Pluripotent Stem Cell-Engineered Cell Sheets Reassembled with Defined Cardiovascular Populations Ameliorate Reduction in Infarct Heart Function Through Cardiomyocyte-Mediated Neovascularization

Abstract

Although stem cell therapy is a promising strategy for cardiac restoration, the heterogeneity of transplanted cells has been hampering the precise understanding of the cellular and molecular mechanisms. Previously, we established a cardiovascular cell differentiation system from mouse pluripotent stem cells, in which cardiomyocytes (CMs), endothelial cells (ECs), and mural cells (MCs) can be systematically induced and purified. Combining this with cell sheet technology, we generated cardiac tissue sheets reassembled with defined cardiovascular populations. Here, we show the potentials and mechanisms of cardiac tissue sheet transplantation in cardiac function after myocardial infarction (MI). Transplantation of the cardiac tissue sheet to a rat MI model showed significant and sustained improvement of systolic function accompanied by neovascularization. Reduction of the infarct wall thinning and fibrotic length indicated the attenuation of left ventricular remodeling. Cell tracing with species-specific fluorescent in situ hybridization after transplantation revealed a relatively early loss of transplanted cells and an increase in endogenous neovascularization in the proximity of the graft, suggesting an indirect angiogenic effect of cardiac tissue sheets rather than direct CM contributions. We prospectively dissected the functional mechanisms with cell type-controlled sheet analyses. Sheet CMs were the main source of vascular endothelial growth factor. Transplantation of sheets lacking CMs resulted in the disappearance of neovascularization and subsequent functional improvement, indicating that the beneficial effects of the sheet were achieved by sheet CMs. ECs and MCs enhanced the sheet functions and structural integration. Supplying CMs to ischemic regions with cellular interaction could be a strategic key in future cardiac cell therapy.

Disclosure of potential conflicts of interest is found at the end of this article.

Pluripotent stem cells, Tissue engineering, Myocardial infarction, Ventricular remodeling

Introduction

Cardiovascular disease remains the leading cause of death worldwide, and this has prompted researches into new therapeutic approaches including cardiac regeneration. With discoveries of various stem/progenitor cell populations possessing cardiogenic or cardioprotective potential, the notion of restorative stem cell therapies has begun to take shape [1–8]. Several clinical trials and preclinical reports have demonstrated that intracoronary or intramyocardial injection of bone marrow (BM)-derived cells (hematopoietic stem/progenitor cells, mesenchymal stem cells, and endothelial progenitor cells) [9–14], skeletal myoblasts [15, 16], or cardiac stem/progenitor cells [17, 18] ameliorate cardiac function following acute and chronic myocardial infarction (MI), even though there is still a deficiency in conclusive results for a full-scale therapy [19–21]. There seems to be little differentiation of transplanted stem cells into mature cardiovascular cell types, suggesting paracrine effects of transplanted cells in which humoral factors favorably affect the injured myocardium, including angiogenesis, apoptosis prevention, and promotion of healing [18, 22–27]. Nevertheless, as previous studies are consequences of injections of heterogeneous cell populations, insights to the cellular and molecular behavior and mechanisms after cell transplantation are still restricted. It would be a breakthrough for further improvement of cardiac cell therapy to understand the role of each cell population as well as the various cellular interactions in the chaos of heterogeneity.

In this regard, it is rational to use defined cell populations, which can be prepared from pluripotent stem cells, to analyze the mechanism of cardiac regeneration. Previously, we have developed a novel embryonic stem (ES) and induced pluripotent stem cell (iPSC) differentiation system that exhibits early cardiovascular developmental processes using Flk1 (also designed as vascular endothelial cell growth factor [VEGF] receptor-2)-positive cells as common cardiovascular progenitors. Three cardiovascular cell types, namely cardiomyocytes (CMs), endothelial cells (ECs), and vascular mural cells (MCs), can be systematically induced and purified with this system [28–31]. These defined cell populations with directed differentiation from ESCs/iPSCs are valuable experimental tools toward cell type-oriented understanding of the underlying mechanisms and the exploration of novel cardiac restoration strategies.

Another factor that has been hampering the elucidation of the mechanisms involved is the poor engraftment of injected cells to the heart [32]. Alternative strategies for cell transplantation are required to facilitate survival of grafted cells. We have developed a culture surface covalently grafted with temperature-responsive polymer poly(N-isopropylacrylamide) that enables the preparation of cell sheets without enzymatic digestion [6, 33–36]. Combining this cell sheet system and our unique ESC/iPSC differentiation system, we can reconstitute and transplant various cell type-controlled sheets with combinations of defined cardiovascular cell populations. This new experimental system enables us to prospectively elucidate the effects, roles, and interactions of each cell type in vitro and in vivo.

