Cardiac repair in a porcine model of acute myocardial infarction with human induced pluripotent stem cell-derived cardiovascular cell populations

Differentiation of hiPSCs into cardiac-lineage CMs, ECs, and SMCs

HiPSCs were reprogrammed from human dermal fibroblasts and engineered to express eGFP (Figure S1) and differentiated into CMs via the Sandwich method (Zhang et al., 2012). Isolated areas of contracting cells typically appeared on day seven of differentiation (Movies S1–S3) and were collected one week later for purification via a micro-dissection and preplating method. Expression of cardiac-specific proteins in differentiated hiPSC-CMs was evaluated on Day 30 after cells started contracting. Nearly all hiPSC-CMs expressed slow myosin heavy chain, α-sarcomeric actin (Figure 1A), and the cardiac-specific myofilament cTnT (Figure 1B). Approximately 20–30% of the hiPSC-CMs expressed the 2v isoform of myosin light chain (MLC2v) (Figure 1B, middle), which is found only in ventricular CMs, and cardiac connexin-43 was present at numerous points of contact between adjacent cells (Figure 1C). The purity of the final hiPSC-CM population was as high as 93% when evaluated via flow cytometry analysis of cTnT expression (Figure 1D–E) and >90% when evaluated via fluorescence immunostaining for cTnT expression (Figure 1F). hiPSC-ECs and -SMCs were generated via established differentiation protocols (Hill et al., 2010; Woll et al., 2008) and expressed EC- and SMC-specific proteins (Figures 1G–1L).

hiPSC-derived cardiac cells engraft and survive following transplantation into a porcine model of myocardial infarction

We then tested our hypothesis that transplantation of three hiPSC-derived cardiac cell types would improve recovery after ischemic myocardial injury in a large-animal model (swine) of ischemia-reperfusion (IR) injury. A total of 92 pigs underwent the IR protocol; 89 pigs survived and were divided into 5 groups (Table S1). Animals receiving CMs, EC, and SMCs (Cell group) or CMs, ECs, SMC, and the fibrin/IGF-1 patch (Cell+Patch group) were injected with 2 million hiPSC-CMs, 2 million hiPSC-ECs, and 2 million hiPSC-SMCs (6 million cells total) directly into the injured myocardium; for animals in the Cell+Patch group, the needle was inserted through an IGF-1–containing fibrin scaffold patch (Figure S2) that had been created over the site of injury. Animals in the Patch group were treated with scaffold patch alone, and both the patch and the cells were withheld from animals in the MI group. Animals in the Sham group underwent all surgical procedures for induction of IR injury, except for the ligation step, and recovered without any experimental treatments.

We then evaluated the engraftment and survival rates of the transplanted cells. Since the transplanted cells were genetically male and engineered to express GFP, while the recipient pigs were female, we performed quantitative PCR (qPCR) assessments for the human Y chromosome. Lineages of surviving cells were determined by staining for expression of the human-specific EC marker CD31 (hiPSC-ECs), for co-expression of GFP and α-smooth-muscle actin (SMA) (hiPSC-SMCs), and for co-expression of GFP and cTnT (hiPSC-CMs).

At four weeks after injury, 4.2±1.1% of transplanted cells survived in animals in the Cells group while 8.97±1.8% survived in the hearts of Cell+Patch animals. In contrast, 3.2±0.4% of the 2 million cells administered to the CM group survived. Substantial proportions of all three transplanted cell types were observed in Cell+Patch hearts: 27.1±5% of surviving transplanted cells were hiPSC-CMs, 34.2±10% were hiPSC-ECs, and 40.5±1% were hiPSC-SMCs. These observations suggest that the IGF-1–containing fibrin patch may have substantially improved the engraftment of all three hiPSC-derived cell lineages. Furthermore, GFP-expressing cells were present in cardiac muscle fibers (Figures 2A), but rarely in fibers composed of porcine cardiomyocytes (Figures S3A–S3D), suggesting that the transplanted hiPSC-CMs may have developed into new muscle fibers rather than incorporating into existing host cardiac muscle. GFP⁺ cells were also identified in capillaries and arterioles (Figures 2B–2C), although the proportion of vessels apparently generated directly from transplanted cells was less than 0.1%.

