Stirred Bioreactors Generated hiPSC-CM CMTs have > 90% CMs Contents

Purified hiPSC-derived “cardiac microtissues” (CMTs) were previously transplanted in rodents showing enhanced retention [20], whereas their single cell counterparts displayed substantial early loss [21]. To validate the translational value of transplanting CMTs and their improved retention profile, we performed intramyocardial injections into the porcine heart and evaluated the acute cellular retention (Fig. 1A and Sup. Figure 1A). First, CMTs were generated in suspension culture by directed hiPSC cardiomyogenic differentiation in stirred bioreactors as previously described [19, 22]. This resulted in a population of spherical CMT of 188 ± 50.76 µm diameter (Fig. 1B, C left panel and Sup. Figure 1B) having a CM content of >90% based on flow cytometry analysis (Sup. Figure 1C). To directly compare retention of similar cell sources, we dissociated the CMTs into a suspension of single cells and smaller aggregates (termed dissociated aggregates; DAGs) of approximately 70 ± 33.15 µm diameter (Fig. 1B, C right panel). For post-transplantation detection and visualization of the injected cells in the myocardium, DAGs and CMTs were fluorescently labelled with carboxyfluorescein succinimidyl ester (CFSE). CFSE staining penetrated the entire CMTs, including the center core without essential toxicity (Sup. Figure 1B) since cell death did not exceed 10% both at 2-h (5.6 ± 1.0%) and 24-h (8.5 ± 0.9%) after staining resembling levels in unstained CMTs (2-h: 4.1 ± 0.9% and 24-h: 7.7 ± 0.95) (Sup. Figure 1D).

Fig. 1

Comparative assessment of retention between CMTs and DAGs. (A) Schematic representation of the acute cell transplantation study. (B) Diameter quantification of CMTs and DAGs. (C) Representative bright field images of CMTs and DAGs and histological analysis of pig hearts injection site area of hearts that received CMTs and DAGs. Scale bar bright field = 200 µm; Scale bar histology = 400 µm. (D) Whole heart myocardial tissue scans of hearts treated with CMTs or DAGs. Grey arrows point towards the fluorescent CFSE signal in heart slices indicating CFSE signal present in hearts receiving CMTs. (E and F) Quantification of relative fluorescence of the whole heart myocardial tissue slices and area (µm2) of CFSE positive signal intensity in slices from hearts transplanted with CMTs and DAGs. **** p < 0.001, diameter of CMTs versus DAGs. Panel (A) was created with BioRender.com. Panel (B): unpaired Welch’s t test (n = 95-142 biological replicates); data are mean ± SD. Panels (D-E): Mann-Whitney test (n = 3-4 biological replicates); data are mean ± SD. CFSE: carboxyfluorescein succinimidyl ester; CMTs, cardiac microtissues; DAGs, dissociated aggregates

Injection of CMTs Results into Higher Acute Retention Compared to DAGs

After assessing the dimensions and viability of DAGs and CMTs, cellular retention of the CMTs was then assessed in an acute experiment in healthy porcine hearts. For this, CFSE-labelled CMTs or DAGs were injected intramyocardially at five different locations of the left ventricle and the tissue was recovered 10 min later (Fig. 1A, Sup. Figure 1 A and Sup. video 1). Whole-heart myocardial tissue analysis revealed CFSE signal in hearts injected with CMTs (Fig. 1D, left panel). In contrast, cross sections from hearts transplanted with the dissociated suspension displayed almost no fluorescence (Fig. 1D, right panel). Interestingly, CFSE signal of the DAGs in the transplanted hearts was still more prominent than previously observed upon injections of single-cell suspensions of bone marrow-derived stromal cells [7, 8].

