Aims Myocardial cell replacement therapies are hampered by a paucity of sources for human cardiomyocytes and by the expected immune rejection of allogeneic cell grafts. The ability to derive patient-specific human-induced pluripotent stem cells (hiPSCs) may provide a solution to these challenges. We aimed to derive hiPSCs from heart failure (HF) patients, to induce their cardiomyocyte differentiation, to characterize the generated hiPSC-derived cardiomyocytes (hiPSC-CMs), and to evaluate their ability to integrate with pre-existing cardiac tissue.
Methods and results Dermal fibroblasts from two HF patients were reprogrammed by retroviral delivery of Oct4, Sox2, and Klf4 or by using an excisable polycistronic lentiviral vector. The resulting HF-hiPSCs displayed adequate reprogramming properties and could be induced to differentiate into cardiomyocytes with the same efficiency as control hiPSCs (derived from human foreskin fibroblasts). Gene expression and immunostaining studies confirmed the cardiomyocyte phenotype of the differentiating HF-hiPSC-CMs. Multi-electrode array recordings revealed the development of a functional cardiac syncytium and adequate chronotropic responses to adrenergic and cholinergic stimulation. Next, functional integration and synchronized electrical activities were demonstrated between hiPSC-CMs and neonatal rat cardiomyocytes in co-culture studies. Finally, in vivo transplantation studies in the rat heart revealed the ability of the HF-hiPSC-CMs to engraft, survive, and structurally integrate with host cardiomyocytes.
Conclusions Human-induced pluripotent stem cells can be established from patients with advanced heart failure and coaxed to differentiate into cardiomyocytes, which can integrate with host cardiac tissue. This novel source for patient-specific heart cells may bring a unique value to the emerging field of cardiac regenerative medicine.
Induced pluripotent stem cells
Recent advancements in stem cell biology and tissue engineering have paved the way for the introduction of a new discipline in biomedicine: regenerative/reparative medicine. One of the goals of this emerging cardiovascular discipline is to repopulate scar tissue with new contractile cells in an attempt to assist the failing heart.1,2 This exciting approach has been hampered, however, by the paucity of cell sources for human cardiomyocytes.
The derivation of human embryonic stem cells (hESCs)3 provided a possible solution to this cell-sourcing problem because of their capacity to differentiate into cardiomyocytes4,5 and their ability to improve cardiac function in animal models of myocardial infarction.6,7 Yet, a major obstacle to the clinical utilization of these cells is their restricted availability and the anticipated immune rejection following transplantation of allogeneic cells.
An important breakthrough was made with the introduction of the human-induced pluripotent stem cells (hiPSCs) technology.8–10 The hiPSCs were derived by the reprogramming of adult somatic cells with a set of transcription factors, yielding pluripotent cells closely resembling hESCs. More recent studies demonstrated the ability to differentiate hiPSCs into the cardiac lineage11–13 and to develop in vitro models of inherited cardiac disorders.14–16
While the aforementioned studies confirmed the ability to derive hiPSCs from young or healthy individuals, it is still not clear whether such lines could be established, and differentiated into cardiomyocytes also in elderly and diseased patients. To address this important issue, we aimed to derive hiPSCs from ischaemic cardiomyopathy patients suffering from advanced heart failure (HF) (representing the prototype candidates for future autologous myocardial cell-replacement procedures). Our results show the ability to efficiently derive hiPSC lines from such patients and to differentiate them into cardiomyocytes. The generated hiPSC-derived cardiomyocytes (hiPSC-CMs) were demonstrated to functionally integrate with pre-existing cardiac tissue in vitro and to engraft, survive, and integrate structurally with the host myocardium following in vivo transplantation.
Generation and cardiomyocyte differentiation of patient-derived human-induced pluripotent stem cells
The study was approved by the Rambam Medical Center Helsinki committee. Patient-derived hiPSC clones were established by retroviral delivery of three reprogramming factors (Sox2, Klf4, and Oct4) followed by valproic acid (VPA) treatment as previously described.14 To induce cardiomyocyte differentiation, hiPSCs were dispersed into small clumps with collagenase-IV (Life-Technologies, 1 mg/mL), cultivated in a suspension for 10 days as embryoid bodies (EBs), and plated on 0.1% gelatin-coated culture dishes.13
Generation of transgene-free heart failure-human-induced pluripotent stem cells
A single polycystronic lentiviral vector containing the four-factor ‘stem cell cassette’ (STEMCCA),17 which can be excised after integration using Cre-recombinase, was used for transduction of the fibroblasts (kind gift from Prof. Mostoslavsky). The STEMCCA cassette was co-transfected with the plasmid Gag-Pol and the helper plasmid encoding VSVG into HEK293T cells to produce viruses. After infection, cells were cultured in ES medium starting from 1 day post-infection and treated with 0.9 mM VPA for 2 weeks.
