Cardiomyocytes From CCND2-Overexpressing Human Induced-Pluripotent Stem Cells Repopulate the Myocardial Scar in Mice: A 6-month Study

Abstract

BACKGROUND:

Cardiomyocytes that have been differentiated from CCND2-overexpressing human induced-pluripotent stem cells (hiPSC-CCND2^OE CMs) can proliferate when transplanted into mouse hearts after myocardial infarction (MI). However, it is unknown whether remuscularization can replace the thin LV scar and if the large muscle graft can electrophysiologically synchronize to the recipient myocardium. Our objectives are to evaluate the structural and functional potential of hiPSC-CCND2^OE CMs in replacing the LV thin scar.

METHODS:

NOD/SCID mice were treated with hiPSC-CCND2^OE CMs (i.e., the CCND2^OE group), hiPSC-CCND2^WT CMs (the CCND2^WT group), or an equal volume of PBS immediately after experimentally-induced myocardial infarction. The treatments were administered to one site in the infarcted zone (IZ), two sites in the border zone (BZ), and a fourth group of animals underwent Sham surgery.

RESULTS:

Six months later, engrafted cells occupied more than 50% of the scarred region in CCND2^OE animals, and exceeded the number of engrafted cells in CCND2^WT animals by ~8-fold. Engrafted cells were also more common in the IZ than in the BZ for both cell-treatment groups. Measurements of cardiac function, infarct size, wall thickness, and cardiomyocyte hypertrophy were significantly improved in CCND2^OE animals compared to animals from the CCND2^WT or PBS-treatment groups. Measurements in the CCND2^WT and PBS groups were similar, and markers for cell cycle activation and proliferation were significantly higher in hiPSC-CCND2^OE CMs than in hiPSC-CCND2^WT CMs. Optical mapping of action potential propagation indicated that the engrafted hiPSC-CCND2^OE CMs were electrically coupled to each other and to the cells of the native myocardium. No evidence of tumor formation was observed in any animals.

CONCLUSIONS:

Six months after the transplantation, CCND2-overexpressing hiPSC-CMs proliferated and replaced more than 50% of the myocardial scar tissue. The large graft hiPSC-CCND2^OE CMs also electrically integrated with the host myocardium, which was accompanied by a significant improvement in LV function.

1. Introduction

Mammalian cardiomyocytes exit the cell cycle shortly after birth and lose the ability to proliferate, which likely explains why the cardiac tissue lost to myocardial infarction (MI) or other forms of heart disease cannot be adequately replaced via endogenous mechanisms of myocardial repair [1, 2]. Researchers have attempted to improve myocardial recovery by delivering cardiomyocytes and other types of cells to the damaged region [3, 4]. However, the proportion of cells that are retained and survive at the site of administration is exceptionally small, as shown in murine, swine and nonhuman primate models [5–7]. Thus, the results from clinical trials generally suggest that cell therapies provide no impressive functional and structural benefits [8]. The functional improvements observed in animal studies are largely attributed to paracrine mechanisms that are induced by the transplanted cells [5, 9], rather than to replacement of the myocardial scar with functioning cardiac muscle (i.e., remuscularization).

D type cyclins regulate the G1-to-S phase transition of the cell cycle, and a transgene overexpressing CCND2 (cyclin D2) activates the cell cycle in cardiomyocytes [10, 11]. hiPSC-derived cardiomyocytes (hiPSC-CMs) gradually exit the cell cycle when cultured for an extended period [12–15]. If cell cycle activity can be preserved (or re-activated) [16], then even a small number of engrafted hiPSC-CMs may proliferate and eventually remuscularize the infarcted region. We have tested this strategy previously by overexpressing the cell cycle regulator CCND2 in hiPSC-CMs, and then injecting the cells directly into the hearts of mice after left anterior descending coronary artery (LAD) ligation [12]. Our results indicated that the number of CCND2-overexpressing hiPSC-CMs (hiPSC-CCND2^OE CMs) increased, while infarct sizes declined, between the first and fourth weeks after MI and treatment, and that infarct sizes at week 4 were smaller in animals treated with hiPSC-CCND2^OE CMs than in animals treated with hiPSC-CMs that expressed normal levels of CCND2 (hiPSC-CCND2^WT CMs). However, the LAD ligation resulted in a large piece of LV scar in both groups at 4 weeks, and significant remuscularization of LV scar did not occur [12]. The present study examines whether the continuous proliferation of the grafted hiPSC-CCND2^OE CMs results in the LV remuscularization of LV scar over a 6-month period, and if so, whether there is an electrophysiological integration between the large graft and recipient heart, and if there is a risk of tumor formation by prolonged myocyte proliferation.