This study demonstrates that pluripotent stem cell-derived cell sheets as cardiac tissue have a distinct potential to ameliorate cardiac dysfunction following MI. We also demonstrate the possible mechanisms for cardiac functional improvement following transplantation with cell tracing and cell type-controlled sheet analyses. We indicate novel roles of CMs and the valuable cellular interactions in functional amelioration after cell transplantation, suggesting a strategic key role for future cell therapy.

Materials and Methods

Detailed methods and associated references are provided in Supporting Information.

Mouse ESC Culture

Two mouse ESC sublines from E14tg2a cell line were used. EMG7 mouse ESC line that carries mouse α-myosin heavy chain (MHC) promoter-driven EGFP gene was used for differentiation of CMs [28, 30]. EStTA-ROSA mouse ESC line was used for differentiation of ECs and MCs [29].

Cell Differentiation

Induction of and sorting for Flk1⁺ cells were performed as previously described [28, 30]. In brief, mouse ESCs were cultured in differentiation medium (DM) [alpha minimum essential medium (GIBCO, Grand Island, NY, http://www.invitrogen.com/site/us/en/home/brands/Gibco.html) supplemented with 10% fetal bovine serum (FBS) and 5.5 mmol/l 2-mercaptoethanol] without leukemia inhibitory factor (LIF) on gelatin-coated dishes for 96–112 hours. For the differentiation of CMs, purified Flk1⁺ cells were plated onto mitomycin-C-treated confluent OP9 cell sheets (MMC-OP9) and cultured in DM to induce CM differentiation (Supporting Information Fig. S1A). Cyclosporin-A (3 μg/ml) was added to Flk1⁺ cell culture to promote CM differentiation [37]. For the differentiation of ECs or MCs, purified Flk1⁺ cells were plated onto gelatin-coated dishes and then cultured with DM in the presence of VEGF₁₆₅ (50 ng/ml) and 8-bromoadenosine-3′: 5′-cyclic monophosphate sodium salt (8bromo-cAMP) (0.5 mmol/l) (Supporting Information Fig. S1B) [38].

Flow Cytometry Analysis and Cell Sorting

After 4 days culture of Flk1⁺ cells on MMC-OP9, cultured cells were harvested and subjected to cell sorting with FACS (fluorescence-activated cell sorting) (Aria II, BD Biosciences, Franklin Lakes, NJ, http://www.bdbiosciences.com/home.jsp). Viable green fluorescent protein (GFP)-positive cell population was evaluated and sorted as differentiated CMs (Supporting Information Fig. S1A) [30]. ESC-derived ECs and MCs were selectively induced and harvested on the third day of Flk1⁺ cell culture on gelatin-coated dishes with VEGF and 8bromo-cAMP (Flk1-d3) [28, 38].

Mouse ESC-Derived Tissue Sheet Formation

Flk1⁺ cells induced from EStTA-ROSA cells were plated onto a gelatin-coated 12-multiwell UpCell at 2.5–4.0 × 10⁴ cells per well with 1 ml of DM containing VEGF₁₆₅ (50 ng/ml) and 8bromo-cAMP (0.5 mmol/l) to selectively induce ECs and MCs on UpCell. After 3 days of culture, purified cell suspension [5.0 × 10⁵ CMs and 5.0 × 10⁵ bulk harvest of Flk1-d3 (ECs and MCs)] was plated onto vascular cells on UpCell (i.e., purified C + E + M cells onto E + M cells) (Supporting Information Fig. S1D, S1E). After 4 days of culture, cells were moved to room temperature. Within 15–30 minutes, cells detached spontaneously and floated in the media as monolayer cell sheets (C+E+M sheet) (C: CMs, E: EMs, M: MCs).

Animals

Male athymic nude rats aged between 10 and13 weeks were used for transplantation. All protocols were performed in accordance with the guidelines for Animal Experiments of Kyoto University, which conforms to Japanese law and the Guide for the Care and Use of Laboratory Animals.

Animal Model Preparation and Transplantation

MI model rats were generated as previously described [39, 40]. The rats whose hearts revealed less than 40% of left ventricular (LV) fractional shortening (FS) with echocardiogram on the sixth day of MI induction were enrolled in further experiments. One week after MI induction, each rat was randomly assigned to one of the two groups: transplantation (Tx) group and sham group (Supporting Information Fig. S2C, S2D; Supporting Information Video 3). In Tx group, three-layered cell sheet was put on the area and spread manually to make whole MI area covered by the sheet. The sheet could be stably placed onto the epicardium without sutures. The chest was closed 15 minutes after transplantation. In sham group, the chest was closed 15 minutes after thoracotomy without cell sheet transplantation.

