Engineered heart tissue graft derived from somatic cell nuclear transferred embryonic stem cells improve myocardial performance in infarcted rat heart.

The concept of regenerating diseased myocardium by implanting engineered heart tissue (EHT) is intriguing. Yet it was limited by immune rejection and difficulties to be generated at a size with contractile properties. Somatic cell nuclear transfer is proposed as a practical strategy for generating autologous histocompatible stem (nuclear transferred embryonic stem [NT-ES]) cells to treat diseases. Nevertheless, it is controversial as NT-ES cells may pose risks in their therapeutic application. EHT from NT-ES cell-derived cardiomyocytes was generated through a series of improved techniques in a self-made mould to keep the EHTs from contraction and provide static stretch simultaneously. After 7 days of static and mechanical stretching, respectively, the EHTs were implanted to the infarcted rat heart. Four weeks after transplantation, the suitability of EHT in heart muscle repair after myocardial infarction was evaluated by histological examination, echocardiography and multielectrode array measurement. The results showed that large (thickness/diameter, 2-4 mm/10 mm) spontaneously contracting EHTs was generated successfully. The EHTs, which were derived from NT-ES cells, inte grated and electrically coupled to host myocardium and exerted beneficial effects on the left ventricular function of infarcted rat heart. No teratoma formation was observed in the rat heart implanted with EHTs for 4 weeks. NT-ES cells can be used as a source of seeding cells for cardiac tissue engineering. Large contractile EHT grafts can be constructed in vitro with the ability to survive after implantation and improve myocardial performance of infarcted rat hearts.

Introduction

Myocardial infarction (MI) leads to cardiomyocyte loss, ventricular remodelling and consequent impairment of myocardial function. Current therapeutic approaches have limited effects in attenuating disease progression. The only successful treatment is heart transplantation. Yet it is used for late-stage patients only and constrained by the shortage of organ supply [1-3]. Recent advancements in stem cell biology, cell therapy and tissue engineering have paved the way for establishing a novel discipline in regenerative medicine. The concept of developing cell-replacement strategies has gained increasing attention [4-7]. Such strategies aim to repopulate scar tissue with new contractile cells in an attempt to restore function in the failing heart.

Two principal strategies were proposed for myocardial repairs: the cell therapy and the tissue-engineering approach. The former focuses on the direct cell transplantation into the dysfunctional myocardial areas [4-7]. The latter attempts to generate a tissue-engineered muscle construct followed by in vivo engraftment of the engineered tissue [8, 9]. Despite progress in the former strategy, problems such as the low rate of retention and survival still exist [10, 11]. Notwithstanding initial donor cell retention and survival following in vivo engraftment can be improved by using scaffolds, it still falls short for the injected cells to regenerate cardiac tissues to fit the shape of the infarcted area (especially when the distribution of injected cells were uneven) [12–15, 16]. However, it may theoretically allow complete replacement of diseased myocardium or reconstitution of cardiac malformations to implant engineered heart tissue (EHT) [16]. In recent years, progress is made in generating EHT. Among the studies in heart tissue engineering, various sources of seeding cells were used to construct EHT, including terminal differentiated, mesenchymal stem (MSCs) and embryonic stem (ES) cells [7, 8, 16–18]. ES cells have many advantages in direct differentiation into cardiomyocytes [19], capable of integrating with the host heart and improving electrical conduction [20, 21]. Therefore in theory, ES cells could potentially provide an unlimited supply of cardiomyocytes for cell therapy to regenerate functional myocardium. However, the immunological rejection after the ES cells transplantation makes the application of cell-replacement strategies difficult.

The immunological rejection in cell-replacement strategies can be avoided by using patient-specific cells derived from the induced pluripotent stem (iPS) cells [22] or the nuclear transferred embryonic stem (NT-ES) cells [23]. Although iPS cells can shun ethical controversies, the virus integration and oncogenes application hamper further clinical application. Therapeutic cloning has been successfully performed in mice and non-human primates, whereby somatic cell nuclear transfer is used to generate customized ES cells from differentiated somatic cells of specific individuals [23, 24]. In theory, the NT-ES cells carried the same genome as the donor somatic cells. After directed induction, the differentiated cells could rescue the damaged tissues without immune rejection [25]. However, the persistence of abnormalities in cloned animals has doubted whether ES cells derived by somatic cell nuclear transfer may pose risks in their therapeutic application. [26]. Although it has been demonstrated that the transcriptional profiles and developmental potentials of ES cells derived from cloned blastocysts are identical to those of ES cells derived from fertilized blastocysts (F-ES) [27], there is no report focusing on their ability in engineered tissue construction and damaged tissue repairs up to date. The results from the studies shall provide more sufficient support to the notion that the ES cells derived from cloned blastocysts have a strong therapeutic potential.

