none have yet demonstrated the formation of a bona fide epicardial compartment. The present invention provides cardiac organoids showing selforganization of highly functional ventricular myocardium and epicardium, which were thus called epicardioids. As evidenced in the disclosed Experiments, time course single-cell genomics in epicardioids revealed principles of human epicardial biology, including lineage heterogeneity and functional crosstalk with other cardiac cell types. In addition, the herein disclosed epicardioids represent an advanced system to model multicellular mechanisms of heart disease.

The invention is described in detail below. All embodiments can be combined with each other, except where otherwise stated.

The inventions disclosed herein do not concern processes for modifying the germline genetic identity of human beings. The inventions disclosed herein do not concern uses of human embryos for industrial, commercial or any other purposes. The inventions disclosed herein do not concern the human body at the various stages of its formation. The pluripotent stem cells used herein for the production of the epicardioids and thus used in the inventive methods are not obtained by means of a process in which human embryos are destroyed.

Epicardioids

The invention relates to a heart tissue model named epicardioid. Such epicardioid has a three- dimensional, sphere-like shape comprising two different compartments, the inner myocardial compartment forming the "core" of the epicardioid, and the outer epicardial compartment, wrapped around the core, i.e., forming an envelope of the core. Each compartment comprises different cell layers (see Fig. 1). The outer epicardial compartment comprises from the outside to the inside a layer of epicardium comprising KRT18+ mesothelial epicardial cells, and a layer of subepicardium comprising Vim+ epicardial-derived cells (EPDCs).

The inner myocardial compartment comprises from the outside to the inside a layer of compact-like myocardium comprising densely packed cTnT+ myocardial cells, and a layer of trabecular-like myocardium comprising looser packed cTnT+ myocardial cells compared to the compact-like myocardium.

All cells present in the epicardioid have been developed from one initial in vitro cell population. This means, that no cells were added to the epicardioid at a later point after the initial cell population has been seeded for epicardioid generation according to the inventive methods. Therefore, a disclosed epicardioid is no artificial composition of different pre-existing cell types put together in vitro but resembles naturally developed embryonic heart tissue. cTnT, also known as "TNNT2", is the sarcomeric marker cardiac troponin T. Vim is the abbreviation for the mesenchymal marker vimentin. Although Vimentin is an established mesenchymal marker, it can be expressed in the epicardium.

The initial in vitro cell population is an in vitro pluripotent stem cell population comprised in cell culture medium, wherein the pluripotent stem cell population is preferably an induced pluripotent stem cell population, more preferably a human induced pluripotent stem cell population, or wherein the pluripotent stem cell population is preferably an embryonic stem cell population, more preferably a human embryonic stem cell population. In another embodiment, the pluripotent stem cell population is a human induced pluripotent stem cells reprogrammed from patients with genetic or complex cardiac disorders, including congenital heart diseases. In a preferred embodiment, the epicardioid has been developed from one single population of human induced pluripotent stem cells.

The trabecular-like layer may also be seen as a trabecular-like inner core of the three-dimensional epicardioid.

The cells of the outer epicardial compartment may express, in addition to KRT18, the epicardial transcription factors TBX18.

The cells present in the layer of compacted myocardium preferably comprise a significantly higher proportion of Ki67 ⁺ proliferating cardiomyocytes than the inner myocardium. Significantly higher in this specific case means that the proportion of Ki67+ cells in the outer compact layer is between about 100%-350%, more specifically between 129%-309% higher than the proportion of Ki67+ cells in the inner trabecular-like layer.

The inner myocardial compartment may further comprise in both layers non-cardiomyocyte cells comprising cardiac fibroblasts, vascular smooth muscle cells and endothelial cells. The endothelial cells may be CD31 ⁺ .

The two compartments of the epicardioid preferably match the multilayered structure of the ventricular epicardium and myocardium of early human embryos.

Use of epicardioids

The epicardioids disclosed herein may be used as a heart tissue model. Therefore, the present invention also relates to the use of the disclosed epicardioids as a heart tissue model.

Epicardioids are formed through processes that mimic the development of the left ventricle during embryonic development, including the specification of TBX5+ cells corresponding to first heart field (FHF) progenitors. To our knowledge, they are also the first human system showing the existence of juxta-cardiac field (JCF) cells recently discovered in the mouse as an early progenitor of both myocardium and epicardium (Tyser et al., (2021), "Characterization of a common progenitor pool of the epicardium and myocardium", Science, 80:eabb2986; Zhang et al., (2021) "Unveiling Complexity and Multipotentiality of Early Heart Fields", Circ. Res. 129:474- 487). The presence of the epicardium in epicardioids promotes morphological, molecular, and functional patterning of the myocardium that was previously not achieved in cardiac in vitro models. Specifically, as evidenced herein in the Examples, the epicardium stimulates cardiomyocyte proliferation via cell-cell interactions that have been described in human embryos.

Epicardioids therefore offer a unique platform to investigate mechanisms underlying epicardial lineage formation and the functional crosstalk between the epicardium and myocardium (e.g. via transcriptomic and proteomic analyses of epicardioids at relevant time points, lineage tracing studies with reporter hPSC lines, generation of epicardioids from genetically engineered hPSC lines to modify specific processes).

Notably, insights from these studies could lead to novel strategies to re-activate the epicardial capacity to stimulate cardiomyocyte proliferation, which is lost in adults, to replace cardiomyocytes lost after injury (e.g. myocardial infarction).

Therefore, in one aspect, the herein disclosed epicardioids are used for studying heart diseases. The inventive epicardioid may be used to recapitulate specific normal or abnormal conditions of the heart, such as diseases, illnesses or injuries, as well as a recovery from such conditions.

In one embodiment, the heart disease is a left ventricular heart disease.

In another preferred embodiment, the herein disclosed epicardioids are used for preclinical testing of drugs for the treatment of heart diseases. In another preferred embodiment, the herein disclosed epicardioids are used for preclinical testing of compounds, drugs or other substances for the treatment of heart diseases. Such heart diseases may be selected from hypertrophic, fibrotic, infective, or arrhythmic or other heart diseases. In even another embodiment, the epicardioids are used for preclinical testing of drug-induced arrhythmias.

In an embodiment, the compound to be tested is added to the culture during culture days 1 to 13. In another embodiment, the compound to be tested is added on or after day 13. In a preferred embodiment, the compound to be tested is added on or after day 15 of culture. In an embodiment, the compound is added once, in another embodiment, the compound is added more than once, for instance, each day, each second day or each third day of the culture.

The compound to be added may be a drug, a pharmaceutical, a small molecule, an antibody, a non-pharmaceutical therapeutic agent comprising genetic material (DNA or RNA) and CRISPR/Cas9 components or other gene editing or modulation agents. CRISPR/Cas9 components may be added in form of mRNA, DNA, or proteins which are preferably delivered to the epicardioid or early-epicardioid via viral vectors or lipid particles. Thus, the treatment of the heart disease may be a gene therapy or gene modulation treatment. Therapeutic gene modulation differs from gene therapy in that gene modulation seeks to alter the expression of an endogenous gene whereas gene therapy concerns the introduction of a gene whose product aids the recipient directly.

Moreover, also encompassed by the invention is the use of epicardioids generated from human induced pluripotent stem cells reprogrammed from patients with genetic or complex cardiac disorders, comprising congenital heart diseases, for studying the mechanisms of disease and for testing drugs to rescue disease phenotypes. The epicardioids can thus be used for testing genetic alterations, for instance by observing the effects of mutated (e.g., disease-related mutations), suppressed or over-expressed genes during heart development wherein the testing comprises the production of a heart tissue model (epicardioid) according to the disclosed method, wherein the cells have a suppressed candidate gene or overexpress a candidate gene or have a mutation in a gene and comparison of the development and/or functionality of the heart tissue model with the development of a heart tissue model not having such genetic alteration. Mutation, overexpression or suppression of expression can be done with any known method in the art, such as suppression by gene knock out, siRNA inhibition or CRISPR/Cas-based inactivation. Overexpression can be done by introduction of a transgene or applying a gene activator that leads to upregulation of expression of a gene.

In another embodiment, the herein disclosed epicardioids are used for investigating disease states and their treatment by application of pharmacological stimuli to epicardioids. In a preferred embodiment, the herein disclosed epicardioids are used to study left ventricular hypertrophy, wherein the epicardioids are treated with endothelin-1.

Candidate drugs or substances identified in epicardioids would need to be tested in an animal model of cardiac disease or injury (e.g. myocardial infarction). This may be achieved by delivery of recombinant proteins or small molecules, or by gene activation or inhibition via CRISPR/Cas9 technology. Different methods of delivery could be envisioned, including intrapericardial, intramyocardial, and endovascular injection and engraftment of a patch (e.g. collagen, fibrin) loaded with appropriate factors at the site of injury.

In addition, single-cell genomics (scRNA/ATAC-seq) combined with lineage tracing revealed that the disclosed epicardioids are formed by two main populations of progenitors: first heart field (FHF) cells differentiating into cardiomyocytes and juxta-cardiac field (JCF) cells giving rise to both cardiomyocytes and epicardial cells, respectively. A subset of epicardial cells then undergoes epithelial-to-mesenchymal transition to become epicardium-derived cells (EPDCs) that migrate into the myocardium and differentiate into fibroblasts, mural cells (smooth muscle cells and pericytes) and cardiomyocytes. Importantly, the inventors could show that the epicardial populations emerging within epicardioids better recapitulate the transcriptional profile of their in vivo counterparts compared to epicardial cells obtained in 2D. Therefore, epicardioids can be used as "living factories" for cell products with high biological functionality. In other words, epicardioids and early-epicardioids disclosed herein may be used as a cell source for cell-based therapeutic applications. Defined cell populations may be isolated and purified from the epicardioid or early-epicardioid. Cell populations to be isolated from a full mature epicardioid after culture day 13 comprise: JCF-derived cells comprising epicardium cells and cardiomyocytes, epicardium-derived cells comprising EPDCs, fibroblasts, mural cells (vascular smooth muscle cells and pericytes), and cardiomyocytes, or endothelial cells. JCF-derived epicardial cells isolated from an epicardioid are promising candidates for cell therapy, as they also possess myocytic differentiation potential.

The fate potential of the JCF cells makes them a highly potent candidate for cell therapy: proliferative pre-JCF cells may invade the injured tissue before repopulating it with cardiomyocytes and epicardial cells, the latter becoming a source of both coronary vessel components and supportive signaling factors. Therefore, cell populations may not only be isolated from the mature epicardioid, but also from the early-epicardioid starting from about day 4 of the epicardioid culture to about day 12 of the epicardioid culture. Cell populations to be isolated from an early-epicardioid at days 4 to 12 of the epicardioid culture thus comprise: FHF cells, pre-JCF cells, JCF cells, JCF-derived cells comprising epicardium cells and cardiomyocytes, and EPDCs.