For Cre-recombinase-mediated vector excision, hiPSCs were harvested with 0.05% trypsin/EDTA (Invitrogen), re-suspended in PBS (1 × 106cells), and transfected by electroporation with pCAG-Cre-EGFP (10 mg). Cre-recombinase eGFP-expressing cells were selected from a single-cell suspension by FACS sorting (FACSAria, BD-Biosciences) 72 h after electroporation. Selected cells were re-plated at low density in ES medium containing the ROCK inhibitor (Sigma).
Specimens were fixed with 4% paraformaldehyde, permeabilized with 1% Triton-X-100 (Sigma), blocked with 5% horse serum, and incubated overnight at 4°C with primary antibodies targeting: Tra-1-60, Oct-4, connexin-43 (Cx43) (all from Santa-Cruz), Nanog (Abcam), TRA-I-81, human mitochondria antigens (Millipore), cardiac troponin-I (cTnI, Chemicon), SSEA-4, cardiac troponin-T (cTnT) (R&D), and sarcomeric-α-actinin (Sigma). Preparations were incubated with secondary antibodies for 1 h and examined using a laser-scanning confocal microscope (Zeiss LSM-510-PASCAL).
Undifferentiated hiPSCs were injected subcutaneously to immunodeficient SCID-beige mice. Teratomas, developing 4–8 weeks after injection, were harvested, cryosectioned (10 µm), and H&E stained.
Genomic DNA (1 µg) was bisulfite converted with the Methylamp-DNA-Modification-kit (Epigentek) and amplified using Faststart Taq-polymerase (Roche). PCR products were TA-cloned into pTZ57R/T plasmid (Fermentas). Inserts were sequenced with M13 universal primers. Primer sequences appear in Table 1.
List of primers used for RT–PCR, quantitative-PCR, and bisulfite sequencing
Primers for RT–PCR and quantitative-PCR
Primers for bisulfite sequencing
Karyotype analysis was performed using the standard G-banding chromosome analysis at the institution's cytogenetic laboratory.
Gene expression studies
Undifferentiated hiPSCs and beating EBs (30–40 days) were frozen in liquid nitrogen. RNA was isolated using the RNeasy-plus mini-kit (Qiagen). Reverse transcription into cDNA was conducted using a high capacity cDNA-reverse-transcription kit (Applied-Biosystems). The PCR program used was 3 min at 93°C, 30 s at 93°C, 30 s at 60°C, and 30 s at 72°C (30 cycles) using DreamTaq™ DNA Polymerase (Fermentas).
SYBR green quantitative-PCR studies were performed in triplicates using SYBR Green PCR Master Mix and the ABI-7000 Sequence Detector (Applied-Biosystems). Samples were cycled 40 times (2 min at 50°C, 15 min at 95°C) followed by 40 cycles of 15 s at 95°C, 30 s at 60°C, and 30 s at 72°C. The threshold cycle (CT) was calculated under default settings for the real-time sequence detection software. Expression levels were normalized to β-tubulin transcript levels. Primers for RT–PCR and quantitative-PCR are detailed in Table 1.
Multi-electrode array recordings
Contracting EBs were microdissected and plated on fibronectin-coated MEA plates (multi-channel systems). Local activation times (LATs) were determined at each of the 60 electrodes and used to generate colour-coded activation maps as described.18
Primary monolayer cultures of neonatal rat (Sprague-Dawley) ventricular myocytes (NRVMs) were prepared as described18 and cultured on MEA plates (2 × 106cells/mL). Contracting EBs were then added to the NRVM cultures.