Transplantation of hiPSC-derived cardiac cells improves cardiac function and bioenergetics after MI

Measurements of cardiac function were evaluated one and four weeks after injury and cell transplantation. Left-ventricular ejection fraction (LVEF) was significantly better in Cell+Patch group animals than in MI and Patch animals (Figure 3A) at four weeks. Systolic thickening fractions that reflect regional myocardial contractility, were significantly greater in the infarct zone (IZ) and at the border zone (BZ) of the infarct in hearts from Cell+Patch animals than in MI hearts at both week 1 and week 4 timepoints (Figure 3B). Infarct size was significantly smaller in Cell+Patch hearts than in MI hearts (p<0.05) at Week 1, while the difference between Cell+Patch and Patch hearts approached statistical significance (p=0.051) (Figure 3C). Measurements of regional wall stress performed 4 weeks after injury were significantly lower in Cell+Patch hearts than in Patch hearts (Figure 3D). Together, these results show that transplantation of multiple hiPSC-derived cardiac lineages, in combination with IGF-1/fibrin patch, improves cardiac function following MI.

We then determined whether these functional improvements in Cell+Patch animals were accompanied by improved myocardial bioenergetics and efficient ATP utilization. Myocardial ATP hydrolysis rates were measured in vivo with the ³¹P MRS-MST method (Xiong et al., 2013). Measurements were performed at the BZ of the infarct under both baseline and elevated cardiac workstates in response to catecholamine stimulation.

Myocardial PCr/ATP ratios that reflect mitochondrial energetic effciancy, were significantly higher in animals from the Cell+Patch group than in MI animals under baseline conditions (Figure 3E), while the rate of ATP hydrolysis was significantly higher in Cell+Patch group hearts than in MI and Patch hearts during high cardiac workload (Figure 3F); measurements in Patch and MI animals were similar under both conditions. Collectively, these observations indicate that the transplantation of hiPSC-derived cardiovascular cells improved LV pump function and myocardial energy metabolism while reducing infarct size.

Combined hiPSC-CM and patch treatment does not induce ventricular arrhythmias

Arrhythmogenesis is a primary risk associated with pluripotent cell-derived CM therapy for treatment of cardiac disorders. In a recent study, significant ventricular arrhythmias were detected following transplantation of hESC-CMs in a non-human primate model of MI (Chong et al., 2014). Thus, we examined whether the hiPSC-CM patch transplantation protocol used here might be associated with onset of cardiac arrhythmia. Ischemia-reperfusion injury was induced in 16 pigs who then received the fibrin/IGF-1 patch alone (Patch group), the patch and 10 million hiPSC-CMs (Patch+CMs), or neither experimental treatment (Table S1). ECGs were recorded continuously for 4 weeks afterward using an implantable loop recorder. Severe arrhythmias and ST elevations were observed in all animals during coronary artery occlusion and reperfusion, but no animal in any treatment group developed spontaneous arrhythmia during the four -week follow-up period. Furthermore, no evidence of ventricular tachycardia (VT) or ventricular fibrillation (VF) was induced in response to programmed electrical stimulation (PES) in hearts of Patch+CM group animals, even when the most aggressive stimulation protocol was applied. Thus, the combined Patch+hiPSC-CM administration protocol used for the experiments described in this report does not appear to impair the electromechanical stability of swine hearts.

Transplantation of hiPSC-derived cardiac cells reduces cardiomyocyte apoptosis and stimulates Nkx2.5 expression after IR injury

We next assessed how Cell+Patch treatment improves cardiac function. One possibility is through induction of cytoprotective mechanisms. To evaluate this possibility, we measured apoptosis in the hearts of animals sacrificed three days after IR injury and treatment (Figures 4A–G). At the BZ of the infarct, apoptotic (i.e., TUNEL⁺) cells were significantly less common in animals from the Cell+Patch group than in MI and Patch animals, both in the total cell population (Figure 4F) and specifically among CMs (i.e., cTnI⁺ cells) (Figure 4G). Furthermore, the proportion of CMs that expressed Nkx2.5 (Figures 4H–N), which has been shown to protect CMs from oxidative stress (Toko et al., 2002), was significantly greater in Cell+Patch animals than in MI animals at Week 4 (p<0.05), and the difference between Cell+Patch and Patch animals approached statistical significance (p=0.056), at Week 1 after IR injury (Figure 4N). Nkx2.5 expression in MI hearts increased at Week 4, but remained significantly lower than in Cell+Patch hearts. Thus, transplantation of hiPSC-derived cells increased cell survival during the first few days after IR injury, and is associated with upregulation of Nkx2.5 expression. The decline in apoptosis could also result from improvements in wall stress and decreased fibrosis. Consistently, examination of Masson-Trichrome–stained heart sections from animals sacrificed at week 4 showed thicker subepicardium in the region where the patch was applied (~2.44 mm) than in animals from the Sham (0.16 mm) or MI (0.88 mm) groups (Figure S4).