Early-epicardioids may thus be used as a source for FHF cells, proliferative pre-JCF cells, JCF cells, and JCF-derived epicardial cells. Pre-JCF cells or JCF-derived cells obtained from an epicardioid disclosed herein may be for use in cell therapy. Cells isolated from early-epicardioids thus include FHF cells (isolatable at about days 4 to 5 of culture), pre-JCF cells (isolatable at about days 4 to 5 of culture), epicardial cells (isolatable at about days 10 to 12 of culture), EPDCs (isolatable at about days 10 to 12 of culture), FHF- and JCF-derived cardiomyocytes (isolatable at about days 7 to 12 of culture) and endothelial cells (isolatable at about days 5 to 12 of culture).

A further aspect of the invention thus relates to a cell population isolated from the disclosed epicardioids for use in therapy. Additionally, the invention relates to a cell population isolated from the disclosed early-epicardioid culture for use in therapy. In an embodiment, the cell population isolated from the disclosed epicardioids is for use in the treatment of heart diseases such as myocardial infarction, heart failure, and acquired or inherited forms of cardiomyopathy (including hypertrophic cardiomyopathy, dilated cardiomyopathy, arrhythmogenic cardiomyopathy, Duchenne muscular dystrophy, restrictive cardiomyopathy). In an embodiment, the cell population may be FHF cells (expressing NKX2.5 and TBX5), pre-JCF cells (expressing ISL1), JCF cells (expressing HAND1, HOXB6 and/or MAB21L2), mesothelial epicardial cells (expressing KRT18, KRT19, and/or CDH1), EPDCs (expressing VIM), fibroblasts (expressing COL1A1 and/or TNC), vascular smooth muscle cells (expressing RGS5 and/or MYH11), pericytes (expressing MCAM) , cardiomyocytes (expressing TNNT2), endothelial cells (expressing CD31 and/or CDH5). Methods of producing heart tissue models

The inventive tissue models, i.e., early-epicardioids or epicardioids are the results of the inventive methods disclosed herein, or can be used in the inventive methods and for the inventive uses.

All cell comparisons herein are under standard conditions for cell cultures for maintaining cells expressing relevant distinctive markers or signature marker expression, in particular so that the markers or marker expression can be determined as a cell-intrinsic property. Such conditions usually comprise a medium consisting of all required nutrients for maintenance under ambient pressure and at the cells' physiological temperature.

The inventive heart tissue model is artificial and grown in culture using natural principles of development. By following the inventive method steps the heart tissue model in form of an epicardioid is generated by self-organization driven by growth factor signaling.

The method of producing a three dimensional heart tissue model comprises the following consecutive steps:

First pluripotent stem cells are provided and seeded in an initial culture medium on day -1 of the culture.

The pluripotent stem cells (PSCs) are preferably from a mammal. Preferably the PCSs are human PSCs or non-human primate PSCs, or PCSs from a rodent, mouse, hamster, pig, cat, dog, horse, or cow. The PSCs may be from a cell line or a cell culture. Preferably, the PSCs are induced pluripotent stem cells (iPSCs). iPSCs stem from differentiated cells that were again turned pluripotent, such as by treatment with the Yamanaka factors. Such iPSCs are preferably human iPSCs (hiPSCs). Usage of iPSCs allows for the investigation of cardiac development of specific individuals, which may or may not have genetic aberrations that may affect heart development. The iPSCs may therefore be developed from cells of healthy or diseased subjects, preferably humans, wherein a diseases subject for instance suffers from a heart disease, in particular a genetic heart disease. Alternatively, the pluripotent stem cells are embryonic stem cells (ESCs), preferably human embryonic stem cells (hESCs) of a hESC cell line. Of note, when using ESCs, no embryos have been destroyed.

The PSCs are grown in a medium before they are provided for the steps of the inventive method. This initial culture medium is preferably Essential 8 (E8) medium. Preferably, the cells are cultured on Geltrex-coated plates in E8 medium.

At day -1 of the culture, the PSCs are detached from the culture plate, washed and counted. This PSC cell culture is then provided for the first step of the inventive method. The cells are seeded in an initial culture medium immediately after cell counting, i.e., on day -1. The initial culture medium used for cell seeding is preferably an E8 medium comprising thiazovivin. Preferably, the E8 medium comprises about 1 to about 5 pM thiazovivin, more preferably about 2 pM thiazovivin, most preferably 2 pM thiazovivin.

In an embodiment, the PSCs are seeded in low attachment culture vessels. Such vessels are preferably U-shaped or V-shaped. In a preferred embodiments the PSCs are seeded in wells of a U-shaped 96-well plate. Low attachment culture vessels prevent the PSCs from attaching to the surface of the bottom of the vessels and thus supports the cells to grow in spheres, i.e. it supports a 3D culture. Low attachment surfaces are preferably hydrophilic, neutrally charged or non-ionic. They may have a hydrogel layer or coating that repels cell attachment, forcing cells into a suspended state. In preferred embodiments, the vessels are coated with a low attachment agent or low cell adherence agent. Such agent is for instance poly-HEMA. In a preferred embodiment, the PSCs are seeded in U-shaped wells of a poly-HEMA-coated U-shaped 96-well plate.

A preferred starting cell number seeded to generate the heart tissue models of the present invention is between about 30,000 and about 40,000 cells, i.e., about 30,000, about 30,500, about 31,000, about 31,500, about 32,000, about 32,500, about 33,000, about 33,500, about 34,000, about 34,500, about 35,000, about 35,500, about 36,000, about 36,500, about 37,000, about 37,500, about 38,000, about 38,500, about 39,000, about 39,500, or about 40,000 cells. Too low numbers may lead to structures that do not start spontaneously beating, indicating failed cardiac differentiation. Of note, epicardioids of the present invention start spontaneously beating around day 8 of culture. On the other hand, structures generated from too high numbers may not acquire the correct morphology in the early stages of epicardioid formation.

In a preferred embodiment, 30.000-40.000 cells are seeded in 150 pl of culture medium in a U- shaped well of a poly-HEMA coated 96-well plate.

Day -1 means about 20-28 hours, preferably about 24 hours, before starting the cell differentiation steps.

On day 0 of the culture, the culture medium is replaced with a first basal differentiation medium.

Day 0 means 20-28 hours after seeding the PSCs. The culture is preferably replaced with the first basal differentiation medium about 24 hours after seeding, more preferably 24 hours after seeding.

The first basal differentiation medium preferably comprises bone morphogenetic protein 4 (BMP4), Activin A, fibroblast growth factor beta (bFGF, also called FGF-P or FGF2), a PI3-Kinase inhibitor, preferably LY-29004, and a Wnt activator, preferably CHIR-99021.

The first basal differentiation medium is preferably based on a basic basal differentiation medium, which is further supplemented with the components of the first basal differentiation medium. The basic basal differentiation medium preferably comprises a standard medium suitable for PSC differentiation cell culture, such as DMEM. Preferably DMEM is mixed about 1:1 with IMDM. Preferably the basic basal differentiation medium further comprises BSA, chemically defined lipid concentrate, transferrin, a-monothioglycerol. Preferably the basic basal differentiation medium further comprises about 1% chemically defined lipid concentrate, about 0,05% transferrin, about 0,04% a-monothioglycerol.

The first basal culture medium used for the first medium replacement step of the method, may be the basic basal differentiation medium described above, which is further supplemented with preferably BMP4, Activin A, bFGF, LY-29004, CHIR-99021, more preferably with about 10 ng/mL BMP4, about 50 ng/mL Activin A, about 30 ng/mL bFGF, about 5 pM LY-29004, and about 1,5 pMCHIR-99021.

On day 2 of the culture, the culture medium is replaced with a second basal differentiation medium comprising retinoic acid. The retinoic acid is preferably present at a concentration of 0,3- 0,75 pM, preferably of about 0,5 pM, preferably of 0,5 pM.

Day 2 of the culture means about 40 to 50 hours after the seeding of the cells, preferably, day 2 means about 40-42 hours after seeding.

The second basal differentiation medium preferably further comprises insulin, BMP4, bFGF, and a Wnt antagonist, preferably IWP2, more preferably about 10 pg/mL insulin, about 10 ng/mLBMP4, about 8 ng/mL bFGF and about 5 pM IWP2.

Preferably, the second basal differentiation medium is the basic basal differentiation medium further supplemented with about 0,5 pM retinoic acid, about 10 pg/mL insulin, about 10 ng/mLBMP4, about 8 ng/mL bFGF and about 5 pM IWP2.

This culture medium is replaced with fresh second basal differentiation medium each 24 hours until day 6 of the culture.

Next, on day 6 of the culture, the medium is replaced with a third basal differentiation medium.

Day 6 of the culture means about 134 to 144 hours after the seeding of the cells, preferably, day 6 means about 136-138 hours after seeding.

The third basal differentiation medium preferably comprises insulin, BMP4, and bFGF, more preferably about 10 pg/mL insulin, about 10 ng/mLBMP4, and about 8 ng/mL bFGF.

Preferably, the third basal differentiation medium is the basic basal differentiation medium further supplemented with about 10 pg/mL insulin, about 10 ng/mLBMP4, and about 8 ng/mL bFGF. This culture medium is replaced with fresh third basal differentiation medium on day 7 of the culture.

On day 8 of the culture, the spheroids are embedded into a gel comprising collagen type I. For this, a pipet tip may be used to transfer the spheres into molds in the gel. The molds are subsequently filled with collagen-l solution which was then allowed to solidify. This gel sheets are then transferred to maintenance medium, i.e., they are surrounded by maintenance medium.

Day 8 of the culture means about 184-194 hours after seeding of the cells, preferably about 184- 188 hours after seeding of the cells.

The maintenance medium preferably comprises insulin and VEGF, more preferably, about 10 pg/mL insulin and about 100 ng/mL VEGF.

Preferably, the maintenance medium is the basic basal differentiation medium further supplemented with about 10 pg/mL insulin and 100 ng/mL VEGF.

The maintenance medium is replaced with fresh maintenance medium every 2-3 days after day 8 of the culture.

Optionally, the gel sheets floating in maintenance medium are placed on a rocking shaker on day 10 of the culture.

Day 10 of the culture means about 230-240 hours after seeding the cells, preferably about 230- 236 hours after the seeding of the cells.

As a last step, the spheroids, which are now called epicardioids can be isolated from the gel sheets for further use. Isolation of epicardioids can take place starting on day 13 of the culture.

