Combined Treatment of Human iPSC-derived Cardiomyocytes and Endothelial Cells Regenerate the Infarcted Heart in Mice and Non-human Primates

Abstract

BACKGROUND:

Remuscularization of the mammalian heart can be achieved after cell transplantation of induced pluripotent stem cell (iPSC)-derived cardiomyocytes (CMs). However, several hurdles remain before implementation into clinical practice. Poor survival of the implanted cells is related to insufficient vascularization, and the potential for fatal arrhythmogenesis is associated with the fetal cell-like nature of immature CMs.

METHODS:

We generated three lines of hiPSC-derived endothelial cells (ECs) and hiPSC-CMs from three independent donors and tested hiPSC-CM sarcomeric length, gap junction protein and calcium handling ability in coculture with ECs. Next, we examined the therapeutic effect of co-transplantation of hiPSC-ECs and hiPSC-CMs in NOD-SCID mice undergoing myocardial infarction (MI, n ≥14). Cardiac function was assessed by echocardiography whereas arrhythmic events were recorded using 3-lead EKG. We further used healthy non-human primates (n=4) with cell injection to study the cell engraftment, maturation and integration of transplanted hiPSC-CMs, alone or along with hiPSC-ECs, by histological analysis. Lastly, we test the cell therapy in ischemic-reperfusion (IR) injury in non-human primates (n=4, 3, and 4 for EC + CM, CM, and control, respectively). Cardiac function was evaluated by echocardiography and cardiac MRI while arrhythmic events were monitored by telemetric EKG recorders. Cell engraftment, angiogenesis, and host-graft integration of human grafts were also investigated.

RESULTS:

We demonstrate that human iPSC-ECs promote the maturity and function of human iPSC-derived CMs in vitro and in vivo. When co-cultured with ECs, CMs showed more mature phenotypes in cellular structure and function. In the mouse model, co-transplantation augmented the EC-accompanied vascularization in the grafts, promoted the maturity of CMs at the infarct area and improved cardiac function after MI. Furthermore, in non-human primates, transplantation of ECs and CMs significantly enhanced graft size and vasculature, improved cardiac function after IR.

Conclusions:

These results demonstrate the synergistic effect of combining iPSC-derived ECs and CMs for therapy in the post-MI heart, enabling a promising strategy toward clinical translation.

INTRODUCTION

The inability of the adult human heart to regenerate or to compensate for lost cardiomyocytes after injury contributes to the progression of heart failure, a leading cause of death worldwide. To address this challenge, cell therapy opens the door to replenish the loss of cardiomyocytes and may restore cardiac function after injury. Improvement in cardiac function in preclinical and clinical studies using cell sources derived from adult tissues has been attributed to paracrine and immunomodulatory effects rather than remuscularization. In contrast, recent studies using human pluripotent stem cell (PSC)-derived cardiomyocytes have shown promising preclinical results in remuscularizing damaged heart tissue^1-5. However, major barriers to clinical development remain, including the immunogenicity of the grafts, low engraftment efficacy, and induced ventricular arrhythmias.

A major challenge for embryonic stem cell (ESC)-derivatives is susceptibility to host immune rejection⁶. While most current studies utilize immunosuppressive drugs or immunodeficient animals to reduce rejection, cardiomyocytes (CMs) generated from autologous induced pluripotent stem cells (iPSCs) may provide a safer and more efficient alternative to ESC-CM⁷. Successful regeneration of non-human primate hearts with allogenic transplantation with iPSC-CMs has been reported⁸. This demonstrates a potential advantage in selecting iPSC-CMs as the superior treatment in overcoming hurdles in cardiac cell replacement therapy.

The presence of ventricular tachycardia following PSC-CM transplantation has been attributed to intrinsic automaticity of the immature CMs in the grafts. Despite the potential benefits of stem cell-derived CMs, another major challenge is that the transplanted cells remain immature in morphology and function⁹. These shortcomings have encouraged extensive research in methods to promote CM maturation. Among the numerous methods documented, a recent study showed that ESC-CM co-transplantation with ESC-epicardial cells promoted vascular density within grafts¹⁰, suggesting that vascularization is crucial to functional myocardial development. The mere engraftment of CMs does not guarantee the generation of functioning tissue, as lack of oxygen and nutrient supply into the target area reduces survivability¹¹. Thus, a successful stem cell-based heart regeneration treatment would require the maturation of CMs and revascularization simultaneously.

Endothelial cells are the most abundant cell type in the heart¹², which communicate and interplay with CMs through both direct cell-cell contact and paracrine effects^13,14. Previous studies have shown that ECs promote CM survival and organization in culture^15,16. In the present study, we hypothesized that ECs promote CM maturation in vitro and when co-transplanted in vivo, promote CM maturation and graft vascularization, which ultimately lead to restoration of cardiac function. Here, we report that co-transplantation with isogenic human iPSC-derived ECs not only promotes CM maturation but also enhances blood vessel formation, as well as improves cardiac regeneration after myocardial infarction (MI) in mice and non-human primates. Furthermore, we assessed the bidirectional effect between ECs and CMs to determine the mechanism by which the benefits of this treatment occur.

METHODS

The data, analytic methods, and study materials are available to other researchers for purposes of reproducing the results or replicating the procedure.