The results of the current investigation demonstrate the feasibility of a concept of promoting myocyte proliferation to remuscularize hearts suffering from acute myocardial infarction. However, the practical technical challenges such as: the effective cardiac delivery of cells, viral transduction to modify genes, and therefore the long duration period for FDA approval, etc. will have to be overcome before this genetically modified hiPSC approach can be started in clinical trials.

2. Material and methods

A detailed description of the experimental procedures used in this investigation is provided in the Online Data Supplement. All experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Alabama at Birmingham and performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH publication No 85–23). Data were presented as mean ± SEM, and significance (P<0.05) was determined via the Student’s t-test for comparisons between two groups or one-way analysis of variance for comparisons among three or more groups.

3. Results

3.1. CCND2 overexpression significantly increased the long-term (6 month) potency of transplanted hiPSC-CMs for myocardial repair in a murine MI model.

MI was surgically induced in the hearts of NOD/SCID mice by permanently ligating the left anterior descending coronary artery; then, animals in the CCND2^OE group were treated with hiPSC-CCND2^OE CMs, animals in the CCND2^WT group were treated with hiPSC-CCND2^WT CMs, and animals in the MI group were treated with an equal volume of PBS. The cells (or PBS) were administered via direct intramyocardial injection to three sites (1×10⁵ cells/site, 3×10⁵ cells/animal); one site was located in the infarcted region, and two were located in the region surrounding the infarct. A fourth group of animals, the Sham group, underwent all surgical procedures for MI induction except the ligation step and recovered without any experimental treatment.

Echocardiographic assessments (Fig. 1A) of left-ventricular ejection fraction (EF) (Fig. 1B) and fractional shortening (FS) (Fig. 1C) were equivalent in all groups before MI induction or sham surgery, and six months afterward, measurements were significantly lower in the three groups that underwent MI surgery than in Sham animals. However, EF and FS at month 6 were markedly greater in CCND2^OE animals than in animals from the CCND2^WT or MI groups. Histological assessments (Fig. 2A) of infarct size (Fig. 2B) and left ventricular (LV) anterior wall thickness (Fig. 2C) were also significantly better in CCND2^OE animals than in CCND2^WT or MI animals, but measurements in the CCND2^WT and MI groups did not differ significantly for any parameter (EF, FS, infarct size, or LV wall thickness).

Fig. 1. Six months after MI induction, cardiac function significantly improved in CCND2^OE animals than in CCND2^WT animals.

MI was surgically induced in mice; then, animals in the CCND2^OE group were treated with hiPSC-CCND2^OE CMs, animals in the CCND2^WT group were treated with hiPSC-CCND2^WT CMs, and animals in the MI Only group were treated with an equal volume of PBS. Animals in the Sham group underwent all surgical procedures for MI induction except the ligation step and recovered without any experimental treatment. (A) Echocardiographic assessments of (B) left ventricular ejection fraction and (C) fractional shortening (C) were performed before surgery (Pre-S) and six months afterward (Post-S). *P<0.01 versus Sham, ^†P<0.01 versus MI Only, ^‡P<0.05 versus CCND2^WT.

Fig. 2. Infarct sizes and ventricular wall thicknesses significantly improved in CCND2^OE animals than in CCND2^WT animals.

Animals in the Sham, MI Only, CCND2^WT, and CCND2^OE groups were sacrificed six months after MI induction or Sham surgery; then, left ventricles were harvested, sectioned, and stained with fast green to identify functional myocardium and with sirius red to identify scar tissue (n = 12 per experimental group, scale bar = 1 mm). (B) Infarct size, and (C) left-ventricular anterior wall thickness were quantified and expressed as a percentage. *P<0.01 versus Sham, ^†P<0.01 versus MI, ^‡P<0.01 versus CCND2^WT.