Relative fluorescence intensity quantification of the cardiac tissue slices showed a non-significant, increase in fluorescence intensity in the CMT transplanted hearts (Fig. 1E). The moderate increase was probably caused by a combined effect of the relative thickness of heart slices (~ 3 mm), a high degree of auto-fluorescence of the tissue, and limited laser penetration of the Typhoon laser-scanner. Using high-magnification fluorescence microscopy, the injections sites transplanted with CMTs clearly exhibited higher CFSE signals (Fig. 1C, left panel). Quantification of these injection sites showed a trend towards more CFSE positive fluorescent area in the CMT group compared to DAGs (CMTs: 13.5 × 103 ± 4.8 × 103 µm2versus DAGs: 3.1 × 103 ± 0.6 × 103 µm2; p = 0.057) (Fig. 1F). Comparable results were observed when assessing the total CFSE fluorescence intensity (CMT group: 6.6 × 106 ± 2.1 × 106 vs. DAG group: 2.1 × 106 ± 0.5 × 106, p = 0.114) (Sup. Figure 1E). Herewith, we concluded that transplanting CMTs is favorable over smaller or even single cell injections regarding cellular retention.

Adequate Systemic Levels of Immunosuppressants Were Achieved Using a Clinically Relevant Triple-Drug Immunosuppressive Regimen

Xenograft rejection is an additional major concern when human-derived cells are tested in clinically relevant large-animal models. To reach stable, systemic immunosuppressive concentrations, equivalent to levels established for human patients, pigs were subjected to a triple-drug immunosuppressive regimen consisting of tacrolimus, azathioprine and methylprednisolone (Fig. 2A). Starting doses of immunosuppressive drugs, equivalent to commonly used concentrations in clinical practice, were not sufficient to reach target concentrations in plasma and required an increasing dose until sufficient levels were reached (Fig. 2B and Sup. Table 2). Blood levels showed that all treated animals achieved circulating tacrolimus levels of 5–20 µg/L after increasing the dose to 1 mg/kg/day (Fig. 2B and Sup. Table 2). As an inactive prodrug, azathioprine requires conversion into the active metabolite 6-thioguanine nucleotides (6-TGNs) to exert its function [23]. One out of three animals reached the aimed human target range of 100–200 pmol/8 × 108 RBC for the active metabolite 6-TGN, whereas the other two animals approached levels at the borderline when administered 7.0 mg/kg/day (Fig. 2B and Sup. Table 2). Circulating methyl-prednisolone levels could technically not be measured, and therefore, dosing remained constant (1.5 mg/kg/day).

Fig. 2

Overview of immunosuppressive study, drug concentrations, and biochemistry in pigs. (A) Study design and timeline of the immunosuppressive study. (B) Plasma tacrolimus and azathioprine 6-thioguanine nucleotides (6-TGN) concentrations during increased drug dosing and after human donor-cell transplantation. (C) Serum concentrations of routine clinical chemistry parameters for kidney and liver function at start of immunosuppressive treatment (baseline) and termination. Azathioprine 6-methylmercaptopurine nucleotides (6-MMPs) levels were also measured during increasing drug dosing and after xenogeneic cell transplantation. Panels (B–C): No-IS (n = 2 biological replicates); IS (n = 3 biological replicates). Data are individual values ± SD. ALAT, Alanine Aminotransferase; ASAT, Aspartate Aminotransferase; IS, immunosuppression; i.v., intravenous; RBC, red blood cells; t, timepoint; w, week; No-IS, non-immunosuppressed; IS, immunosuppressed; ɣ-GT, Gamma-Glutamyl Transferase

High levels of immunosuppressive drugs are also associated with adverse effects, such as the increased risk of hepatoxicity in humans via the formation of 6-methylmercaptopurine nucleotides (6-MMPs) when treated with azathioprine [23]. In consequence, clinical chemistry parameters were monitored to ensure doses remained below toxic reference values (Fig. 2C and Sup. Table 3). Renal and hepatic enzymes were examined in the immunosuppressed pigs, revealing values within the physiological range at baseline and termination, similar to control animals, indicating that the immunosuppressive cocktail did not negatively affect kidney and liver function (Fig. 2C and Sup. Table 3).

An evaluation of an adequate immunosuppression was further assessed in vitro in isolated PBMCs from transplanted animals (with and without immunosuppression). PBMCs were stimulated with ConA and IL-2 for three days. Upon stimulation, PBMCs formed proliferative leukocyte clusters resulting in higher numbers of total cells and specifically for CD3+, CD4+ and CD8+ subsets after normalizing to unstimulated controls, significantly increasing in CD8+ (p = 0.0109), as analyzed with flow cytometry (Fig. 3A-B and Sup. Fig. S2). PBMCs from immunosuppressed pigs were much less susceptible to stimulation (non-immunosuppressed data not shown), but a significant increase in CD8+ cells was also observed after normalization to unstimulated controls (p = 0.0232) (Fig. 3C). PBMC activation and proliferation of non-immunosuppressed control animals was typically obtained by normalizing to stimulated animals, and subsequently repressed by adding tacrolimus in a dose equivalent to the in vivo concentration (20 µg/L) (Fig. 3D), while a higher dose (100 µg/L) did not further reduce proliferation (Fig. 3D).