Non-human primate surgery

Fifteen male rhesus macaques (Macaca mulatta) (Table S1) were obtained from the Wisconsin National Primate Research Center and animal protocol has been approved by the experimental animal committee of UW-Madison (IACUC number G006084-A07). For feasible study, four animals were subjected to direct cell transplantation without myocardial injury. Immunosuppression was achieved with three drugs. Tacrolimus was given i.m. to maintain serum trough levels of 15 ng/mL every day from day 7 until non-human primates were euthanized. Methylprednisolone was given i.v. 30 mg/kg on the day before cell transplantation followed by maintenance doses of i.m. 6 mg/kg for 2 d, and then 3 mg/kg thereafter until macaques were euthanized. Lastly, Abatacept 12.5 mg/kg was administered subcutaneously on the day before cell transplantation and every 2 weeks thereafter. For all major surgeries and procedures, non-human primates were anaesthetized with intra-muscular injection of ketamine and midazolam, intubated, and ventilated using isoflurane to maintain anesthesia. At day 0, non-human primates underwent left-sided thoracotomy and the anterior surface of the heart was exposed to visualize the mid-left anterior descending (LAD) coronary artery. Cryopreserved ECs and CMs were quickly thawed in 37°C water bath, and cells were resuspended in DMEM-HG supplemented with 10% FBS for centrifugation. Afterward, cells were resuspended in a smaller volume of DMEM-HG with 10% FBS medium for trypan-blue viability test. Viable cells were more than 95% and 80% for iPSC-ECs and iPSC-CMs, respectively. A desired number of cells were mixed into the same tube and subsequent wash step using DMEM-HG supplemented with 10% FBS was performed. The cell pellet was resuspended in 7-factor pro-survival cocktail modified based on Liu, Y.W. et al⁴. In brief, the pro-survival cocktail consists of 30 % Matrigel in FBS (vol/vol) supplemented with: 10 μM ZVAD-FMK/Caspase Inhibitor (Calbiochem/ EMD-Millipore); 50 nM TAT-BH4 / BCL-XL (Calbiochem/EMD-Millipore); 200 nM Cyclosporine A (Sandimmune/Novartis); 50 μM Pinacidil (Sigma); 100 ng/ml IGF-1 (Peprotech); 100 ng/ml VEGF-1 (Peprotech); and 100 ng/ml PDGF-BB (Peprotech). Both EC + CM (2.5 × 10⁸ cells + 2.5 × 10⁸ cells) suspended in PSC and CM (2.5 × 10⁸ cells) suspended in PSC were delivered intra-myocardially into the myocardium near to LAD coronary artery via 4 injections of 375 μl each using a 29-gauge injection needle. All macaques received both EC+CM and CM-only treatment by injecting the cells at different sites of the heart at a distance of more than 1 cm away. Two macaques received EC+CM treatment at the base of left ventricular myocardium and received CM treatment at the apex of left ventricular myocardium and vice versa for the other two macaques.

Six animals were involved in functional study under ischemic reperfusion injury. Five days before 1^st intervention, animals were treated with 24 mg/kg anti-arrhythmic drug amiodarone, 9.7 mg/kg for four days post-surgery, and then 4.85 mg/kg for another two days. To induce myocardial ischemia, percutaneous coronary intervention or thoracotomy was performed at day −28, mid-LAD coronary artery was occluded for 90 minutes followed by reperfusion and ischemia was confirmed by ST-segment elevation on EKG. In subjects R10098 and R14104, the percutaneous coronary intervention was performed following procedures previously described⁴. In other subjects having small vessels or unclear LAD, thoracotomy was performed. Loop recorder (Medtronic Reveal LINQ) was subcutaneously implanted after completion of the thoracotomy. Immunosuppression was given to animals receiving cell transplantation (EC+CM and CM-only) by three drugs as mentioned above. At day 0, non-human primates underwent thoracotomy and infarct was visualized near LAD coronary artery. We followed the same cell preparation protocol as described above for macaques without injury. EC + CM (5 × 10⁸ cells + 5 × 10⁸ cells) suspended in PSC was transplanted via 4 injections of 375 μl each using a 29-gauge injection needle. Thoracotomy was not applied on control animals at day 0.

Animals were regularly monitored by laboratory and Primate Center staff. Euthanasia was induced by i.v. injection of pentobarbital after sedation with ketamine or under anesthesia with ketamine, isoflurane, and a constant rate infusion of fentanyl. The EC + CM -treatment group was allocated in an unblinded and non-randomized manner. Non-human primate hearts fixed in 4% paraformaldehyde were dissected to remove the atria and right ventricle before cross-sections were obtained by sectioning parallel to the short axis at 4 mm thickness. The tissue slices were then processed and embedded in paraffin, and five micrometer sections were cut for staining.

Statistical analysis

All values are expressed as mean ± SEM. Normality was tested using the Shapiro-Wilk test. Data that did not pass the test were reanalyzed using nonparametric tests (unpaired t-test with Mann-Whitney test, paired t-test with Wilcoxon matched-pairs signed rank test, one-way ANOVA with Kruskal-Wallis test, and two-way ANOVA with Friedman test). For those that passed the Shapiro-Wilk test were further tested if they had equal variances using F test and Brown-Forsythe test for data analyzed by unpaired t-test and one-way ANOVA, respectively. Data that had equal variances were analyzed by ordinary t-test or ANOVA, and those that did not were reanalyzed by t-test or ANOVA with Welch’s correction. For data that met the assumptions of normal distribution and equal variances, multiple comparisons were done with Tukey adjustment. For data that did not meet normal distribution, Dunn’s multiple comparisons test was used. As for data that did not have equal variances, Dunnett’s T3 multiple comparisons test was used, with individual variances computed for each comparison. All statistical analyses were performed using GraphPad Prism software, with the threshold for significance level set at P < 0.05.