If the maintenance medium is replaced all 2-3 days of the culture after day 8, the epicardioids can be cultured for longer time, for instance until day 100 of the culture. Preferably, epicardioids are isolated between days 13 and 100 of the culture, more preferably, between days 15 and 40 of the culture.

During embryonic cardiac development, the epicardium secretes factors that stimulate cardiomyocyte proliferation, thereby promoting the formation of a compact layer that is essential for proper cardiac function. By following the disclosed method, this is recapitulated in the end product of the disclosed method, which is a heart tissue model named epicardioid, which consistently forms an inner myocardial core (the trabecular-like layer) an outer region of higher cardiomyocyte density of approximately 50 pm width (the compact-like layer).

The molecular processes taking place during initial epicardioid formation when following the herein disclosed method have been analyzed based on single-cell RNA sequencing at days 2, 3, 4, 5, 7, 10 and 15 (see Fig. 2A,B). This revealed early induction of cardiac mesoderm (marked by MESP1) followed by cardiac progenitor cells predominantly corresponding to the first heart field (marked by TBX5) (Fig. 2C). Cardiomyocytes started expressing the ventricular marker MYL3 from day 7; by day 15 most cardiomyocytes matched to a human ventricular signature (co-expression of MYH7, MYL2, S100A4, VCAN, MASP1) (Fig. 2C,D). On day 15, there was also a cluster matching a human epicardial signature (co-expression of SLILF1, TM4SF1, CCDC80, S100A10, CFI) (Fig. 2D).

In earlier epicardioids, a pro-epicardial progenitor population could be identified which corresponds to the juxta-cardiac field reported in mice as a common progenitor or the myocardium and epicardium (marked by HAND1, MAB21L2, H0XB6, H0XB5, BNC2) - this is the first time that this population has been described in a human system (Fig. 2D).

Based on these observations the present invention also relates to a method for screening or testing a candidate compound on its effects on heart development and/or functionality.

This screening method is based on the inventive culture method disclosed herein for generating a heart tissue model by performing the disclosed method steps while treating the cells with the candidate compound and comparing development of the heart tissue model with development and/or functionality of a heart tissue model that was not treated with the candidate compound. For a successful analysis all culture steps should be the same except the treatment with a candidate compound or drug.

The method may also comprise a step of causing an injury to the heart tissue model and optionally further a recovery from such injury. Agents or compounds may be tested to cause tissue injury or a diseased state of the tissue. Alternatively, such compounds may be tested for their effect during an injury or their effect on the healing process.

For getting insights into effects of the certain compound in early cell differentiation, heart tissue models can be isolated before having completed all method steps. Therefore, the spheres present on day 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 of the culture, preferably present at any day prior to day 10 of the culture, can be isolated for further analysis. These isolated spheres are herein also referred to as heart tissue models, or "early-epicardioids". The term "epicardioid", however, is only used for heart tissue models which are isolated only after the completion of all method steps of the method disclosed herein. So there are several possible end-products called "heart tissue model" or "early-epicardioid" but only one end-product called "epicardioid", wherein the epicardioid is isolated at day 13 or later, preferably at day 15 or later.

Methods of Isolation and purification of cells from the epicardioid or early-epicardioid

Epicardioids are generated according to the methods disclosed herein. To analyze and/or use cells generated in the still developing early-epicardioid or the developed mature epicardioid, the culture of the spheres is stopped at a desired time point by isolation of the spheres from the culture and dissociation of the spheres into its single cells.

Dissociation of epicardioids and pre-epicardioids, respectively, into single cells is performed by enzyme treatment, preferably by protease treatment, more preferably by papain treatment. Other dissociation techniques are known to the skilled person, too, such as mechanical dissociation. Preferably, a papain solution comprising 20 U/mL papain and 1 mM L-cysteine in PBS ^/_ is used for dissociation of the spheres.

The number of pooled epicardioids and the dissociation time is adapted to the stage of development (see, Table 1). Briefly, a 2x papain solution consisting of 40 U/mL papain (Worthington Biochemical Corporation; LS003124) and 2 mM L-cysteine (Sigma-Aldrich; C6852) in PBS-/- was incubated 10 min at 37°C to activate the papain before 1:2 dilution in PBS-/- to obtain the lx solution. Spheroids were then removed from the collagen gel if necessary and washed twice with 2 mM EDTA in PBS-/-. They were then dissociated in 750 pL lx papain solution at 37°C and 750 rpm on a Thermomixer (Eppendorf, Germany). The enzymatic reaction was stopped with 750 pL stop solution consisting of 1 mg/mL trypsin inhibitor (Sigma-Aldrich; T9253) in PBS-/-. After pipetting up and down approximately 30 times to obtain a single cell suspension, cells were passed through a 40 pm strainer washed with 5 mL 1% BSA (Gibco; 15260037) in PBS- /-. After centrifugation for 3 min at 200 g, cells were resuspended in 500 pL 0.5% BSA in PBS-/- for counting with Trypan blue. For samples exceeding 15% cell death, dead cells were immediately depleted using the Dead Cell Removal Kit (Miltenyi Biotec; 130-090-101) according to the manufacturer's instructions before further processing. Cells from the same cell suspension may then be used in therapeutic applications and/or the treatment of a heart disease, or were then used for scRNA-seq and scATAC-seq analysis.

Table 1: Preferable dissociation times

Purification of cell populations is preferably performed by flow-based cell separation or by magnetic cell separation. Both techniques use cell markers for the enrichment of the desired cell population. Fluorescent markers may be cell surface markers or cell intrinsic markers. In an embodiment, a fluorescent reporter is expressed under the control of a cell type-specific gene promoter. The isolated cells are then purified by washing cells. The cells are subsequently stored in a cell culture medium suitable for the respective further use of the cells. The isolated and purified cells are either a pure cell population (positive selection) such as JCF cells or epicardial cells, or a mixture of cardiac cells. Such mixture is provided by depletion (negative selection) of undesired cells, such as endodermal cells.

Preferable markers to be used for cell isolation are provided in the following Table:

Table 2: Cell markers

Cells obtained from epicardioids

The present invention relates as well to cells obtained from epicardioids or early-epicardioids. Such cells are either one purified and defined cell population such as JCF cells, or epicardioids or a mixture of cardiac cells.

Cells obtained from epicardioids or early-epicardioids may be further used for their regeneration ability in injured heart tissue. Alternatively, the cells may be used in therapy, preferably in the treatment of a heart disease.

Heart tissue model obtained by the inventive method

The present invention also relates to a heart tissue model obtained by the herein disclosed inventive method(s). The present invention therefor provides a heart tissue model which has been prepared according to the disclosed method.

In preferred embodiments, the heart tissue model is an epicardioid, as described herein, comprising its characteristic structure in terms of cellular compartments.

In another embodiment, the heart tissue model is an early-epicardioid, which has been isolated from culture before day 10, for example, at day 1, 2, 3, 4, 5, 6, 7, 8 or 9 of the culture, wherein the culture protocol does not differ from the method disclosed herein. Isolating early-epicardioids may be used for analysis of cell development at different stages of cell differentiation.

EXAMPLES

The catalog number of commercially available reagents and consumables is indicated; it belongs to the routine work of the skilled person that all reagents mentioned herein could likely be replaced with equivalent products from other companies without affecting the outcome. All steps are performed using sterile technique. hPSC culture

Different cell lines have been used for the Examples disclosed herein. The following hiPSC lines were used: hPSCreg MRI0003-A (hiPSCl), MRI001-A (hi PSC2), and MRI003-A-6 (AAVS1-CAG-VSFP; hiPSC3). For ESCs, the hESC cell line HES-3 (hPSCreg ESI Ble003) was used.

Prior to epicardioid differentiation, hPSCs are maintained on Geltrex-coated (Gibco; A14133-02) 35 mm tissue culture plates (Falcon; 353001) in Essential 8 medium (E8; Gibco; A1517001). Cells are passaged at a ratio of 1:14-1:18 every 4 days with 0.5 mM EDTA in PBS without Ca ²⁺ or Mg ²⁺ (PBS ^_/ “; Gibco; 10010023). The ROCK inhibitor Thiazovivin (TV; Sigma-Aldrich; SML1045) is added at a concentration of 2 pM for 24 h after passaging to promote survival. Example 1: generation of epicardioids

Preparation of basal medium

The basal differentiation medium is prepared by mixing 247.36 mL DMEM/F-12 with GlutaMAX (Gibco; 31331028), 237.36 mL IMDM (Gibco; 21980032), 5 mL chemically defined lipid concentrate (Gibco; 11905031), 10 mL IMDM containing 10% BSA (GE Healthcare; K41-001), 250 pL transferrin (Roche; 10652202001), and 20 pL a-monothioglycerol (Sigma-Aldrich; M6145). The basal medium can be stored for up to 2 months at 4°C. Complete differentiation media are prepared maximum 2 days in advance and stored at 4°C. Media containing retinoic acid are protected from light. Poly-HEMA-coated plates are prepared by adding 50 pL polyHEMA solution consisting of 5% poly-HEMA (Sigma-Aldrich; P3932) in 95% ethanol (Merck; 1009831000) per well of a U-shaped 96-well plate (Falcon; 351177) and letting them air dry overnight.

Epicardioid differentiation and maintenance

On day -1, hPSCs having reached ~80% confluency are washed with 1 mL 0.5 mM EDTA in PBS ^/_ and dissociated for 6-8 min with 1 mL of the same solution at room temperature (RT) (until cells can easily be detached from the plate). After removing the solution, 1 mL E8 medium containing 2 pM TV is added to the plate to detach the cells from the plate by gently pipetting up and down 10-15 times. The cell suspension is transferred to a 50 mL tube (Greiner Cellstar; 227261) and diluted with 1 mL E8 + TV. After counting cells with a Neubauer chamber, 30,000-40,000 cells are then seeded into poly-HEMA-coated U-shaped 96-well plates in a total volume of 150 pL Essential 8 medium with 2 pM Thiazovivin. To ensure that a single spheroid will be formed per well, the plate is centrifuged for 1 min at 300 g at RT before placing in the incubator under standard conditions that are maintained during the entire culture (37°C, 5% CO2).

On day 0 (roughly 24 h after seeding cells on day -1), the entire medium is replaced with 150 pL basal medium supplemented with 10 ng/mL BMP4 (R&D; 314-BP), 50 ng/mL Activin A (Sigma- Aldrich; SRP3003), 30 ng/mL FGF2 (R&D; 233-FB-025/CF), 5 pM LY-29004 (Tocris; 1130), and 1.5 pM CHIR-99021 (Axon Medchem; 1386).