In this study, we produced engineered spontaneously contracting heart tissue constructs in vitro from the mouse NT-ES cell-derived cardiomyocytes. After 7 days of static and mechanical stretching, respectively, the EHTs were implanted to the infarcted rat heart. Four weeks after transplantation, the suitability of EHT in heart muscle repairs after MI was evaluated by histological examination, echocardiography and multielectrode array (MEA) measurement.

Materials and methods

Mouse NT-ES cell culture

NT-1 ES cell line was used in the following experiments of this study. Derivation of NT-ES cell lines from cloned blastocysts and pluripotency analysis of NT-ES cells were described in our recent studies, which showed that the NT-ES cell line – NT-1 (proved to be pluripotent) could differentiate into cardiomyocytes in vitro [28]. In the present study, NT-1 ES cells (about 20 passages) were used. (Details of NT-ES cell culture medium and differentiation medium can be found in the Supporting Information.)

NT-EB production in slow-turning lateral vessel (STLV) rotating bioreactor

Nuclear transferred embryoid body (NT-EB) in STLV was prepared by a method described previously by our laboratory [28]. In brief, the NT-ES cells were dissociated with 0.25% trypsin (Sigma-Aldrich, St. Louis, MO, USA) + 0.04% ethylenediaminetetraacetic acid (Sigma-Aldrich) at 37°C for 1 to 2 min., and 0.4–1 × 105 cells/ml in culture medium without leukaemia inhibitory factor were transferred into a 250 ml STLV (Synthecon, Inc., Houston, TX, USA) for cell expansion and EB formation. The rotating speed was 10 rpm for the first 12 hrs and was then gradually adjusted to 20 rpm. A total of 50% of the medium was replaced by fresh medium daily. After 3 days in the STLV, 1% ascorbic acid was added into the medium. Two days later, NT-EBs were transferred onto gelatine-coated plates (one to three NT-EBs/cm2). After cultivation in differentiation medium for additional 14 days, the cultures were then examined for the presence of beating cells and subjected to analysis of cardiomyocytes by immunostaining of cardiac troponin T (cTnT; for details, see Supporting Information [Table S1, Fig. S1]).

Percoll enrichment of cardiomyocytes

Differentiated NT-ES cell clusters containing beating cells were dissociated, resuspended in differentiation medium and loaded onto a Percoll (Amersham, Uppsala, Sweden; Pharmacia, Piscataway, NJ, USA) gradient for enrichment of cardiomyocytes as previously described [18, 29]. In brief, Percoll was diluted in a buffer containing 20 mmol/l HEPES and 150 mmol/l NaCl. The gradient consisted of a 40.5% Percoll layer over a layer of 58.5% Percoll. After centrifugation at 1500 × g for 30 min., cell layers could be observed. Cells at different fractions were collected, washed, resuspended in the differentiation medium, plated into chamber slides and cultured for an additional 3 days for immunocytochemical staining (for details, see Supporting Information). A mixture of fractions IV and V was collected for the following experiments of this study (Fig. S2).

EHT construction

The casting moulds were prepared as described previously [30]. A total of 1.0 3 107 freshly isolated cardiomyocytes derived from mouse NT-ES cells were mixed with 1 ml mixture of liquid collagen type I prepared from rat tails, a basement membrane protein mixture (Matrigel; Becton Dickinson Biosciences, San Jose, CA, USA), and concentrated serum-containing culture medium (2× Dulbecco’s modified essential medium [DMEM], 20% foetal bovine serum [FBS], 200 U/ml penicillin and 200 μg/ml streptomycin); pH was neutralized by titration with 0.1 M NaOH. The reconstitution mix was pipetted into casting moulds and incubated for 45 to 60 min. at 37°C and 5% CO2 to allow hardening of the reconstitution mixture [31]. Thereafter, 1.5 ml serum-containing culture medium (DMEM, 10% FBS, 100 U/ml penicillin and 100 mg/ml streptomycin) was added to each dish. After 7 days in culture, EHTs were transferred into a modified stretch device and hooked on the working rods for unidirectional stretch (10%, 2 Hz) for an additional 7 days. Culture medium was changed 12 hrs after EHT casting and then every other day while the culture was maintained in casting moulds. Transferred into the stretch device, the culture medium was changed daily.

