Engineered Heart Tissues formed with Cardiac Progenitors and Differentiated Cardiomyocytes Exhibit Similar Physiologic Properties at Differentiation-Matched Timepoints

Abstract

Congenital heart diseases (CHD), including single ventricle heart defects such as hypoplastic left and right heart syndromes, remain a leading cause of neonatal death and long-term morbidity. Regenerative medicine approaches hold great therapeutic promise for treating single ventricle disease, specifically through the use of human pluripotent stem cell-derived cardiomyocytes (iPSC-CM) to generate pulsatile conduits capable of growing and developing over time within the recipient. However, current strategies for rapidly fabricating large-scale engineered heart muscle to create such conduits face limitations, including the shear stress generated during most bioprinting processes along with harsh enzymatic treatments required for initial singularization of cells prior to bioprinting, which together can compromise cell viability and downstream tissue function. Here, we explored the use of induced pluripotent stem cell-derived cardiovascular progenitors (iPSC-CVP) as an alternative to fully differentiated cardiomyocytes as a potential cell source for future biomanufacturing efforts. We demonstrate that iPSC-CVP can be used to form functional engineered heart tissues with similar electrophysiological properties to tissues formed from fully differentiated iPSC-CM, while also being more amenable to enzymatic dissociation and mechanical manipulation. Our results suggest that iPSC-CVP may be an ideal cell population for future efforts in biofabrication of contractile structures such as engineered heart muscle and pulsatile conduits.

Introduction

Congenital heart diseases (CHD) are the most prevalent birth defect and often result in neonatal death without corrective surgeries. While lifesaving, these surgeries can leave patients with suboptimal hemodynamics due to abnormal cardiac anatomy, resulting in reduced quality of life and high mortality rates^{1, 2}. Regenerative medicine approaches have been proposed as an alternative treatment paradigm. Human embryonic stem cells (hESC) or induced pluripotent stem cell (iPSC) derived cardiomyocytes (hESC-CM/iPSC-CM) can be generated en masse^3-6 and directly injected into the myocardium to induce repair^{7, 8}, or can be implanted using heart muscle grafts or patches^9-11. The relative functional immaturity of iPSC-CM and hESC-CM^{12, 13} remains a significant challenge in these regenerative approaches. In particular, because of their electrophysiologic immaturity, pluripotent stem cell derived cardiomyocytes have been found to be arrhythmogenic when delivered to the myocardium in pigs¹⁴ and primates⁷, strongly suggesting these cells could cause arrhythmia in patients as well. A multitude of studies have focused on applying epigenetic^{15, 16}, chemical^17-19 and biophysical^{20, 21} stimuli toward the goal of maturing human pluripotent stem cell derived cardiomyocytes to overcome this fundamental obstacle.

One CHD where regenerative medicine approaches are being investigated is single ventricle heart defects, including hypoplastic left heart syndrome (HLHS) and hypoplastic right heart syndrome (HRHS), where children are born with one functioning ventricle. Children born with single ventricle heart defects undergo a series of surgeries that bypass venous return from the heart directly into the lungs, with the one functioning ventricle acting as the systemic ventricle. However, the long-lasting effects of this passive blood flow of deoxygenated blood to the lungs without a pulsatile right ventricle include multi-organ failure, most seriously of which is liver failure, due to lifelong elevated venous pressures. Patients with a surgical “Fontan circulation” often succumb in the third or fourth decade unless they receive both a heart and liver transplant. Regenerative medicine approaches including a pulsatile conduit²² - which could be used in place of non-contractile cavopulmonary grafts to restore pulsatile pulmonary flow – to treat single ventricle heart defects, have been suggested to offer a path leading a longer-lasting functional cure^23-27. Ideally, a pulsatile conduit would preclude the need for children to undergo a Fontan circulation and thus prevent damage to other organs. In this scenario, the conduit would also need to have the capacity to grow with the child. While promising, the aforementioned challenges with iPSC-CM functional maturity^{12, 28} limit the performance of the current generation of pulsatile conduits²⁴.

Several approaches have been used to create cardiac patches/grafts and pulsatile conduits. Of these, mold-casting^{9, 11}, and bio-printing^{26, 29} are especially popular in the field. Maintaining high cell viability is a central requirement for tissue biofabrication, which presents challenges to using stress-sensitive, non-proliferative cells such as iPSC-CM. For example, in extrusion bioprinting, cells within the bioink experience shear stress, which increases as the nozzle diameter decreases and bioink viscosity increases³⁰. On the other hand, within light-based bioprinting, cells experience thermal and radiative stress³¹. Aside from the stresses generated during the bioprinting process, additional chemical and mechanical stress is generated during dissociation of differentiated iPSC-CM into single cells to create bioinks. This is due to buildup of extracellular matrix (ECM), particular by fibroblast-like bioproducts of the differentiation, which becomes especially intensive during later stages of iPSC-CM differentiation^32-34. This extra ECM significantly increases incubation times required for enzymatic dissociation of these cells, often resulting in a requirement for more vigorous mechanical trituration, further compounding difficulties with cell viability.

In this study, we sought to address the possibility of using multipotent iPSC-CM progenitors, rather than fully-differentiated iPSC-CM, for engineered tissue formation. We selected these progenitors partly because they occur early in the iPSC-CM differentiation process, before cells have produced extensive ECM. During the process of in vitro cardiomyocyte differentiation from human pluripotent stem cells, differentiating cells pass through multiple transient states including a cardiovascular progenitor (e.g. iPSC-CVP) stage³⁵ that is denoted by expression of markers including ISL1 and Nkx2.5^{36, 37}. iPSC-CVP are a transient, multipotent population with the ability to commit to cardiomyocyte, endothelial and smooth muscle cell-like cells^{35, 38}. Like iPSC, iPSC-CVP are easily dissociated enzymatically³⁹, allowing these cells to be manipulated in culture with minimal risk of damage due to excessive enzymatic digestion. However, unlike pluripotent stem cells, iPSC-CVP have not demonstrated risks of forming teratomas in vivo, and in fact, transplantation of iPSC-CVP has led to substantial cardiac repair^{40, 41}.