Between 40 and 42 h after day 0 (referred to as day 2), the medium is replaced with 150 pL basal medium supplemented with 10 pg/mL insulin (Sigma-Aldrich; 19278), 10 ng/mL BMP4, 8 ng/mL FGF2, 5 pM IWP2 (Tocris; 3533), and 0.5 pM retinoic acid (Sigma-Aldrich; R2625), which is refreshed every 24 h on days 2 to 5.

On day 6, the medium is replaced with 150 pL basal medium supplemented with 10 pg/mL insulin, 10 ng/mL BMP4, and 8 ng/mL FGF2, which is refreshed 24 h later on day 7.

On day 8, the spheroids are embedded into a gel containing collagen type I. For this, the edges of 15 mm x 15 mm x 5 mm molds (Tissue-Tek: 4566) are trimmed with scissors to fit into the wells of a 6-well culture plate (Nunc; 140675) and the molds are sterilized with 70% ethanol. A cut 1,000 p.L pipet tip is then used to transfer the spheroids into the molds placed in the 6-well plate. Any transferred medium is then removed and the molds are filled with 400 pL collagen I solution consisting of 2.17 mg/mL collagen I (Corning; 354249), 20% distilled water (Gibco; 15230162), 5% lOx DPBS (Gibco; 14080055), and 8.3 mM NaOH in EB20 medium consisting of DMEM/F-12 with 20% fetal bovine serum (Sigma-Aldrich; F7524), 1% non-essential amino acids (Gibco; 11140050), 1% Penicillin-Streptomycin-Glutamine (Gibco; 10378016), and 0.1 mM (3-mercaptoethanol (Sigma-Aldrich; M7522) prepared on ice. After briefly distancing the spheroids from each other using a pipette tip and closing the lid of the 6-well plate, they are placed at 37°C for 30 min until the collagen solidifies. The gel sheets obtained are then carefully removed from the molds by adding 1 mL maintenance medium consisting of basal medium supplemented with 10 pg/mL insulin, 0.5% Penicillin-Streptomycin (Gibco; 15140-122) and 100 ng/mL VEGF (R&D; 293-VE-010) per well and slowly pipetting 1 mL medium under the sheets to make them slide into the well.

On day 10, the medium is refreshed and the plates are placed on a rocking shaker (Assistant, Germany) at 40 rpm to improve oxygen and nutrient diffusion.

For long-term culture, the maintenance medium is refreshed every 2-3 days.

This protocol has been successfully applied to several iPSC and ESC lines, comprising hPSCreg MR1003-A (hiPSCl), MR1001-A (hiPSC2), MR1003-A-6 (AAVS1-CAG-VSFP; hiPSC3), and HES-3 (hPSCreg ESIBISe003; hESC) (see, Fig. 3).

Example 2: Harvesting epicardioids

Epicardioids are harvested by gently pushing them out of the collagen sheet into the culture medium, for example with a pipette tip. They can then be transferred to other containers using a pipette, e.g. for fixation or dissociation (see Materials and Methods).

Example 3: Epicardioids recapitulate the molecular, morphological, and functional patterning of the ventricular wall

To further investigate the composition of epicardioids, we performed whole-transcriptome analysis by single-cell RNA-sequencing (scRNA-seq) at days 15 and 30 (Fig. 4E). Unsupervised cluster analysis revealed that the most abundant cells were ventricular cardiomyocytes, which showed a higher ratio of adult to fetal cardiac troponin I isoforms (TNNI3/TNNI1) as well as increased expression of Ca2+ handling genes (ATP2A2, SLC8A1, PLN) at day 30 compared to day 15, indicating progressive maturation (Fig. 4F,G; Fig. 7A). Interestingly, there was a small cardiomyocyte cluster (cluster 8) showing high SHOX2 and HCN4 expression, suggesting pacemaker identity (Fig. 7B). By comparison with recently available sequencing datasets from human fetal epicardium, we identified three clusters sharing epicardial markers (Fig. 8A). The first (cluster 11) comprised mesothelial epicardial cells, which expressed epithelial genes and showed progressive upregulation of specific human epicardial markers (TNNT1, C3, CALB2) (Fig. 4H; Fig. 8B). The second largest cluster (cluster 5) contained EPDCs expressing markers of EMT and mesenchymal epicardium (Fig. 4H; Fig. 8C). It also comprised epicardium-derived cardiac fibroblasts and vascular smooth muscle cells (Fig. 4H; Fig. 8D,E). This cluster showed no notable differences between the two time points, suggesting a continuous capacity of the mesothelial epicardium to undergo EMT and further differentiation (Fig. 4H; Fig. 8C-E). The third epicardial cluster (cluster 10) resulted from the segregation of highly proliferative cells of both mesothelial and mesenchymal identity (Fig. 8F). Finally, a cluster of endothelial cells (cluster 12) was identified (Fig. 9).

We further investigated the heterogeneity of epicardial cells (clusters 5, 10, and 11) by performing subclustering and inferring cellular dynamics based on the kinetics of gene expression via RNA velocity (Fig. 5A,B). This revealed two mesothelial populations (subclusters 9 and 5; KRT19, CDH1) with heterogeneous expression of the canonical epicardial markers WT1, TBX18, BNC1, TCF21, SEMA3D, and ALDH1A2, as described in vivo and in vitro (Fig. 5C,D). We also identified EPDCs in various stages of differentiation based on the expression levels of EMT genes and well-established markers of epicardial derivatives. We could distinguish between uncommitted EPDCs (subclusters 0, 6), EPDCs differentiating into fibroblasts (subclusters 2, 7; TNC, COL1A1), and EPDCs differentiating into mural cells, comprising vascular smooth muscle cells and pericytes (subclusters 8 and 10; RGS5, PDGFRB), with subcluster 8 specifically expressing pericyte-related genes (MCAM, KCNJ8) (Fig. 5D). Interestingly, subcluster 1 contained EPDCs expressing both fibroblast (COL1A1) and cardiomyocyte markers (TNNT2, ACTN2), and the latter were further upregulated in subcluster 4, suggesting myocytic differentiation (Fig. 5D). Epicardial cells with high expression of cardiac sarcomeric genes have not yet been reported in hPSC-based 2D epicardial differentiation models. Indeed, the transcriptional signatures of 2D epicardial cells described by Gambardella et al. (doi: 10.1242/dev.174441) specifically marked mesothelial cells of subclusters 5 and 9 (BNCl ^high signature) and EPDCs of subclusters 0 and 2 (TCF21 ^high signature) (Fig. 6A,B). Moreover, specific markers of fetal human epicardium found in subclusters 5 and 9 (TNNT1, SFRP2, SFRP5), including some shared by fetal and adult epicardium (LRP2, CALB2, C3), were low or absent in 2D cells (Fig. 6C), suggesting further epicardial development in the 3D environment of epicardioids.

Inferring cell-cell communications from the scRNA-seq dataset revealed ample interaction between epicardial cells and other cell types in epicardioids (Fig. 10A). The volume of signaling from mesothelial epicardium to cardiomyocytes noticeably declined over time, however, hinting to a progression towards the quiescent state characteristic of adult epicardium (Fig. 10A). Epicardial signals known to stimulate cardiomyocyte proliferation in the mouse, such as the secretion of IGF2 and fibronectin binding to integrin pi, as well as the CXCL12-CXCR4 axis promoting coronary angiogenesis in zebrafish were detected (Fig. 10B). Epicardioids also recapitulated ligand-receptor interactions described between epicardial cells and cardiomyocytes (NRP2_VEGFA) or endothelial cells (MIF_TNFRSF10D) in human embryos (Fig. 10B). Although most interactions were common between mesothelial epicardium (cluster 11) and epicardium- derived cells (cluster 5), signals specifically described in adult mesothelial epicardium, such as binding of vitronectin (VTN) to otv|31 integrin, were exclusive to the mesothelial cluster (Fig. 10B, Fig. 11).

In vivo, epicardial stimulation of cardiomyocyte proliferation contributes to the formation of a subepicardial compact myocardium that is molecularly and functionally distinct from the trabecular myocardium facing the ventricular lumen. Opposing gradients in the expression levels of genes enriched in human compact (FTH1, FTL, COL3A1) and trabecular (COL2A1, TTN, MALAT1) myocardium suggested an equivalent patterning in epicardioids, as we could distinguish between clusters with compact, trabecular, and intermediate transcriptional profiles (Fig. IOC). This was further corroborated by the presence of an approximately 50 pm-wide zone of higher cardiomyocyte density underneath the epicardial layer from day 15, which mimicked the morphology of human fetal tissue (Fig. 3A-C). Of note, this was not observed in spheroids differentiated without RA (noRA) lacking the epicardial layer (Fig. 12B,C). Consistent with the molecular mechanisms described above, the compact layer of epicardioids had a significantly higher proportion of Ki67+ proliferating cardiomyocytes than the inner myocardium (Fig. 12D,E).

It was next evaluated if this morphological patterning was associated with regional differences in cardiomyocyte function. In the heart, cardiomyocytes closest to the epicardium generate shorter action potentials (AP) than those in the middle of the ventricular wall, an evolutionary conserved feature increasing the efficiency of ventricular contraction, which is known as the transmural voltage gradient. To assess AP dynamics across the myocardial compartment in epicardioids, we generated them from hPSCs constitutively expressing a FRETbased voltage sensor knocked into the AAVS1 safe harbor locus (AAVS1-CAG-VSFP; Fig. 12F). Optical measurements of AP duration across 250 pm-thick slices prepared by vibratome sectioning revealed significantly shorter durations to 50% and 90% repolarization in cardiomyocytes of the subepicardial layer compared to the inner myocardium at day 35 (Fig. 12G,H). As excitation-contraction coupling is dependent on the intracellular dynamics of calcium, calcium imaging was performed with the fluorescent indicator Fluo-4 (Fig. 121). This showed a corresponding pattern of shorter durations to 50% and 90% peak decay in the subepicardial layer compared to the inner layer (Fig. 12J). Neither of these functional gradients were observed in age-matched noRA spheroids, confirming that they are not intrinsic properties of cardiac spheres (Fig. 12G,H,J).

We then sought to investigate the interaction between epicardial IGF2 and myocytic IGF1R, which was identified as the primary driver of myocardial compaction in rodents but has not yet been studied in a human system. After confirming the protein expression of IGF2 and IGF1R in epicardial cells and cardiomyocytes, respectively (Fig. 13A), we treated epicardioids with increasing concentrations of the small molecule IGF1R inhibitor Linsitinib from days 11 to 15 (Fig. 13B). In DMSO-treated controls, immunostaining for the cell cycle activity marker Ki67 ³⁸ at day 15 revealed significantly higher cardiomyocyte proliferation in the compact outer myocardium (OM) compared to the inner myocardium (IM), in line with higher mitotic activity in the compact layer during development (Fig. 13C,D). Linsitinib treatment dramatically reduced the percentage of proliferating cardiomyocytes in both layers at every concentration applied (Fig. 13C,D). This was associated with a decrease in the cardiomyocyte density ratio between the OM and IM, indicating a failure of subepicardial compaction (Fig. 13E). The opposite effects were observed when treating noRA spheroids, which lack the epicardial layer, with recombinant human IGF2 (Fig. 13F). IGF2 treatment caused a dose-dependent increase in proliferation in the OM, which was associated with a significantly higher cardiomyocyte density ratio between the OM and IM (Fig. 13G-I). This suggested that IGF2 was sufficient to induce myocardial compaction in the absence of an epicardium.