RESULTS

Co-culture with hiPSC-ECs promotes hiPSC-CM maturity

To develop isogenic CMs and ECs from hiPSCs, we induced CM differentiation according to Sharma, A. et al.¹⁶, with an average purity of 96.2 ± 0.7% cardiac troponin I (cTnI) positive CMs (Figure S1A and S1B), and modified an established EC differentiation protocol (Figure S1C)¹⁷. We characterized the purity of ECs by the expression of CD31 and CD144, which showed 91.4 ± 0.7% double positive cells measured by flow cytometry (Figure S1D). Light microscopy imaging and immunofluorescence labeling revealed cobblestone patterns in morphology (Figure S1E) and presence of low-density lipoprotein uptake in the cytoplasm (Figure S1F), which are defining characteristics of ECs.

With a well-characterized EC culture, we sought to investigate the effect of ECs on CM sarcomere length in co-culture and transwell environments. The mean sarcomere length of co-cultured CMs was longer compared to CM culture alone (Figure 1A, 1B, and Figure S1G-S1I). In addition, co-culture of ECs and CMs increased the expression of gap junction protein connexin 43 polarized at the cell border (Figure S1J). ECs and CMs were cultured in a transwell system to determine if direct cell-to-cell contact is essential for increased sarcomere length in CMs. There was no difference in sarcomeric length between CM + EC in transwell cultures and CM alone (Figure S1K-S1N), and no significant difference in connexin 43 expression and distribution in CM (Figure S1O), indicating that ECs promote CM maturation through cell-to-cell contact rather than paracrine effects, which is consistent with our previous findings¹⁸.

Figure 1. Co-culture with ECs promotes the maturity of CMs.

A, Sarcomeric structure in EC+CM co-culture and CM alone groups. Scale bars, 20 μm. B, Quantification of sarcomeric length. C, Representative regular calcium transients of CMs co-cultured with ECs or alone. D, Amplitude of calcium transients. E, Slope of calcium upstroke. F, Slope of calcium downstroke. G, Tau decay of calcium transients. H, Maximum diastolic potential of EC+CM co-culture and CM alone groups. I, Ultrastructural images of CM sarcomeric structure (upper panel). Arrowhead, Z disk. Arrow, I band. Ultrastructural images of mitochondria in CMs (lower panel). Scale bars, 1 μm. Statistical significance was determined by two-tailed unpaired t-test in B, D, F and H and determined by the Mann-Whitney test in E and G. n=28 independent samples for EC+CM and n=28 independent samples for CM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Next, calcium handling tests were performed on CMs to further validate the enhanced maturity due to ECs. Live-cell imaging showed that both co-cultured CMs and CM alone have regular calcium transients (Figure 1C), while co-cultured CMs had a greater amplitude, a sharper slope of calcium upstroke and more rapid calcium downstroke with faster tau decay compared to CM alone (Figure 1D-1G), indicating a faster calcium handling response could be achieved in the presence of ECs. We also tested the calcium handling ability of mature CMs with the other two isogenic iPSC-derived CM and EC cell lines and showed similar results (Figure S1P-S1T).

To further characterize the maturity of CMs after co-culture, the electrophysiological properties and cellular features were examined. Whole cell patch clamp recordings indicated a lower maximum diastolic potential in co-cultured CMs compared to CM culture alone (Figure 1H). Using transmission electron microscope, ultrastructure imaging further revealed that co-cultured CMs consisted of myofibrils with properly aligned z-disks and I-bands. In addition, the mitochondria in co-cultured CMs have a long and slender morphology with well-formed cristae within the inner membrane of mitochondria compared to CMs only (Figure 1I). Collectively, these results demonstrate that ECs promote CM maturation in co-culture, both structurally and functionally, via direct cell-to-cell contact.

Co-transplantation of hiPSC-CMs and hiPSC-ECs improves cardiac function after infarction in mice

We next examined the effects of co-transplantation of CMs and ECs on cardiac function in animal models. MI was induced in NOD-SCID mice, followed by cell injection into the infarct and the border zone. A timeline of experimental design including the cell type, cell number, and vehicle control is outlined in Figure 2A. There was no significant difference in animal mortality among the groups with or without cell injection, suggesting co-transplantation does not cause extra burden to the hearts (Figure 2B). Next, echocardiography revealed significantly impaired cardiac function two days after MI surgery in all treatment groups (Figure 2C). Twenty-one days after MI, cell transplantation groups displayed a preservation or improvement of cardiac function (Figure 2C and 2D). Notably, co-transplantation of ECs and CMs most significantly improved fractional shortening (FS) and ejection fraction (EF) of the hearts compared to vehicle control, EC alone, CM alone, or double CM alone in comparable total cell number (Figure 2D, S2A, and S2B). Data consistency was confirmed and validated through Bland-Altman plots, which display the mean difference as well as the difference in ejection fraction between the independent measurements of two blinded investigators (Figure S2C). When we looked at the changes of FS and EF between day 2 and day 21, EC + CM exhibited significantly improved cardiac function (Figure 2E and Figure S2D). Left ventricular end systolic dimension (LVESD) and end systolic volume (ESV) of EC + CM were significantly decreased compared to vehicle control group (Figure S2E and S2F). Similarly, left ventricular end diastolic dimension (LVEDD) and end diastolic volume (EDV) tended to be smaller in co-transplantation group compared to vehicle control (Figure S2G and S2H).

Figure 2. Co-transplantation of ECs and CMs restores cardiac function and exhibits lower arrhythmic events.