Female Sprague-Dawley rats (250–300 g) were anaesthetized (ketamine/xylazine), intubated, and ventilated. The hiPSC-derived cardiomyocytes (1.5 × 105) were pre-labelled with a fluorescent cell tracer (Qtracker, Invitrogen) and injected into the rat's left ventricular myocardium (n= 3). To prevent graft rejection, animals were treated with cyclosporine-A (15 mg/kg/day) and methylprednisolone (2 mg/kg/day). Hearts were harvested 7–10 days following grafting, frozen in liquid nitrogen, and cryo sectioned (10 µm) for histological examination.
Results are reported as mean ± SEM. To compare between the different hiPSC clones (for differentiation efficiency and beating rates), one-way ANOVA was used followed by post hoc Bonferroni analysis. Drug effects on beating rates were analysed by paired T-tests. All statistical methods were two sided. The statistical software used for analysis was Sigma-Stat (version 3.5). P< 0.05 was considered statistically significant.
Derivation and characterization of the heart failure-human-induced pluripotent stem cells
Dermal fibroblasts were obtained from two patients with advanced HF due to ischaemic cardiomyopathy (both males, ages 51 and 61). The fibroblasts were reprogrammed to generate the HF-hiPSCs by retroviral infection of three reprogramming factors: Oct4, Sox2, and Klf4. Control hiPSCs lines were generated by reprogramming of foreskin fibroblasts of a healthy individual.
The reprogramming efficiency was evaluated by immunostaining for the pluripotency marker Tra-1–60 after 21 days. The efficiency ranged between 0.014 and 0.02% (resulting in 15.5 ± 2.1 colonies/105cells). This efficiency is higher than previously reported for the reprogramming of healthy fibroblasts with the retroviral or lentiviral three-factor approaches.19 This improved efficiency can be explained by the supplementation of the chromatin modifier valproic acid, previously shown to enhance hiPSC reprogramming.20
For each patient, ten HF-hiPSC clones were generated, two of which (HF2-5, HF2-8, HF3-2, and HF3-3) were continuously propagated, characterized, and used for cardiomyocyte differentiation. All hiPSC clones exhibited the characteristic hESC morphology, expressing the pluripotent markers NANOG, OCT4, Tra-I-60, Tra-I-81, and SSEA4 (Figure 1A, Supplementary material online, Figure S1A), and maintained a stable karyotype. Quantitative real-time PCR analysis revealed the re-expression of the endogenous pluripotency genes OCT4, NANOG, and SOX2 in the HF-hiPSCs, at similar levels to control hiPSCs (Figure 1B). Next, bisulfite genome sequencing demonstrated that the NANOG promoter was hypomethylated in both HF and control hiPSCs, in contrast to the hypermethylated state in the parental fibroblasts (Figure 1C, Supplementary material online, Figure S2A). Finally, injection of control and HF-hiPSCs to immunodeficient SCID-beige mice led to the formation of teratomas, containing advanced tissue derivatives of all the three germ layers (Figure 2, Supplementary material online, Figure S2B). Thus, our data show that HF-hiPSCs are properly reprogrammed and pluripotent.
Heart failure-human-induced pluripotent stem cells derivation and characterization. (A) Phase-contrast and immunostainings of healthy control, HF2-8, and HF3-3 human-induced pluripotent stem cells colonies for the pluripotency markers: OCT4, NANOG, SSEA4, TRA-1-60, and TRA1-81. (scale bars: 20 µm). (B) Quantitative real-time PCR showing reactivation of the endogenous OCT4, SOX2, and KLF4 genes in control and heart failure human-induced pluripotent stem cell clones (HF2-5, HF2-8, HF3-2, and HF3-3) in comparison with the parental fibroblasts. Values were normalized to β-tubulin and presented as mean ± SEM. Expression values are relative to the corresponding dermal fibroblasts. (C) Bisulfite sequencing of the NANOG promoter in control-human-induced pluripotent stem cell, HF2-8-hiPSCs, HF3-3-hiPSCs, and parental fibroblasts. Open and closed circles indicate unmethylated and methylated CpGs, respectively. (D) H&E staining of teratomas formed following injection of control and heart failure human-induced pluripotent stem cells into SCID-beige mice. Note the presence of pigmented epithelium (ectoderm), gastrointestinal epithelium (endoderm), and hyaline cartilage (mesoderm) (scale bars: 200 µm).