Echocardiogram and Catheterization

Transthoracic echocardiogram was performed with a Vivid 7 system (GE Healthcare, Waukesha, WI, http://www3.gehealthcare.com/en/Global_Gateway) provided with an 11-MHz imaging transducer (GE 10S ultrasound probe, GE Healthcare). Echocardiographic measurements were performed as previously described [39, 40]. LV pressure-volume loop analysis with cardiac catheterization was performed as previously described [41]. In brief, right internal carotid artery was exposed and cannulated with a conductance- and pressure-measuring catheter transducer (SPR-869; Millar instruments, Houston, TX, http://millar.com/), which was then advanced into the LV. Pressure-volume loops were recorded with or without preload reduction by inferior vena cava compression via midline laparotomy.

Species-Specific Fluorescent In Situ Hybridization Analysis

The fluorescent in situ hybridization (FISH) probes that recognize and hybridize with sequence repeats specific for each animal species were arranged by Chromosome Science Labo (Sapporo, Japan, http://www.chromoscience.ip/index_e.html) [42, 43]. The nucleotide probes were applied to the fixed and pretreated sections, denatured, followed by hybridization. Additional immunofluorescent staining for cardiac troponin-T (cTnT) was performed to the FISH samples.

RNA Extraction and Quantitative Reverse-Transcription Polymerase Chain Reaction

Total RNA was extracted from cell sheets using RNeasy (QIAGEN, Venlo, The Netherlands, http://www.qiagen.com/default. aspx), according to the manufacturer's instructions. Reverse transcription was performed with the SuperScript III first-strand synthesis system (Invitrogen, Eugene, OR, http://www.invitrogen. com/site/us/en/home.html). Quantitative PCR was performed using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com/absite/us/en/home.html) and StepOnePlus system (Applied Biosystems).

Statistical Analysis

The data were processed using Dr. SPSS II software for windows (version 11.0.1J, SPSS Inc, Chicago, IL, http://www-01.ibm.com/software/analytics/spss/). Comparisons between two groups were made with unpaired or paired t test. Comparisons between more than two groups were made with one-way analysis of variance followed by Tukey's test as post hoc. Values are shown in mean ± SD. p values <.05 were considered significant.

Results

Construction and Characterization of Cardiac Tissue Sheet

First, we attempted to fabricate cardiac tissue consisting of mouse ESC-derived CMs, ECs, and MCs. CMs were differentiated and purified as previously reported [30, 37]. Briefly, when Flk1⁺ cells derived from mouse ESCs carrying cardiac-specific GFP gene are cultured on OP9 stroma cells, spontaneously beating CMs (GFP⁺) can be induced and subsequently purified by FACS. The purity of sorted GFP⁺ CMs was 96.0% ± 2.0% in this study (Supporting Information Fig. S1A). On the other hand, ECs and MCs were selectively induced as previously reported [28, 29]. When Flk1⁺ cells are cultured on gelatin-coated dishes with VEGF, almost all of the cells that appear are either ECs or MCs. Vascular endothelial (VE)-cadherin⁺ ECs were 31.8% ± 14.8% on average in this study. The remainder indicated MC population (approximately 68%) (Supporting Information Fig. S1B, S1C). We reassembled these defined cardiovascular populations on 12-multiwell temperature-responsive culture dishes (UpCell; CellSeed, Tokyo, Japan) (Fig. 1; see detailed protocol in Methods; Supporting Information Fig. S1). Temperature reduction successfully provided reassembled cardiac tissue sheets that consisted of CMs (40.2% ± 8.5% of total cells), ECs (8.4% ± 5.7%), and MCs (51.4% ± 9.2%) (C + E + M sheets; n = 24, FACS analyzed) (Fig. 1A, 1B; Supporting Information Fig. S1, Supporting Information Video 1). Total cell count of the sheet was 2.62 ± 0.74 × 10⁵ on average (n = 24).

1

Construction and characterization of cardiac tissue sheets. (A): Schematic diagram of cardiac tissue sheet formation. Defined cardiac cell populations (CMs, ECs, and vascular MCs) are systematically differentiated from pluripotent stem cells, sorted, and reassembled to construct tissue sheets. (B–F): Tissue sheet components. (B): Flow cytometry analysis (n = 24). Upper panel: α-MHC-GFP⁺ CMs in total cells from sheets. Lower panel: CD31⁺ ECs in GFP⁻ non-CM population. GFP⁻/CD31⁻ cells correspond to MCs. Representative plot data are shown. (C, D) Cross-sections. (C): Upper panel: H&E staining showing cell appearance of the sheet. Lower panel: Sirius red staining showing intact extracellular matrix. (D): Upper panel: cTnT staining (brown). Lower panel: two-photon laser imaging microscopy. Double staining for cTnT (CMs, red) and VE-cadherin (ECs, green). (E, F) Immunostaining of sheets for cTnT (red), VE-cadherin (green), and DAPI. Low-magnification view with full-focused fluorescent imaging microscopy (E) and high-magnification view with two-photon laser imaging microscopy (F) showing distribution of CMs (cTnT⁺), ECs (VE-cadherin⁺), and MCs (cTnT⁻/VE-cadherin⁻). (G–I): Electrophysiological study for extracellular potentials. (G): Tissue sheet set on the electrodes (black dots). (H): Extracellular potentials of each microelectrode. (I): Propagation map. Electrical propagations are visualized as a contour image. Regular and unidirectional electrical conduction is observed. Scale bars = 200 μm in (E) and 50 μm in (F). Abbreviations: αMHC, α-myosin heavy chain; αSMA, α-smooth muscle actin; ch, channel; CM, cardiomyocyte; cTnT, cardiac troponin-T; DAPI, 4,6-diamidino-2-phenylindole; ECs, endothelial cells; ES, embryonic stem; GFP, green fluorescent protein; H&E, Hematoxylin and Eosin; iPS, induced pluripotent stem; MC, mural cell; SSC, side scatter; VE-cadherin, vascular-endothelial cadherin.