In vitro evaluation of EHTs

Seven days after culturing and static stretching in casting moulds, the EHTs were strong enough to allow for mechanical stretching. After additional 7 days, histological sections were prepared. Cardiomyocytes as well as non-cardiomyocytes in the EHTs were examined by immunohistochemical staining. Cell viability in EHTs was assessed by Live/Dead Viability/Cytotoxicity assay (Molecular Probes, Inc., Eugene, OR, USA) (for details, see Supporting Information).

Myocardial infarction and EHT grafting

We generated MIs through ligation of the left coronary artery (6–0 prolene surture) in sodium pentobarbital-anesthetized (1.5%) female Sprague-Dawley rats (250 ± 10 g) as described previously [15, 32] (for details, see the Supporting Information). Animals meeting the echocardiographic inclusion criterion (fractional shortening < 40%) were used in the following experiments. Two weeks after MI, EHTs were sutured (6–0 prolene surture) onto a recipient heart (n = 10). We placed the sutures in healthy myocardium adjacent to the visible infarct scars to arrange the centre of the EHTs just above the latter. The non-infarcted group (n = 10) serving as sham-operation control underwent thoracotomy and cardiac exposure without coronary ligation (suturing without tying the left anterior descending coronary artery), and the untreated MI group (n = 10) serving as negative control was produced by ligation of coronary without implantation of EHT. While, non-contractile grafts (NCG) were produced by formaldehyde (4%) fixation of EHT [16]. Implantation of NCG served as tests for effects of NT-ES derived cardiomyocytes (n = 8). For immunosuppression all rats received immunosuppressants (mg/kg body weight/day cyclosporin A, 5; methylprednisolone, 5) daily by subcutaneous injection.

In vivo evaluation of EHTs

Four weeks after transplantation, the suitability of EHT in heart muscle repairs after MI was evaluated by echocardiography, MEA measurement and histological examination, (for details, see the Supporting Information).

Statistical analysis

Results are reported as mean ± standard deviation. anova and Tukey’s multiple comparison tests were used to determine possible significant differences in the cardiac function between groups. anova with Mann-Whitney U-test was used to determine possible significant differences in the total activation time between groups. Significance was accepted at P< 0.05.

Results

Construction of the EHT

A total of ∼ 8.7 × 107 cells containing more than 49% cTnT+ cardiomyocytes were procured (see the Supporting Information). To improve the size and quality of EHTs, we used a self-made mould during the reconstruction. We have reported previously that the four glass columns standing in the moulds could provide static stretch and prevent the collagen/Matrigel from gradual contraction during the cultivation. Then helped to avoid thickening the construction and facilitate the survival of the cells in the centre of the reconstructed tissues [30]. In this study, the EHTs condensed after 1 day of culture and contracted gradually (Fig. 1A). The EHTs without the static stretching will contract into contact lenses-like structures in 7 days. They were unable to be transferred into the stretch device for mechanical stretching (Fig. 1D). Spontaneously contracting cell clusters could be observed in 4–9 days. The beating clusters increased in size with the time and contracted independently from each other. The coherent contraction of the complete EHT was achieved 7 days after mechanical stretching. Over time, contractions became more regular, and the beating rates could reach 1.1 Hz and even 1.2 Hz at day 10. After 7 days of mechanical stretching, the EHT beat regularly and unidirectionally along the direction of stretching (Fig. 1B). Upon removal from the stretching device, the EHT (diameter, ≍10 mm; thickness, 2–4 mm) beat continuously at ≍1.1 Hz for >10 days (Fig. 1C).