A potential obstacle to using iPSC-CVP rather than iPSC-CM for engineering heart muscle is the chance that these progenitors could differentiate toward undesirable cell types, or form heart tissue with suboptimal function. To assess this possibility, we formed micro-heart muscle arrays (μHM^42-44) with either iPSC-CM or iPSC-CVP (Fig. 1a). We assessed the dynamics of μHM formation and compared the electrophysiologic properties of μHM formed from iPSC-CM or iPSC-CVP . Interestingly, at a common differentiation timepoint (differentiation day 30), μHM seeded with iPSC-CVP had similar morphology and calcium handling properties when compared to those seeded with committed cardiomyocytes. Optical mapping studies likewise demonstrated similar conduction velocity, regardless of the differentiation stage at which μHM were formed. Finally, single cell RNA sequencing showed that μHM derived from iPSC-CVP and iPSC-CM were comprised of a comparable final cell composition at cardiac differentiation day 30, suggesting that iPSC-CVP have sufficient cues to commit to the cardiomyocyte lineage within a high-density tissue format. These results suggest that it is feasible to form engineered heart tissues from cardiac progenitors rather than fully differentiated cardiomyocytes. This methodologic strategy may prove advantageous in applications like biofabrication, which often requires extensive chemical and mechanical cellular manipulation, including shear, thermal, and radiative stresses³¹.

Figure 1.

Schematic of the study timeline. μHM are seeded either with multipotent iPSC-CVP (differentiation day 5), or fully differentiated iPSC-CM (differentiation day 15), then matured in situ to a common timepoint, differentiation day 30, for assessing tissue physiology.

Methods

Human Induced Pluripotent Stem Cell Culture and Cardiomyocyte Differentiation

Cardiomyocytes were derived from hiPSC (Wild-type C, WTC; Coriell Repository # GM2525) harboring a single-copy of CAG-driven GcaMP6f^{45, 46}) via small molecule driven changes in Wnt signaling as originally described by Palecek et. al^{4, 43, 44}. Briefly, iPSCs were maintained in E8 (Essential 8) pluripotency media on tissue culture plastic coated with GelTrex^™. iPSCs were passaged using gentle dissociation reagent (PBS supplemented with 0.5mM Ethylenediamine tetraacetic acid, EDTA, and 1.8g/L sodium chloride⁴⁷). For differentiation, iPSCs were plated into E8 media with 10 μM ROCK Inhibitor (Y27632) at a density of 37,500 cells/cm² and cultured until they achieved 70-80% confluency. At this point, day 0 of the differentiation, media was switched to RPMI1640 (Roswell Park Memorial Institute Medium 1640) with B27i (B27 supplement without insulin) and 6μM CHIR99021. 48 hours later, Wnt signaling was inhibited using 5 μM IWP-2 in RPMI with B27i. Between days 0-4 of differentiation, cultures were supplemented with 150μg/mL L-ascorbic acid. From day 6 onward, cells were cultured in RPMI 1640 with B27c (B27 complete supplement). Beating cells were typically observed by differentiation day 10.

Flow Cytometry Analysis

Cardiomyocyte purity was quantified using flow cytometry^{4, 44}. Briefly, cells were dissociated at either differentiation day 5 (iPSC-CVP, dissociated with Accutase^™) or differentiation day 15 (iPSC-CM, dissociated with 10x TrypLE Select). 2 million cells were fixed (4% paraformaldehyde in PBS). Following fixation, cell pellets were stained for cardiac troponin (cTnT, 2.5μg/mL mouse anti-cTnT, clone 13-11) in blocking buffer (PBS containing 4% normal goat serum and 0.5% wt/v saponin), followed by Alexa Fluor 488 conjugated goat anti-mouse IgG (10μg/L) in blocking buffer. Following staining, cells were analyzed on an ACEA NovoCyte flow cytometer (San Diego, CA). iPSCs were used as negative controls for TNNT staining, and data were analyzed using NovoExpress software, ver 1.6.2.

Fabrication of Micro-Heart Muscle Devices

Micro-heart muscle (μHM) forming devices were prepared in 35 mm tissue culture dishes by attaching 1mm-thick stencils containing “dog-bone” shaped through-holes to the tissue culture surface^{43, 48}. First, 3D-printed masters (FormLabs gray resin) with positive features (the opposite of the through-holes shown in Figure 1) designed using SolidWorks (Dassault Systemes SE, Velizy-Villaczoublay, France) using Hydrogel Assisted Stereolithographic Elastomer (HASTE) prototyping⁴⁹. Briefly, the resin 3D prints were replica molded into 1.5% wt/v agar. The agar negatives were then used to cast poly(dimethylsiloxane) (PDMS) (10:1 ratio of Sylgard 184 base:crosslinker) replicas of the original 3D print. PDMS was crosslinked overnight at 37°C, removed from agar, and post-crosslinked at 60°C overnight. PDMS replicas of the 3D print were next oxidized (Harrick plasma, Ithaca, NY; pressure/flow rate: 560-680 mTorr) for 90 seconds at 30W, and then exposed to vapors of Trichloro (1H, 1H, 2H, 2H – perfluorooctyl) silane for 72 hr. Passivated PDMS replicas of the 3D print were then used to cast Sylgard 184 into 1mm-thick stencils containing dogbone-shaped through-holes (Fig. 1). Following crosslinking of the PDMS sheet with an array of through-holes, individual dogbone molds (Fig. 1) were cut and autoclaved.