Fig. 3

Porcine PBMCs proliferation profiles after in vitro stimulation. (A) Light microscopy representation of non-stimulated (left) or ConA- and IL-2-stimulated porcine clustered-PBMCs (right; yellow arrows) after three days. Scale bar micrograph = 1000 µm; scale bar zoom-in = 250 µm. (B) Ratio of total cell proliferation and T cell subsets of non-stimulated and stimulated porcine PBMCs after three days. (C)Ratio of total cell proliferation and T cell subsets of non-stimulated or stimulated porcine PBMCs from immunosuppressed (IS) animals. (D) Ratio of total cell proliferation and T cell subsets of stimulated porcine PBMCs from non-IS animals incubated with in vivo-relevant (Tacro +, 20 ng/L) or high concentrations of tacrolimus (Tacro ++, 100ng/L). Panels (A–C): unpaired t test in IS (n = 3 biological replicates) and non-IS (n = 2 biological replicates). In (B), * p < 0.05, CD8+ proliferation in stimulated versus non-stimulated PBMCs. In (C), * p < 0.05, CD8+ proliferation of stimulated versus non-stimulated PBMCs from IS pigs. Data are mean ± SD. IS (-), non-immunosuppressed; IS (+), immunosuppressed; stim (-), non-stimulated; stim (+), stimulated

Limited Immune Cell Infiltrate and Enhanced Human Donor-Cell Survival in Triple-Drug Immunosuppressed Animals

Once relevant triple-drug immunosuppression levels were reached, we implanted high-purity CMTs (Sup. Figure 3) using a GFR Matrigel® carrier in the abdominal area of both immunosuppressed and non-immunosuppressed control pigs (Sup. Figure 4 A and Sup. video 2).

Two weeks after transplantation, the Matrigel® plugs showed macroscopic ulcerous lesions in non-immunosuppressed animals, as confirmed by histology (Sup. Figure 4B). Meanwhile, only moderate immune cell infiltrates were observed in the immunosuppressed pigs (Fig. 4A, right panel) and this infiltration was mostly detected surrounding the suture material (Fig. 4A, asterisks) used to demarcate the transplantation site. The effects of our immunosuppression regimen were further verified by visualizing total cellular infiltration in HE stainings (Sup. Figure 4C-G), which was significantly reduced in immunosuppressed pigs, both in Matrigel®-only (p = 0.0381) and CMT-transplanted (p = 0.0095) pigs, compared to non-immunosuppressed animals (Fig. 4B).

Fig. 4

In vivo detection of CMTs and host cell infiltrate upon transplantation. (A) Hematoxylin/Eosin staining (HE) of sectioned Matrigel® plugs followed by full scan microscopy visualized the immune cell infiltrate in the CMT-transplanted areas (dashed circle). Plugs isolated from immunosuppressed (IS) pigs revealed limited infiltrate compared to plugs from non-immunosuppressed (No-IS) animals. Scale bar top section = 2 mm; Scale bar magnification = 20 µm. (B) Assessment of total cell infiltration via nuclear count quantification in HE stainings after plug implantation. (C) Anti-human HLA class I staining (brown) of plugs from IS pigs demonstrated acceptance of the CMTs two weeks after transplantation. Scale bar top section = 2 mm; Scale bar magnification = 20 µm.(D) Specific host cell infiltrates in CMT-loaded Matrigel® plugs isolated from No-IS pigs and IS ones. Scale bar = 20 µm. (E) Quantification of CD3+ cell infiltrates in implanted plugs. Panels (B) and (E): Mann-Whitney test in No-IS versus IS (n = 2 technical replicates per animal). In (B), * p < 0.05, cellular infiltration in Matrigel® only from No-IS versus IS pigs and ** p < 0.01, cellular infiltration in CMTs from No-IS versus IS pigs. In (E), ** p < 0.01, CD3+ infiltrate in Matrigel® plugs from No-IS versus IS pigs. Data are mean ± SD. Asterisk, insoluble suture material hole; black square, zoomed area; dashed line, location of Matrigel® plug; IS (+), immunosuppressed; No-IS (-), non-immunosuppressed