Fig 1

Macroscopy of EHT. (A) The EHTs condensed after 1 days of culture and contracted gradually. (B) Mechanical stretching of EHTs. (C) Microscopic view of spontaneously contracting EHT constructed by seeding cardiomyocytes derived from mouse NT-ES cells into liquid type I collagen. (D) The EHTs without the static stretching will contract into contact lenses-like structures in 7 days. Bars 5 = mm (A, C, D), 1 cm (B).

Sections of paraffin-embedded EHTs revealed that the complex of multicellular aggregates in the EHT patch was formed. The cells in the EHTs had the tendency of spread along the long axis of stretching, although many cells were haphazardly arranged. The cells had oval or round nuclei. The pyknotic nuclei could be found frequently, too. In addition, the cells exhibited less clear cross-striation, scant cytoplasm, high ratio of nuclei and plasma, and less organized distribution and compact overall structure (Fig. 2A). In comparison with mature adult rat myocardium (Fig. 2C), histological features of cells in EHTs resembled those of myocytes within immature neonatal rat more, especially in the high ratio of nuclei and plasma (Fig. 2B) [33]. In order to investigate the cell viability in EHTs, a Live/Dead Viability/Cytotoxicity assay was performed. The results showed that the cell viability of EHTs in vitro was 92 ± 4.2% (see Supporting Information [Fig. S3]). These results illustrate the survival and rearrangement of the NT-ES-derived cells in the EHTs, a prerequisite for cardiac tissue engineering.

Fig 2

Histological and immunohistochemical examination of EHTs in vitro. (A–C) Examination of EHT and natural cardiac muscle. Photomicrographs of haematoxylin–eosinstained paraffin sections from EHTs showed formation of complexes of multicellular aggregates and longitudinally oriented cell bundles (A). In comparison with mature adult rat myocardium (C), the histological morphology of the EHTs resembles that of immature neonatal cardiac tissue, especially in the high ratio of nuclei and plasma (B). (D–K) Immunohistochemical examination of EHTs. The EHTs showed immunoreactivity for a variety of cardiac markers, including the cardiac transcription factors GATA-4 and Nkx2.5, cTnT and a-sarcomeric actinin. GATA-4 and Nkx2.5 immunoreactivity were found in nuclei of a large number of cells in EHTs, indicating the presence of NTES-derived cardiomyocytes (D, E). The cTnT+ cells formed cell strands and interconnected cell bundles in most parts within the EHT patch (F). The α-sarcomeric actinin also expressed in the EHTs. However, the sarcomeric organization was obscure, indicating a low degree of development (G). Connexin43 (protein that expected to align following mechanical conditioning) immunostaining demonstrated the development of gap junction between the cells in EHTs (H). The vWAg+ cells indicated the spontaneous formation of vessel-like structures within the EHT patch (I). Fibroblasts were well distributed in the EHTs (J). Neurons scattered throughout EHTs (K). There was no positive signal in the negative control in all staining. Figure 2L showed the representative negative control image. Bars = 20 μm (F), 50 μm (A–E, G–L).

Cell type identification in EHT

The EHTs showed immunoreactivity for a variety of cardiac markers, including the cardiac transcription factors GATA-4 and Nkx2.5, cTnT and a-sarcomeric actinin. The transcription factors GATA-4 and Nkx2.5 are expressed in pre-cardiac mesoderm and persist in the heart during development. GATA-4 and Nkx2.5 immunoreactivity were found in nuclei of a large number of cells in EHTs, indicating the presence of NTES-derived cardiomyocytes (Fig. 2D, E). The percentages of GATA-4+ nuclei and NKX2.5+ nuclei were determined by dividing the number of GATA-4+ nuclei and NKX2.5+ nuclei by the total number of nuclei counted in each section, respectively. The results showed that the EHTs contained 51.6 ± 6.5% GATA-4+ cells, and 53.2 + 8.7% NKX2.5+ cells. The cTnT1 cells formed cell strands and interconnected cell bundles in most parts within the EHT patch (Fig. 2F). The α-sarcomeric actinin also expressed in the EHTs. However, the sarcomeric organization was obscure, indicating a low degree of development (Fig. 2G). Connexin43 (protein that expected to align following mechanical conditioning) immunostaining demonstrated the development of gap junction between the cells in EHTs (Fig. 2H). Besides myocytes, other types of cells –e.g. endothelial cells (positive for vWAg), fibroblasts (positive for smooth muscle α-actin) and neurons (positive for β-III tubulin) – were also detectable in the EHT patch (Fig. 2I–K). The vWAg+ cells indicated the spontaneous formation of vessel-like structures within the EHT patch (Fig. 3I). Fibroblasts were well distributed in the EHTs (Fig. 2J). Neurons scattered throughout EHTs (Fig. 2K). There was no positive signal in the negative control in all staining. Figure 2L showed the representative negative control image.