We next focused on epicardial NRP2, predicted to interact with ligands from different cell types in epicardioids and human embryos. NRP2 is widely expressed in the heart and was first implicated in embryonal neuronal guidance and angiogenesis. It was also found to be upregulated in the epicardium following cardiac injury in adult zebrafish, but its role in the developing epicardium is not clear. In epicardioids, NRP2 protein was detected in the mesothelial epicardium and in cardiomyocytes but not in cells in the subepicardial space (Fig. 14A). To perturb NRP2 activity, we treated epicardioids with a blocking anti-NRP2 antibody from days 11 to 15 (Fig. 14B). The highest concentration led to a significant thickening of the epicardial layer by day 15 compared to DMSO-treated controls (Fig. 14C,D). In these samples, the CDH1 ⁺ BNC1 ⁺ mesothelial layer was disrupted, and it lay in close contact with the myocardium while large numbers of mesenchymal cells were oriented outwards (Fig. 14E). This reflected an increased percentage of CDH1- TWIST1 ⁺ cells and a decrease of CDH1 ⁺ TWIST1- cells but no change in proliferation compared to controls (Fig. 14F-H), suggesting that the thickening of the epicardial layer was caused by excessive epicardial EMT - leading to partial loss of the mesothelial layer - rather than hyperproliferation. The inverted orientation of cells additionally pointed to a defect in cell migration, a process that is regulated by NRP2 signaling in tumor cells of several cancer types. Of note, immunodetection of the NRP2 blocking antibody after treatment indicated its predominant localization in the epicardial layer, suggesting that it did not reach the myocardium (Fig. 141).

Example 4: Epicardioids formed by early segregation of first heart field and juxta-cardiac field progenitors

Animal models have shown that the epicardium is formed by cells of the proepicardium, a transient structure located at the venous pole of the looping-stage heart, but the ontogeny of proepicardial precursors in humans is still unclear. To dissect the developmental processes taking place in epicardioids, we performed scRNA-seq in parallel with chromatin accessibility profiling via scATAC-seq at days 2, 3, 4, 5, 7, 10, and 15.

Combined analysis of the 31 cell clusters obtained by scRNA-seq revealed early induction of cardiac mesoderm followed by the emergence of cells expressing markers of first heart field progenitors (FHF; TBX5, SFRP5), which mainly contribute to the left ventricle in the mouse (Fig. 15A, 16A-C). Of note, cells expressing markers specific of anterior second heart field precursors (SHF; TBX1, FGF8) - which generate the right ventricle and the outflow tract - were absent (Fig. 16D). Differentiated ventricular cardiomyocytes were detected from day 7 and showed progressive upregulation of the mature ventricular marker MYL2 (Fig. 15A; Fig. 16A,B). In parallel to cardiac development, we observed a small contribution of endodermal cells that was largely lost by day 15 (Fig. 15A; Fig. 16A,B). Epicardial cells and their derivatives first emerged on day 10 and expanded on day 15 (cluster 23; Fig. 15A; Fig. 16A,B). Interestingly, they were preceded by cells closely matching the transcriptional signature of the juxta-cardiac field (JCF; HAND1, HOXB6, MAB21L2), a common progenitor of myocardium and epicardium recently described as a subset of the FHF in the mouse and not yet identified in humans (clusters 21 and 27; Fig. 15A,B, Fig. 16A,B). Computational reconstruction of the developmental trajectories of cells from day 2 to day 15 suggested that the epicardial lineage as well as a proportion of cardiomyocytes were descendants of these cells, with a small population of cardiomyocytes appearing to derive from epicardial cells themselves (Fig. 15C, Fig. 17). It also suggested a distinction between bipotent JCF (for epicardial and myocardial lineages) and unipotent JCF restricted to the myocytic fate (Fig. 15C). Both populations were downstream of the same putative "pre-JCF" precursors (cluster 11), which appeared on days 4 and 5 at the same time as classical FHF cells (Fig. 15A,C, Fig. 16A). Remarkably, pre-JCF cells expressed the multipotent cardiovascular progenitor marker ISL1 (25) but not the myocytic marker NKX2.5 (Fig. 18A). This allowed us to follow their physical segregation from ISL1+/NKX2.5+ FHF progenitors, which was clearly visible by day 5, with pre-JCF cells already concentrated at the outer layer (Fig. 18B). High expression of ISL1 was maintained in the JCF at days 7 and 10, followed by a decrease in the epicardial cluster at day 15 (Fig. 19A). This reflected a maintenance of ISL1 solely in the mesothelial epicardium and a loss of expression in subjacent EPDCs (Fig. 15D). As ISL1 expression was not previously reported in the epicardium (human or otherwise), ISL1 immunostaining was performed in human fetal heart tissue to verify this finding. This revealed that epicardial cells were indeed positive for ISL1 at 5 weeks post conception, but expression was completely lost by 6 weeks (Fig. 15D). We also observed loss of ISL1 in epicardial cells of 30-day-old epicardioids, suggesting equivalent expression dynamics (Fig. 19B).

To gain deeper insight into the specification of JCF progenitors in epicardioids and identify putative regulatory programs that could underlie divergence of bipotent and unipotent pools, chromatin accessibility profiles were analyzed via scATAC-seq at days 3, 4, and 5. Combined cluster analysis revealed four main cell types showing dynamic progression over time: cardiac mesoderm (clusters 0 and 12), endoderm (clusters 4, 5, and 10), cells with myocytic commitment (clusters 1, 3, and 11), and cells with (pre-)JCF commitment (cluster 2, 8, 9) (Fig. 15E, Fig. 21A). The latter were identified by screening for concomitant chromatin accessibility of 54 genes defining pre-JCF and early JCF cells in our scRNA-seq dataset (Fig. 15F). Differential enrichment analysis of accessible transcription factor (TF) motifs among clusters revealed a separation within the (pre-)JCF population, with clusters 2 and 8 sharing common motif patterns with cluster 12, while cluster 9 showed similarity to the myocytic cluster 3, suggesting related developmental trajectories (Fig. 15G, Fig. 21B). We identified TF motifs showing differential accessibility between clusters 2/8 and 9 and investigated the activity of potential downstream targets of these TFs by inferring gene regulatory networks from the scRNA-seq data using SCENIC. Cluster 9 showed increased accessibility for genes with motifs bound by cardiomyocyte-related TFs such as MEIS1, which displayed regulon activity in both FHF clusters and a subset of the pre-JCF (Fig. 15G,H, Fig. 24A). By contrast, clusters 2 and 8 presented increased accessibility for genes with motifs bound by TFs showing regulon activity exclusively in the pre-JCF, such as TFAP2A (Fig. 15G,H, Fig. 24B). Analysis of peaks at cell type specific loci further indicated that cluster 9 had open chromatin at the transcription start sites of myocytic (NKX2.5) but not epithelial (KRT7) genes, while clusters 2 and 8 had open chromatin for both, denoting potential for both cardiomyocyte and epicardial specification (Fig. 151).

To functionally validate the fate potential of different progenitor populations in epicardioids, we generated a hiPSC reporter line in which a FRT-flanked neomycin cassette blocking transcription of a pCAG-driven fluorescent reporter (mKate2) fused to an HA-tag was knocked into the AAVS1 safe harbor locus (Fig. 22A-F). For JCF lineage tracing, epicardioids from this reporter line were transduced with a lentiviral vector encoding an inducible flippase (FLP ^ERT2 ) under the control of the MAB21L2 promoter at day 3, and 4-hydroxytamoxifen (4-OHT) was applied at the beginning of the JCF stage (days 7-8) (Fig. 23A,C). In the absence of a reliable antibody against MAB21L2, we performed co-staining of the HA-tag and ISL1 after 48 hours to confirm successful labeling of ISL1 ⁺ JCF cells located at the outer layer (Fig. 23B). On day 12, immunofluorescence analysis revealed HA-tag-positive cardiomyocytes and epicardial cells (both mesothelial and EPDCs), confirming the dual fate potential of JCF cells as seen in the mouse (Fig. 23C). Having also observed MAB21L2 ⁺ cells in clusters categorized as FHF cells, we alternatively applied 4-OHT at the corresponding stage (days 4-5) (Fig. 23D). In this case, 78.4% of HA-tag-positive cells at day 12 were cardiomyocytes, compared to 54.7% when applying 4-OHT at the JCF stage (Fig. 23C,D). Considering the close relationship between the JCF and the FHF in the mouse, it is unclear whether (some of) these cells descended from classical FHF cells expressing MAB21L2 or if there exists an early JCF population with higher commitment to the myocytic lineage.

Taken together, our findings suggest that JCF cells are not a uniform population with regards to lineage potentiality, but further clonal analyses will be required to resolve this aspect. Example 5: Epicardioids allow investigation of the fate potential of human epicardial cells

Beyond their embryonic origin, there are still many open questions concerning the molecular and functional heterogeneity of epicardial cells, which have important implications on potential epicardial re-activation as a therapeutic target. It is debated whether the lineage fate of EPDCs is pre-determined at the (mesothelial) epicardial stage or if specification occurs after EMT. Moreover, it is not known whether there are subsets of epicardial cells that are permissive for opposing cell fates prior to their lineage commitment. Inferring the trajectory of the epicardial clusters in our scRNA-seq dataset at days 7, 10, and 15 revealed diverging fates for mesothelial epicardial cells (KRT19, I SL1) downstream of JCF cells (Fig. 25A,B). There was a clear segregation between epicardial cells maintaining their mesothelial identity at day 15 and those giving rise to EPDCs eventually differentiating into fibroblasts and smooth muscle cells (Fig.25B,C). Surprisingly, there was also an entirely distinct branch showing myocytic commitment, which supported the existence of epicardium-derived cardiomyocytes in our system (Fig. 25C).

To verify the lineage potential of mesothelial epicardial cells in our system, we generated epicardioids from the FLP/FRT-based hiPSC reporter line (Fig. 22) and transduced them with a lentiviral vector encoding flippase under the control of the CDH1 promoter at day 15 (Fig. 23E,G). After 72 h, we detected HA-tag ⁺ /CDHl ⁺ cells, indicating correct labeling of the mesothelial layer (Fig. 23F). On day 24, immunofluorescence analysis revealed HA-tag-positive smooth muscle cells and fibroblasts but also cardiomyocytes (Fig. 23G), supporting the fate potentiality previously inferred from gene expression (Fig. 5A,D).