The sheet consisted of three to four cell layers with intact stratified structure of collagen adjacent to the cell component (Fig. 1C). Immunostaining for CMs (cTnT⁺) and ECs (VE-cadherin⁺) showed an almost even distribution of CMs, ECs, and MCs (cTnT⁻/VE-cadherin⁻) in the sheet (Fig. 1D–1F). Cell surface potentials showed continuous, regular, and unidirectional electrical propagation in the sheet without ectopic foci (Fig. 1G–1I; Supporting Information Video 2).

Cardiac Tissue Sheet Transplantation Ameliorates Cardiac Function

Three cardiac cell sheets were piled up to form a three-layered cell sheet and transplanted to rat subacute MI model (Supporting Information Fig. S2; Supporting Information Video 3). Cell sheet transplantation (sham operation, n = 9; C + E + M sheet transplantation, n = 9) was performed 1 week after MI induction. All rats survived for 4 weeks of observation periods after transplantation. We evaluated cardiac functions with echocardiogram (n = 9 each) (Fig. 2A–2E; Supporting Information Video 4) [39, 40] and LV pressure-volume loop study with catheterization (n = 8 each) (Fig. 2F, 2G) [41]. Transplantation group revealed restoration of anterior wall contraction (Fig. 2A). Transplantation significantly increased FS and fractional area change compared to those of pretreatment phase as well as sham operation group (Fig. 2B, 2C). The systolic thickening of infarct wall significantly recovered to almost normal value (1.6 ± 0.2 vs. 1.6 ± 0.3 [normal value (n = 26) vs. sheet transplantation (4 weeks, n = 9)]) (Fig. 2D). Increase in diastolic area of LV showed a tendency to be decreased (Fig. 2E). LV pressure-volume loop study can measure LV performance independent from loading conditions or heart rate [41]. End-systolic elastance (Ees) reflecting systolic function was significantly higher at 4 weeks after transplantation (Fig. 2F, 2G). Time constant (Tau), an index for diastolic function, was, however, not significantly improved (Fig. 2G). Transplantation limited the extent of fibrosis and thinning of the infarct wall at 4 weeks after transplantation (Fig. 2H, 2I). The improvement of systolic function in echocardiogram was stably maintained at 3 months after transplantation (sham operation, n = 3; C + E + M sheet transplantation, n = 3) (Supporting Information Figs. S2D, S3). No tumor formation was observed after the cardiac tissue sheet transplantation throughout all experimental periods (n = 12). All these findings indicate that cardiac cell sheet transplantation successfully improves LV systolic function and attenuates LV remodeling following MI.

2

Cardiac functions after cardiac tissue transplantation to rat MI. (A–F): Echocardiogram (n = 9). (A): Representative M-mode image. Note that infarct anterior wall started to move 2–4 weeks after transplantation (Tx). (B): FS. (C): FAC. (D): Systolic thickening of infarcted wall. (E): Diastolic area of LV (fold increase to PreTx). (F, G): LV pressure-volume (P-V) loop study 4 weeks after Tx (n = 8). (F): Representative P-V loops. Slopes indicate Ees (arrows). (G): Ees: End-systolic elastance (left panel). Tau: time constant (right panel). (H, I): Sirius red staining at Tx4w. (H): Representative sections. Reduction of fibrotic area and wall thinning after Tx indicates attenuation of LV remodeling (arrows). (I): The ratio of MI length of total length (left panel) and wall thickness of infarct wall (average thickness) (right panel) (25 sections). *, p < .05; **, p < .01; and ***, p < .001 (unpaired t test). ^†, p < .05 and ^‡, p < .01 (vs. PreTx, paired t test). PreTx, Pretransplantation; Tx2w, Tx4w, 2 and 4 weeks after transplantation, respectively. Scale bars = 1 mm. Abbreviations: FAC, fractional area change; FS, fractional shortening; LV, left ventricle; MI, myocardial infarction.