Sterile tissue culture polystyrene dishes were coated with bovine plasma fibronectin (20 ug/mL in PBS)⁴³. In parallel, the sterilized tissue-forming stencils were placed in a covered glass beaker and soaked with methanol. Coated tissue culture substrates were washed 3 times with PBS, followed by a single wash with methanol, and stencils were attached to the exposed tissue culture plastic. The dishes were then placed in a 60°C oven to let the methanol evaporate for at least 8 hours, physically adsorbing the stencils to the tissue culture plastic^{48, 50}. The assembled deviceswere then disinfected via exposure to sterile 70% ethanol. After complete evaporation of the ethanol, devices were coated in a sterile solution of 1% wt/v Pluronics F127 in PBS to minimize cell adhesion to the stencil. Devices finally washed and maintaind in sterile PBS at 4°C for up to one week, until the time of cell seeding.

Micro-Heart Muscle Formation

Micro-heart muscle arrays (μHM) were formed by loading a high density of iPSC-CVP or iPSC-CM into the dogbone-shaped recesses of the PDMS stencils atop the FN-coated tissue culture substrates (Fig. 1). In both cases, cells were enzymatically dissociated and passed through a 40μm mesh filter to produce a single-cell suspension.

For iPSC-CVP isolation, iPSC-CM differentiations were dissociated and singularized by exposure to Accutase^™ on day 5 of differentiation (Fig. 1). Enzyme activity was quenched with KO DMEM (Knockout Dulbecco’s modified Eagle medium) supplemented with 20% FBS (fetal bovine serum), 1% Non-Essential Amino Acids, 1% GlutaMax and 1% sodium pyruvate. iPSC-CVP were then resuspended into cell resuspension media: embryoid body 20 media (EB20: KO DMEM containing 20% FBS, 1% NEAA (Non-essential amino acids), 1% GlutaMax, 1% Sodium Pyruvate, 150 ug/mL L-ascorbic acid, 4 ug/mL vitamin B12, and 3.2 ug/mL penicillin) supplemented with 10 μM Y27632⁴⁸. Stencils were seeded with 210,000 cells in a volume of 3μL and placed into the incubator (37°C, 5% CO₂) for 2 hours to allow initial cellular agglomeration before the additional cell resuspension media to cover the tissues. Upon observation of beating (24 – 48 hours post cell seeding), media was switched from cell resuspension media to standard hiPSC-CM media (RPMI/B27c). Media was then exchanged every 2-3 days until the tissues reached cardiomyocyte differentiation day 30 (day 25 of tissue; Fig. 2a). A subset of iPSC-CVP from each differentiation were allowed to continue differentiation until day 15 to allow tissue formation at this later timepoint, and for assessing the ultimate success of the differentiation. Only iPSC-CVP from differentiation batches that went on to yield a minimum of 50% cTnT positive cells at day 15 of differentiation were used for subsequent analysis of tissue formation and physiology (Fig. 1).

Figure 2. Study Timeline.

A.) Schematic of the study timeline. iPSC-CVP (differentiation day 5) or iPSC-CM (differentiation day 15) were enzymatically singularized and used to form tissue, which was then cultured until cardiomyocyte differentiation day 30 B.) Representative FACs analysis of the cardiomyocyte marker cTnT. C.) Quantification of the percentage of cTnT positive cells shows no discernable expression of cTnT in iPSC-CVP, while most cells at the iPSC-CM stage are cTnT+. Error bars: SD, n = 4 independent differentiation batches. *** p < 10^−3,, unpaired 2-way t-test.

For iPSC-CM isolation, cardiomyocytes at day 15 of differentiation were dissociated and singularized by exposure to 10x TrypLE Select for 15 min (37°C, 5% CO₂). Cells were gently triturated before the reaction was quenched with cell suspension media. iPSC-CM were then seeded into dogbone molds using the same procedures and cell density used for iPSC-CVP. Spontaneous beating was typically observed within 48 hr of tissue seeding, at which point the media was changed to RPMI/B27c. The formation and stability of iPSC-μHM were monitored until day 30 of the differentiation (day 15 of tissue).

These differentiation timepoints were chosen to compare μHM made with multipotent iPSC-CVP versus μHM made from fully differentiated iPSC-CM. Day 5 of differentiation represents the earliest timepoint at which the cell population is primarily composed of multipotent cardiovascular progenitors. By day 15 of differentiation, the majority of cells are committed cardiomyocytes, leading to robust population-level expression of cardiac troponin T (cTnT; Fig. 2b). Prior studies show that at the day 15 timepoint, few detectable multipotent progenitors remain, and, instead, cTnT negative cells at this timepoint tend to be CD90 positive fibroblast-like stromal cells⁵¹. Within μHM, we allowed the differentiation as allowed to continue to a common timepoint, day 30. This common timepoint was chosen because iPSC-CM continue to mature over time in culture¹⁴.

Morphology Measurements of Micro-Heart Muscles

Brightfield micrographs obtained on an Evos XL cell imaging system (Thermo Scientific) were used to monitor μHM formation and compaction. The extent and dynamics of μHM formation were quantified (ImageJ) via: 1) the overall available area of the mold occupied by the tissue; 2) the overall length of the mold occupied by tissue, and 3) the width of the tissue within the thinner rectangular “shaft” region that connects tissue within the square-shaped “knobs” on either end.