We next investigated whether the mechanisms of fate determination vary between different epicardium-derived cell types by studying chromatin accessibility patterns associated with epicardial differentiation in the scATAC-seq data of days 7, 10, and 15 (Fig. 26A). Subclustering of cells of the JCF and epicardial lineage, identified via the corresponding scRNAseq signature, revealed four cell populations (Fig. 25D, Fig. 26B). Analysis of differential gene activity allowed us to distinguish between epicardial cells associated with EMT (cluster 0), EPDCs and their derivatives (cluster 1), muscle-fated epicardial cells and EPDCs (cluster 2), and mesothelial JCF/epicardium (cluster 3) (Fig. 26C). Aligning cells in pseudotime based on inferred gene activity along the differentiation trajectory revealed that JCF cells on day 7 as well as epicardial cells with mesothelial identity from day 10 and 15 located at the beginning of the trajectory, with no progression in pseudotime, suggesting that they may represent a maintained progenitor pool. This was further supported by high activity of ISL1 and other genes involved in self-renewal (YAP1) (Fig. 25E,F, Fig. 27). A vast majority also presented high accessibility for all three lineage-specific markers, namely TNNT2 (cardiomyocyte), MYH11 (smooth muscle), and DDR2 (fibroblast), indicating that they are indeed a multipotent population (Fig. 25E-H, Fig. 27). A similar mesothelial subset, which showed lower accessibility for fibroblast genes, appeared transiently at day 10 (Fig. 25E-H, Fig. 27). Cells further along the trajectory showed progressive loss of CDH1 activity, indicating that they undergo EMT. This was associated with dynamic changes in the activity of lineage-specific markers resulting in EPDCs maintaining activity only for one or two lineage markers towards the end of the trajectory, reflecting a more restricted differentiation potential (Fig. 25E-H, Fig. 27). Interestingly, a reduction in TNNT2 activity could also be observed from day 7 to day 15 in the mesothelial epicardial cells located at the beginning of the trajectory, which was paralleled by an increase of DDR2 accessibility (Fig. 25G,H). This points to a progressive loss of cardiomyocyte potential of the mesothelial cells in favor of the fibroblast fate in this population over time, which is in line with the lack of cardiomyocyte potential of adult human epicardium. Overall, our data does not support the existence of discrete subsets of embryonic epicardial cells restricted to a single lineage before EMT, but rather advocates a model of dynamic fate specification over time.

Example 6: Epicardioids model multicellular processes of left ventricular hypertrophy and remodelling

Cardiac stressors such as high blood pressure or valvular heart disease can lead to left ventricular hypertrophy (LVH) and fibrosis, a maladaptive remodeling of the myocardium that increases patients' risk for heart failure and life-threatening arrhythmia. Current 2D in vitro models largely recapitulate the myocytic features of LVH, but fail to account for the pivotal role of fibrosis in the progression towards heart failure. Hypothesizing that the 3D multilineage architecture of the herein disclosed epicardioids could resolve this gap, 1-month-old epicardioids were treated with endothelin-1 (ET1), a potent vasoconstrictor known to induce hypertrophy in vivo and in vitro. ET1 triggered a dose-dependent upregulation of myocytic hypertrophy markers (NPPA, NPPB, ACTA1, MYH7/MYH6) and an increase in cardiomyocyte size (Fig. 27A,B). The concomitant upregulation of ECM genes (COL1A2, COL3A1, FN1, POSTN) suggested that epicardioids do have the capacity to mount a fibrotic response (Fig. 27C). This was corroborated by abundant ECM deposition in the subepicardial space and the emergence of ot-SMA+ myofibroblasts (Fig. 27D,E). Calcium imaging in ETl-treated epicardioid slices additionally revealed cardiomyocyte dysfunction across the myocardial layers, including frequent arrhythmic events and decreased calcium transient amplitudes, two well-established features of failing hearts (Fig. 27F-H).

Example 7: Epicardioids recapitulate the myocardial hyperproliferation and fibrosis typical of the cardiomyopathy associated with Noonan syndrome

To test the capacity of epicardioids to model congenital myocardial fibrosis, we used hiPSCs from a Noonan syndrome patient who presented with severe LVH and myocardial fibrosis at birth (Fig. 28A). The hiPSC-derived cardiomyocytes from this patient displayed cell cycle defects leading to hyperproliferation rather than a classical hypertrophic phenotype when cultured in 2D. The same was observed in patient-specific epicardioids: they did not have larger cardiomyocytes compared to healthy controls and did not upregulate hypertrophy markers but showed increased cardiomyocyte proliferation across the myocardial layers (Fig. 28B-D). We additionally observed an upregulation of ECM genes and the appearance of areas containing large numbers of fibroblasts and SMA ⁺ myofibroblasts as early as day 15, indicating that the cellular environment of epicardioids is indeed permissive to fibrotic changes associated with developmental defects (Fig. 28E-G).

Example 8: Isolation of cells from epicardioids (and pre-epicardioids)

Epicardioid dissociation is performed using 2x papain solution. A detailed protocol is described in the Materials and Methods section.

Treatment with papain solution achieves epicardioid dissociation with >90% cell survival. Dissociation is performed around days 7-10 and 15-30 of differentiation to obtain JCF and epicardial cells, respectively. Two strategies may be followed for cell enrichment:

1) To obtain pure cell populations (of JCF cells or epicardial cells), flow-based or magnetic separation is performed using surface markers. For instance, CDH1 may be used for epicardial cell marking and MAB21L2 or ITGA8 for JCF cell marking, respectively. This strategy can be used for the isolation of any desired cell population described herein based on the specific cell markers.

2) Alternatively, any contaminating endodermal cells may be removed from the dissociated epicardioid cells by magnetic depletion based on markers FGFR4 or CDH6 (negative selection) resulting in a cell suspension comprising all cardiac cells. These may include, depending on the time point of (early-)epicardioid dissociation: FHF cells (expressing NKX2.5 and TBX5), pre-JCF cells (expressing ISL1), JCF cells (expressing HAND1, HOXB6 and/or MAB21L2), mesothelial epicardial cells (expressing KRT18, KRT19, and/or CDH1), EPDCs (expressing VIM), fibroblasts (expressing COL1A1 and/or TNC), vascular smooth muscle cells (expressing RGS5 and/or MYH11), pericytes (expressing MCAM), cardiomyocytes (expressing TNNT2), and endothelial cells (expressing CD31 and/or CDH5)

Example 9: Ex vivo use of epicardioid-derived cell populations in injury models

The regenerative potential of JCF cells and their progeny may be tested using ex vivo injury models.

To track cells, epicardioids are generated from hiPSCs expressing a fluorescent reporter (AAVS1- CAG-eGFP; hPSCreg ID MRIi003-A-8). Experiments are based on pig and non-human primate heart slices cultured in biomimetic chambers, as previously reported (Poch et al. (29022), "Migratory and anti-fibrotic programs define the regenerative potential of human cardiac progenitors", Nature Cell Biol., 24:659-671).

Two models are used based on pig or non-human primate heart slices:

1) Acute injury: a defined area of tissue will be destroyed using radiofrequency ablation.

2) Chronic injury: long-term culture of native tissue (>3 weeks) will lead to progressive cell death and loss of contractility, imitating heart failure.

Both acute injury and chronic injury models use the tissue slices referred to above. This refers to approx. 300 pm slices of live pig or non-human primate myocardial tissue prepared through vibratome sectioning. These slices are then cultured ex vivo in biomimetic culture chambers where they are exposed to mechanical and electrical stimulation (as described in Poch et al., (2022)).

In both cases, candidate cells are seeded onto injured slices for instance by using bioprinting and high-resolution imaging is used to track their behavior (survival, proliferation, migration) along with their effect on injury resolution (remuscularization, scarring, contractility). Bioprinting refers to the use of a 3D printer to apply a pluronic frame on the heart tissue slice, which allows seeding of cells in a specific area. The person skilled in the art is, however, aware of other possibilities of cell seeding. Candidate cells comprise any cardiac cell population isolated from the (early- Jepicardioid as described in Example 7. Preferably, the candidate cell populations used are JCF cells, epicardial cells, or all cardiac cell types. Single-cell spatial transcriptomics is used to analyze dynamic cellular states, de novo cell differentiation, and cell-cell signaling in response to injury.

Example 10: In vivo use of epicardioid-derived cells

Based on insights from the ex vivo models, the in vivo regenerative potential of candidate cells is tested using pig models of cardiomyopathies and acute or chronic ischemic heart injury, as previously described (Poch et al. (2022)). To track cells, epicardioids are generated from AAVS1- CAG-eGFP hiPSCs (for immunofluorescence analyses) or AAVS1-CAG-DTPA-R hiPSCs encoding a novel PET tracer (for in vivo imaging). Readouts include hallmarks of injury resolution (remuscularization, neovascularization, scarring, inflammation) and heart function (monitored by cMRI). An important area of investigation is the potential arrhythmogenicity of injected cells, a key aspect for future clinical application. In the case of successful heart regeneration, further investigation of the mode of action using single-cell spatial transcriptomics is followed. Moreover, different modes of in vivo delivery of cells should be tested in pigs or NHPs; ideally, the procedure should be minimally invasive and maximize the survival and integration of cells. This will lay the groundwork for future clinical trials, e.g. using JCF or epicardial cell transplantation to prevent heart failure progression in patients following myocardial infarction. The therapeutic effect of cells isolated from early-epicardioids or epicardioids may be enhanced by introducing targeted gene modifications (e.g., to confer improved capacity for survival or differentiation into specific cell types by altering the expression of one or several genes).

MATERIAL AND METHODS

The below disclosed materials and methods are one way of performing the inventive methods and uses. It is clear to the skilled persons that the specific cell types, specific compounds, supplements etc. can be replaced by surrogates from other manufacturers.

Culture of human pluripotent stem cells

Human iPSCs were generated using the CytoTune-iPS 2.9 Sendai Reprogramming Kit (Invitrogen; A16157) as previously described. The following hiPSC lines were used in differentiation experiments: hPSCreg MRI003-A (hiPSCl), MRI001-A (hiPSC2), and MRI003-A-6 (AAVS1- CAG- VSFP; hiPSC3). Alternatively, hESC line HES-3 (hPSCreg ESIBIe003) generated by ES Cell International Pte Ltd in Singapore was used. In addition, Noonan hiPSCs (MRIiO25-A) may be used.

