Induction of Cardiomyocyte-Like Cells in Infarct Hearts by Gene Transfer of Gata4, Mef2c, and Tbx5Novelty and Significance
Rationale: After myocardial infarction (MI), massive cell death in the myocardium initiates fibrosis and scar formation, leading to heart failure. We recently found that a combination of 3 cardiac transcription factors, Gata4, Mef2c, and Tbx5 (GMT), reprograms fibroblasts directly into functional cardiomyocytes in vitro.
Objective: To investigate whether viral gene transfer of GMT into infarcted hearts induces cardiomyocyte generation.
Methods and Results: Coronary artery ligation was used to generate MI in the mouse. In vitro transduction of GMT retrovirus converted cardiac fibroblasts from the infarct region into cardiomyocyte-like cells with cardiac-specific gene expression and sarcomeric structures. Injection of the green fluorescent protein (GFP) retrovirus into mouse hearts, immediately after MI, infected only proliferating noncardiomyocytes, mainly fibroblasts, in the infarct region. The GFP expression diminished after 2 weeks in immunocompetent mice but remained stable for 3 months in immunosuppressed mice, in which cardiac induction did not occur. In contrast, injection of GMT retrovirus into α-myosin heavy chain (αMHC)-GFP transgenic mouse hearts induced the expression of αMHC-GFP, a marker of cardiomyocytes, in 3% of virus-infected cells after 1 week. A pooled GMT injection into the immunosuppressed mouse hearts induced cardiac marker expression in retrovirus-infected cells within 2 weeks, although few cells showed striated muscle structures. To transduce GMT efficiently in vivo, we generated a polycistronic retrovirus expressing GMT separated by 2A “self-cleaving” peptides (3F2A). The 3F2A-induced cardiomyocyte-like cells in fibrotic tissue expressed sarcomeric α-actinin and cardiac troponin T and had clear cross striations. Quantitative RT-PCR also demonstrated that FACS-sorted 3F2A-transduced cells expressed cardiac-specific genes.
Conclusions: GMT gene transfer induced cardiomyocyte-like cells in infarcted hearts.
Cardiomyocytes are terminally differentiated cells with limited regenerative capacity in adult heart. After myocardial infarction (MI), cardiac fibroblasts (CF), which account for more than half of the cells in heart, proliferate, synthesize extracellular matrix, and undergo fibrosis, leading to cardiac dysfunction.1,2 The large population of endogenous CF might be a potential source of cardiomyocytes for regenerative applications if they could be readily and reliably reprogrammed into functional cardiomyocytes.
We recently found that a combination of 3 cardiac-specific transcription factors, Gata4, Mef2c, and Tbx5 (GMT), could reprogram mouse postnatal CF and dermal fibroblasts directly into functional cardiomyocytes in vitro.3 The induced cardiomyocytes (iCMs) were similar to neonatal cardiomyocytes in global gene expression profile and their ability to contract spontaneously. Zhou et al4 demonstrated that adenoviral gene transfer of the 3 β-cellenspecific transcription factors into mouse pancreas reprogrammed pancreatic
In This Issue, see p 1107
Editorial, see p 1108
exocrine cells into functional β cells. Finally, local delivery of MyoD, a master regulator of skeletal myocytes, into cryo-injured hearts induced skeletal muscle differentiation in healing heart lesions.5 In this study, we investigated whether gene transfer of GMT into infarcted hearts could similarly induce cardiomyocyte generation in vivo.
Transgenic mice overexpressing green fluorescent protein (GFP) under the control of an α-myosin heavy chain promoter (αMHC) were generated as described previously.3 Eight-week-old male wild-type (WT) ICR and nude mice were obtained from CLEA Japan (Tokyo, Japan). The Keio University Ethics Committee for Animal Experiments approved all experiments in this study.
Construction of Retroviruses
The pMXs retroviral vectors containing GFP, DsRed, Gata4, Mef2c, or Tbx5 were generated as described.3 The 3F2A construct was designed and generated by a custom gene synthesis service with codon optimization (Genscript Japan, Tokyo, Japan). 3F2A was subcloned into the pEnter-d-TOPO vector by PCR and subsequently cloned into pMXs-Gw using the Gateway system (Invitrogen) to generate pMXs-3F2A. The pMXs retroviral vectors were transfected into Plat-E cells with Fugene 6 (Roche) to generate retroviruses. Virus-containing supernatants were collected after 48 hours.