Survival of Transplanted CMs In Vivo

Next, to precisely evaluate the cellular and molecular mechanisms of transplantation, we tried to trace the fate and distribution of transplanted cell sheets using species-specific (SS)-FISH analysis, which can distinguish mouse donor cells from rat recipient cells [42, 43]. A pilot study, just after transplantation of a single-layered cell sheet to normal rat heart, evidently demonstrated a single array of mouse nuclei on the very surface of rat heart, certifying the validity of this method (Fig. 3A). We performed quantification of the engrafted CMs with double staining of cTnT (immunostaining; for mouse and rat CMs) and SS-FISH after cell sheet transplantation to MI model. Whereas obvious cTnT⁺ CM graft area with mouse nuclei on the surface of the infarct heart was observed 1 day after the transplantation, the grafted CM area started to diminish within several days. Very little CM clusters remained at 4 weeks after transplantation (Fig. 3B, 3C). No mouse-nuclei-positive cells were observed outside of these cluster areas, suggesting that most of the engrafted mouse cell populations had disappeared. The discrepancy between functional improvement and CM engraftment suggests that the improvement of cardiac function is not mainly mediated by direct contribution of CMs but other indirect roles, possibly paracrine effects, of the sheet at the early stage of transplantation.

3

Tracing of transplanted cells with species-specific fluorescent in situ hybridization (SS-FISH). (A): SS-FISH analysis for mouse (yellow) and rat nuclei (magenta). Just after transplantation of mouse monolayer cell sheet. Right panel: higher magnification view of boxed area. (B, C): Cell tracing after transplantation. (B): Double staining for cTnT (immunostaining, red), SS-FISH (mouse nuclei, yellow), and DAPI. Dotted squares indicate engrafted clusters. Insets: higher magnification views. (C): Quantification of engrafted area (mouse nuclei⁺/cTnT⁺ area [mm²], n = 10): 0.55 ± 0.36 (Tx-d1), 0.42 ± 0.16 (Tx-d3), 0.20 ± 0.05 (Tx-d7), and 0.0023 ± 0.0008 (Tx-d28). **, p < .01 (vs. Tx-d1, unpaired t test). Tx-d1, d3, d7, and d28: 1, 3, 7, and 28 days after Tx, respectively. Scale bars = 100 μm in (A) (left panel), 20 μm in (A) (right panel), and 200 μm in (B). Abbreviations: cTnT, cardiac troponin-T; DAPI, 4,6-diamidino-2-phenylindole.

Neovascularization Following Cardiac Tissue Sheet Transplantation

Neovascularization is one of the major phenomena in paracrine effects reported in BM cell transplantation [25]. We examined neovascularization after cardiac tissue transplantation with von Willebrand factor (vWF) staining for ECs and found that capillary density was significantly increased within both peri-MI and central-MI (more prominent in peri-MI) area 4 weeks after transplantation (Fig. 4A, 4B) (peri- and central-MI are illustrated on Supporting Information Fig. S4A). We further examined the kinetics and cellular origin of the neovascularization. One day after transplantation, vWF⁺ cells were distributed within the engrafted area but not evident outside of this area (data not shown). At 3 days after transplantation, prominent accumulation of vWF⁺ cells was observed in the proximity of the engrafted area (Fig. 4C; Supporting Information Fig. S4B). SS-FISH revealed that the accumulated cells originated from the recipients but not from the grafts (Fig. 4D). At 7 days after transplantation, capillaries with lumen that originated from the recipients were generated in the area around the graft (Fig. 4C). These results indicate that cardiac tissue sheet transplantation potently induces the accumulation of endogenous vascular cells and capillary formation around the engrafted sheet within the early stages of transplantation.

4

Neovascularization after transplantation. (A, B): Capillary formation at Tx-d28. (A): Double staining for vWF (ECs, green) and cTnT (cardiomyocytes [CMs], red) at peri-MI and central-MI areas (Supporting Information Fig. S4A). Note that newly formed capillaries are clearly observed in transplantation group (dotted circles). (B): Quantification of capillary density (capillary number per square millimeter). Peri-MI area (left panel) and central-MI area (right panel) (15 views each). **, p < .01 and ***, p < .001 between corresponding values (unpaired t test). (C, D): Kinetics (C) and origin (D) of neovascularization. (C): Double staining for vWF (green), cTnT (red), and DAPI. Note that prominent vWF⁺ cell accumulation in the proximity of grafted CMs (Tx-d3, upper panel) and induction of vascular structures with lumen are observed (arrows, Tx-d7, lower panel). (D): Triple staining for vWF, cTnT, and species-specific fluorescent in situ hybridization (mouse nuclei, yellow) (Tx-d3). Most of the accumulated vWF⁺ cells are negative for mouse nuclear staining (arrows). Inset: higher magnification view. (E): Double staining for VEGF (green) and cTnT in tissue sheet in vitro. VEGF expression is restricted to cTnT⁺ area. Dotted lines: boundary of cTnT-positive and negative area. Scale bars = 100 μm in (A, C, D) (main panel) and (E) (left panel), 50 μm in (D) (inset). Abbreviations: cTnT, cardiac troponin-T; DAPI, 4,6-diamidino-2-phenylindole; MI, myocardial infarction; VEGF, vascular endothelial cell growth factor; vWF, von Willebrand factor.