Cardiomyocyte differentiation of the human-induced pluripotent stem cell-cardiomyocytes. (A) Immuostaining of control and heart failure-human-induced pluripotent stem cell-cardiomyocytes for sarcomeric-α-actinin, cTnI, and cTnT (scale bars: 50 µm). (B) Semi-quantitative RT–PCR studies revealing the expression of NKX2-5, MYH-6, MYH-7, MLC-2v, TNNI, and MYL7 by heart failure and control human-induced pluripotent stem cell-cardiomyocytes. Undifferentiated human-induced pluripotent stem cells, in contrast, robustly expressed OCT4 and NANOG. (C) Comparison of cardiomyocyte differentiation of control and heart failure human-induced pluripotent stem cells. Shown are average number of beating EBs/plate from two clones of each patient (HF2-5, n= 4; HF2-8 n= 8; HF3-2, n= 5; HF3-3 n= 6) and healthy control human-induced pluripotent stem cells (n= 4). (D) Real-time quantitative PCR showing comparable expression levels of cardiogenic markers (MYH6, MYH7, TNNI, ACTC, GATA4) in heart failure and control human-induced pluripotent stem cell clones.
The two HF-hiPSC lines were coaxed to differentiate into cardiomyocytes using the EB differentiation system (Supplementary material online, Movie S1). Immunostaining confirmed the cardiac phenotype of enzymatically dispersed cells derived from these contracting EBs. As depicted in Figure 2A, both control and HF-hiPSC-CMs demonstrated a typical striated pattern following staining for the sarcomeric proteins cTnI, cTnT, and α-actinin. Similarly, the gene expression analysis of differentiating hiPSC-CMs (Figure 2B) confirmed the expression of cardiac-specific transcription factors (NKX2–5) and structural genes including cTnI (TNNI), α and β myosin heavy chains (MYH6, MYH7), ventricular myosin light chain (MLC-2V), and myosin light chain 2 atrial isoform (MYL7). This cardiac gene-expression profile was similar in both healthy and HF-hiPSC-CMs but was undetectable in undifferentiated hiPSCs, where the expression of pluripotency genes (OCT4 and NANOG) was dominant.
We next sought to compare the cardiac differentiation capacity of the HF and control hiPSCs. As appreciated in Figure 2C, the average number of beating EBs/plate after 30 days of differentiation was not significantly different between all HF-hiPSC clones and control hiPSCs (P= 0.673). To strengthen these results, we used quantitative real-time PCR for selected cardiac genes (GATA4, MYH6, MYH7, TNNI, and ACTC) as an additional method to compare the cardiomyocyte yield. These studies also revealed comparable differentiation efficiencies between controls and HF-hiPSC clones (Figure 2D).
Functional characterization of heart failure-human-induced pluripotent stem cell-cardiomyocytes
A multi-electrode array (MEA) mapping system was used to study the electrophysiological properties of the HF-hiPSC-CMs. Microdissected hiPSC-derived cardiac tissues were cultured on the MEA plates (Figure 3A). The extracellular potentials recorded from electrodes underlying these beating EBs (Figure 3B) were used to generate detailed activation maps (Figure 3C). These maps revealed the development of a functional cardiac syncytium, with the activation wavefront propagating from the pacemaker area (red) to the latest (dark blue) activation sites (Figure 3C). When evaluating beating frequencies, we found no significant differences between the different HF and control hiPSC-CMs (n= 10, P= 0.28; Figure 3D).
Multi-electrode array recordings. (A and B) Extracellular recordings (B) from the heart failure-human-induced pluripotent stem cell-derived cardiac tissue cultured on the MEA plate (A). (C) The resulting activation map. (D) Comparison of beating rates between control and heart failure human-induced pluripotent stem cell-cardiomyocytes.
We next assessed the response of HF-hiPSC-CMs to neurohumoral stimulation. Administration of the β-agonist isoproterenol (1 µM) led to a comparable increase in the beating frequency of control, HF2-8, and HF3-3 hiPSC-CMs (by 37 ± 10, 48 ± 6, and 55 ± 8%, respectively; n= 5; Figure 4A). Similarly, application of the muscarinic agonist carbamylcholine (1 µM) led to comparable negative chronotropic responses (by 26 ± 4, 45 ± 7, and 32 ± 11%, respectively; n= 5; Figure 4B). The chronotropic changes induced by isoproterenol and carbamylcholine (Figure 4) were statistically significant for each of the hiPSC lines studied but the magnitudes of the effects did not differ between control and HF-hiPSC-CMs (P= 0.63 for isoproterenol; P= 0.26 for carbamylcholine).