Spontaneous Beating and Calcium Handling of Micro-Heart Muscles

Spontaneous beating and calcium dynamics were monitored throughout the course of tissue formation and in situ maturation. Calcium dynamics were visualized using the GCaMP6f reporter genetically encoded into the iPSC line. The tissues were imaged using a 4x objective at 100 frames per second using an Eclipse Ts2R epifluorescence microscope (Nikon; Tokyo, Japan) with a temperature-controlled stage set to 37°C (Tokai Hit; Shizuoka, Japan) and a Hamamatsu ORCA-Flash4.0 V2 digital CMOS camera. Videos were recorded 5 days following tissue formation and then again when cardiomyocytes in the tissues reached differentiation day 30. At the endpoint of differentiation day 30, calcium transient waveform morphology was analyzed using open source Matlab-based software, which reported beat rate, Ca²⁺ transient amplitude, upstroke time and time to decay 75% from peak amplitude (τ₇₅)⁴⁶. Upstroke time and τ₇₅ were beat-rate corrected using Fredericia’s method.

Optical Mapping, Effective Refractory Period and Conduction Velocity

Optical mapping was performed at the end of the study (cardiomyocyte differentiation day 30) using a high-resolution imaging system (MiCAM03 N256 Single Camera System, SciMedia, Costa Mesa, CA). Tissues were stained with the voltage sensitive dye Di-4-ANEPPS in RPMI/B27c for 2 hours, then allowed to recover in RPMI/B27c without Di-4-ANEPPS for 30 minutes. Just before mapping, the media was replaced with a Tyrode solution. Each 35mm dish containing 3-6 μHM was placed onto a temperature-controlled plate (Alpharetta, GA) set to 37°C. All tissues were initially mapped without stimulation to obtain their spontaneous beating rate and activation. Subsequently, tissues were stimulated with a custom-built point stimulator into the knob at one end of the tissue. The stimulation protocol increased voltage in discrete integers until capture at 1 Hz to determine the excitation threshold voltage. All subsequent point stimulation and ERP measurements were done at 1.5 times the excitation threshold voltage. The effective refractory period (ERP) of the tissue was determined by stimulating at an S1 drive train of 1 Hz, followed by a single S2 delivered at a shorter stimulus interval, starting at 950 ms, with decrements of 50 ms. The longest interval at which the tissue failed to capture was recorded as the effective refractory period. Conduction velocity through the shaft was subsequently determined using the open source MATLAB-based Rhythm software package developed originally by Laughner et al⁵². To ensure rigor and reproducibility, tissues were excluded if they exhibited any of the following criteria: 1) conduction block on the activation map, as indicated by the absence of a clear color gradient going left to right through the shaft without sharp line(s) of conduction block or reverse activation within the shaft of the tissue; 2) fewer than 20 vectors generated on the vector map during analysis; or 3) a standard deviation of the conduction velocity that was ≥ 1/3 of the mean conduction velocity (conduction velocities were measured computationally from all vectors on the map).

Single Cell RNA Sequencing

A total of 23 tissues were pooled and dissociated to retrieve single cells. Briefly, μHM were washed twice in PBS, then transferred to microtubes which were precoated with BSA (bovine serum albumin). Next, μHM were resuspended in 10x TrypLE Select and vortexed briefly, followed by incubation at 37°C and 5% CO₂ for 30 minutes, with brief gentle vortexing every 10 minutes. The reaction was then quenched in 2x volume of wash buffer (1:1 mixture of DMEM/F12 to 1% BSA, supplemented with 0.01 mg/ml DNase I from Bovine Pancreas). Cells were singularized with gentle pipetting. Cells were then pelleted at 300 g for 5 minutes and washed twice in wash buffer, followed by filtration with a 40 μm filter to remove debris or cell aggregates. Next, cells were counted, and cell viability confirmed via trypan blue exclusion using a Countess 3 automated cell counter. Cell concentration was then adjusted to 5x10⁵ cells/ml, and cells were stained with 5 μl/mL 7-AAD to identify dead cells. 7AAD negative, viable cells were sorted on a Beckman Coulter MoFlo cell sorter at the Siteman Flow Cytometry Core at the Washington University School of Medicine. Unstained controls were used for sample gating. After sorting, recovered cells were pelleted at 300 g for 5 minutes, followed by resuspension at 700-1000 cells/μl in 1x PBS supplemented with 0.04% BSA.

Library construction and single-cell RNA-sequencing of the singularized μHM were performed by the Genome Technology Access Center at Washington University School of Medicine. cDNA was prepared after the Gel Beads-in-Emulsion generation and barcoding, followed by the GEM-RT reaction and bead cleanup. Purified cDNA was amplified for 11-13 cycles before being cleaned up using SPRIselect beads. cDNA concentration was determined using a Bioanalyzer. GEX libraries were prepared as recommended by the 10x Genomics Chromium Single Cell 3’ Reagent Kits User Guide (v3.1 Chemistry Dual Index) with appropriate modifications to the PCR cycles based on the calculated cDNA concentration. For sample preparation on the 10x Genomics platform, the Chromium Next GEM Single Cell 3’ Kit v3.1, 16 reactions (PN-1000268), Chromium Next GEM Chip G Single Cell Kit, 48 reactions (PN-1000120), and Dual Index Kit TT Set A, 96 reactions (PN-1000215) were used. The concentration of each library was accurately determined through qPCR utilizing the KAPA library Quantification Kit according to the manufacturer’s protocol (KAPA Biosystems/Roche) to produce cluster counts appropriate for the Illumina NovaSeq X Plus instrument. Normalized libraries were sequenced on a NovaSeqX 10B Flow Cell using the 151x10x10x151 sequencing recipe according to manufacturer protocol. Read 1 was trimmed to the 10x Genomics recommendation of 28bp. A median sequencing depth of 50,000 reads/cells was targeted for each Gene Expression Library.