In Vivo Gene Transfer
Pooled virus-containing supernatants were collected, filtered through 0.45-mm pore membranes, concentrated by 10- to 100-fold with centrifugation (8000g for 16 hours), and then resuspended in phosphate-buffered saline supplemented with 4 μg/mL polybrene. To generate our MI model, mice were intubated and anesthetized with isoflurane gas. The chest cavity was exposed by cutting the intercostal muscle, and the left coronary artery was then ligated with a 7–0 silk suture, as described previously.6 Immediately after the coronary artery ligation, 30 μL of virus-containing solution was injected into the boundary between the infarct and border zone at one site with a 32-gauge needle. For infection into dermal fibroblasts, mice were intubated and anesthetized with isoflurane gas, the skin was cut, and 100 μL of virus-containing solution was then injected. In some experiments, virus solution was administered after 3 and 7 days of MI. To determine the effect of immunosuppressants, mice were treated with cyclosporine A (Sigma) by intraperitoneal injection for 7 days at dosages of 10 to 20 mg/kg per day before viral delivery and coronary artery ligation. The cyclosporine A treatment was continued until analyses.7 The surgeon was blinded to the study and mortality after MI was <10%.
Viral Transduction Into Cultured Cardiac Fibroblasts
The αMHC-GFP mice were subjected to MI by coronary artery ligation and euthanized after 3 or 7 days. The hearts were excised and the infarcted area was dissected from surrounding normal myocardium under direct visualization. The explants were minced into small pieces less than 1 mm3 and cultured for 10 days in explant medium (IMDM/20% FBS) on gelatin-coated dishes. Migrated fibroblastic cells were harvested and filtered with 40-μm cell strainers (BD) to avoid contamination with heart tissue fragments. The αMHC-GFP–/Thy1+ cells were FACS-sorted and plated at a density of 104/cm2 for the retrovirus transduction. After 24 hours of infection, the medium was replaced with DMEM/M199 medium and changed every 2 to 3 days.
Genomic DNA PCR
Genomic DNA was extracted from mouse hearts using a standard protocol, and PCR was performed using the following primers: GFP sense, 5′-TGAACCGCATCGAGCTGAAGGG-3′; GFP antisense, 5′-TCCAGCAGGACCATGTGATCGC-3′; GAPDH sense, 5′-AACTTTGGCATTGTGGAAGG-3′; GAPDH antisense 5′-ACA CATTGGGGGTAGGAACA-3′.
FACS Analyses and Sorting
For αMHC-GFP and GFP expression analyses, cells were harvested from cultured dishes and analyzed on a FACS Calibur (BD Biosciences) with FlowJo software. For GFP+ cell sorting, mouse cardiac tissues were digested with collagenase and sorted using a FACS Aria (BD Biosciences). For GFP+/Thy1+ cell expression analyses and αMHC-GFP-/Thy1+ cell sorting, cells were incubated with APC-conjugated anti-Thy1 antibody (eBioscience) for 30 minutes before FACS.3
RNA Extraction and Quantitative RT-PCR
Mock- or GMT-transduced cardiac fibroblasts were collected from culture dishes by incubation with trypsin-EDTA, and the GFP- or 3F2A/GFP-transduced cells in hearts were collected by FACS. RNA was extracted from the cells using Pico-Pure RNA Isolation (Arcturus), and quantitative RT-PCR was performed as described previously,3 using the following TaqMan probes: Actc1 (Mm01333821_m1), Ryr2 (Mm00465877_m1), Nppa (Mm01255748_g1), and Tnnt2 (Mm00441922_m1) (Applied Biosystems). The mRNA levels were normalized by comparison to GAPDH mRNA.
Western Blot Analysis
Lysates were prepared by homogenization of cells in RIPA buffer and run on SDS-PAGE to separate proteins prior to the immunoblot analyses as described.8 After transfer to nitrocellulose membranes, immunodetection was performed with antibodies to Gata4, Mef2c, and Tbx5, followed by the appropriate HRP-conjugated secondary antibodies (Cell Signaling Technology). The antibody-bound proteins were visualized by chemiluminescence detection (ECL, Amersham).