Pharmacological studies. Positive (A) and negative (B) chronotropic responses induced in the control and heart failure-human-induced pluripotent stem cell-cardiomyocytes by adrenergic (isoproterenol, A) and muscarinic (carbamylcholine, B) stimulation respectively. *P< 0.05 when compared with baseline values.
Generation of transgene-free heart failure-human-induced pluripotent stem cell
To derive transgene-free HF-hiPSCs, we initially reprogrammed the HF3 fibroblasts with an excisable single polycistronic lentiviral vector (STEMCCA) encoding all four factors (Oct4, Sox2, Klf4, and c-Myc).17 Fifteen clones were isolated and subjected to the Southern blot analysis to determine the number of genomic integration sites of the stem cell cassette. Eight of these clones contained a single integration site, of which one (HFS-5-hiPSC) was continuously propagated.
The HFS-5-hiPSC line expressed the pluripotent markers NANOG, OCT4, Tra-I-60, Tra-I-81, and SSEA4 (Figure 5A) and maintained a normal karyotype. Next, we aimed to remove the STEMCCA cassette in order to generate HF-hiPSCs free of the integrated transgenes. Since the STEMCCA cassette was designed to contain lox-P sites flanking the reprogramming transgenes, we attempted to excise the integrated vector by expression of Cre-recombinase. To this end, the HFS-5-hiPSCs were transiently transfected with an expression vector encoding Cre-recombinase and EGFP (pCAG-Cre-EGFP). Cre-recombinase expressing cells (identified by eGFP) were then selected from a single-cell suspension by FACS sorting 72 h after electroporation. Selected cells were re-plated at low density and the generated clones were isolated 14 days later. Successful excision of the STEMCCA cassette was confirmed by PCR analysis of genomic DNA using primers targeting the WPRE sequence of the lentiviral vector (Figure 5B).
Generation of transgene-free heart failure-human-induced pluripotent stem cell. (A) Phase-contrast and immunostaining of HFS-5-hiPSC colonies for the pluripotency markers OCT4, NANOG, SSEA4, TRA-1-60, and TRA1-81 (scale bars: 50 µm). (B) Genomic PCR (using primer set flanking the WPRE region of the vector) verifying the excision of the lentiviral vector, originally present in the HFS-5-hiPSC clones. Note that clone Cre-HFS-5-hiPSCs showed no detectable WPRE in the genomic DNA. HF3-hiPSC and HF2-hiPSC lines, generated by retroviral infection were used as negative controls. (C) Semi-quantitative RT–PCR studies demonstrating the expression of pluripotent markers (OCT4 and NANOG) in undifferentiated cells and the expression of cardiac-specific markers (NKX2–5, MYH-6, MYH-7, MLC-2v, TNNI, MYL7) by the differentiating EBs. (D) Positive staining of HFS-5-hiPSC-CMs for α-actinin and cTnT (scale bars: 50 µm). (E) Real-time quantitative PCR showing comparable expression levels of cardiogenic genes (MYH6, MYH7, TNNI, ACTC) during the differentiation of the transgene-free human-induced pluripotent stem cells derived cardiomyocytes (Cre-HFS-5-hiPSC-CMs) and the transgene-containing clone derived cardiomyocytes (HFS-5-hiPSC-CMs).
In a similar manner to the HF-hiPSCs derived by retroviral transduction, the HFS-5-hiPSC line could be differentiated into cardiomyocytes using the EB differentiation method. Semi-quantitative RT–PCR studies (Figure 5C) revealed the expression of cardiac-specific genes by the differentiating EBs. Similarly, immunostaining analysis for cTnT and of sarcomeric-α-actinin confirmed the cardiac identity of the differentiating myocytes (Figure 5D).
Finally, we also compared the cardiomyocyte differentiation capacity of the generated transgene-free hiPSCs with that of the transgene-containing HF-hiPSC clones (prior to Cre-excision). Comparable differentiation potentials were noted in both clones as demonstrated by the similar quantitative expression of cardiac-specific genes (MYH6, MYH7, TNNI, and ACTC) by the differentiating cells (Figure 5E).