3.2. hiPSC-CCND2^OE CMs proliferated after transplantation into infarcted mouse hearts, replaced more than 50% of the myocardial scar, and significantly promoted angiogenesis.

Expression of CCND2 is under the control of α-MHC promoter [12]. Therefore, its expression is limited to hiPSC-derived cardiomyocytes but not non-myocytes. The transplanted cells carried a luciferase reporter plasmid (under the control of SFFV promoter) [12]; thus, engraftment can be evaluated at monthly intervals in living animals via bioluminescence imaging (BLI) (Fig. 3A). The BLI signal (Fig. 3B) and, consequently, the calculated number of engrafted cells (Fig. 3C) were significantly greater in CCND2^OE animals than in CCND2^WT animals at each time point during the 6-month follow-up period. Furthermore, while the BLI signal remained relatively constant for animals in the CCND2^WT group, signal intensity in CCND2^OE animals progressively increased, and the calculated number of engrafted hiPSC-CCND2^OE CMs at month 6 exceeded the total number of cells administered (Fig. 3D). Histological assessments of the expression of human cardiac troponin T (hcTnT) also suggested that the population of engrafted cells was larger in CCND2^OE animals at month 6 than in a subset of CCND2^OE animals sacrificed one month after MI injury and treatment (Fig. 3E).

Fig. 3. Engrafted cells were significantly more common in hearts treated with hiPSC-CCND2^OE CMs than with hiPSC-CCND2^WT CMs.

Six months after MI induction or Sham surgery, animals in the Sham, MI Only, CCND2^WT, and CCND2^OE groups were injected with luciferin, and (A) BLI images were collected 10 min later. (B) BLI signal intensity was compared to a standard curve determined from known quantities of hiPSC-CMs; then, (C) the number of engrafted cells was calculated and (D) presented as a percentage of the total number of cells administered. (E) Heart sections from mice that had been sacrificed 1 and 6 months after MI induction and treatment with hiPSC-CCND2^OE CMs were stained for the presence of hcTnT. *P<0.01 vs Sham, ^†P<0.01 vs MI, ^‡P<0.01 vs CCND2^WT, ^§P<0.05 vs month 1, ^∥P<0.05 vs month 2, ^#P<0.05 vs month 3.

The results from our BLI measurements were corroborated via immunofluorescence analyses of the expression of human cardiac troponin T (hcTnT) and human nuclear antigen (HNA) (Fig. 4A) at month 6. Cells expressing both of the human proteins were 8-fold more common in CCND2^OE animals than in the CCND2^WT group (Fig. 4B), and the number of hcTnT/HNA double-positive cells in CCND2^OE animals exceeded the number of cells administered (Fig. 4C). Engrafted cells were more common in the infarcted zone than in the border zone for animals in both cell-treatment groups (Fig. 4D), and more than 95% of grafted cells were cardiomyocytes, as determined at month 1 and month 6 after transplantation (Fig. 4E). Remarkably, more than 50% of the myocardial scar in CCND2^OE animals had been replaced by engrafted hiPSC-CMs, compared to less than 10% in the CCND2^WT group (Fig. 4F-G). More than 90% of the engrafted hiPSC-CCND2^OE CMs continued to express CCND2 at month 6 (Supplemental Fig. 1A) and contained a single nucleus (Supplemental Fig. 1B). Cells that co-expressed hcTnT and Ki67 (Fig. 5A-B), a marker for cell proliferation, or hcTnT and phosphorylated histone 3 (PH3) (Fig. 5A-C), which is present during the M-phase of the cell cycle, were observed in sections from the hearts of CCND2^OE animals, but were nearly undetectable in hearts treated with hiPSC-CCND2^WT CMs.

Fig. 4. hiPSC-CCND2^OE CMs replaced a substantial portion of the myocardial scar.