Transplantation of hiPSC-derived cardiac cells enhance the vasculogenic response to IR injury

Increased angiogenesis could also contribute to improved ventricular function. We therefore assessed whether transplantation of hiPSC-derived cardiac cells promotes angiogenesis in peri-infarct border zone (BZ) of MI hearts. Four weeks after injury, CD31+vascular structures and arterioles that co-expressed CD31 and SMA were significantly greater in the BZ of Cell and Cell+Patch hearts than in the corresponding regions of MI and Patch hearts (Figure 5). Additionally, vascular density was significantly greater in the CM group than in the MI group. Thus, transplantation of the hiPSC-derived cardiac cells promoted neovascularization, likely by inducing paracrine mechanisms in recipient myocardial tissue (Table S2).

The cell and patch treatments may also have delayed the immune response to IR injury. While infiltration of cells expressing the inflammatory marker CD11b (Figures S5A–S5F) peaked three days after injury in the MI group, this was significantly delayed in the IZ and BZ regions of hearts from Cell+Patch animals (Figure S5G). CD11b⁺ cells were significantly more common in Cell+Patch hearts than in MI hearts at week 1 but not at Week 4, while the CM, Cells, and Patch groups showed a statistically non-significant increase at weeks one and four, suggesting a delayed host immune response following cell transplantation, possibly reducing immune rejection and increasing engraftment and survival of hiPSC-derived cells.

Paracrine factor release from the hiPSC-derived cells

This increase in angiogenesis observed following Cell+Patch treatment could be due to release of paracrine factors, as suggested by studies showing endothelial cells promote cardiomyocyte survival (Lan et al., 2006; Xiong et al., 2012). To determine whether hiPSC-derived vascular cells possess cytoprotective effects similar to those previously reported in similar cell types, hiPSC-CMs were cultured under hypoxic conditions in media collected from the hiPSC-derived vascular cells (i.e., conditioned media) or with unconditioned (basal) media. hiPSC-CMs tended to shrink when cultured with unconditioned media but not when cultured with the conditioned media (Figures 5I–5J). Culture in conditioned media was also associated with significant declines in hiPSC-CM apoptosis (Figures 5K–5L, 5N) and in the amount of cytoplasmic leakage of lactate dehydrogenase (LDH) from hiPSC-CMs (Figure 5M). Thus, hiPSC-derived vascular cells appear to release paracrine factors that protect hiPSC-CMs from hypoxic injury.

We then performed protein array analysis to identify paracrine factors that may be responsible for these in-vitro cytoprotective effects and that might induce repair mechanisms in injured myocardial tissue. A total of 21 factors were detected, and their expression was confirmed in hiPSC-CMs, as well as hiPSC-derived vascular cells (Table S2). Several of the identified factors are known to impede apoptosis (angiogenin, angiopoietin, IL-6, MMP-1, PDGF-BB, TIMP-1, uPAR, and VEGF), induce cell migration or homing (angiogenin, angiopoietin, IL-8, MCP-1, MCP-3, MMP-9, uPAR, VEGF), and promote cell division (angiogenin, angiopoietin, PDGF-BB, VEGF), suggesting multiple paracrine mechanisms through which hiPSC-ECs and –SMCs could promote CM survival and cardiac repair.

Myocardial protein expression after IR injury and cell therapy

Understanding changes in the expression profile of myocardial proteins will help reveal the molecular mechanisms underlying improvements in left ventricular function associated with cell therapy. To provide further insights into these mechanisms, we performed a quantitative proteomics analysis on LV tissues from SHAM hearts and from MI hearts that were treated with or without hiPSC-derived vascular cells (which have both hiPSC-ECs and -SMCs) after IR injury (Figures S6). We confidently identified 66 proteins whose expression pattern was altered in MI hearts but fully or partially restored to normal after cell transplantation (Figure 6). Thus, the functional benefits associated with cell therapy appear to be accompanied by a recovery of the myocardial protein expression profile of the recipient hearts.

Generation and characterization of hiPSC-derived cardiac cells

The hiPSC-CMs, -SMCs, and -ECs used in this study were generated from two hiPSC lines: DriPS16 (Cell+Patch group) and GRiPS (for the CM and CM+EC+SMC groups). Both hiPSC lines were reprogrammed from male, human, neonatal, dermal fibroblasts by transfecting cells with either lentivirus (DriPS16) or Sendai viruse (GRiPS) coding for OCT4, SOX2, KLF4, and C-MYC and then engineered to constitutively express green fluorescence protein (GFP). Cells were cultured in hiPSC growth medium with irradiated mouse embryonic fibroblasts and passaged every 6–7 days (Wilber et al., 2007).

hiPSCs were differentiated into ECs and SMCs as described previously (Hill et al., 2010; Woll et al., 2008). hiPSC-ECs were characterized via the expression of CD31, CD144, and vWF-8 (Xiong et al., 2012). hiPSC-SMCs were characterized via the expression of α-smooth muscle actin (SMA), SM22, and calponin.