In vitro integration with pre-existing cardiomyocyte cultures
The HF-hiPSC-derived cardiac tissues were co-cultured with pre-existing cardiac tissue (NRVMs) (Figure 6A). Within 24–48 h, we could already detect synchronous mechanical activity (Supplementary material online, Movie S2) between the human and rat cardiomyocytes (n= 12). We next utilized the MEA mapping technique to evaluate the functional interactions within the co-cultures. By recording extracellular field potentials (Figure 6B) simultaneously from all electrodes, we were able to create activation maps that showed the development of synchronized activity in all hybrid cultures. Note in the example in Figure 6C that electrical activation initiated within the rat tissue (red) and then propagated to the rest of the co-culture, activating also the HF-hiPSC-CMs. The electrophysiological integration observed was continuous. Hence, electrograms recorded simultaneously from the human (blue electrode) and rat (red electrode) tissues showed synchronized activity and tight temporal coupling (Figure 6D) over several beats.
Functional and structural integration in co-cultures. (A) Co-culturing of the heart failure-human-induced pluripotent stem cell-derived cardiac tissue (arrows) with the NRVMs. (B and C) MEA recordings (B) and the resulting colour-coded activation map (C) showing the development of functional integration, with electrical activity originating (red) in the rat tissue and propagating to the human cardiomyocytes. (D) Extracellular potentials, recorded from two electrodes underlying the heart failure-human-induced pluripotent stem cell-cardiomyocytes (red) and NRVMs (blue), showing synchronized activity. (E) Structural integration between the human-induced pluripotent stem cell-cardiomyocytes and NRVMs. Left panel: DAPI and DIC staining showing the hiPSC-derived cardiac tissue (EB) in the centre and NRVMs in the periphery. Right panel: High-magnification Cx43 immunostaining (red punctuate staining) image of the co-culture. Note the development of gap junctions (yellow arrows) at the interphase between the human-induced pluripotent stem cell-cardiomyocytes and NRVMs. (F) ΔLAT (left) and CL (right) plots showing long-term synchronized activity between the heart failure-human-induced pluripotent stem cell-cardiomyocytes and NVRMs at 1 and 4 days.
Next, immunostaining studies targeting the major gap junction protein Cx43 were performed in the co-cultures. As depicted in Figure 6E, a positive punctuate Cx43 immunosignal (suggesting the development of gap junctions, yellow arrows) was identified at the border between the plated EB (containing the HF-hiPSC-CMs) and the NRVMs.
Finally, to evaluate the long-term coupling within the co-cultures, graphs were created, sequentially plotting the cycle lengths (CLs) of the electrical activity in the HF-hiPSC-derived cardiac tissue vs. that in the NRVMs. Similarly, the activation time difference between the two tissue types (ΔLAT) was measured and plotted for each beat. As can be seen in Figure 6F, the long-term electrical synchronization in the co-cultures was reflected by the establishment of fixed ΔLAT between the EBs and NRVMs (Figure 6E, left) and by convergence of their CL plots (linear correlation, Figure 6E right). This coupling persisted for several days in culture.
In vivo tranplantation of the heart failure-human-induced pluripotent stem cell-cardiomyocytes
We next evaluated the ability of the HF-hiPSC-CMs to engraft in the in vivo heart. Beating EBs were mechanically dissected, enzymatically dispersed into small clusters, and transplanted into the LV myocardium of healthy rat hearts. As can be viewed in Figure 7, the engrafted HF-hiPSC-CMs (pre-labelled with Qtracker, red) were identified within the rat myocardium in all cases, their human origin verified by immunostaining with anti-human mitochondrial antibodies (Figure 7A), and their cardiomyocyte identity confirmed by the positive immunostaining for sarcomeric-α-actinin (Figure 7B). In a similar manner to previous hESC-CMs transplantation studies,6 the hiPSC-CMs at 7–10 days following engraftment were smaller in the size than host rat cardiomyocytes, were arranged isotropically, and still displayed an early-striated staining pattern. Finally, to evaluate for structural coupling between host and donor cells, immunostainings were performed for Cx43 (Figure 7C). The resulting punctuate Cx43-immunosignal (white), identified at the interphase between the HF-hiPSC-CMs and rat cardiomyocytes (red arrows), suggests the formation of gap junctions between host and donor cells.