Tissues from the infarct zone (IZ) and the border zone (BZ) in CCND2^WT and CCND2^OE animals at month 6 after MI and treatment were (A) stained with wheat germ agglutinin (WGA), hcTnT- and HNA-specific antibodies, and DAPI (bar=20 μm). (B-D) Cells that co-expressed hcTnT and HNA were counted in 8 fields per section, 20 sections per mouse, 12 mice per group; then, (B) the number of engrafted (i.e., human-lineage) cells in both zones was determined and (C) expressed as a percentage of the total number of cells administered. (D) Data for the IZ and BZ were quantified separately. (E) The proportion of HNA-positive cells that also co-expressed hcTnT was determined at month 1 and month 6 after treatment, and expressed as a %. (F) Sections were stained with sirius red (scar) and fast green (functional myocardium), and then with antibodies against hcTnT (arrows) and non–species-specific cardiac troponin I (cTnI) to identify engrafted (hcTnT) and both engrafted and native (cTnI) cardiomyocytes; nuclei were counterstained with human nuclear antigen (HNA) and DAPI. (G) The proportion of the scarred region that was occupied by engrafted cardiomyocytes was calculated and expressed as a percentage. *P<0.01 vs CCND2^WT, ^†P<0.01 vs BZ in the same experimental group.

Fig. 5. hiPSC-CCND2^OE CMs continued to proliferate after transplantation and promoted the angiogenic response to MI.

(A) Sections from the hearts of CCND2^OE and CCND2^WT animals were obtained at month 6 after MI and stained for the expression of hcTNT (to identify engrafted hiPSC-CMs), for expression of the proliferation marker Ki67, and for the presence of the M-phase marker phosphorylated histone 3 (PH3); nuclei were identified via DAPI staining or the expression of Nkx2.5. Proliferation was quantified as the percentage of hcTnT-positive cells that also expressed (B) Ki67 and (C) PH3. (D) Sections from the infarct zone with engrafted CCND2^OE and CCND2^WT cells were stained for the expression of hcTnT and the endothelial marker isolectin B4 (IB4), nuclei were counterstained with human nuclear antigen (HNA) and DAPI, and (E) vessel density was quantified as the number of IB4-positive vascular structures per mm². (F) Sections from the border zone were stained for the expression of α-sarcomeric actin (αSA) and IB4; nuclei were counterstained with DAPI, and (G) vessel density was quantified as the number of IB4-positive vascular structures per mm². (H) Measurements in the remote zone (RZ) in the same sections were used as a positive control. *P<0.01 versus CCND2^WT, ^†P<0.01 versus MI.

Vascular structures that expressed the endothelial marker IB4 were significantly more common in the infarcted zone (Fig. 5D-E) and border zone (Fig. 5F-G) of CCND2^OE animal hearts than in the corresponding region of hearts from CCND2^WT animals, while no differences were found in the remote zones of Sham, MI, and CCND2^WT- or CCND2^OE-treated animal hearts (Fig. 5H).

Cardiac hypertrophy was evaluated by histology. Analysis of sections stained for the expression of (non–species-specific) cardiac troponin I (cTnI) indicated that although the cross-sectional surface areas of border zone cardiomyocytes were significantly smaller in CCND2^OE animals than in the MI or CCND2^WT groups, measurements remained significantly larger in all three groups that underwent MI induction than in Sham animals (Fig. 6A-B).

Fig. 6. Engrafted hiPSC-CCND2^OE CMs attenuated cardiomyocyte hypertrophy.

(A) Cardiomyocyte cross-sectional surface areas were measured in sections obtained at month 6 from the border-zone (BZ) of the infarct in MI, CCND2^OE, and CCND2^WT animals, and from the corresponding regions of hearts in Sham animals; sections were stained with wheat germ agglutinin (WGA) and cardiac troponin I (cTnI) to visualize cardiomyocytes, and nuclei were counterstained with DAPI. (B)The minimal diameters of cardiomyocyte fibers (MFD) were quantified. *P<0.01 versus Sham, ^†P<0.05 versus MI, ^‡P<0.05 versus CCND2^WT.

Collectively, these observations suggest that although the initial engraftment rate for both hiPSC-CCND2^OE and hiPSC-CCND2^WT CMs was low with the same viability before intramyocardial injection (Supplemental Fig. 1C), the surviving hiPSC-CCND2^OE CMs proliferated and promoted angiogenesis, which produced new cardiac muscle that gradually replaced the scar tissue in the infarcted region, and attenuated cardiac hypertrophy post MI. Notably, the proliferative activity observed in hiPSC-CCND2^OE CMs was not associated with evidence of cardiac tumorigenesis or an increase in thymoma formation.