A, Schematic of study design and timeline for cardiac function evaluation. Echocardiography was performed 7 days before surgery, 2 days after surgery and 21 days follow-up. Three-lead electrocardiogram was carried out on day 21, before euthanization. EC+CM, double CM (CM x 2), CM, EC, and vehicle, n=19, 17, 19, 15, and 14 animals, respectively. MI, myocardial infarction. B, Kaplan–Meier analysis over 21 days following myocardial infarction and cell treatment. (Mantel–Cox log-rank test; n=19, 17, 19, 15, and 14 animals for EC + CM, double CM, CM, EC, and vehicle, respectively.) C, Fractional shortening was measured by echocardiography at day −7, day 2 and day 21. D, Overlay of fractional shortening. E, Changes of fractional shortening between day 2 and day 21. n=16, 12, 12, 11, and 11 animals for EC + CM, double CM, CM, EC, and vehicle, respectively in C, D, and E. F, Representative ECG recordings showing premature ventricular contraction (red arrows) and ventricular couplet (blue arrows) in mice. G, Pie chart showing proportion of animals with arrhythmic events. (n=6, 6, 8, 6, and 7 animals for EC+CM, double CM, CM, EC, and vehicle, respectively.) Statistical significance was determined by two-way mixed-effects ANOVA with the Geisser-Greenhouse correction and Tukey’s multiple comparisons test in C and D; and was determined by one-way ANOVA with Tukey’s multiple comparisons test in E. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns for not significant.

To detect arrhythmia after transplantation on day 21, 3-lead electrocardiograms were recorded in mice. Representative images of normal sinus rhythm, premature ventricular complexes (PVCs), and ventricular couplets are presented in Figure S2I. PVCs were observed in all groups (Figure 2F). However, the group with transplantation of double amount CMs showed the highest rate of PVCs and ventricular couplets relative to the other groups (Figure 2G). These results suggest that EC + CM co-transplantation may benefit the outcome of arrhythmia compared to delivering a higher number of CMs.

HiPSC-ECs promote vascularization and hiPSC-CM graft survival in mouse MI model

To understand the physiological basis of improvements in cardiac function, histological analysis of cardiovascular tissue after treatment was performed. First, heart sections were subjected to picrosirius red/fast green staining to mark the infarct areas (Figure 3A). It shows that CM alone but not EC alone treatment decreases the infarct size 21 days after MI; however, co-treatment of EC and CM more significantly limited infarct extension and replacement fibrosis (Figure 3B). As the vasculature forming entities, hiPSC-ECs have been shown to promote angiogenesis in multiple wound healing animal models^19-22. To study how ECs benefit CM engraftment, we performed immunohistochemistry using a polyclonal CD31 antibody that can label both human and mouse CD31 to identify the vasculature. We found that both EC + CM and EC alone showed a significant increase in vascular density within the infarct area compared to control and CM alone groups (Figure 3C and 3D). Increment of vascular density was limited in the infarct area and the border zone, but not in the non-infarct remote area (Figure S3A-S3D). By co-staining human mitochondria and actinin, we were able to locate the engrafted human cells (Figure 3E). We found that EC + CM had a higher engraftment ratio (Figure 3F). Within these engrafted animal hearts, EC + CM showed a significant increase of human cell engraftment size compared to the double CM or CM alone groups (Figure 3G), suggesting that ECs enhanced the survival rate of CMs through an increase in angiogenesis. We then measured the vascular density and found that the vascular density was significantly higher in EC + CM co-transplanted hearts (Figure 3H and 3I). By co-staining of human CD31 and Ku80 to reveal engrafted ECs, we confirmed the survival of ECs in the grafts (Figure S3E). Using differential interference contrast (DIC) microscopy, we observed bi-concaved erythrocytes located inside CD31-positive signal (Figure S3F), indicating these EC-derived vessels were perfused via the host coronary circulation system.

Figure 3. Co-transplantation of ECs exerts trophic effects on the grafts in MI mouse model.

A, Representative images of infarct area stained by picrosirius red and fast green. Scale bar, 1 mm. B, Quantification of infarct area. n=16, 12, 12, 11, and 11 animals for EC+CM, double CM, CM, EC, and vehicle, respectively. C, Representative images of vasculature in the infarct area. Scale bars, 40 μm. D, Quantification of vascular density in the infarct area. n=16, 12, 12, 11, and 11 animals for EC+CM, double CM, CM, EC, and vehicle, respectively. E, Representative whole slide scan images of grafts. Scale bars, 1 mm. F, Engrafted animal ratio in each group. (n=16, 12, and 12 animals for EC+CM, double CM, and CM, respectively.) G, Quantification of graft size. n=15, 9, and 9 animals for EC + CM, double CM, and CM, respectively. H, Representative images of graft vasculature. Scale bars, 40 μm. I, Quantification of vascular density per animal. n=15, 9, and 9 animals for EC + CM, double CM, and CM, respectively. J, Representative images of sarcomere structure of grafts. Scale bars, 5 μm. K, Quantification of sarcomeric length per animal. n=15, 9, and 9 animals for EC + CM, double CM, and CM, respectively. L, Representative images of proliferated cells in grafts stained by Ki67 antibody. Scale bars, 40 μm. M, Quantification of proliferation index per animal. n=15, 9, and 9 animals for EC + CM, double CM, and CM, respectively. Statistical significance was determined by Kruskal-Wallis test with Dunn’s multiple comparison test in B, D and G and determined by one-way ANOVA with Tukey’s multiple comparisons test in I, K, and M. *P < 0.05, **P < 0.01, ***P < 0.001, ns for not significant.