In vivo transplantation of the heart failure-human-induced pluripotent stem cell-cardiomyocytes. (A and B) The grafted heart failure-human-induced pluripotent stem cell-cardiomyocytes (pre-labelled with Qtracker, red) were identified within the rat myocardium, their human origin verified by immunostaining with anti-human mitochondrial antibodies (green, A), and their cardiomyocyte phenotype confirmed by the positive immunostaining for sarcomeric-α-actinin (green, B). Note the early striated pattern of the engrafted human-induced pluripotent stem cell-cardiomyocytes (B). Scale bars (A—100 µm, B—50 µm). (C) Development of gap junctions between donor human-induced pluripotent stem cell-cardiomyocytes and host rat cardiomyocytes. Left panel: Immunostaining for sarcomeric-α-actinin (green) identifying the rat host cardiomyocytes (in cross section) as larger and structurally more mature cells (R); when compared with the engrafted human-induced pluripotent stem cell-cardiomyocytes (H), which display an early-striated staining pattern. Note that several of the engrafted cells are labelled with Qtracker (red). Right panel: Immunostaining of the same specimen for Cx43. Note the positive punctuate Cx43 immunosignal (white) between the transplanted human-induced pluripotent stem cell-cardiomyocytes themselves and also at the interface between donor and host cardiomyocytes (red arrows) (scale bars: 50 µm).
Myocardial cell replacement therapies have emerged as novel therapeutic paradigms for myocardial repair but have been hampered by the paucity of sources for human cardiomyocytes, by the lack of direct evidence for functional integration between donor and host cells, and by the anticipated immune rejection associated with allogeneic cell transplantation. In this study, we aimed to target these obstacles by establishing and coaxing the cardiomyocyte differentiation of hiPSCs from patients with ischaemic cardiomyopathy.
Our results show: (i) that hiPSCs can be established from dermal fibroblasts of patients with advanced HF using a three-factor reprogramming approach that does not include cMyc; (ii) that similar transgene-free HF-hiPSCs can be derived using an excisable polycistronic lentiviral vector; (iii) that the HF-hiPSCs can be differentiated into cardiomyocytes with appropriate molecular, structural, and functional properties; (iv) that the reprogramming efficiency, hiPSC cardiomyocyte differentiation capacity, and phenotype properties of the generated hiPSC-CMs are comparable when using fibroblasts from HF patients or healthy foreskin fibroblasts; (v) that the HF-hiPSC-CMs can functionally integrate with pre-existing cardiac tissue in a co-culture model; and (vi) that the HF-hiPSC-CMs can engraft, survive, and integrate structurally with host cardiac tissue following in vivo transplantation.
Differentiation, once believed to be a one-way process, was recently shown to be a dynamic process that can be reversed by transduction of stemness transcription factors into somatic cells.9 The reprogrammed cells, also known as iPSCs, highly resemble ESCs in terms of morphology, gene expression, pluripotency, and epigenetic status. The added value of the iPSCs lies in the unprecedented opportunity to generate patient-specific pluripotent stem cells from adult individuals. In this study, we contributed to this rapidly developing field by showing, for the first time, the ability to establish hiPSCs from ischaemic cardiomyopathy patients with advanced HF, who represent the typical candidates for future autologous myocardial cell replacement procedures. We then demonstrated that the patient-specific hiPSC lines fulfil the criteria defining fully reprogrammed pluripotent hiPSCs and that they can differentiate into bone fide cardiomyocytes.
We next compared the HF-hiPSC-CMs to cardiomyocytes derived from healthy control hiPSCs and found comparable properties in terms of cardiomyocyte differentiation efficiency, beating rates, up-regulation of cardiac genes and down-regulation of pluripotency genes, structural organization, electrical activity, and response to neurohormonal triggers. Moreover, the differentiation system in both hiPSC types was not limited to the generation of isolated cardiomyocytes, but rather a functional cardiac syncytium was generated.