3.3. hiPSC-CCND2^OE CMs became mature during the 6-month period after transplantation.

Previously, we reported that one month after transplantation in MI murine hearts, engrafted CCND2^OE CMs displayed an immature phenotype as demonstrated by disorganized spatiotemporal localization of adherens junctions (N-Cadherin staining) and gap junction (Connexin 43 staining), the lack of significant transverse (T)-tubule formation (Caveolin-3), and the low level expression of cardiac troponin I subunit (cTnI) [12]. Here, we found that 6 months after transplantation, engrafted hiPSC-CCND2^OE CMs displayed patterns of expression for cTnI (Supplemental Fig. 2A), N-Cadherin (Supplemental Fig. 2B), Connexin 43 (Supplemental Fig. 2C), and Caveolin 3 (Supplemental Fig. 2D) that were similar to the patterns observed in cardiomyocytes from the non-infarcted region. The size of grafted hiPSC-CCND2^OE CMs increased (Supplemental Fig. 2E-2F) over the 6 month period, and their nuclei density decreased (Supplemental Fig. 2G). Thus, the hiPSC-CCND2^OE CMs appeared to have matured to a greater extent during this 6-month study than in our previous report [12], which was checked 4 weeks after cell transplantation. However, the engrafted cardiomyocytes (i.e., those that expressed hcTnT and HNA) in CCND2^OE animals were significantly smaller than the native cardiomyocytes in Sham animals (Supplemental Fig. 2E-2F), and the density of cardiomyocyte nuclei was greater in CCND2^OE animals than in the Sham group (Supplemental Fig. 2G). To further characterizd the structure of the grafts. We tested the cell components of the grafted area and found that the ECs, SMCs and fibroblasts all came from the host (HNA negative). The structure components are shown in supplemental Fig. 2H.

3.4. Engrafted hiPSC-CCND2^OE CMs were electrically coupled to the native myocardium.

N-Cadherin (Fig. 7A) and Connexin 43 (Fig. 7B), the two major proteins in the cardiomyocyte intercalated disks, were readily detectable between the engrafted and native cardiomyocytes of CCND2^OE animals. The presence of N-Cadherin and Connexin 43 immune reactivity at junctional complexes between adjacent cardiomyocytes implicated the possibility that the engrafted human myocardium might participate in a functional syncytium with host myocardium. To directly test this hypothesis, the electrical activation of the hearts of cell-treated animals was monitored via optical mapping of membrane potential. Hearts (Fig. 7C) were explanted, Langendorff-perfused, and electrically paced at variable cycle length (CL) while action potentials (APs) were recorded over the surface of the left ventricle with a 16×16 array of photodiodes (Fig. 7D). During pacing at CL of 70 ms, AP durations (to 50% recovery [APD₅₀]) in the non-infarcted regions were 20–40 ms, which is consistent with measurements in normal mouse hearts. However, AP durations were ~4-fold greater (80–160 ms, Fig. 7E) in the infarcted regions of hearts from 6 of 7 CCND2^OE animals and from 3 of 8 CCND2^WT animals. Such AP duration is similar to APD₅₀ measured previously in hiPSC-derived cultured cardiac [17]. No long action potentials were observed in the infarcted hearts of MI or Sham animals. When long APs were present, they occurred over a larger surface area in hearts treated with hiPSC-CCND2^OE CMs (3.4±3.3 mm², range: 0–9.1 mm²) than with hiPSC-CCND2^WT CMs (0.53±0.9 mm², range: 0–2.2 mm²). Long APs from infracted regions contained one large upstroke and a variable number of small spikes, which depended on the pacing CL. At the 70 ms pacing CL (Fig. 7F), each large upstroke was followed by 2–3 small spikes. All small spikes had short durations similar to the AP duration of the host tissue outside of the infracted area, were synchronized with activation of the host heart, and their propagation followed activation of the non-infarcted host tissue (Fig. 7G, beat 2) indicating that these short spikes were APs produced by host cells. The large upstrokes contained 2 rising phases (most obvious at locations 2,3,7,8 in Fig.7F), the first of which had an amplitude similar to amplitudes of the following short spikes, indicating that this first rise was caused by activation of host cells. The second rise was followed by a long AP; therefore, it can be attributed to activation of transplanted cells. Large upstrokes were aligned with every third or fourth stimulus, and this relationship was observed in successive beats producing patterns of 3:1 or 4:1 rhythm transformation. At some locations (sites 6–8), rhythm transformation was unstable due to rapid excitation rate. However, pacing with a longer CL (110 ms) and spontaneous sinus rhythm (CL=170 ms) exhibited stable 2:1 rhythm transformation pattern where every large upstroke was followed by one small upstroke (Supplemental Fig. 3). Propagation of both large and small upstrokes was continuous, and followed activation spread over the non-infarcted region (Fig. 7G). Collectively, these observations indicate that the long APs originated from the transplanted cells, that they were triggered by APs from the non-infarcted region, and consequently, that the transplanted hiPSC-CCND2^OE CMs were electrically coupled to native cardiomyocytes in the non-infarcted region; the differences in APD₅₀ between the transplanted hiPSC-CMs and native CMs can likely be attributed to the cells’ lineages of origin, because human cardiomyocytes and hiPSC-derived myocytes have much longer APD, and beat much slower [17] than murine cardiomyocytes.