Clustering and differential expression analysis was performed using the Seurat v.5.0 package according to provided tutorials (https://satijalab.org/seurat/). Briefly, CellRanger matrices were imported to R. Cells with fewer than 200 or greater than 7500 genes and cells with greater than 20% mitochondrial gene content were excluded from downstream analysis. Normalization was performed using the scTransform workflow. Uniform manifold approximation and projection (UMAP) and differential expression analysis were performed. Cell clusters were identified and annotated utilizing FindMarkers and FeaturePlot functions.

Immunofluorescence Staining and confocal analysis of μHM

On cardiomyocyte differentiation day 30, μHM were fixed (4% paraformaldehyde in PBS) for visualization of sarcomeric -Actinin (cardiomyocyte Z-disk marker) and vimentin (stromal cell intermediate filament marker)⁴². Briefly, tissues were permeabilized with 0.5% Triton X-100 for 10 minutes at room temperature, then washed three times with PBS. The tissues were then blocked using blocking buffer (3% Normal Goat Serum, 3% BSA and 0.5% Triton X-100 in PBS) for 1 hour at room temperature) and probed with primary antibodies against sarcomeric -Actinin (EA-53, 1:500 dilution) and Vimentin (1:200 dilution) over 72 hours at 4°C. After washing primary antibodies with blocking buffer, the tissues were probed with secondary antibodies for 2 hours at room temperature. The secondary antibody solution was removed, and tissues were washed with 0.1% Triton X-100 in PBS. Tissues were then transferred to a glass slide and mounted in Prolong Gold and sealed with a coverslip. Stained tissues were visualized on an Olympus Fluoview 1200 laser scanning confocal microscope. Representative maximum-intensity projections were used to assess gross cellular organization and morphology.

Statistical Analysis

At least 12 μHM obtained from ≥3 independent differentiation batches were used for analysis of tissue formation, calcium handling and conduction. scRNA sequencing and bioinformatic analysis was performed on a minimum of 5800 cells at day 30 of cardiac differentiation (tissue day 25 for tissues form from hiPSC-cardiac progenitors, tissue day 15 for tissues formed from hiPSC-CM). Data visualization and statistical analysis were performed using GraphPad Prism (version) software (μHM morphology and physiology) and Seurat v.5.0 package for R (scRNAseq). Comparisons between tissues made from hiPSC-cardiac progenitors vs. hiPSC-CM were made by 2-way student’s t-test. p values of 0.05 were considered statistically significant.

Results

Micro-Heart Muscle Tissue Self-Assembly and Compaction Is Similar Between iPSC-CVP and iPSC-CM

We first characterized the iPSC-CVP (day 5 of differentiation) and iPSC-CM (day 15 of differentiation) using FACS analysis of cTnT, a marker of differentiated cardiomyocytes. In both cases, cells were enzymatically dissociated and passed through a mesh filter to produce a single-cell suspension. iPSC-CVP were more easily dissociated, requiring less time in enzyme and a lower degree of trituration for singularization when compared with iPSC-CM, consistent with a prior report by Vahdat and colleagues³⁹. Consistent with the iPSC-CVP being a progenitor population, there was no discernable expression of cTnT, whereas iPSC-CM from matched differentiation batches exhibited robust expression with 77±8% cTnT positive (Fig. 2b, c). Neither progenitors nor cardiomyocytes were further purified prior to seeding the devices.

After initial seeding, μHM formed from either iPSC-CVP or iPSC-CM remained grossly intact, with no qualitative differences in cell loss after initial seeding. Overall dynamics of cell compaction and self-assembly into organized tissues, as characterized by measuring the end-to-end distance of the tissue, suggested similar compaction rates of μHM formed from iPSC-CVP versus μHM formed from hiPSC-CM (Fig 3a,b). Likewise, by cardiac differentiation day 30, we observed no significant differences in end-to-end μHM length, shaft width, or area occupied by tissue caused by the cell source used to form μHM (Fig. 3c).These data suggest that tissue self-assembly and compaction dynamics are remarkably similar whether tissues are formed from fully differentiated cardiomyocytes versus committed cardiac progenitors.

Figure 3. Micro-Tissues Seeded with iPSC-CVP and iPSC-CM Demonstrate Similar Morphology.

A.) Representative micrographs of tissue compaction following seeding of either progenitors (top) or cardiomyocytes (bottom). B.) Dynamics of tissue compaction (quantified via analysis of end-to-end distance of the tissue. C-E.) Quantitative analysis of C) end to end distance, D) shaft width, and F) overall tissue-occupied area as a fraction of overall available area at day 30 of cardiomyocyte differentiation. Each data point represents one tissue, from at least 3 independent batches. Differences between groups were not statistically significant (unpaired 2-way t-test, with significance defined as p < 0.05). Error bars: SD. Scale bar: 2mm.