Hearts were fixed in 0.4% paraformaldehyde overnight and then embedded in OCT compound for freezing in liquid nitrogen. Hearts were cut vertically into 7-μm sections to show both ventricles. Sections were stained with primary antibodies against α-actinin (Sigma Aldrich), vimentin (Progen), GFP (Invitrogen), cTnT (Thermo Scientific), collagen1 (Millipore), CD31 (BD Biosciences), and RFP (Rockland) and then with secondary antibodies conjugated with Alexa 488, 546, or 633 and DAPI (Invitrogen). GFP+ cells per area were calculated as the ratio between the number of GFP+ cells and the myocardial area, measured by Image J software. The α-actinin+ or αMHC GFP+ cells relative to the total number of GFP+ or DsRed+ cells were counted in 3 randomly selected fields per section. The data for each mouse were calculated from 20 to 35 sections, and we observed 4 to 5 mice in each group. The measurements and calculations were conducted in a blinded manner. All confocal microscopy was carried out on an LSM 510 META microscope (Carl Zeiss), and Z-stack images were collected according to the standard protocol.
Cells were fixed in 4% paraformaldehyde for 15 minutes at room temperature, blocked, and then incubated with primary antibodies against α-actinin, GFP, cTnT, vimentin, CD31, SMA (Sigma), collagen 1, Mef2c (Aviva Systems Biology), Gata4 (Santa Cruz Biotechnology), ANP (Chemicon), Nkx2.5 (Santa Cruz), or SM-MHC (Biomedical Technologies), with secondary antibodies conjugated to Alexa 488 or 546. Finally, cells were stained with DAPI. The percentage of cells immunopositive for α-actinin, GFP, cTnT, and ANP were counted in 6 randomly selected fields in triplicate, and 500 to 1000 cells were counted in total.
Differences between groups were examined for statistical significance using Student t test or ANOVA. Probability values <0.05 were regarded as significant.
Retroviral Gene Transfer Into Mouse Hearts After MI
Coronary artery ligation was used to generate MI in the mouse (Figure 1A and Online Figure IA). To determine the appropriate vector for targeting resident CF, we injected a GFP retrovirus or adenovirus into the WT mouse hearts immediately after MI. One week later, the GFP retrovirus efficiently infected the infarct region (Figure 1B and Online Figure IB). Immunohistochemisty demonstrated that retrovirus-transduced GFP+ cells in the infarct-border regions were immunopositive for vimentin and collagen 1, markers of fibroblasts, but not for sarcomeric α-actinin, a marker of cardiomyocytes (Figure 1C through 1E and Online Figure IC). In contrast, adenovirus preferentially infected cardiomyocytes (Online Figure ID and E).
Immunohistochemisty and genomic DNA PCR analyses for GFP demonstrated that the retrovirus infection efficiency was significantly increased in a dose-dependent manner (Figure 1F through 1H), and we thereafter used 100-fold concentrated retrovirus for gene transfer. FACS analyses demonstrated that 0.9% of heart cells expressed GFP after 1 week of GFP retrovirus injection and more than half of them expressed Thy-1, a cell surface marker of fibroblasts (Figure 1I).1,9 We confirmed that vimentin+ fibroblasts but not α-actinin+ cardiomyocytes were infected by 100-fold concentrated GFP retrovirus. The z-axis confocal projection demonstrated that vimentin and GFP were expressed in the same cells (Figure 1J and 1K and Online Figure IF). These results indicated that retrovirus infects noncardiomyocytes, mainly CF, in infarcted hearts.