Fig. 7. Engrafted hiPSC-CCND2^OE CMs were electrically integrated with the native myocardium.

Heart sections were obtained at month 6 from CCND2^OE animals and stained for the expression of hcTNT, cTnI, and/or HNA, and for the presence of (A) the adherens-junction protein N-Cadherin (N-Cad) or (B) the gap-junction protein Connexin 43 (Cx43); nuclei were counterstained with DAPI. Boxed regions display native cardiomyocytes (CMs) in the non-infarcted region (top rows), native CMs and hiPSC-CCND2^OE CMs near the border zone (middle rows) and engrafted hiPSC-CCND2^OE CMs in the infarcted zone (bottom rows). (C) The hearts of the animals were harvested immediately after sacrifice at month 6, Langendorff-perfused, stained with voltage-sensitive dye RH-237, and electrically paced via a bipolar electrode. (D) Optical action potentials were recorded over the boxed region of the left ventricle shown in panel C with a 16×16 photodiode array during pacing at a cycle length of 70 ms. The site of coronary-artery ligation is visible as a dark mark on the surface of the heart and as the shaded region in the AP map. (E) Heat map shows spatial distribution of action potential duration (APD₅₀) over the heart surface. (F) Selected traces of action potentials displayed in blue in panel D (numbered from the upper right to the lower left) are shown. Closed red and green dots indicate the beginning and end-points, respectively, used for APD₅₀ calculation. Open red circles depict upstrokes of short action potentials overlaying long action potentials of the engrafted hiPSC-CCND2^OE cells. (G) Action potential propagation (arrows) is illustrated for two consecutive beats that include (beat 1) or do not include (beat 2) the upstrokes of abnormally long action potentials present in engrafted hiPSC-CCND2^OE CMs.

4. Discussion

The proliferative capacity of large-mammalian cardiomyocytes declines precipitously during the perinatal period [18]. However, if MI is induced in 1- to 2-day old swine, the surviving cardiomyocytes can reenter the cell cycle and regenerate the damaged tissue [19]. This observation is consistent with a clinical case report of a newborn infant whose heart completely recovered from a massive MI that occurred at birth [20], and suggests that cardiomyocyte cell cycle re-activation may be a viable strategy for improving recovery from myocardial injury in adult patients. Strategies for re-activating the cell cycle in cardiomyocytes [16] include treatment with hormones [21], growth factors [22], microRNAs [23], or small molecules [24], and we have shown that CCND2 overexpression activates cell cycle progression in hiPSC-CMs, and increases the cells’ potency for myocardial repair over a 4-week period in mice [12]. However, when the reparative potency of cardiomyocytes derived from human embryonic stem cells (hESC-CMs) was evaluated in a murine MI model, the benefits observed four weeks after treatment were no longer apparent at week 12 [5], so the long-term fate of transplanted hiPSC-CCND2^OE CMs remains uncertain. Thus, the results presented here are the first to demonstrate that a small number of engrafted hiPSC-CCND2^OE CMs can proliferate and grow in the hearts of mice with MI, thereby reducing infarct size by more than 50% and improving cardiac function for up to six months after administration.