Co-transplanted hiPSC-CMs and hiPSC-ECs mutually mature each other

We hypothesized that a bidirectional relationship between ECs and CMs may benefit each other by forming functional human myocardial tissues in mice. First, we showed that ECs improved CM maturity as demonstrated by an increased sarcomeric length and promoted sarcomeric alignment in contrast to those in double CM and CM alone groups (Figure S3J and S3K). To test if ECs enhance CM proliferation, Ki67⁺ proliferating CMs were labeled and examined under confocal microscopy. Surprisingly, we found that with EC co-transplantation, CMs showed a significantly lower proliferative index compared to double CM or CM alone groups (Figure 3L and 3M), suggesting that these CMs are more terminally differentiated cells, like adult CMs. This also indicated that larger engraftment was not due to CM proliferation. To test if ECs enhance engraftment through preventing cell apoptosis, we labeled apoptotic cells using TUNEL one day after transplantation. Fewer TUNEL-positive cells in EC+CM-treated mice compared to CM-only mice, suggesting that ECs could attenuate CM death when co-transplanted (Figure S4A and S4B). In addition, we found that increased levels of pan cadherin were expressed and polarized to the intercalated disks in CMs co-transplanted with ECs (Figure S3G). In contrast, a lower expression level and circumferential distribution pattern of pan cadherin were observed in the double CM or CM alone group. Collectively, these data demonstrated that ECs improved CM maturation in the ischemic myocardium.

Previous studies have reported that CMs can promote angiogenesis with new vascular cells originating from pre-existing ECs^23,24, we thus wondered whether CMs could also enhance EC angiogenesis when they were injected together. Using a human specific CD31 antibody, we found that the density of human iPSC-EC-derived vessels was indeed higher in EC + CM mice than EC alone mice (Figure S4C and S4D). We then examined whether co-transplanted CMs could promote vascular maturity by recruiting pericytes or smooth muscle cells to coat the vessels. We found that smooth muscle actin (SMA)-positive cells were clearly observed in the infarct area in both the EC + CM and EC alone groups with higher density in the EC + CM co-transplantation group (Figure S4E and S4F). Furthermore, within the grafts, SMA-positive vessels were observed in EC+CM group but not in double CM or CM alone groups (Figure S3H). Under DIC microscopy, we further observed that these SMA-positive vessels were perfused by the host as erythrocytes can be detected inside the vessels (Figure S3I). These results suggest that CMs facilitate not only angiogenesis, consisting of both transplanted ECs and endogenous ECs, but also maturation of the neo-vessels.²⁵

In summary, Figure S5 provides our working model of cardiac function improvement following EC and CM co-transplantation in mice. Upon co-transplantation with CMs, ECs significantly promoted vascularization in the infarct area and within the grafts. These promoted vascular network graft survival, resulting in a greater graft size. Positive correlation was shown between vascular density and graft size (r = 0.3599, P = 0.1876 for EC + CM; r = 0.4848, P = 0.1859 for double CM; and r = 0.6357, P = 0.0658 for CM). Greater graft size ultimately led to improvement of cardiac function as shown by positive correlation between changes of FS and graft size (r = 0.358, P = 0.1902 for EC + CM; r = 0.6934, P = 0.0383 for double CM; and r = 0.1441, P = 0.6913 for CM). Overall, this figure demonstrates the important role of EC co-transplantation on vascular networks and graft survival, which helps to restore cardiac function in an MI NOD-SCID mouse model.

Co-transplantation of hiPSC-CMs and hiPSC-ECs improves cardiac function after ischemic reperfusion injury in non-human primates through augmenting growth of vascularized human myocardium

To gauge the translational potential of our study, we performed a feasibility test to examine the engraftment of CMs and ECs in 4 healthy non-human primates. We developed an immunosuppressive therapy to prevent graft rejection caused by injection of human cells in the primates, which started 7 days before cell injection and continued to 28 days after cell injection. On the surgery day, animals received intramyocardial injection of EC + CM or CM alone in the left ventricular free wall with two EC + CM injections and two CM alone injections for each heart (Figure 4A). Twenty-eight days after cell injection, animals were euthanized and the hearts were prepared for histological analysis. Using whole slide scanning, we found that there was no significant difference in the graft sizes between the EC + CM and CM alone subjects (Figure 4B, 4C, and Figure S6A). We hypothesized that cell survivability was enhanced in both treatments under the healthy myocardium with sufficient blood supply. To test this, we co-stained CD31 to quantify vascular density within the grafts. Interestingly, we found a significant increase in vascular density within the EC + CM grafts compared to CM alone grafts (Figure 4D, 4E, and Figure S6B), suggesting that ECs promote angiogenesis in the grafts. With these data, we further hypothesized that ECs may improve the maturity of CMs in vivo. We examined the maturity of engrafted CMs in these primates by immunochemistry to measure the adherens junction pan cadherin and sarcomeric length. We found that most pan cadherin was located at intercalated disks in EC + CM co-transplanted grafts (Figure 4F and Figure S6C). Moreover, higher frequency of host-graft structure integration was found in EC + CM co-transplanted grafts (Figure 4G and Figure S6D). Also, longer sarcomeric length and larger diameter were observed in EC + CM co-transplanted grafts in all 4 animals compared to CM alone (Figure 4H, 4I, and Figure S6E, and S7A). Additionally, we found that a significantly lower proliferation index, less MLC2a expression, and more cTnI expression in EC + CM grafts, compared to CM alone grafts (Figure 4J, 4K, and Figure S6F, S7B, and S7C), indicating these co-transplanted CMs were more mature. This might indicate that proliferation in CM alone was a driver to increased graft size in healthy myocardium. Taken together, all data suggested that ECs promoted CM maturity in a non-human primate model, as they did in mice.

Figure 4. Co-transplantation of ECs promotes CM maturity in healthy non-human primates.

A, Schematic of study design. B, Representative whole heart slide images of grafts. C, Quantification of graft size. D, Representative images of graft vasculature. Scale bars, 40 μm. E, Quantification of vascular density per animal. F, Representative images of adherens junction expression within grafts. Scale bars, 40 μm. G, Representative images of CM integration with host myocardium. Yellow arrows, pan-cadherin at intercalated disk of CMs. Scale bars, 20 μm. H, Representative images of sarcomeric length in grafts. Scale bars, 10 μm. I, Quantification of sarcomeric length per animal. J, Representative images of proliferated cells in grafts. Scale bars, 40 μm. K, Quantification of proliferation rate per animal. Statistical significance was determined by two-tailed paired t-test in C, E, I, and K. n=4 samples for EC+CM engraftment and n=4 samples for CM engraftment. *P < 0.05, **P < 0.01, ns for not significant.