Two weeks post-transplantation, we deemed the immunosuppressive treatment successful after detecting the human xenografts in examined animals, which displayed HLA class I-positive cells in the CMT group (Fig. 4C). Extensive histological characterization revealed an infiltration of CD3+ T cells in CMT-transplanted animals (Sup. Figure 4H-L), with almost an absence of macrophages (CD68+), natural killer (NK, CD56+) and plasma (CD138+) cells (Fig. 4D). Further quantification of the CD3+ cellular infiltrate in both CMT and empty Matrigel® plugs, showed a reduction in CMT transplants, albeit only significant in Matrigel® plugs (p = 0.0095) (Fig. 4E). Jointly, this indicates that our immunosuppression regimen appears to be suitable for successful xenogeneic cell transplantation in pigs.

CMTs Show Long-Term Retention and Survival in the Infarcted Porcine Heart

The central hypothesis of this study was to address if CMTs show an improved long-term retention. To assess the long-term retention and survival of the CMTs in the context of cardiac disease, we induced a myocardial ischemia/reperfusion injury in eight pigs by placing an occluding coronary angioplasty balloon mid-LAD for 90 min, followed by reperfusion (Fig. 5A). Optimized triple immunosuppressive treatment was initiated two weeks post-infarction and animals were transplanted with hiPSC-CM constructs two weeks thereafter (Fig. 5A and Sup. Figure 5A). Prior to CMTs transplantation, all treated pigs reached the target range for tacrolimus (> 5 µg/L), although none showed azathioprine’s active metabolite 6-TGN levels within the therapeutic window (> 100 pmol 6-TGN/8 × 108 RBC) (Sup. Figure 5B and Sup. Table 4). Immunosuppression levels steadily increased during follow up, reaching clinical human target ranges in two out of eight animals at the time of termination (Sup. Table 4). All serum concentrations of creatinine, gamma-GT (γ-GT), aspartate aminotransferase (ASAT) and alanine aminotransferase (ALAT) measured prior to termination were comparable to baseline values (Sup. Figure 5C and Sup. Table 5), indicating immunosuppression did not affect kidney and liver function.

Fig. 5

CMTs detection in the infarcted myocardium four weeks after xenotransplantation. (A) Schematic representation of the long-term CMTs retention study in infarcted and immunosuppressed pigs. (B) Presence and absence of human cells in the infarcted porcine heart four weeks after xenotransplantation of cardiac CMTs or vehicle, respectively. Indicated by human-specific markers Ku80 nuclear antigen (pink) and HLA Class I (pink). Scale bar micrograph = 75 µm. (C) Cross-section of CMTs-recipient heart co-stained with human nuclear marker Ku80 (pink) or HLA Class I (pink) and cardiac marker desmin (blue) visualizes large areas of grafted tissue within the infarcted area. Scale bar heart section = 5 mm; scale bar micrograph = 75 µm. Micrographs are representative images from CMT-transplanted (n = 1 biological replicate), vehicle (n = 1 biological replicate) and human control (n = 1 biological replicate) hearts. Panel (A) was created with BioRender.com. CMTs, cardiac microtissues; I, infarcted myocardium; R, remote myocardium

At the day of the transplantation, pigs were randomized, and we delivered intramyocardially a dose of CMT equivalent to about ~ 5 × 107 cell in 1 mL PBS/vehicle via intramyocardial injection in the border zone region. Specifically, three batches of CMTs were produced consisting of 144.6 ± 81.5 µm diameter (Sup. Figure 6A and Sup. Figure 6B) with a CM purity of 80–95%, as determined by low cytometry for cardiac troponin T (cTnT), myosin 4 (MF20) and sarcomeric α-actinin (SA) (Sup. Figure 6C).