Next, to elucidate the molecular mechanism of the neovascularization, we examined protein secretion of some angiogenic factors such as tumor necrosis factor-alpha, insulin-like growth factor 1, VEGF, interleukin-6, basic fibroblast growth factor (bFGF), interferon-gamma, epidermal growth factor, leptin, and hepatocyte growth factor in the culture supernatant of cardiac tissue sheet. Among them, VEGF production was much more prominent than the other factors, suggesting the critical role of VEGF (Supporting Information Fig. S4C). To identify the cell populations responsible for VEGF production, we performed immunofluorescent staining of cardiac tissue sheet. VEGF staining was highly restricted to cTnT⁺ CM area (Fig. 4E), indicating that CMs are the main source of VEGF among sheet-composing cell types.

Prospective Examination of Functional Roles of CMs with Cell Type-Controlled Tissue Sheets

Our ESC system possesses distinct advantages from previous cell transplantation studies, in which we can prospectively evaluate the functional roles of each cell population by cell type-controlled sheets with various combinations of cardiovascular cells. VEGF secretion from C + E + M sheet in vitro was 13.3-fold higher (Fig. 5A), and Vegf164 gene expression in vitro was 21.5-fold higher (Fig. 5B) than those from E + M sheet (ECs and MCs) (Supporting Information Fig. S5A), respectively. These results confirm the result of immunostaining for VEGF (Fig. 4E). Next, we performed transplantation of E + M sheets (three-layered) and compared the capillary densities and cardiac functions to those of C + E + M sheet transplantation group. Transplantation of E + M sheets did not induce significant increase in capillary density (Fig. 5C). Echocardiogram revealed that the improvement of systolic function was diminished in E + M sheet transplantation (Fig. 5D). These results indicate that sheet CMs are essential to achieve cardiac functional improvement in the infarct heart.

5

Cell type-controlled sheet analyses. (A): ELISA for VEGF secretion (picogram per10⁴ cells) in culture supernatants of C + E + M and E + M sheets. (B): Quantitative PCR for VEGF mRNA from C + E + M and E + M sheets. Vegf164 (VEGF) expressions were normalized by Actb (β-actin) expression. Expressions in E + M sheets were set as 1.0. (C, D): Transplantation of sham operation (n = 9) versus C + E + M sheets (n = 9) versus E + M sheets (n = 3) (Tx-d28). (C): Capillary density (capillary number per square millimeter). Peri-MI area (left panel) and central-MI area (right panel) (15 views each). (D): Echocardiogram data (fold increase vs. PreTx). FS (left panel), FAC (middle panel), and systolic thickening of infarct wall (right panel). *, p < .05; **, p < .01; and ***, p < .001 (unpaired t test). C: CMs, E: ECs, M: MCs. Abbreviations: FAC, fractional area change; FS, fractional shortening; MI, myocardial infarction; N.S., not significant; VEGF, vascular endothelial cell growth factor.

Roles of ECs and MCs in Cardiac Tissue Sheet Transplantation

Cellular interactions among CMs and nonmyocytes are considered to be involved in myocardial tissue formation and function [44, 45]. Coculture of neonatal rat ECs and CMs is reported to enhance secretion of angiogenic factors such as VEGF [35]. To estimate the role of ECs in the cardiac tissue sheet function, we generated sheets with CMs and MCs (C + M sheet) without ECs and examined the effect of ECs on VEGF secretion (Supporting Information Fig. S5B). CMs in C + M sheet showed a significantly lower positive rate of VEGF (Supporting Information Fig. S5C, S5D) and VEGF secretion in culture supernatant (0.33-fold increase), compared with those in C + E + M sheet (Supporting Information Fig. S5E), indicating that existence of ECs in the tissue sheets promotes VEGF production from CMs.

We also found significant roles of MCs through trying to generate cell sheet without MCs. When various amounts of purified CMs were plated onto the 12-multiwell UpCell (2.5 × 10⁵, 5.0 × 10⁵, or 1.0 × 10⁶ per well) to form pure CM sheet without MCs, pure CMs failed to form sheet structure regardless of the plated cell counts. Sheet formation was successfully recovered by plating the mixture of CMs and MCs onto UpCell dishes (Supporting Information Fig. S5F), suggesting that MCs are essential to form an integrative cell sheet structure. These results indicate that ECs and MCs play critical and beneficial roles in cardiac tissue sheet formation and function through cellular interaction.