In vitro cytoprotection assays are detailed in Supplemental Experimental Procedures

Paracrine factor release from cultured hiPSC-CMs, ECs and SMCs hiPSC-CMs (3.5×10⁵) and hiPSC-SMCs (3.5×10⁵) were cultured in 6-well plates, washed 3 times with DPBS, and then cultured in 1 mL RPMI basal medium; hiPSC-ECs (4×10⁵) were cultured in 1 mL EBM2 basal medium. The medium was collected 48 hours later and analyzed with the Human Angiogenesis Antibody Array G1 Array (RayBiotech, USA). Measurements were corrected for background signals via control assessments with basal media that had been exposed to identical conditions.

Synthesis of IGF-containing microspheres and patch manufacture are detailed in Supplemental Experimental Procedures.

Porcine IR injury model and treatment

Experiments were performed in female Yorkshire swine (~13 kg, 45 days of age, Manthei hog farm, Elk River, MN, USA) (Xiong et al., 2012). The experimental protocol was approved by the IACUC of the University of Minnesota, and all experimental procedures were performed in accordance with the Animal Use Guidelines of the University of Minnesota and consistent with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH publication No 85–23).

A total of 108 pigs underwent the ischemia reperfusion (IR) protocol (Table S1). Ninety-two pigs were used in the first part of the study: 2 pigs died of ventricular fibrillation during occlusion, and 1 died of cardiac arrhythmia one week after IR injury while the MRI data were being collected. The remaining 89 pigs were divided into 6 groups. Animals in the CM+EC+SMC and Cell+Patch groups were treated by injecting 2 million hiPSC-CMs, 2 million hiPSC-ECs, and 2 million hiPSC-SMCs (6 million cells total) directly into the injured myocardium; for animals in the Cell+Patch group, the needle was inserted through an IGF-1–containing fibrin patch that had been created over the site of injury. Animals in the Patch group were treated with the IGF-1–containing patch alone, and both the patch and the cells were withheld from animals in the MI group. Animals in the SHAM group underwent all surgical procedures for the induction of IR injury except for the ligation step and recovered without any of the experimental treatments. 16 pigs were used in loop recorder study. The Patch+CM group used in the arrhythmogenesis experiments exposed to a protocol of fibrin patch enhanced delivery of ten million hiPSC-CMs on surface of the injured myocardium (Table S1).

Patch application was performed by suspending 5 mg of microspheres (loaded with 2.5 μg IGF-1) in 1 mL fibrinogen solution (25 mg/mL); then, the fibrinogen solution was co-injected with 1 mL thrombin solution (80 NIH units/mL, supplemented with 2 μL 400 mM CaCl₂ and 200 mM ε-aminocaproic acid) into a 2.3-cm diameter plastic ring that had been placed on the epicardium of the infarcted region to serve as a mold for the patch; the mixture usually solidified within 30 seconds (Xiong et al., 2012). Cells were suspended in 1 mL MEM and administered via 10 intramyocardial injections (0.1 mL/injection).

Cardiac MRI and MR Spectroscopy are detailed in Supplemental Experimental Procedures

The ECG monitoring and programmed electro-stimulation physiology studies

The implantable loop recorders (Medtronic-Reveal, MN, USA) were placed in the left paraspinal area inferior to the angle of the scapula in the subcutaneous plane. It was sutured in the place where the best electrograms were obtained and there was no evidence of myopotential noise. It was programmed in the conventional manner to document VT and asystole. The loop recorder was interrogated at the time of explantation when the animals were sacrificed 4 weeks after the cell therapy.

The programmed electro-stimulation physiology studies (PES study) were done at the time of sacrifice in four weeks. The PES study was done from the epicardium in an open chest fashion. The PES study was done from two sites: one close to the infarct and one remote from the infarct. The study was done with a Medtronic screw lead in the epicardium and the Bard system was used for stimulation. It was done at two cycle lengths at 400 ms and 300 ms drive trains. Four additional stimuli were given till effective refractory period (ERP) was reached or 160 ms.

hiPSC-EC, -SMC, and -CM engraftment rate and immunohistochemical evaluations are detailed in Supplemental Experimental Procedures

Materials and methods for proteomics are detailed in Supplemental Experimental Procedures.

Statistical analysis

Results are presented as mean±standard error of the mean (SEM). Comparisons among groups were analyzed for significance with one-way analysis of variance (ANOVA). A value of p<0.05 was considered significant. Results identified as significant via ANOVA were re-analyzed with the Tukey correction. Statistical analyses were performed with SPSS software (version 20).