More than 90% of transplanted cells either exit the site of administration or are lost to the cytotoxic environment of infarcted tissue within three weeks of delivery [5]. This low rate of engraftment is considered the primary limiting factor of the effectiveness of cell therapy [25, 26]. Engraftment increased significantly by injecting the cells through a patch of biomaterial to improve retention at the administration site [27], and by promoting the survival of engrafted cells via the co-administration of free-radical scavengers [28] or a pro-survival cocktail [29, 30]. However, none of these reports demonstrate a significant remuscularization of the LV scar, and the long-term engraftment rates with transplanted hiPSC-CMs have been somewhat inconsistent. Funakoshi, et al. [13], reported that the number of engrafted cells increased during the first two months after transplantation, whereas the engraftment rate appeared to decline in studies by Ong, et al [31]. Both investigations were performed in mice, and engraftment was evaluated via luciferase activity; thus, we corroborated the results from the in-vivo luciferase measurements reported here by evaluating the number of transplanted cells that were present in histological sections from the hearts of CCND2^OE mice. The results from both analyses indicated that the number of engrafted hiPSC-CCND2^OE CMs increased substantially from month 1 to month 6 after MI injury and treatment. It is somewhat surprising to observe that the engrafted hiPSC-CCND2^OE CMs were more common in the infarcted zone than in the border zone (Fig. 4D). Although the underlying mechanism remains elusive, the finding confers significant impact as the significant remuscularization of the LV scar for the first time. It has been shown that transplanted hiPSC-CMs activate the paracrine mechanisms that protect native cells from early onset of cell death, and enhance the angiogenic response to MI injury in mice [32–34]. The higher engraftment rate observed in hiPSC-CCND2^OE CM-treated hearts may lead to an increased activation of paracrine mechanisms, such as the increase in angiogenesis (Fig. 5D-H), which further contribute to the improvements in LV remodeling and LV chamber function. It is worthy to note that, comparing to engrafted hiPSC-CCND2^OE CMs at month 1 [12], the cell cycle of these cells at month 6 significantly decreased as demonstrated by the prevalence of Ki67+ hiPSC-CCND2^OE CMs (25.06±4.64% vs. 0.65±0.05%) and PH3+ hiPSC-CCND2^OE CMs in CCND2^OE animals (4.21±1.32% vs. 0.06±0.01%; Fig. 5A-C). The loss of their cell cycle activity is accompanied with a gradual maturation of these cells over a 6-month period in vivo (Supplemental Fig. 2). Our previous observation indicated that hiPSC-CMs gradually lose their telomerase activity during 6 months of culture [12]. It is possible that reduced telomerase activity may contribute to the loss of self-renewal capacity of transplanted hiPSC-CCND2^OE CMs in 6 month grafts. Further experiments are warranted to demonstrate time course of the cell dividing and telomerase activity of hiPSC-CCND2^OE CMs. The normal cardiomyocytes go on for one more cell cycle within the first few days after birth [35]. Further myocyte karyokinesis results in a majority of binucleated cardiomyocytes [36]. The frequency of mononuclear cardiomyocytes in adult hearts varies among mammalian species, and ~78% of cardiomyocytes in postnatal human hearts are mononucleated [37, 38]. The majority (~98%) of engrafted hiPSC-CCND2^OE CMs contained just one nucleus at 6 months post-differentiation (Supplemental Fig. 1B). It was reported that mononuclear cardiomyocytes are capable of entering into the cell cycle more efficiently than binucleated cardiomyocytes in mice [39]. However, further studies are warranted to confirm the correlation between the status of cardiomyocyte nucleation and their regenerative potency.