Retrovirus-Infected Cells Were Reduced in Immunocompetent Mice But Maintained in Immunosuppressed Mice After MI
To determine the time course of viral gene expression, we performed immunohistochemistry after GFP retroviral delivery into the WT mouse hearts. GFP+ cells were apparent at 1 week but were reduced 2 weeks after MI and barely detectable after 4 weeks (Figure 2A and 2B). GFP retrovirus injected into the WT hearts at 3 and 7 days after MI also revealed few GFP+ cells by 2 weeks after the injection in either case (data not shown). To investigate the effect of a reduced immune response, we injected GFP retrovirus into immunosuppressed nude mouse hearts immediately after MI. Notably, GFP expression remained stable 2 weeks later and was present up to 3 months after injection (Figure 2A and 2B). We also analyzed virus DNA insertion in WT and nude mouse hearts by genomic DNA PCR at 1, 2, and 4 weeks after injection of GFP retrovirus. The GFP DNA was present at 1 week but reduced at 2 and 4 weeks after injection into WT mouse hearts, in accordance with the immunohistochemistry data. In contrast, GFP DNA was clearly detected in the nude mouse hearts at 1, 2, and 4 weeks after the injection (Figure 2C). To also investigate the effect of immunosuppressive treatment on the maintenance of virus-transduced cells, we injected GFP retrovirus into the WT hearts under treatment with an immunosuppressant, cyclosporine A.7 Despite increasing the dosage of cyclosporine A up to toxic levels, GFP+ cells were still reduced by 2 weeks after the retroviral injection (Figure 2D through 2F). These results suggested that the retrovirus efficiently infected CF until 1 week after MI, but thereafter the transduced cells and viral gene expression were reduced in immunocompetent mice.
Cardiac Gene Induction by Gata4/Mef2c/Tbx5 Transduction In Vitro
To address whether GMT could induce cardiac gene activation in CF, we used αMHC promoter-driven EGFP (αMHC-GFP) transgenic mice, in which only cardiomyocytes express GFP3 (Figure 3A). Immunopositivity for αactinin but negative staining for vimentin and CD31 demonstrated that cardiomyocytes expressed αMHC-GFP in the adult mouse hearts (Figure 3B and Online Figure IIA). We also confirmed that dissociated αMHC-GFP+ cells comprised cardiomyocytes but not smooth muscle cells, endothelial cells, or fibroblasts (Online Figure IIB), as demonstrated in the previous report.3
For the in vitro experiments, we obtained Thy1+/αMHC-GFP− cells from the infarct region at 3 and 7 days after MI by FACS sorting (Figure 3C). These cells expressed fibroblast markers, vimentin, and collagen 1 but not cardiomyocyte (α-actinin, cTnT, and Nkx2.5), smooth muscle cell (SM-MHC), or endothelial cell (CD31) markers (Figure 3D and Online Figure IIC). The antibody immunoreactivities were confirmed in a positive control (Online Figure IID). The transduction efficiency of CF obtained from the infarcted hearts was reduced to 30% to 60%, compared with mt90% in CF from intact hearts (Online Figure IIE). We next transduced CF with a pool of GMT or DsRed retrovirus to determine the cardiac induction. αMHC-GFP+ cells were not detected in CF infected with DsRed, but 3% to 7% of GMT-transduced cells expressed αMHC-GFP after 1 week of culture (Figure 3E). Quantitative RT-PCR demonstrated that cardiomyocyte-specific genes Actc1 (cardiac α-actin), Nppa (natriuretic peptide precursor type A), Ryr2 (ryanodine receptor 2), and Tnnt2 (cardiac troponin T) were significantly upregulated in GMT-transduced cells (Figure 3F). Immunocytochemistry revealed the expression of cardiac proteins, α-actinin, cardiac troponin T, and atrial natriuretic peptide after 4 weeks of infection, and the induced cardiomyocytes had striated muscle structures (Figure 3G through 3I). These results indicate that GMT transduction induced cardiac gene expression in fibroblast cells obtained from infarcted hearts.
Cardiac Gene Activation by Gene Transfer of Gata4/Mef2c/Tbx5 In Vivo
To determine cardiac induction in vivo, we then injected either the mixture of GMT and DsRed or DsRed retrovirus alone directly into the αMHC-GFP mouse hearts after MI. All measurements were performed in a blinded fashion. Cells in fibrotic tissues injected with DsRed did not express αMHC-GFP, whereas approximately 3% of DsRed+ cells coinfected with GMT and DsRed expressed αMHC-GFP after 1 week of gene transfer (Figure 4A and 4B). Three-dimensional analyses (Z-stack imaging) confirmed that DsRed and αMHC-GFP were expressed in the same cells (Figure 4C). These results indicate that gene transfer of GMT induced cardiac gene activation in a subset of cells in infarcted hearts.