To evaluate the potential of co-transplanting ECs and CMs as a regenerative therapy to treat ischemic heart injury, we induced ischemic reperfusion (I/R) injury in non-human primates for 28 days and treated them with EC + CM, CM alone, or vehicle (Figure 5A). Cardiac function was evaluated by echocardiography at day −28, day 0 and day 28. All macaques had comparable decrease in cardiac function four weeks after I/R injury at day 0 (Figure S8A-S8F). CM alone treatment preserved cardiac function four weeks after cell transplantation, evidenced by 1.917% ± 4.117% (P = 0.6657, ns) change in EF between day 0 and day 28. Notably, EC + CM treatment significantly elevated EF by 8.893% ± 1.326% (P = 0.0005, ***) (Figure S8A, statistics not shown on the overlay figure). EC + CM treated group had a significantly greater increase in EF and FS compared to vehicle group, whereas CM alone was not significantly different from control (Figure 5B and 5E). Analysis of ventricular size shows that EC + CM treatment significantly decreased in change of ESV and tended to decrease in change of EDV, compared to control (Figure 5C and 5D), similar to left ventricular dimension (Figure 5F and 5G). To evaluate the therapeutic effect of combining ECs to CM transplantation, we compared EC + CM treatment to CM alone and found that EC + CM tend to greater improve cardiac function and prevent LV dilation at end-systole (Figure 5B-5G). Cardiac magnetic resonance imaging also supported that EC + CM treatment had better improvement in cardiac function and decrease in ESV, compared to control animals (Figure S9).

Figure 5. Co-transplantation of ECs and CMs restores cardiac function after ischemic-reperfusion injury in non-human primates.

A, Schematic of study design. Cardiac magnetic resonance imaging was performed before the day of cell injection and at 28 days follow-up. Echocardiography was performed before I/R surgery, before cell injection and before termination. Electrocardiogram was obtained using a loop recorder throughout the experiment. I/R, ischemic-reperfusion. B, Difference in EF between day 0 and day 28 for each macaque. C, Difference in ESV between day 0 and day 28 for each macaque. D, Difference in EDV between day 0 and day 28 for each macaque. E, Difference in FS between day 0 and day 28 for each macaque. F, Difference in LVESD between day 0 and day 28 for each macaque. G, Difference in LVEDD between day 0 and day 28 for each macaque. n=4 macaques for EC+CM engraftment, n=3 macaques for CM engraftment and n=4 macaques for control. H, Sustained ventricular tachycardia (VT) episodes monitored by loop recorder. I, Sustained VT duration recorded by loop recorder as min per day (m/d). J, Non-sustained VT episodes monitored by loop recorder. K, Non-sustained VT duration recorded by loop recorder as m/d. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test in B through G.*P < 0.05, **P < 0.01, ns for not significant.

To monitor cardiac health, electrocardiogram was obtained using a loop recorder throughout the experiment. Two animals were excluded from this electrocardiogram study, with one CM alone animal suffered from continuous diarrhea and one control animal missed the loop recorder device. Both EC + CM and CM alone treated groups showed incidence of arrhythmia after cell transplantation as evidenced by onset of ventricular tachycardia (VT) at day 0 (Figure 5H-5K). We observed that arrhythmias of most EC+CM animals peaked in the early seven days and subsided at later timepoint, whereas arrhythmias continuously emerged in CM-only animals. Quantification shows that sustained VT episodes were significantly higher in cell-transplanted groups (EC+CM and CM), compared to control (Figure S8G). We also observed a trend of fewer sustained VT episodes in EC+CM animals compared to CM-only animals but not significantly different (P = 0.48, ns), possibly due to low sample size. There is no difference in sustained VT duration, and non-sustained VT episodes and duration among groups (Figure S8H-S8J). We treated animals with amiodarone before and after 1^st intervention to prevent arrhythmia induced by I/R surgery, however, the effect of amiodarone on arrhythmia at later timepoint could be excluded due to the undetectable plasma level before cell injection in all animals except for R17023 which had 21.33 ng/mL Amiodarone, a value close to limit of detection (18 ng/mL), detected before day 0 (Table S1). Collectively, the results indicated that co-transplantation of ECs and CMs effectively rescued heart dysfunction and prevented ventricular dilation after an I/R injury in non-human primates. It might be accompanied by sustained ventricular tachycardia events that would decrease over time.

To identify the infarct area, we collected the hearts at day 28 after cell injection and stained the heart sections with picrosirius red/fast green and found that animals with EC + CM co-transplantation tend to had smaller infarct size compared to CM alone and significantly smaller than control (Figure 6A and 6B). We labeled human grafts and quantified the graft size in all animals. Both cell-transplanted groups had identifiable graft islands. EC + CM tended to result in greater graft size than CM alone (Figure 6C and 6D). Next, we sought to investigate if the transplanted cells improved cardiac function in I/R non-human primates through augmenting vascularization of engrafted myocardium that successfully integrated with host myocardium. Vasculature was labeled with a CD31 antibody and the density was quantified (Figure S10A). Compared to CM alone and control animals, we found vessels to be significantly more abundant in EC + CM treated animals in the infarct area (Figure S10B) and border zone (Figure S10C) but not in the remote area (Figure S10D). Within the grafts of cell treated animals, we also observed a significantly higher degree of neo-vascularization in co-transplanted animals (Figure S10E and S10F). For CM maturation, we found significantly greater sarcomeric length (Figure S11A and S11B), polarized adherens junction protein pan-cadherin in hiPSC-CMs within the human grafts, and more importantly, a host-graft integration of pan-cadherin in EC + CM animals, which was less frequently found in CM alone (Figure S11C and S11D). Overall, these results demonstrated that the EC + CM therapy effectively improved post-I/R cardiac function in non-human primates through enhancement of vascularization in the infarct area with highly matured and host-integrated hiPSC-CM grafts.