Calcium Handling is Similar in μHM Formed From iPSC-CVP and iPSC-CM

Prior studies have indicated minimal occurrence of cardiac arrhythmias in primate hearts injected with iPSC-CVP⁴¹, whereas hearts injected with pluripotent stem cell derived cardiomyocytes exhibited initial signs of arrhythmia (e.g. ventricular tachycardia)⁵³. However, it is unclear whether engineered tissues and/or grafts formed from iPSC-CVP would be similarly free form arrhythmias or pro-arrhythmic behavior. Thus, we analyzed calcium transient propagation 5 days after cell seeding (day 10 of differentiation for tissues seeded with iPSC-CVP, and day 20 of differentiation for tissues seeded with iPSC-CM). At this timepoint, we observed a high occurrence of re-entrant spiral waves in tissues made from iPSC-CVP (Fig. 4), with an average of 53±35% μHM seeded from progenitors from each differentiation batch exhibiting re-entrant behavior. In contrast, no re-entry was observed in μHM 5 days after seeding with iPSC-CM. Interestingly, despite the markedly high incidence of re-entrant waves in iPSC-CVP derived μHM at this early timepoint, they completely resolved by day 30 of differentiation (Fig. 4). The spiral waves are likely to result from slower conduction velocity of the iPSC-CVP derived μHM within the first several days after tissue formation, which is corroborated by the much longer time required for wave propagation in these μHM (Fig. 4). These data are consistent with our prior observations of μHM formed from iPSC-CM, which become more organized, and exhibit improved conduction velocity over time after cell seeding⁴³.

Figure 4. Early-stage calcium propagation abnormalities resolve in iPSC-CVP seeded micro-tissues.

Calcium video analysis of tissues seeded as iPSC-CVP at five-day interval reveals a unique observation of re-entrant spiral waves for tissues seeded as iPSC-CVP after five days of culture, which subsequently resolves. MATLAB generated activation videos and maps were provided for quantitative supplement. Scale bar = 1000 μm.

Analysis of Calcium Transients obtained from Spontaneous Calcium Imaging

Detailed analysis of calcium-handling kinetics within tissues at the terminal timepoint (day 30 of cardiomyocyte differentiation) demonstrated no differences in spontaneous beat rate, upstroke duration, or decay kinetics in μHM seeded with iPSC-CVP versus those formed with iPSC-CM (Fig. 5). We did observe a slight, albeit statistically significant increase in Ca²⁺ transient amplitude in μHM seeded with progenitors as opposed to those seeded with iPSC-CM (Fig. 5). Overall, these data suggest that, although tissues formed from iPSC-CVP initially exhibit poor organization of calcium handling, cells within these tissues eventually differentiate into cardiomyocytes which exhibit robust calcium handling.

Figure 5. Similar calcium transient dynamics in μHM at cardiac differentiation day 30 regardless of population used to seed tissues.

A-D) Calcium transients from spontaneously contracting μHM at differentiation day 30 were assessed for a) spontaneous beat rate, b) peak amplitude (GCaMP6f fluorescence), c) upstroke duration, and d) time to decay 75% from peak to baseline (τ₇₅). Each data point represents one μHM from at least 3 independent batches. *p< 0.05, unpaired 2-way t-test. Error bars: SD.

Conduction Properties of Micro-Heart Muscles

μHM formed from iPSC-CVP, and iPSC-CM have similar conduction velocity and effective refractory period at cardiac differentiation day 30

Tissues formed from iPSC-CVP had a slight, but statistically significant, reduction in excitation threshold voltage as compared to μHM formed from iPSC-CM (Fig. 6a). This suggests that tissues made from progenitors have increased excitability as they require less voltage for depolarization, reflective of a more mature differentiation state with upregulation of the sodium current. Functional upregulation of sodium current is a well-established indicator of cardiomyocyte maturation. In more mature phenotypes there is a transition from calcium dependent depolarization to sodium dependent action potential⁴³. In contrast, we observed no significant change in the effective refractory period (ERP), a measure of repolarization (Fig. 6b).

Figure 6. Micro-Tissues Seeded with iPSC-CVP and iPSC-CM Demonstrate Similar Electrophysiologic Properties.

A-B.) Voltage of capture and effective refractory period (ERP) (respectively) were obtained during point stimulation. C.) Vector maps were generated by MATLAB during the process of calculating conduction velocity. D.) Conduction velocity values obtained from MATLAB to determine the speed of conduction throughout the tissue. E.) Activation maps were also generated by MATLAB to determine the isochronal profiles of the conduction. F.) Additionally, we calculated the wavelength of the conduction profile, as a product of the conduction velocity and the effective refractory period. This predicts whether re-entry of the signal is possible. Analysis is based on data from independent timepoint studies (n= 5 biological replicates). ***p<10^−3, unpaired 2-way t-test.

Another important metric of tissue maturation is conduction velocity, the speed of electrical propagation through heart tissue^{52, 54}. Optical mapping of point stimulated tissues revealed that μHM formed from iPSC-CVP were similar in their conduction properties to μHM formed from iPSC-CM aged to the same differentiation timepoint, similar to what was seen with the calcium transient analysis. Vector maps (Fig. 6c) demonstrated organized, linear propagation, with conduction velocities shown in Fig. 6d. Activation maps likewise suggested against qualitative changes in action potential propagation in μHM formed from iPSC-CM versus iPSC-CVP (Fig. 6e). Tissue wavelength is calculated by multiplying the effective refractory period by the conduction velocity^{21, 22}. Sustained reentry will only be possible if the pathlength of the reentrant circuit is greater than the tissue wavelength, so increasing tissue wavelength by increasing either conduction velocity or refractory period reduces the likelihood of forming reentrant circuits and developing reentrant arrhythmias²². Since ERP and conduction velocities were similar between μHM made from progenitors versus cardiomyocytes, it follows that wavelength was likewise similar (Fig. 6f). Like our findings on calcium propagation and calcium handling kinetics in tissues at day 30 of cardiomyocyte differentiation, these data suggest that tissues formed from iPSC-CVP versus iPSC-CM have very similar physiologic function.