Fig 3

Echocardiographic effects of EHT graft on ventricular function. Four weeks after EHT transplantation, there was a marked attenuation of an increase in LVEDD (A) and LVESD (B) of the hearts receiving EHT versus NCG and MI controls (P < 0.01). There was a marked attenuation of the decrease in FS (C) and EF (D) in hearts receiving EHT versus NCG and MI controls (P< 0.01).

Graft effects on ventricular function

To determine the physiological consequences of implanting EHT into infarcted hearts, we compared the functional outcome with EHTs to those observed with sham-operated non-infarcted and untreated infarcted rats. A third control of NCG implantation was performed to specifically test effects of the NT-ES-derived cells in improving the heart function.

Four weeks after EHT transplantation, all three infarction groups showed ventricular dilation by echocardiography. All infarction groups showed increased LVEDD and LVESD compared with the non-infarcted group (P < 0.01). Yet there was a marked attenuation of an increase in LVEDD and LVESD of the hearts receiving EHT versus NCG and MI controls (P, 0.01) (Fig. 3A, B). LVEDD in EHT-engrafted hearts was 8.2± 0.3 mm versus 9.2 + 0.5 mm in NCG, 9.2 ± 0.4 mm in MI rats. LVESD in EHT-engrafted hearts was 5.9 ± 0.3 mm versus 7.8 ± 0.4 mm in NCG, 8.0 ± 0.2 mm in MI rats. There was no difference in LVEDD and LVESD among NCG and MI control groups at 4 weeks (P > 0.05).

Left-ventricular systolic function (measured by fractional shortening and ejection fraction) declined significantly in infarction groups compared with that in the non-infarcted group (P < 0.01). There was a marked attenuation of the decrease in FS and EF in hearts receiving EHT versus NCG and MI controls (P, 0.01) (Fig. 3C, D). FS in EHT-engrafted hearts was 25.9 ± 1.3%versus 15.1 ± 1.9% in NCG, 13.1 ± 0.9% in MI rats. EF in EHT-engrafted hearts was 61.7 ± 2.6%versus 38.1 ± 3.2% in NCG, 34.6 ± 2.0% in MI rats. There was no difference in FS and EF among NCG and MI control groups at 4 weeks (P > 0.05).

Structural integration and integrity of EHT in vivo

Four weeks after grafting, the EHTs could be identified by their appearance and location. Haematoxylin and eosin stained paraffin sections showed that the putative EHT-derived tissue (diameter, 510 ± 26 μm) covered the infarcted myocardium above (Fig. 4A). Immunohistochemistry staining showed the formation of cardiac muscle-like structure of the putative EHTs above the infarcted area (Fig. 4B). The formation of blood vessel like structure in the putative EHT was observed by vWAg immunohistochemistry staining (Fig. 4C). Whether the vessels were of donor origin or had a role in de novo vascularization remains to be investigated. The connexin43± area proved structural intercellular connections inside the putative EHTs and with the putative host native myocardium (Fig. 4D). However, further studies are needed to evaluate whether true electrical coupling occurred.

Fig 4

Histological evaluation of EHT grafts at 4 weeks. (A, H and E) staining showed the formation of thick cardiac muscle on top of the infarct scar. Inset shows the corresponding boxed area magnified threefold. (B) cTnT immunohistochemistry staining. (C) vWAg immunohistochemistry staining. (D) Connexin43 immunohistochemistry staining. Bars 5 0.5 mm (A), 50 mm (B, D), 100 mm (C).