Discussion

In this study, we showed therapeutic potential of cardiac tissue sheets from pluripotent stem cells and novel cellular mechanisms of cardiac functional improvement with a prospective strategy by reassembling defined cardiac cell populations (Supporting Information Fig. S6). Cardiac tissue sheet transplantation distinctively restored cardiac function after MI mainly through the suppression of LV remodeling with the induction of neovascularization in the proximity of the sheet. CMs played a central role for the functional restoration not through direct contribution but through the induction of neovascularization. Cellular interaction among CMs, ECs, and MCs efficiently enhanced structural integration and function of the cell sheets. Understanding the roles of each transplanted cell population, especially the new role of CMs, would provide a valuable strategic principle in future therapy for cardiac restoration.

In this study, we used SS-FISH, which can distinctively stain and discriminate mouse and rat nuclei. This method is amenable to sensitive and specific tracing of transplanted mouse ESC derivatives, avoiding the problem of transgene silencing during ESC differentiation, and false positives of transplanted cell marking by cell fusion phenomena [42, 43]. The results clearly showed that even though apparent improvement of cardiac function was induced and sustained long after transplantation, the majority of the grafted cells diminished within the earlier period. Moreover, massive neovascularization originating from the recipients was induced in the proximity of the transplanted sheets, indicating that graft-elicited neovascularization is critical for suppression of LV remodeling and following functional restoration. LV remodeling is a complex alteration in ventricular architecture after MI. During scar formation and tensile strength increase, the infarct region becomes thinner and elongates due to continuous expansive stimulation, termed “infarct expansion” (Supporting Information Video 4) [46–48]. Infarct expansion finally results in LV remodeling with dilated LV lumen and fibrous tissue deposition. It is also said that the viable myocardium in the infarct border zone is significantly hypertrophied following infarction. The impaired capillary network that cannot support the demand of hypertrophied myocardium accelerates infarct expansion [46]. Thus, a sufficient blood supply to the border zones is critical to prevent infarct expansion. Indeed, induction of neovascularization by BM-derived cell transplantation to MI model was reported to attenuate infarct expansion and LV remodeling [25]. Predominant neovascularization in border zone following cardiac tissue sheet transplantation should contribute to suppress infarct expansion and LV remodeling through attenuation of capillary insufficiency.

Some previous studies with stem cell transplantation, including cell sheet experiments (stem cell antigen 1-positive cardiac progenitor cells or adipose tissue-derived mesenchymal stem cells), showed functional restoration after MI [6, 9, 19–21, 23, 25, 36]. Nevertheless, their precise cellular mechanisms still remain unclear [26, 27] possibly because of the heterogeneity of the consisting cell populations of the sheet with various lineages or differentiation stages. How can we achieve a more profound elucidation of the mechanism, which is hampered due to the heterogeneity of the transplanted cells? Recently, Yoon et al. tackled this issue through the injection of genetically modified lines of BM mononuclear cells for acute MI model, followed by selective cell depletions of CMs (αMHC⁺), ECs (endothelial nitric oxide synthase positive), or MCs (SM22α⁺) 2 weeks after injection [49]. The results that elimination of ECs or MCs induced a significant deterioration of cardiac function suggested the importance of angiogenesis in functional restoration after BM cell therapy. As this study is retrospective cell depletion 2 weeks after the BM cell transplantation, it would be difficult to determine the initial situation of cell contribution, that is, how much donor cell populations actually existed and distributed as CMs, ECs, or MCs. The roles and interactions of each cell population in cardiac restoration are still unclear. To provide further mechanistic insights into cardiac restoration after cell transplantation, here we performed a new attempt of prospective analysis using cell type-controlled transplantation with reassembly of defined cell populations. We succeeded in identifying that CMs play central roles in functional restoration mainly through neovascularization, and interactions among cardiovascular cells induce substantial enhancement of the angiogenic function. Our study thus can set a strategic principle for cardiac restorative therapy after MI; that is, to maximize the supply of CMs with cellular interaction to ischemic area, which can rescue missing endogenous CMs through induction of neovascularization (Supporting Information Fig. S6). In most of the current stem cell therapies, differentiation efficiency to CMs was rather low. Increasing the amount of CMs in cell therapy should be important not only for direct contribution of CMs but also for further efficient neovascularization.