3.4. Single Cell RNA Sequencing

Single cell RNA sequencing demonstrates similar cell identity and transcriptomic state between iPSC-CVP and hiPSC-CM at differentiation day 30

iPSC-CVP are a transient, multipotent population with the ability to commit to cardiomyocyte, endothelial, and smooth muscle cell-like cells. Differentiation efficiency from this stage into hiPSC-CM under uninterrupted 2D culture conditions is relatively high; however, it is not known whether cellular dissociation and reseeding into engineered tissue models disrupts this differentiation trajectory. We therefore asked how the terminal cell populations from μHM seeded with hiPSC-CVP compared to those seeded with hiPSC-CM. Based on our findings demonstrating similar overall tissue morphology and electrophysiologic phenotype in μHM formed from cardiovascular progenitors versus cardiomyocytes, we hypothesized iPSC-CVP have the necessary cues at this early stage of differentiation to continue to differentiate into cardiomyocytes, and that this would result in similar cellular composition of μHM at a common differentiation timepoint (day 30) regardless of the cell seeding population. To address this hypothesis, we performed single cell RNA-sequencing on μHM at day 30 of cardiomyocyte differentiation. A total of 23 μHM each from the hiPSC-CVP and hiPSC-CM groups were dissociated, and viable cells were isolated by sorting for 7AAD negative population by flow cytometry, followed by 10x Genomics Chromium Single Cell library preparation and Next Generation sequencing. Unsupervised clustering, integration, and gene expression was performed with Seurat. After QC and filtering, our final dataset included 5854 cells from μHM seeded with hiPSC-CVP and 6359 cells from μHM seeded with hiPSC-CM.

Uniform manifold approximation projection (UMAP) comparison of local and global relationships suggested a strong overlap of cellular composition between tissues seeded as iPSC-CVP or iPSC-CM (Fig. 7a). We identified 16 cell clusters, each present across both of our experimental conditions (Fig. 7b). We next labeled these cell populations based on expression of the following cell-specific markers: TNNT2 (cardiomyocytes), COL1A1 (fibroblasts), VIM (fibroblasts), PECAM1 (endothelial cells), SOX2 (pluripotency marker), ACTA2 (vascular smooth muscle cells and immature cardiomyocytes), TAGLN (smooth muscle cells), PDGFRB (pericytes), EFEMP1 (epicardial cells), and POU5F1 (pluripotency marker) (Fig. 7c). Of our 16 clusters, nearly all expressed high levels of the cardiomyocyte marker TNNT2: 12 represent cardiomyocyte subpopulations, with an additional 3 populations identified as immature cardiomyocytes which retain proliferative capacity based on expressing both TNNT2 and TOP2A. This corresponded to the high frequency of cardiac troponin positive cardiomyocytes identified at day 15 of the differentiation for the iPSC-CVP and iPSC-CM populations used in this study (Fig. 2c) and suggests high differentiation efficiency regardless of seeding timepoint.

Figure 7. Single Cell RNA Sequence analysis of Cellular Composition in μHM formed from progenitors versus differentiated cardiomyocytes.

A.) Combined Uniform Manifold Approximation Projection (UMAP) of cells recovered from micro-tissues seeded as iPSC-CVP (orange) and iPSC-CM (blue) after quality control and filtering shows a strong global overlap between cell populations recovered from tissues seeded as iPSC-CVP or iPSC-CM. B.) UMAP clustering (left) and quantification of cell populations (right) shows strong similarities in cell populations recovered in tissues seeded as iPSC-CVP or iPSC-CM. C.) Expression heat maps of cell markers for expected or possible cell types define the cell populations recovered. Regardless of cellular differentiation stage at the time of tissue seeding, most cells recovered at the end of the study are defined by positive expression of TNNT2 indicating the majority of cells are cardiomyocytes, with a small population defined by high VIM and COL1A1 expression representing fibroblast-like stomal cells.

We also identified one cluster of cells with high VIM and COL1A1 expression, representing a fibroblast-like stromal population, which is expected to be present at low levels after hiPSC differentiation into hiPSC-CM. In contrast, markers for endothelial cells (PECAM1) and smooth muscle cells (ACTA2) were largely absent in μHM-constituent cells at differentiation day 30. Populations expressing ACTA2 along with high levels of TNNT2 likely represent ACTA2⁺ cardiomyocytes rather than true smooth muscle cells and are present only in small numbers. Overall, the proportion of each cell population and the ratio of cardiomyocytes to fibroblasts recovered from day 30 tissues was grossly similar regardless of whether tissues were seeded with iPSC-CVP or iPSC-CM. This demonstrates a high degree of similarity in cell identity and transcriptomic state between the cell populations at differentiation day 30, despite seeding μHM at the two different aforementioned timepoints. These results strongly suggest that the default differentiation mode of cardiomyocytes³⁵ was maintained regardless of whether cells were seeded into tissues at day 5 (iPSC-CVP stage) or day 15 (iPSC-CM stage) of differentiation, despite observation of iPSC-CVP multipotency in prior studies^{35, 38}.

Confocal micrographs of μHM whole-mount stained for cardiomyocytes (sarcomere -actinin, ACTN2) and stromal cells (vimentin, VIM) confirmed that not only was the cellular composition of μHM grossly similar at differentiation day 30 regardless of whether tissues were formed from iPSC-CVP, or iPSC-CM, overall cellular organization and sarcomere quality were grossly similar as well (Fig. 8). Consistent with data on tissue physiology, these observations strongly suggest that the cellular composition and ultimate function of μHM is similar at a common differentiation timepoint regardless of whether the tissues are formed from iPSC-CVP or iPSC-CM.

Figure 8. Immunofluorescent staining of tissues seeded with iPSC-CVP or iPSC-CM.