Human pluripotent stem cells were cultured on Geltrex-coated plates (Gibco; A14133-02) in Essential 8 medium (Gibco; A1517001) containing 0.5% Penicillin/Streptomycin (Gibco; 15140- 122). Cells were passaged every 4 days with 0.5 mM EDTA (Invitrogen; AM92606) in PBS without Ca2+ or Mg2+ (PBS-/-; Gibco; 10010023).

Immunofluorescence analysis

For the production of cryosections, spheroids were washed with DPBS and fixed with 4% PFA (Sigma-Aldrich; 158127) for 1 h at RT. After washing 3 times with DPBS, they were kept in 30% sucrose at 4°C overnight. They were then embedded in a solution of 10% sucrose and 7.5% gelatin in DPBS before freezing in a 2-methyl-butane bath (Sigma-Aldrich; M32631) cooled with liquid nitrogen and transferring to -80°C. Cryosections prepared with a Microm HM 560 cryostat (Thermo Fisher Scientific, Germany) were transferred onto poly-L-lysine slides (Thermo Fisher Scientific; J2800AMNT) and stored at -80°C before immunofluorescence analysis.

For immunostaining, live cells were washed with DPBS and fixed with 4% PFA for 15 min at RT, while cryosections were fixed with 4% PFA for 10 min at RT. After washing 3 times with DPBS, cells or cryosections were permeabilized with 0.25% Triton X-100 (Sigma-Aldrich; X100) in DPBS for 15 min at RT. After washing another 3 times with DPBS, samples were blocked with 3% BSA in PBST, consisting of 0.05% Tween 20 (Sigma-Aldrich; P2287) in DPBS, for 1 hour at RT. Primary antibodies were then added at the indicated dilutions in 0.5% BSA in PBST and incubated overnight at 4°C. After washing 3 times for 5 min (cells) or 5 times for 10 min (cryosections) with PBST, appropriate secondary antibodies diluted in 0.5% BSA (Sigma-Aldrich; A9647) in PBST were added for 1 hour (cells) or 2 hours (cryosections) at RT, protected from light. After repeating the previous washing steps, Hoechst 33258 (Sigma-Aldrich; 94403) was added at a final concentration of 5 pg/mL in DPBS for 15 min at RT, protected from light. Samples were mounted with fluorescence mounting medium (Dako; S3023) and stored at 4°C until imaging with an inverted or confocal laser scanning microscope (DMI6000B and TCS SP8; Leica Microsystems, Wetzlar, Germany).

Single-cell genomics

Cell preparation

Epicardioids were dissociated to single cells using papain, adapting the number of pooled epicardioids and dissociation time to the stage of development. Briefly, a 2x papain solution consisting of 40 U/mL papain (Worthington Biochemical Corporation; LS003124) and 2 mM L- cysteine (Sigma-Aldrich; C6852) in PBS-/- was incubated 10 min at 37°C to activate the papain before 1:2 dilution in PBS-/- to obtain the lx solution. Spheroids were then carefully pushed out of the collagen gel using 10 pL pipette tips if necessary and washed twice with 2 mM EDTA in PBS- /-. They were then dissociated in 750 pL lx papain solution at 37°C and 750 rpm on a Thermomixer (Eppendorf, Germany). The enzymatic reaction was stopped with 750 pL stop solution consisting of 1 mg/mL trypsin inhibitor (Sigma-Aldrich; T9253) in PBS-/-. After pipetting up and down approximately 30 times to obtain a single cell suspension, cells were passed through a 40 pm strainer washed with 5 mL 1% BSA (Gibco;15260037) in PBS-/-. After centrifugation for 3 min at 200 g, cells were resuspended in 500 pL 0.5% BSA in PBS-/- for counting with Trypan blue. For samples exceeding 15% cell death, dead cells were immediately depleted using the Dead Cell Removal Kit (Miltenyi Biotec; 130-090-101) according to the manufacturer's instructions before further processing. Cells from the same suspension were used for single-cell RNA-seq and singlecell ATAC-seq as described below.

Single-cell RNA-seq

After dissociation, samples were processed for single-cell RNA sequencing with a targeted cell recovery of 8,000. To generate Gel Bead-ln-EMulsions (GEMs) and single-cell sequencing libraries, the Chromium Single Cell 3' GEM, Library & Gel Bead Kit v3 (lOx Genomics; PN-1000092), the Chromium Chip B Single Cell Kit (lOx Genomics; PN- 1000073), and the Chromium i7 Multiplex Kit v2 (lOx Genomics; PN-120262) were used for samples d2-dl5 and the Chromium Next GEM Single Cell 3' Library & Gel Bead Kit v3.1 (1000128, lOx Genomics), Chromium Single Cell G Chip Kit (1000127, lOx Genomics), and Single Index Kit T Set A (PN-1000213, lOx Genomics) were used for the d30 sample. Quality control of cDNA samples was performed on a Bioanalyzer (Agilent, Germany) using a High Sensitivity DNA kit (Agilent; 5067-4626). Library quantification was performed with the KAPA quantification kit (KAPA Biosystems; KK4824) following the manufacturer's instructions. Libraries were pooled and sequenced using a NovaSeq SI flow cell (Illumina, San Diego, CA) with 150 bp paired-end reads with 28 cycles for readl, 91 cycles for read2, 8 cycles for i7, and 0 cycles for i5, and with a read depth of at least 25-30,000 paired-end reads per cell.

The Cell Ranger pipeline (v6.1.1) was used to perform sample demultiplexing, barcode processing and generate the single-cell gene counting matrix. Briefly, samples were demultiplexed to produce a pair of FASTQ files for each sample. Reads containing sequence information were aligned using the reference provided with Cell Ranger (v6.1.1) based on the GRCh37 reference genome and ENSEMBL gene annotation. PCR duplicates were removed by matching the same UMI, lOx barcode and gene, and were collapsed to a single UMI count in the gene-barcode UMI count matrix. All the samples were aggregated using Cell Ranger with no normalization and treated as a single dataset. The R statistical programming language (v3.5.1) was used for further analysis. Count data matrix was read into R and used to construct a Seurat object (v4.0.1). The Seurat package was used to produce diagnostic quality control plots and select thresholds for further filtering. Filtering method was used to detect outliers and high numbers of mitochondrial transcripts. These pre-processed data were then analyzed to identify variable genes, which were used to perform principal component analysis (PCA). Statistically significant PCs were selected by PC elbow plots and used for UMAP analysis. Clustering parameter resolution was set to 1 for the function FindCI ustersf) in Seurat. For sub-clustering analysis, we used the clustree package (vO.4.3). All DEGs were obtained using Wilcoxon rank sum test using as threshold p-value < 0.05. We used adjusted p-value based on Bonferroni correction using all features in the dataset. For cell type-specific analysis, single cells of each cell type were identified using FindConservedMarkers function as described within the Seurat pipeline. Analysis of cell-cell interactions was performed with CellPhoneDB v2.0. To infer the developmental lineage relationships between cells in our study, we used the R package URD (version 1.1.1). Marker genes in each branch were defined using markersAUCPR() function (adjusted p-val <0.05). For all the gene signatures analyzed we used a function implemented in yaGST R package (https://rdrr.io/github/miccec/yaGST/). Gene regulatory network (GRN) analysis was performed using the R-package SCENIC (Single Cell rEgulatory Network Inference and Clustering) vl.2.4 and the command line interface (CLI) of the python implementation pySCENIC v0.11.2. The raw, filtered count matrix containing only cells from Day 3 to Day 5 was extracted from the Seurat object. The matrix was pre-filtered and genes with at least 450 counts present in at least 150 cells, equal to at least 3 UMI counts in 1% of the cells, were used as input for the CLI. The human motif collection v9 and the cisTarget databases for hgl9 were used in the pipeline and downloaded from https://resources.aertslab.org/cistarget/. For analysis of the 2D epicardium scRNA-seq dataset from Gambardella et al., we downloaded the raw data from https://www.ncbi. nlm.nih.gov/geo/query/acc.cgi?acc=GSE122827. Single-cell ATAC-seq

After dissociation, nuclei isolation for single-cell ATAC sequencing was performed following recommendations of lOx Genomics. Briefly, ~500,000 cells from each sample were transferred to a 1.5-mL microcentrifuge tube and centrifuged at 300 g for 5 min at 4 °C. The supernatant was removed without disrupting the cell pellet, and 100 pL of chilled Lysis Buffer (10 mM Tris-HCI pH 7.4; 10 mM NaCI; 3 mM MgCI2; 0.1% Tween-20; 0.01% Nonidet P40 Substitute; 0.01% Digitonin and 1% BSA) was added and pipette-mixed 10 times. Samples were then incubated on ice for 30- 120 seconds (the optimal incubation time was optimized in advance for each time point). Following lysis, 1 mL of chilled Wash Buffer (10 mM Tris-HCI (pH 7.4); 10 mM NaCI; 3 mM MgCI2; 0.1% Tween-20 and 1% BSA) was added and pipette-mixed. Nuclei were centrifuged at 500 g for 5 min at 4 °C, the supernatant was removed without disrupting the pellet, and nuclei were resuspended in the appropriate volume of chilled Diluted Nuclei Buffer (lOx Genomics) to obtain a nuclei concentration suitable for a target nuclei recovery of 8,000.

Samples were then processed using the Chromium Next Single Cell ATAC Library & Gel Bead Kit vl.l (1000175, lOx Genomics), Chromium Single Cell H Chip Kit (1000161, lOx Genomics) and Chromium Single Index Kit N, Set A (PN-1000212; lOx Genomics) to generate Gel Bead-ln- EMulsions (GEMs) and single-cell ATAC sequencing libraries. Libraries were pooled and sequenced using a NovaSeq SI flow cell (Illumina, San Diego, CA) with 150 bp paired-end reads with 50 cycles for readl and read2, 8 cycles for i7, and 16 cycles for i5, and with a read depth of at least 25- 30,000 paired-end reads per cell. Sequencing raw data were processed using the Cell Ranger ATAC algorithms (vl.2.0). Before alignment to the human reference genome the ATAC-seq sequences were quality-checked using FastQC. The parameters evaluated were: (i) total number of reads, (ii) sequencing length distribution, (iii) sequence quality per base and (iv) duplication level. All the samples were aggregated using Cell Ranger ATAC with no normalization and treated as a single dataset. The R statistical programming language (v3.5.1) was used for further analysis. Count data matrix was read into R and used to construct Signac object. The ATAC-seq peaks were subsequently called using the Signac pipeline. To create a gene activity matrix, we extract gene coordinates using FeatureMatrixf) function extending them at the promoter region.