Because retrovirus-transduced cell numbers were reduced after 2 weeks of gene transfer in immunocompetent mice, we used nude mice for further analyses. We injected either a pool of GMT and GFP or GFP alone into the nude mouse hearts after MI. The cells transduced with GFP did not express α-actinin, whereas approximately 1% of GFP+ cells cotransduced with GMT and GFP expressed α-actinin after 2 weeks (Figures 4D and 5G). We found that approximately 15% of the induced α-actinin+/GFP+ cells had striated muscle structures (Figures 4E and 5H). Three-dimensional analyses confirmed that GFP and α-actinin were expressed in the same cells (Figure 4F). Despite an incubation period extending up to 1 month, most α-actinin+/GFP+ cells were smaller than endogenous ventricular cardiomyocytes and did not have clear cross striations.
Induction of Cardiomyocyte-Like Cells in Infarcted Hearts by a Single Polycistronic Vector Expressing Gata4/Mef2c/Tbx5
It is possible that cells transduced in vivo by 3 separate vectors might not receive all 3 genes at sufficient levels for full cardiac reprogramming. To address this, we generated a polycistronic retrovirus expressing GMT at near equimolar levels from the same promoter using “self-cleaving” 2A peptides (3F2A) (Figure 5A). Western blotting demonstrated that all 3 genes (Gata4/Mef2c/Tbx5) were expressed in the cells transfected with 3F2A (Figure 5B). In addition, approximately 70% of fibroblasts transduced with 3F2A retrovirus expressed both Gata4 and Mef2c, suggesting concomitant expression of GMT in fibroblasts (Figure 5C and 5D). We then injected 3F2A and GFP retroviruses directly into the nude mouse hearts after MI. Four weeks after the gene transfer, we found α-actinin+/GFP+ cells in the infarct area, of which 30% showed cross striations (Figure 5E, 5G, and 5H). GFP and α-actinin were expressed in the same cells by 3-dimensional analyses (Figure 5F). In addition, cTnT, a specific marker of cardiomyocytes, was expressed in the 3F2A-transduced cells (Figure 5I). Next, to determine the cardiac gene induction more quantitatively, we FACS sorted GFP+ cells from infarcted hearts after 1 week of gene transfer with 3F2A/GFP or GFP and analyzed mRNA expression in the GFP+ cells (Figure 5J). Quantitative RT-PCR demonstrated that cardiac-specific genes Actc1, Nppa, and Tnnt2 were significantly upregulated in 3F2A/GFP-transduced cells compared with GFP alone, suggesting that cardiac genes were induced by the in vivo 3F2A gene transfer (Figure 5K). Importantly, we did not find tumor formation in the transduced mouse hearts. Together, these results demonstrate that a subset of cells in cardiac granulation tissue is differentiated into cardiomyocyte-like cells by GMT, provided that the 3 genes are appropriately introduced.
Our results provide evidence that GMT gene transfer induces cardiomyocyte-like cells in infarcted hearts and that expression of GMT via a polycistronic vector enhances cardiac differentiation. Several lines of evidence suggest that the αMHC-GFP+/DsRed+ cells and α-actinin+/GFP+ cells that we describe in the current study are not residual cardiomyocytes but new cardiomyocyte-like cells induced by the GMT gene mix. First, we found that retrovirus infecting only proliferating cells transduced noncardiomyocytes, mainly fibroblasts, in the infarcted heart region, consistent with previous reports.10–12 Second, control heart showed no αMHC-GFP+/DsRed+ cells or α-actinin+/GFP+ cells. Third, high in vitro efficiency of cardiac induction from isolated CF supported the interpretation that endogenous fibroblasts generate iCMs. Finally, cardiac-specific genes were significantly upregulated in FACS-sorted 3F2A/GFP-transduced cells compared with GFP alone by quantitative RT-PCR. Nevertheless, we cannot completely exclude the possibility that rare cardiac progenitors were the origin of iCMs in vivo. To address this, we tried to analyze “fibroblast-lineage tracing mice” obtained by crossing “fibroblast-specific Cre” transgenic mice and indicator mice. However, we could not use the mice due to leaky EGFP expression in the cardiomyocytes (data not shown). We also injected retroviral GMT or 3F2A with GFP into mouse skin to convert dermal fibroblasts into cardiac cells. However, we did not observe induction of cardiac-specific genes in the transduced skin cells (Online Figure III). There are several possibilities for this result. Although we injected more retrovirus into the skin, infection efficiency was lower in the dermal fibroblasts compared with CF in infarcted hearts. Different cell types, tissue permeabilities, and injury models could account for such a difference. Moreover, skin cells might be more resistant to cardiac conversion than CF, as demonstrated in vitro.3,13–15 Given that high transduction efficiency and addition of other factors to GMT would be beneficial for cardiac conversion in vitro, other injury models, modification of the gene delivery system, and optimization of reprogramming factors might be needed for efficient cardiac induction in skin.11,15,16