Representative images of nuclei (DAPI/blue), sarcomeres (Sarcomeric alpha actinin/green), and fibroblasts (vimentin/red) of tissues seeded with either A.) iPSC-CVP or B.) iPSC-CM. Scale bars: 50μm.

Discussion

In this study, we assessed the possibility of forming engineered heart muscle from iPSC-CVP (day 5 of cardiac differentiation) as compared to the standard method of using iPSC-CM (day 15 of differentiation). We characterized the effects of cell population at time of seeding through comparative analysis of tissue morphology, calcium-handling dynamics, electrophysiology, final cell composition, and immunostaining. Surprisingly, we observed a high degree of similarities in compaction dynamics and morphology of the tissues seeded as iPSC-CVP and iPSC-CM (Fig. 3). This suggests that the tissue stability was not affected by the population of cells that were seeded into tissue constructs. Analogously, scRNAseq demonstrated similar overall final cellular composition (Fig. 7) and optical mapping indicating similar conduction velocity and ERP by cardiac differentiation day 30, regardless of whether μHM were formed from iPSC-CVP for iPSC-CM (Fig. 6).

Interestingly, spontaneous propagation of the electrical impulse was different between the iPSC-CVP and iPSC-CM five days after seeding, with re-entrant spiral waves present in μHM formed from iPSC-CVP (Fig. 4). This may stem from the very low conduction velocity of tissues containing progenitor cells, with a reduced tissue wavelength capable of supporting reentrant circuits in our small tissue constructs. Anatomical circuits longer than the wavelength of electrical impulse cause an excitable gap between the head and tail of the circulating wave, which can result in re-entrant spiral waves⁵⁵. By the later differentiation timepoint (cardiac differentiation day 30), these re-entrant spiral waves disappear coincident with increased conduction velocity with continued in situ tissue maturation. This continuous maturation of tissue electrophysiology is consistent with prior observations by our team⁴³ and others¹³. Altogether, these results suggest that grafts formed from iPSC-CVP might require pre-culture before transplantation in vivo if there was a requirement for the graft to be functional at the time of implantation, but after this initial maturation time, tissues formed from iPSC-CVP would function very similarly to tissues formed from iPSC-CM.

We note that several studies reporting on electrophysiology measurements, including our recent work⁴³, show higher conduction velocities compared to those report here, ranging from 7 to 14 cm/sec in the absence of extra measures taken to mature cardiomyocytes⁵⁶. Given the similar levels of organization observed within our μHM (Fig.8), which are both lower than what we previously observed in μHM with geometry more conducive to high cell alignment (lower shaft width and higher shaft length), it is likely that the incomplete alignment of cells in our tissues resulted in slower conduction⁴³. Future studies could discern whether biophysical stimulation of iPSC-CVP versus iPSC-CM results in differential maturation capacity.

Our findings that tissues made from iPSC-CVP were overall virtually indistinguishable in composition, morphology and physiology formed from tissues made from iPSC-CM is consistent with the high efficiency of our cardiac differentiation protocol (Fig. 2), and prior work by others. For example, Bartulos et al. used mouse ISL1⁺ iPSC-CVP for in vivo cardiac repair and found that while some of these cells attained an endothelial fate, the vast majority differentiated into cardiomyocytes⁵⁷. Moreover, using scRNAseq, Galdos et al. found that the vast majority of cells that undergo small-molecule Wnt manipulation based differentiation (as we used here) go through a first heart field intermediate stage and proceed from there to become left ventricular-like cardiomyocytes⁵⁸. Other work suggests that without the additional manipulation (e.g. chemical activation/inhibition or addition of growth factors), iPSC-CVP commit to the cardiac lineage and further differentiate into committed cardiomyocytes by day 7 of the differentiation process^{37, 59, 60}.

Altogether, our findings suggest that it is neither beneficial nor detrimental to use iPSC-CVP for fabricating micro-tissues like μHM. Because iPSC-CVP can be more easily manipulated (e.g. enzymatically singularized), and retain expansion capacity, this may make this cell population ideally suited to use as a replacement for iPSC-CM in future therapeutic applications. Prior work has demonstrated an ability to stabilize the transient iPSC-CVP state⁶¹, which could allow clonal expansion of these cells to further diminish the chance of delivering tumorigenic, non-differentiated iPSC. Prior work also suggests the potential ability to cryopreserve iPSC-CVP may be superior than the ability to cryopreserve iPSC-CM^{62, 63} which may be further advantageous for cell banking prior to bioprinting. Altogether, these studies, together with our findings here, suggest the potential to use iPSC-CVP as a substitute for iPSC-CM in bioprinting applications.

While our findings are promising, there are several limitations to the present work that merit future investigation. First, although our μHM system allowed us to obtain robust electrophysiology data from the tissues, we do not directly report contractility data. While calcium flux can serve as a predictive surrogate for twitch force, future studies could compare contractile function of tissues made from cells at different stages of the cardiac differentiation. Second, although the present study and prior work⁵⁸ strongly suggests that the default differentiation mode for iPSC-CVP is cardiogenesis, it is possible that in the context of bioprinting or cell encapsulation, the biomaterials themselves, or growth factors that tend to adsorb to those materials, could influence the differentiation potential of iPSC-CVP. For example, alginate gels can adsorb heparin-binding growth factors like VEGF⁶⁴. Finally, future studies could assess practical considerations of deploying iPSC-CVP for biofabrication toward the end of cardiac regenerative medicine, including the robustness of these cells after cryopreservation, and maturation of iPSC-CVP derived tissues in the setting of chemical and/or biophysical stimuli aimed at inducing maturation, in vitro and in vivo.