Endothelin-1 treatment

To induce hypertrophy, epicardioids were treated with 25 nM or 50 nM endothelin-1 (Sigma- Aldrich; E7764) in maintenance medium for 6 days, replacing the medium every day. At the end of the treatment, they were dissociated with TrypLE Express (Gibco; 12605010) for 15 min at 37°C for RNA extraction or fixed with 4% PFA for 1 hour for cryosectioning. For cell size measurements, ETl-treated epicardioids and untreated controls were dissociated to single cells with papain as described above and re-seeded at a density of 25,000 cells/cm ² coated with 2 pg/cm ² fibronectin (Sigma-Aldrich; F1141). After 4 days, cells were fixed for immunofluorescence staining for cTnT as well as the desmosomal marker pla kophili n-2 (PKP2) to visualize cell membranes. The area of cardiomyocytes was quantified in ImageJ (National Institutes of Health).

Linsitinib/NRP2 antibody/IGF2 treatment

(Early-)Epicardioids were treated with 0.25 pM, 0.5 pM or 1 pM Linsitinib (Tocris; 7652) or 200 pg/mL or 500 pg/mL NRP2 blocking antibody (R&D; AF2215) in maintenance medium on days 11, 12, 13, and 14. Spheroids differentiated without retinoic acid were treated with 25 ng/mL, 50 ng/mL or 100 ng/mL recombinant human IGF2 (R&D; 292-G2) in maintenance medium on days 11, 12, 13, and 14. DMSO was used as a vehicle control.

Vibratome sectioning

To prepare live sections, spheroids were removed from the collagen gel and placed in 4% agarose (Biozym; 840004) in sterile DPBS ^+/+ . Once the agarose had solidified, it was trimmed down to a block of approximately 1 cm x 1 cm x 1 cm with a scalpel and 250 pm-thick slices were cut with a vibratome (VT1200S, Leica Biosystems, Germany) in a DPBS bath following the manufacturer's instructions. The spheroid slices were then kept in maintenance medium for 3-5 days before functional assays.

Optical action potential measurements

For optical action potential measurements, 250 pm-thick slices of spheroids derived from the AAVS1-CAG-VSFP hiPSC line (hPSCreg MRI003-A-6) were transferred to Tyrode's solution (135 mM NaCI, 5.4 mM KCI, 1 mM MgCI ₂ , 10 mM glucose, 1.8 mM CaCI ₂ and 10 mM HEPES, pH 7.35) before imaging at 100 fps on an inverted epifluorescence microscope (DMI6000B, Leica Microsystems) equipped with a Zyla V sCMOS camera (Andor Technology, Germany). The VSFP was excited at 480 nm and the emitted GFP and RFP fluorescence signals were separated using an image splitter (OptoSplit II, Cairn Research, UK). The fluorescence of regions of interest relative to background regions was quantified in ImageJ (National Institutes of Health) and subsequent analysis was performed in RStudio using custom-written scripts to determine the duration at 50% (APD50) or 90% repolarization (APD90). APD50 maps were generated by aligning the split image stacks with a custom algorithm in MatLab (The MathWorks Inc), denoising them with the CANDLE algorithm and calculating the ratio between the two. For each action potential, the APD was calculated directly based on the amplitude on each pixel.

Calcium imaging

For calcium imaging, 250 pm-thick spheroid slices were loaded with 1 pM Fluo-4-AM (Thermo Fisher Scientific; F14201) in Tyrode's solution (135 mM NaCI, 5.4 mM KCI, 1 mM MgCI ₂ , 10 mM glucose, 1.8 mM CaCI ₂ and 10 mM HEPES, pH 7.35) containing 0.01% Pluronic F-68 (Gibco; 24040- 032) for 50 min at 37°C. The solution was replaced with Tyrode's solution for 30 min at 37°C for de-esterification of the dye before imaging at 100 fps on an inverted epifluorescence microscope (DMI6000B, Leica Microsystems) equipped with a Zyla V sCMOS camera (Andor Technology, Germany). Pacing was performed with field stimulation electrodes (RC-37FS; Warner Instruments) connected to a stimulus generator (HSE Stimulator P; Hugo-Sachs Elektronik) providing depolarizing pulses at the indicated frequencies. The fluorescence of regions of interest relative to background regions was quantified in ImageJ (National Institutes of Health) and subsequent analysis was performed in RStudio using custom-written scripts to determine the transient duration at 50% (TDso) or 90% decay (TD90).

Quantitative real-time PCR (qPCR)

Total RNA was isolated from cells using the Absolutely RNA Microprep kit (Agilent; 400805) and cDNA was prepared using the High Capacity cDNA RT kit (Applied Biosystems; 4368814) according to the manufacturers' instructions. Quantitative real-time PCR was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems; 4368706) on a 7500 Fast Real-Time PCR instrument (Applied Biosystems, Germany). The mRNA expression levels of genes of interest were quantified relative to GAPDH expression using the ACt method.

Measurement of cardiomyocyte size

For cell size measurements, epicardioids were dissociated to single cells with papain as described above and re-seeded at a density of 25,000 cells/cm ² on plates coated with 2 pg/cm ² fibronectin (Sigma-Aldrich; F1141). After 4 days, cells were fixed for immunofluorescence staining for cTnT and the desmosomal marker plakophili n-2 (PKP2) to visualize cell membranes. The area of cardiomyocytes was quantified in ImageJ (National Institutes of Health).

Lineage tracing

Generation of the AAVSl-CAG-FRTt-f tanked STOP-mKate2-HA reporter line

To construct the donor plasmid pAAVSl-CAG-FRT-flanked STOP-mKate2-HA-polyA, the pCAFNF- GFP plasmid (Addgene #13772) was digested with Spel and Sall, and the CAG-FRT-flanked STOP cassette (CAG promoter and neomycin resistance gene flanked by FRT sites) was cloned into the pAAVSl-Nst-MCS vector (Addgene #80487), which was digested with Spel and Sall. The Simian virus 40 PolyA (Sv40-polyA) signal was then amplified by PCR from the pCAFNF-GFP plasmid using primers containing Pad- at 5' and EcoRI-restriction sites at 3' and introduced into the pAAVSl- CAG-FRT-flanked STOP plasmid, digested with Pad and EcoRI. The mKate2 coding sequence fused to an HA-tag was amplified by PCR from the p3E-mKate2-HA no-pA plasmid (Addgene #80810) as a template and inserted into Swal-Pacl sites on the pAAVSl-CAG-FRT-flanked STOP-polyA plasmid.

Healthy control hiPSCs (hPSCreg MRI003-A) (lxlO ⁶ ) were nucleofected with 1 pg pXAT2 plasmid (Addgene #80494) containing sequences for an AAVS1 locus specific sgRNA [GGGGCCACTAGG GACAGGAT] and the Cas9 nuclease, and 3 pg donor construct (pAAVSl-CAG-FRT-flanked STOP- mKate2-HA-polyA) following the Lonza Amaxa 4D Nucleofector protocol for human stem cells. Cells were subsequently plated onto Matrigel-coated (BD; 354277) 6 well-plates (Nunclon; 150687) in mTeSRl (Stemcell Technologies; 05854) with 10 pM Thiazovivin. 24 h later, and every day afterwards, the medium was replaced with fresh mTeSRl. Three days after nucleofection, 150 pg/ml neomycin (Gibco; 10131) was added into the mTeSRl for selection for 2 weeks. When the hiPSC colonies were large enough, cells were dissociated with Accutase (Thermo Fisher Scientific; A11105-01) and replated for single clone expansion at low density (1,000 cells per 10 cm Matrigel-coated dish). Single clones were then picked for PCR genotyping and further expansion into wells of a Matrigel-coated 96-well plate (Nunclon; 161093). The genotype of the selected clones was verified by PCR screening and confirmed by Sanger sequencing (Eurofins MWG Operon)

Karyotype analysis after editing was performed at the Institute of Human Genetics of the Technical University of Munich using G-banding (20 metaphases counted). Three out of ten potential off-target sites predicted by the CRISPOR tool (https://crispor.tefor.net) were amplified and verified by Sanger sequencing. To verify correct reporter expression, positive hiPSCs clones (lxio ⁶ ) were nucleofected with 3 pg pCAGGS T2A FLPo plasmid (containing the coding sequence of puromycin in frame with FLPo, Addgene #124835). Three days after nucleofection, antibiotic selection with 0.2 pg/ml puromycin (Calbiochem; 540411) was induced for 10 days. Cells were then fixed and immunostained with an anti-HA-tag antibody.

Generation of lentiviral CDH1- and MAB21L2-promoter reporter constructs and lineage tracing of JCF and mesothelial epicardium

For the generation of the lentiviral transfer vector carrying a flippase under control of the human ~1.37 kb CDH1 promoter, RFP from the lentiviral pHAGE-Ecadherin-promoter-RFP plasmid (Addgene; 79603) was replaced by a flippase (FLP) from the plasmid pCAGS-T2A-FLP (Addgene; 123845). Lentiviral transfer vectors carrying a tamoxifen-inducible flippase under the control of the human ~1.88 kb MAB21L2 promoter (chr4:150, 581, 151-150, 583, 029) were synthetized by Vectorbuilder.

Lentiviruses were produced in HEK293T cells by transient co-transfection of the lentiviral transfer vector, the CMVDR8.74 packaging plasmid and the VGV.G envelope plasmid using Fugene HD (Promega; E2311). Viral supernatants were harvested after 48 hours and used for infection of epicardioids derived from the AAVSl-CAG-FRT-flanked STOP-mKate2-HA reporter hiPSCs in the presence of 8 pg/ml polybrene (Sigma-Aldrich; 107689).

For lineage tracing of JCF cells, (early-)epicardioids were infected at day 3 with the MAB21L2- promoter-FLP ^ERT2 lentivirus and 2.5 pM 4-OHT (Sigma-Aldrich; H6278) was applied at days 4 and 5 or d7 and 8 to induce flippase expression. They were then harvested at day 8 or day 12 for immunofluorescence analysis. For lineage tracing of mesothelial epicardial cells, epicardioids were infected at day 15 with the CDHl-promoter-flippase lentivirus and harvested at day 18 or day 24 for immunofluorescence analysis.

Statistics

Statistical analysis was performed with GraphPad Prism version 9.1.0. (La Jolla California, USA). Box-and-whiskers plots indicate the median, 25th and 75th percentile, with whiskers extending to the Sth and 95th percentiles; bar graphs indicate the mean ± SEM and all data points, unless otherwise indicated. Normally distributed data from two experimental groups were compared by Student's t-test, otherwise a Mann-Whitney-Wilcoxon test was applied. Normally distributed data from more than two experimental groups were compared using one- or two-way analysis of variance (ANOVA). In the case of multiple comparisons, an appropriate post hoc test was applied as indicated. A p-value < 0.05 was considered statistically significant.