Figure 6. Co-transplantation of ECs reduces infarct size in non-human primates with ischemic-reperfusion injury.

A, Representative images of infarct area stained by picrosirius red and fast green. Scale bar, 10 mm. B, Quantification of infarct area. n=4 macaques for EC+CM engraftment, n=3 macaques for CM engraftment, and n=4 macaques for control. C, Representative whole heart slide images of grafts. D, Quantification of graft size. n=4 macaques for EC+CM engraftment and n=3 macaques for CM engraftment. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test in B and was determined by unpaired t test in D. *P < 0.05, ns for not significant.

DISCUSSION

The inability of myocardium to regenerate hinders self-recovery from ischemic heart injury. To tackle this world-leading cause of death, studies aiming to develop cell therapies by injecting hiPSC-derived CMs in the hope of heart remuscularization are thriving and showing promising results. However, the maturity of injected cells and their integration with host cells remain major concerns. In this study, we used isogenic CMs and ECs derived from hiPSCs and showed that co-transplantation with ECs augmented CM graft vascularization and CM maturation which boosted graft-host integration and ultimately improved cardiac function in both murine and non-human primate models.

The effect of ECs in promoting hPSC-CM maturity in vitro has been reported previously. For example, Vuorenpaa et al. reported the effect of HUVECs and human foreskin fibroblasts in promoting hPSC-CM maturity, and Pasquier reported that the use of HUVECs benefits hESC-CMs. In contrast to these studies, the current study, using ECs derived from the same human iPSC line as CMs, would be more clinically feasible because such an approach would lower the complexity of immunogenicity in cell therapy. Using the same cell line to differentiate various cell types, Giacomelli et al. demonstrated the formation of microtissues by combining hiPSC-CFs and hiPSC-ECs with hiPSC-CMs to improve their maturity. In the current study, we found that ECs sufficiently promote CM maturity. This difference between the two studies might be because of the differences in EC isolation protocol, the ratio between ECs and CMs, and the co-culture time period. Furthermore, we showed the efficacy of ECs in improving CM maturity in vivo and the therapeutic effect of EC+CM combined treatment. The effect of human PSC transplantation for the treatment of ischemic heart injury has been studied in a variety of animal models including rats²⁶, guinea pigs², pigs³, and non-human primates^4,5. Most of these studies showed improvement in cardiac function after one to three months after PSC-CMs injection through remuscularization in the infarcted myocardium. However, poor survival of injected CMs, due to insufficient vascularization within the grafts, is a major hurdle to overcome before this cell therapy can be clinically applicable. To address this issue, some studies introduced the use of hESC-derived epicardial cells¹⁰, rat primary aortic ECs²⁷, or rat adipocyte-derived microvessels²⁷ in addition to CMs in rodent models; or hiPSC-derived ECs and smooth muscle cells in combination with hiPSC-CMs in a porcine model²⁸. These advanced therapies displayed an improvement in CM maturity and survival rate that yield better cardiac outcomes. Our current study further used a more phylogenetically proximate animal model to demonstrate the efficacy of hiPSC-EC in combination with isogenic hiPSC-CMs in treating ischemic heart injury, which was developed based on a thorough investigation of in vitro and in vivo mouse models. Previous studies have demonstrated the efficacy of hPSC-CM graft in restoring heart function in non-human primates with subacute myocardial infarction by injecting cells two weeks after cardiac injury^4,5. In the current study, we further investigated the efficacy of hiPSC-CM transplantation, adding combined treatment with ECs, in a more chronic setting by injection cells four weeks after cardiac injury, to increase clinical relevance.

In the mice receiving cell injection after MI, 15 out of 16 mice receiving EC + CM had a positive signal of human graft 21 days after surgery, whereas the grafts survived in only three-quarters of the mice when injected with CM only. Doubling the amount of CMs injected had no effect on lifting the transplantation success rate. Moreover, in animals with positive human graft signal, those receiving EC + CM had a two-fold larger graft size than those receiving the same number of CMs alone and even those with double amount of CMs. This indicates that co-injection with ECs effectively enhances CM survival rate while increasing injected CM number does not. Another intriguing thing is that injecting ECs alone could improve cardiac function despite the lack of a functioning graft. From the histological results, we found injection of ECs alone effectively enhances vessel density in the infarct region, which might help in reducing post-infarction cell death and thereby limiting fibrotic area. However, co-injecting CMs with ECs outperformed ECs alone treatment by enhancing neo-vascular density, restricting fibrosis, and better-restoring cardiac function without increasing the chance of arrhythmic incidence. Therefore, these results indicate that although injecting ECs alone readily benefits post-infarction heart repair, compensating for the loss of myocytes by adding CMs is still required to maximize the therapeutic effects.