Immunodeficient mice are a useful xenograft model for preclinical studies. The NOD/SCID mice were derived from hybridization of SCID mice and NOD/Lt mice, and characterized for the severe combined immune deficiency with an absence of three major immune cells (T-cell, B-cell, NK-cell), and circulating complements, macrophage cells and antigen processing cells [40]. Xenografts are more successful in NOD/SCID mice since their immune system is more defective than orthotopic models. Using this model, we have demonstrated that activation of the cell cycle in grafted hiPSC-CMs via CCND2 overexpression (hiPSC-CCND2^OE CMs) leads to large grafts, which repopulates more than 50% of LV scar in MI mice. To guide the use of this strategy in the clinics, future comparative studies of the regenerative potency of hiPSC-CCND2^OE CMs in immunocompetent large animals such as pigs or NHP models, shall be tested.

Lasting improvement in the performance of infarcted hearts is unlikely to occur unless the myocardial scar is replaced with functional cardiac muscle [41]. Some degree of remuscularization (~10% of the infarct zone) [42] has been reported in primates after treatment with hESC-CMs; however, animals in both the cell-treatment and control groups developed arrhythmic complications, which are perhaps the most critical concern associated with cardiac cell therapy. The arrhythmias were partially attributable to the size of the infarcts, which were larger than those in a previous primate study [43]; nevertheless, monkeys with smaller infarcts also developed arrhythmias after hESC-CMs administration, but not after treatment with the control vehicle. The number of cells administered (0.75–1.0×10⁹ per animal) was exceptionally large in both studies, which led to regions of engrafted cells that were up to 10-fold larger than those observed in most rodent models [43]. Furthermore, clusters of cardiomyocytes may be less electromechanically integrated when generated via the aggregation of injected cells than via the self-replication of a parent cell, which may explain, at least in part, how engrafted cells could remain electromechanically coupled to the native myocardium of the CCND2^OE animals studied here, despite occupying more than 50% of the infarct zone. Arrhythmias could also be caused by the different action potential durations in engrafted hiPSC-CMs and native murine cardiomyocytes, but this difference can be attributed to the cells’ species of origin and is likely to be less prominent in patients or animals with human-like heart rates.

Limitations.

Normal cardiac muscle is characterized by collagen fibers that are oriented in a unique direction depend on the specific position from epicardium to endocardium and from base to apex. Along these extracellular matrix, cardiomyocytes and vessels form the functional syncytium that are important basis for the normal myocardial perfusion, electro-mechanical coupling, as well as LV rhythmic contraction. The significant regenerated muscle grafts in the present study (Fig. 4 and Fig. 7), although are supported by the rich spouting of the preexisting vessels (Fig. 5), are by no means equal to the naïve cardiac muscles in terms of normal myocardial fiber direction, coronary perfusion and ECM. These structural characteristics of muscle grafts are important and warrant future studies.

In conclusion, the results of the present study demonstrate for the first time, that “remuscularization” of hearts with myocardial infarction can be achieved by turning back the cell cycle of cardiomyocytes. The activation of cardiomyocyte proliferation by overexpressing CCND2 in the hiPSCs before differentiation can result in replacement of more than 50% of the myocardial scar over the ensuing six months. The large graft of hiPSC-CCND2^OE CMs also electrically integrated with the recipient myocardium, which was accompanied by significant improvements in LV function. The future preclinical studies using large animal model with non-viral approaches to induce reactivation of myocyte cell cycle, are warranted for achieving the ultimate clinical objective of remuscularization of LV infarcts in patients with post infarction LV remodeling.

HIGHLIGHT:

• We have demonstrated that remuscularization of a LV infarct can be achieved by turning back the cell cycle activity of cardiomyocytes.

• The activation of cardiomyocyte proliferation by overexpressing CCND2 in can result in replacement of more than 50% of the myocardial scar over the ensuing six months

• The grafted hiPSC-CMs electromechanically integrate with the host myocardium, suggesting an ultimate clinical objective of remuscularizing the injured heart in patients with myocardial infarction.

Abbreviations

hiPSCs

Human induced pluripotent stem cells

hiPSC-CMs

Human induced pluripotent stem cell-derived cardiomyocytes

hiPSC-CCND2^OE

CMs hiPSC-CMs overexpressing CCND2

hiPSC-CCND2^WT

CMs hiPSC-CMs with endogenous levels of CCND2

IZ

Infarct zone

BZ

Border zone

RZ

Remote zone

LAD

Left anterior descending artery

LV

Left ventricular

MI

Myocardial infarction