In addition to neovascularization, paracrine effects after cell therapy were reported to be involved in cell survival, migration, extracellular matrix regulation, and so on. We thus further investigated other humoral factor expressions in the cell sheets (Supporting Information Fig. S7). Among the factors we tested, the following molecules showed more than a fivefold increase of mRNA expression in C + E + M sheet than that in E + M sheet: Sfrp1 (secreted frizzled-related protein 1; 8.21), Fgf2 (bFGF; 6.26), and Timp1 [tissue inhibitor of metalloproteinase-1 (TIMP1); 5.73]. Tmsb4x (Thymosin β4) and Vcam1 [vascular cell adhesion molecule 1 (VCAM1)] expressions were also enhanced in C + E + M sheet. bFGF is well known to possess broad biological effects including angiogenesis, cardiac stem cell survival, and growth [17]. SFRP and soluble VCAM1 were reported to induce angiogenesis and suppress cell death [6, 50]. Thymosin β4 promotes cardiac cell migration and survival [51]. Thymosin β4 is also reported to promote induction of de novo CMs from an epicardial origin of the progenitor population in adult mouse heart after injury [52]. TIMP1 was reported to suppress apoptosis and LV remodeling after MI [53, 54]. All these factors (and others) should have a combinatorial effect on functional advantage after cardiac tissue transplantation. Detailed studies for functions and contributions of each factor would provide further understanding of the molecular mechanisms for functional improvement after cell therapy.

As a future perspective, improvements and further exploration would be expected as follows. First, a more efficient survival of transplanted cardiac tissue sheets is expected. In this study, we performed the sheet transplantation at 1 week after MI induction (subacute stage). This should be the optimal transplantation timing, avoiding damage of transplanted cell with acute inflammation and obtaining sufficient tissue repair before fibrotic scar formation in chronic stage [55, 56]. However, sufficient long-term engraftment was not achieved in this study. This may be due to xenotransplantation between mouse cells and nude rats. Although we did not observe apparent immune responses after the sheet transplantation, a more stringent control with immunosuppressants or severer immunodeficient rat models, such as X-linked severe combined immunodeficiency (X-SCID) rats [57], may improve the survival. Otherwise, since more cells that survived were observed in peri-MI than central MI region, the severe ischemic condition at central-MI areas may not be suitable for sheet survival. Novel techniques increasing blood flow in the graft should be applied, such as prevascularization in three-dimensional tissue formation [35, 58, 59] and/or vascularized flap grafts. Furthermore, the presence of epicardium may become a physical barrier for the engraftment to host myocardium with this sheet transplantation. Arrhythmogenic potential of the engrafted CMs would be another obstacle for the clinical application. Considering these points, although the sheet transplantation is a potent option for the treatment of heart failure, further exploration for the feasibility of this new therapeutic modality would be required.

Second, the extension of the technology to iPSCs should be explored. We have already succeeded in generating cardiac tissue sheets using mouse iPSCs with the same method as that of ESCs (Supporting Information Fig. S8). We recently reported functional CM induction from human iPSCs [60]. We further established a more efficient CM induction as well as purification methods in human iPSCs [61]. EC and MC induction from human iPSCs has already been reported [62]. Technological basis to generate human cardiac tissue sheets from iPSCs has thus been established. Transplantation experiments in various animal models would prove the feasibility of human cardiac tissue sheet strategy for clinical use.

Finally, an exploration of cell-free therapy is expected. Here, we showed that cardiac tissue sheets mainly act through CM-originated paracrine effects of VEGF and other factors (Supporting Information Fig. S7) [18, 22–25]. When we could spatially and temporally reconstitute all these paracrine functions of the cardiac tissue sheets by defined humoral factors, it should be possible to reproduce the same therapeutic outcome by a cell-free factor delivery system. Such cell-free method would be more easily and broadly applicable to clinical use than cell therapy.

Conclusions

Here, we have successfully demonstrated the potentials and novel mechanisms of pluripotent stem cell-derived cardiac cell therapy with new experimental approaches. This study would provide a hallmark for cell therapy with pluripotent stem cells and strategic principle for future cardiac restoration therapy.

Acknowledgements

We thank Dr. S. Yamanaka (Center for iPS Cell Research and Application, Kyoto University) for mouse iPS cells (20D-17). We thank Dr. M. Takahashi for critical reading of the manuscript. This work was supported by research grants from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, the Ministry of Health, Labor, and Welfare, Japan (to J.K.Y. and T.I.), the New Energy Industrial Development Organization of Japan, the Project for Realization of Regenerative Medicine (to J.K.Y.), the fellowship from the Japan Society for the Promotion of Science (to K.Y., H.U., and G.N.), Invited Research Project of Transnational Research Center, Kyoto University Hospital (to A.M. and R.S.), and Japan Heart Foundation Young Investigator's Research Grant (to K.Y.).

Disclosure of Potential Conflicts of Interest

T.S. is a consultant for CellSeed, Inc. T.O. is an investor in CellSeed, Inc. and an inventor/developer designated on the patent for temperature-responsive culture surfaces. The authors indicate no competing financial interests.