Electrical coupling of EHT grafts to host myocardium

We assessed electrical coupling of EHTs to the host myocardium 4 weeks after engraftment by MEA. The total activation time on the region of MI, right, left and posterior segments of the investigated hearts were assessed. Sham-operated MI rats (n = 6) showed the expected delay of total activation time in MI segments. In contrast, the total activation time was normal in EHT grafted hearts (n = 6), indicating undelayed anterograde coupling of EHTs to the host myocardium (Fig. 5A). Acidification of the perfusion buffer to pH 6.7 led to uncoupling of EHTs from the native myocardium, proving direct but subnormal graft-host coupling (Fig. 5B).

Fig 5

Electrical coupling of EHT grafts to host myocardium. (A) Total activation time on the region of MI, right, left and posterior segments of the investigated hearts. (B) EHT could be uncoupled after acidification of the hearts. * P< 0.05 versus sham-operated MI rats by anova with Mann-Whitney U-test.

Discussion

Cardiac tissue engineering approaches are designed to repair damaged heart tissues through the assembly of cells onto biomaterial scaffolds and their implantation into the areas of lesion [34]. Through this technology, functional EHTs can be manufactured to replace native ones. Despite the progress in the cardiac tissue engineering, challenges still remain for future clinical applications, such as immunological rejection after implantation and to construct EHTs at a size with contractile features supporting failing hearts [16]. In this study, we attempted to tackle both issues. By using the NT-ES cells as the source of seeding cells, the EHTs have the potential to solve immune rejection after transplantation. However, risk concerns of using NT-ES cells must be addressed first. By using the self-made mould, we were able to construct EHTs at a size with contractile features. This study is the first to report in vitro EHT constructs from NT-ES cell-derived cardiomyocytes. The data showed that EHTs derived from NT-ES cells integrated, electrically coupled to host myocardium, and exerted beneficial effects on left ventricular function of infarcted rat heart upon transplanting EHTs. They demonstrated that NT-ES cells could be used as a source of seeding cells for cardiac tissue engineering. Large contractile EHT grafts can be constructed in vitro with the ability to survive after implantation and improve myocardial performance of infarcted rat hearts. Our approaches to construct EHTs had the potential for the therapy of heart regeneration.

A qualified EHT construction requires careful calculation and optimization at each step of cardiac tissue engineering, especially in the preparation of seeding cells. We demonstrated the derivation of the EHTs through improved techniques to produce NT-EBs, cardiogenic differentiation and enrichment of cardiomyocytes from differentiated cell mixtures in large amounts. In our previous study, we demonstrated that the NT-EBs formed in STLV bioreactor were more uniform in size [28]. In addition, the STLV-produced NT-EBs differentiated into cardiomyocytes more efficiently (Online Supporting Information). These results suggested that STLV bioreactor provided a more ideal culture condition by facilitating the formation of better quality NT-EBs.

Purification of NT-ES-derived cardiomyocytes is vital in using these cells in tissue engineering or transplantation. This is because there is a mixture of non-cardiac cell types in the cells dissociated from NT-EBs. In our previous works, Percoll enrichment method proved successful in enriching cardiomyocytes from a mouse ES cell line to compare with neonatal rat cardiomyocytes [29] and to construct circular engineered cardiac tissues [18]. This study also evidenced the effectiveness of this strategy. After Percoll enrichment, about 8.7 × 107 cells, more than 49% of these cells expressed cTnT (Supporting Information), were collected. With no genetic alteration, the cardiomyocytes (derived from NT-ES cells) enriched from this system shall be more promising in tissue engineering applications.

EHT patches need to be engineered at a size with contractile features supporting failing hearts in order to repair MIs or correct heart defects. In the previous studies, the size, thickness and shape of constructed EHTs were unsatisfactory because of insufficient metabolic supply. Zimmermann et al. stacked five circular EHTs crosswise on a custom-made holding device after 7 days in culture and circular EHTs fused and formed synchronously contracting multiloop EHTs (diameter, 15 mm; thickness, 1–4 mm) [16] to increase graft size and facilitate transplantation. Albeit a bold attempt to construct EHTs at a size, the method was too complicated. In this study, we used our own mould to avoid the contraction of EHTs during culture and then enhance the exchange of nutrition. The mould could provide static stretch to the EHTs simultaneously. The static stretch may also exert positive effect on the EHTs before the mechanical stretch. Under these conditions, the EHT (diameter, <10 mm; thickness, 2–4 mm) beat continuously at ≍1.1 Hz for >10 days upon the removal of the stretching device (Fig. 1C). This study made bold moves to report the addition of static stretch in the construction of EHTs.