Four weeks post-transplantation, all animals showed complete transmural infarcts as indicated by picrosirius red staining (representative image, Sup. Figure 6D). Interestingly, in three out of four histologically examined animals, human HLA class I-positive cells were located within the infarcted area (Fig. 5B, C and Sup. Figure 7), indicating CMT retention and cell survival up to four weeks in the recipient hearts (Fig. 5B and C). In fact, the contoured islands found in the largest graft were also positively stained by the human Ku80 nuclear antigen. Co-staining of Ku80 and HLA Class I with desmin confirmed that the identified cell clusters belonged to human muscle cells (Fig. 5C) that remained CMs after being transplanted (Sup. Figure 8). More importantly, large grafts were readily identified in one animal with islands measuring more than half a centimeter in their greatest dimension and, in total, occupying about 10% of the infarcted area. This finding suggested that CMTs reside within the scar tissue and cluster together. When examining other hearts, we could only trace back HLA class I-positive cells in two more pigs (out of four) and the dimensions of these grafts were considerably smaller (pig #1: ~ 5 mm versus pig #3: ~ 800 µm versus pig #6: ~ 300 µm) (Sup. Figure 7).

To confirm the presence of CMTs and better understand cell biodistribution, one of the porcine hearts was used to generate a 3D replica of the grafts (Fig. 6 and Sup. video 3). The islands of cells detected were on the order of ~ 1 cm (Fig. 6A, right panel), similar in size to what human-specific markers revealed via histology in one animal (Fig. 5C). After discriminating the fluorescent signal, cardiac grafts accounted for 5.26% of the total scar area (Fig. 6D), representing approximately 200 µL of the segmented structures in the heart (Fig. 6D), and potentially reflecting the reduced number of total cells initially transplanted (5 × 107 cells/animal). Looking closely at the distribution of the CMTs in this 3D reconstruction (Fig. 6B), CMTs appeared to be engrafted at the injection sites within the scar tissue, and in clusters bigger than their original injection dimensions, similarly to what histology suggested. These results demonstrate that CMTs were retained long-term, showing the tendency to cluster together, and were readily visible within the infarct scar four weeks post-transplantation.

Fig. 6

3D visualization of fluorescently labeled CMTs in an infarcted pig. (A) Representative macroscopic images of transplanted areas in a heart, from base to apex (top to bottom). Reflection images (left panel) display autofluorescence from the remodeled tissue, together with the fluorescently labelled cells. Upon excitation, xenotransplanted CMTs (right panel) stand out from the remodeled and remote myocardium, indicating localization of xenografts (dashed line). Scale bar = 2 cm. (B) Stack of 3350 cross-sectional images of a transplanted heart enabling 3D reconstruction of the transplanted heart. (C) Total volume quantification of left ventricle, infarcted tissue, and labelled-CMTs and (D) relative contribution of graft in the scar area of the 3D reconstructed myocardial mesh. Representative images and data are single values from a CMT-transplanted heart (n = 1 biological replicate). CMTs, cardiac microtissues

Finally, despite the optimized immunosuppression regimen applied, we wanted to confirm the potential immunogenic reaction of the grafts four weeks post-transplantation. HE staining confirmed a general cell infiltration in the myocardium, which was prominently found surrounding the grafts, in contrast to animals treated with vehicle (Sup. Figure 9A). As a matter of fact, this cellular infiltration positively stained for a pan-leukocyte marker CD45 and CD107a which is expressed on activated macrophages, NK cells and CD8+ cytotoxic T cells (Sup. Figure 9B). Coincidentally, three of the CMT recipients that presented long-term engraftment and a positive immune infiltration surrounding the CMTs did not reach azathioprine target levels at termination (Sup. Table 4). Nonetheless, human reference values were used for the immunosuppression treatment and thus it should not necessarily question the efficacy of the immunosuppression regimen.

To evaluate the reorganization, possible integration, and maturation stage of the CMTs within the native porcine myocardium four weeks after transplantation, intercalated disk protein N-cadherin was used. A robust expression of N-cadherin was present in human hearts, being much lower and disorganized in the in vivo long-term transplanted porcine heart. Moreover, the lack of circumferential disposition of the intercellular junctions suggested cardiomyocyte immaturity (Sup. Figure 9C) [24].