The ‘healthy’ cardiogenic phenotype observed in the HF-hiPSC-CMs is not surprising given that the disease process in these HF patients was acquired. Hence, the patients studied did not have any abnormal genetic background that could impact the properties of the reprogrammed fibroblasts or their differentiated cardiomyocyte progeny. This is in contrast to recent hiPSC studies in inherited monogenic cardiac disorders (such as long QT syndrome)14–16 in which hiPSC-CMs were established from dermal fibroblasts carrying the mutation, consequentially leading to a disease-specific abnormal phenotype.
Another important finding of the current study is the ability to produce hiPSCs and to differentiate them into cardiomyocytes without c-Myc. One of the major hurdles that may hamper the introduction of the hiPSC technology into the clinic is the potential risk for tumorgenicity.21 This concern stems from the initial requirement for the use of oncogenic transcription factors (such as cMyc) for reprogramming, the random transgene insertion into the host's genome, and the use of viral vectors. Okita et al.,22 for example, reported that the chimeras and progeny derived from mouse iPSCs frequently showed tumour formation; and that in these tumours, retroviral expression of cMyc was reactivated. In contrast, when cMyc-free-iPSC lines were used, no tumours were found in the offspring.
Beyond the use of a cMyc-free reprogramming approach, several strategies are being developed to generate transgene-free hiPSCs, which are further expected to decrease the tumorogenic risk as well as potential alterations in hiPSC characteristics because of residual transgene expression. Such strategies can be grouped into: (i) methods that use non-integrating DNA vectors such as adenoviruses,23 conventional plasmids,24 and minicircles; (ii) DNA-free methods such as by the administration of synthetic-modified mRNA25 or protein delivery;26 and (iii) methods that use integrating but excisable systems in which the transgenes are removed from the genome through the piggyBac transposon/transposase system27 or by the Cre/loxP approach.17
In the current study, we used the latter strategy to derive transgene-free hiPSCs from the HF patients. Our protocol consisted of reprogramming the HF fibroblasts into hiPSCs with an excisable single polycistronic lentiviral vector (STEMCA) that contained all four factors, screening to select the hiPSC clones that possess a single integration site, and subsequent excision of the vector by transient introduction of Cre-recombinase.
Another important finding of this study is the ability of the HF-hiPSC-CMs to integrate electrically and mechanically with pre-existing cardiac tissue. Numerous reports suggest that cell transplantation can improve the cardiac performance in animal models of myocardial infarction.1,2 However, it is not entirely clear whether this functional improvement is due to the direct contribution to contractility by the transplanted myocytes or by alternative indirect mechanisms such as attenuation of the remodelling process, amplification of an endogenous repair process, or induction of angiogenesis. True systolic augmentation would require structural, electrophysiological, and mechanical coupling of donor and host tissue so that the transplanted cell graft would participate actively in the synchronous contraction of the ventricle.
The observations that HF-hiPSC-CMs could couple with host cardiomyocytes and generate a single functional syncytium in vitro and to form a gap junction with host cardiomyocytes in vivo are in agreement with previous studies showing the ability of cardiomyocytes derived from embryonic hearts or ESCs to couple both in vitro and in vivo with host cardiomyocytes.18,28–30 This capability may be crucial for future use of these cells not only for the treatment of HF but also for conduction system repair (biological pacemaker approach).18
Nevertheless, long-term application of the hiPSC-CMs rely more on the prospect of in vivo use. In this study, we demonstrated that HF-hiPSC-CMs can survive following transplantation, form stable cell grafts, and establish electromechanical junctions with host cardiomyocytes in the healthy rat heart. Future studies will have to evaluate whether similar engraftment could also occur in small and large animal models of cardiac injury and whether such an engraftment could lead to a functional benefit. To conduct such studies, however, more efficient cardiomyocyte differentiation systems and scaling up procedures should be developed (as already reported for hESCs)7,31,32 to derive a clinically relevant number of purified cardiomyocytes.
This work was supported in part by the Israel Science Foundation (1449/10, Mina and Oto Shpirman fund); by the European Research Council (ERC-2010-StG-260830-Cardio-iPS); and by the J&J-Technion research grant; and by the Nancy and Stephen Grand philanthropic fund.
Conflict of interest: none declared.
We thank Dr Sara Selig and Dr Annie Rebibo-Sabbah for their help in bisulfite sequencing and southern blot analyses. We thank Dr Edith Suss-Toby for her assistance with confocal imaging and Dr Ofer Shenker and Yaakov Sakoury for their assistance with flow cytometry.