Engraftment and functional integration of EHTs in the infarcted heart is critical for successful regenerative therapy. In this study, we took care to study functional consequences of EHTs transplantation by histological examination, echocardiography and MEA measurement. The results demonstrated that EHTs integrated and electrically coupled to host myocardium and exerted beneficial effects on the left ventricular function of infarcted rat heart. Four weeks after grafting, the connexin43+ area demonstrated structural intercellular connections inside the EHTs and with the host native myocardium (Fig. 4D). A possible explanation for the structural integration suggested that the procedures of transplantation damage EHTs and host myocardium and thereby destroy physiological barriers [16]. NCG seemed to exert some beneficial effects on ventricular ejection fraction, possibly resulting from ventricle-stabilizing effects. A marked attenuation of the increase in LVEDD/LVESD and decrease in FS/EF in hearts receiving EHT versus NCG and MI controls (P< 0.01) (Fig. 3) suggested that the EHT engraftment improves the heart function in lieu of scar stabilization. The results form MEA measurement indicated the coupling of EHTs to the host myocardium (Fig. 5). The results demonstrated that NT-ES cells could be used as a source of seeding cells for cardiac tissue engineering. Our approach to construct EHTs has the potential for the therapy of heart regeneration.

The vascularization of the construct remains a major challenge for future studies of myocardial tissue engineering. In our study, liquid collagen was used as a scaffold. In the EHTs, the gelled collagen still allows for mass transfer and medium diffusion through the construct. This was further improved during the mechanical stretch. Additionally, the EHTs not only contained cardiomyocytes but represented organoids because of the presence of non-cardiac cell types –e.g. endothelial cells, fibroblasts, neurons (Fig. 2). The presence of vascular endothelial cells and fibroblasts in the construct may engender vasculature formation [35]. These factors may constitute the major (if not the sole) reason that we could generate EHTs with a thickness up to 4 mm. It is intriguing that formation of blood vessel-like structure in the EHT was observed 4 weeks after transplantation (Fig. 4C). Whether the vessels were of donor origin or had a role in de novo vascularization remains to be investigated.

The study has limitations. First, mouse EHTs were xenotransplanted into rat myocardium because rat MI model was more appropriate compared with mouse model because (1) the relatively large heart would facilitate the transplantation and observation of EHTs and (2) the surgery was relatively easier to increase the survival chances of animals. No obvious immune rejection was observed 4 weeks after EHTs transplantation. Nonetheless, allotransplantation or autologous transplantation will be more useful practically for therapeutic cloning purpose. Second, the tumour formation was not assessed systematically after EHT transplantation. Because NT-ES cells are pluripotent stem cells, the possibility of inducing tumour formation has to be evaluated when the cells are transplanted in vivo. In this study, no tumour formation was observed in the histopathological sections of infarcted myocardium 4 weeks after EHTs transplantation. Exhaustive long-term investigations remain to be studied to ensure the safety of potential therapy for EHT transplantation. Third, in vivo visualization and quantification of the engrafted tissue were not performed in animal study to investigate the integration and cell migration or fusion. In this study, in order to determine the therapeutic effect of EHTs derived from NT-ES cells, no NT-ES cells were genetically labelled for tracking. EHTs could only be identified by their appearance and location in this study. Without the markers to identify graft cells, it is difficult to identify the graft from residual host muscle. So these structures could only be identified as the putative graft or host. More appropriate studies need to be undertaken for infect NT-ESC with luciferase or GFP labelled slow virus or to label NT-ESC with ferric oxide nanoparticle.

Nevertheless, this study is the first to describe the EHT construction from NT-ES cells and transplanted EHTs in infarcted rat heart for therapeutic application. Our results strongly suggest that NT-ES cells can be used as a source of seeding cells for tissue engineering, especially for those tissues or organs based on cell types difficult to obtain elsewhere, e.g. cardiac tissue. In addition, the EHTs increased the size and facilitated transplantation, with a strong potential for the therapy of heart regeneration.