In both rodent and non-human primate models, it has been shown that injected hPSC-CMs, having a fetal characteristic in its structure and function within one month after transplantation, required three months to become more mature and display more adult phenotypes in vivo^4,27. For example, an about 10% increment of ejection fraction was reported after the delivery of 7.5 x 10⁸ day 21-differentiated ESC-CMs two weeks after cardiac injury⁴. Whereas an increase of approximately 2% of ejection fraction was observed after transplantation of 5 x 10⁸ day 16-differentiated iPSC-CMs four weeks post injury in the current study. Our study showed positive results in CM alone treatment, in line with the previous study. However, there are lots of factors that might affect the outcome such as number of cells delivered, cell lines, CM differentiation protocol (e.g., modified pro-survival cocktail), and surgent carried out the animal experiments. Furthermore, we found that co-injection with ECs boosted CM maturation, evidenced by longer sarcomeric length compared to CMs alone and the presence of polarized pan-cadherin with graft-host integration one month after cell injection in the non-human primate model. The greater reduction in time required for hPSC-CMs to grow into functioning CMs allows the approach reported to be more efficient and economic. In addition, tumorigenicity has been a major concern for PSC-derived cell therapies due to its high proliferation capacity²⁹. As the inverse relationship between CM proliferation and maturation has been well demonstrated^30,31, our findings of lower proliferation index in EC-co-treated CMs indicate that the CMs are more mature. Reduction in CM proliferation index could also be achieved by long-term treatment^4,27. This suggests that EC + CM co-injection therapy is safer than CMs alone therapies by reducing the time highly proliferative CMs remain in the recipients.

Another potential benefit of EC + CM co-treatment is the reduction in occurrence of arrhythmia in MI mice receiving cell therapy. We found that every one in four mice treated with CMs alone would have PVC even 21 days after treatment. Doubling the amount of CMs doubled the PVC incidence and even caused ventricular couplet. However, co-treating ECs and CMs in comparable total cell numbers largely reduced the arrhythmic incidence. However, due to limitations in mouse-applicable devices and detection time, we could not track every arrhythmic event throughout the experiment. To confirm that ECs are beneficial not only for preventing arrhythmia incidence at termination but also for reducing arrhythmic events during the process of cardiac repair, long-term ECG recording should be implemented in future studies. Using non-human primates, we showed that ECs promote neo-vascularization and CM maturation in the human graft that could integrate into the host myocardium. Furthermore, we showed that EC + CM treatment is effective in restoring cardiac function in non-human primates with an I/R injury. Non-human primates receiving this cell therapy had vascularized grafts consisting of well-aligned and rod-shaped CMs that integrate with the host myocardium. The main bottleneck associated with bringing this therapy to the clinic is arrhythmogenesis. In fact, different animal protocols including species, infarct size, treatment interval, and different cell protocols including cell line, cell number, differentiation method, and purity, and even detection limit of arrhythmia detection devices could affect arrhythmogenesis and observation. Our treatment paradigm of administering ECs and CMs together might reduce this risk as shown by subsided arrhythmias in the co-transplanted group at later time point, however, future studies with prolonged time points and a larger biological sample size would be needed to understand the long-term risk of arrhythmogenesis.

In summary, we demonstrated that co-transplantation of EC + CM enhances CM structural and functional maturity in vitro and in vivo. The mechanisms of EC + CM benefiting cardiac repair after ischemic heart injury through promoting angiogenesis, CM survival, and CM maturation were fully investigated in a MI mouse model. In healthy non-human primates, co-transplantation with EC augmented graft neo-vascularization and enhanced CM maturation. With an I/R injury, non-human primates receiving EC + CM co-transplantation had improved cardiac function with low frequencies of ventricular tachycardia, accompanied by more highly matured CMs and vascularized grafts that integrated with the host myocardium. Overall, we proposed a novel cell therapy using isogenic hiPSC-CMs and hiPSC-ECs for the treatment of ischemic heart disease that holds hope for patients to receive grafts in an efficient and safe way. With the reported approach, we also demonstrate the feasibility of biomanufacturing cells in Taiwan. These cells can be safely shipped internationally and delivered effectively, indicating that international distribution could meet the urgent needs of patients worldwide.

Limitation of study

There are limitations in the current study that need to be addressed. One of the major benefits in EC + CM co-injection is that ECs promote neo-vascularization within the grafts. The neo-vessels were perfused by the host circulating system, evidenced by the presence of bi-concaved disc-shaped erythrocytes in the lumen of neo-vessels. With a human specific CD31 antibody, we identified hiPSC-EC-derived vessels in the mouse model. However, it cannot distinguish human and macaque CD31. Further study will be needed to determine the fate of the injected ECs in non-human primate. Due to high mortality after IR surgery and the second thoracotomy in NOD-SCID mice, we used MI model and deliver the cells immediately post-MI without performing a second thoracotomy. After getting positive results from EC+CM treated mice and in healthy macaques, we then moved on to the more clinically relevant macaque model of IR. In the non-human primate models, further studies would be required to understand the effect of ECs alone and vehicle control groups as well as longer time-periods with varying doses of CMs, including a CM alone group with comparable total cell number to EC+CM group. In addition, we only used male animals in the in vivo studies. Further studies investigating female non-human primates will be important to demonstrate if there is a gender difference in therapeutic efficacy.

Clinical Perspective.

What Is New?

• For the first time, co-transplantation of human iPSC-ECs with human iPSC-CMs derived from the same iPSC line was demonstrated to improve cardiac function in mice and non-human primates following ischemic heart injury.

• Transplantation of iPSC-ECs and iPSC-CMs was superior to iPSC-CMs alone in therapeutic efficacy with improved cardiac function, augmentation of neo-vascularization, and enhancement of CM maturation.

What Are the Clinical Implications?

• Human iPSC-ECs are a promising cell source to augment remuscularization by iPSC-CMs in patients with ischemic heart injury.

• These findings identify a novel strategy to promote heart regeneration using a combination of multiple iPSC-derived cardiac cell types for optimal treatment of ischemic heart injury.

Nonstandard Abbreviations and Acronyms

hiPSC

human induced pluripotent stem cell

CM

cardiomyocyte

EC

endothelial cell

PSC

pluripotent stem cell

ESC

embryonic stem cell