During the review of this article, 2 other studies of in vivo cardiac reprogramming by transcription factors were published by 2 other independent groups.10,11 As concluded in our present study, Qian et al10 found that nonmyocytes in infarcted hearts were converted into cardiomyocyte-like cells by GMT retroviral gene transfer. In addition, Song et al11 reported that adding Hand2 to GMT (GHMT) converted CF into functional cardiomyocyte-like cells more efficiently than GMT alone in vitro and in vivo.11 Although all 3 studies (ours and the 2 others10,11) demonstrated in vivo cardiac reprogramming, the approaches used to address this issue differ. The other 2 groups used mainly fibroblast-lineage tracing mice to demonstrate cardiac conversion from CF, whereas we took an alternative approach of cotransduction with GMT plus marker genes into immunosuppressed nude mice, to determine cardiac induction among the retrovirus-infected cells.
Qian et al10 showed that approximately 35% of cardiomyocytes in the border/infarct zone were newly generated iCMs derived from CF using fibroblast-lineage tracing mice, and half of these cells had well-organized sarcomeric structures. They also transduced a marker gene with GMT and demonstrated that the cardiac conversion rate relative to the total number of GMT-infected cells was 10% to 15%, which was higher by more than 10-fold compared with our study. Their iCMs also exhibited functional characteristics of adult ventricular CMs, including cellular contraction, electrophysiological properties, and functional coupling to other cardiac cells. Notably, retroviral GMT gene transfer into mouse infarcted hearts significantly improved cardiac function and reduced the fibrotic area at 2 and 3 months after MI. Overall, the reprogramming efficiency and induction of mature iCMs by GMT were higher than those achieved in our study, which is possibly attributable to the different mouse strains, transgene expression levels, and viral titers between studies. We used immunosuppressed nude mice to demonstrate cardiac conversion because retrovirus-infected cells were dramatically reduced after 2 weeks of MI in immunocompetent mice. Consistent with this, Byun et al12 also demonstrated that retrovirus lacZ-infected cells were reduced from 14% of heart cells at 1 week to 4% at 4 weeks in immunocompetent rat cryo-injured hearts.16 Our results that viral DNA levels were reduced in WT mouse but maintained in nude mouse hearts suggest that a T cell–mediated immune response contributed to the loss of retroviral-infected cells in immunocompetent animals. It is possible that Qian et al10 achieved higher infection efficiency of fibrotic cells, because they showed that reprogrammed cells remained after 4 weeks of infection in the lineage tracing immunocompetent mice. However, it is not clear how many transduced cells remained after 4 weeks in their study, and further investigation might be needed to clarify this point.
Song et al11 showed that 2% to 6% of cardiomyocytes in the border/infarct area were newly generated cardiomyocyte-like cells with clear striations and functional properties similar to endogenous ventricular cardiomyocytes, using fibroblast-lineage tracing mice. However, because they did not transduce a marker gene with GHMT, the cardiac conversion rate relative to the total number of GHMT-infected cells remains unclear. Song et al11 also demonstrated that cell fusion events were unlikely for iCMs generated using αMHC-MerCreMer/Rosa26-LacZ mice, in which endogenous cardiomyocytes could be chased by pulse-labeling with tamoxifen. Finally, they demonstrated that cardiac function improved from 2 weeks and was sustained until at last 12 weeks by GHMT gene transfer in the MI mouse. The ejection fraction was also increased by 2-fold in GHMT-treated mice compared with controls by MRI (57% versus 28%), and the scar size was decreased to half (20% versus 40%). However, the mechanisms of such functional improvements were undetermined.
We found that separate GMT gene transfer induced mostly immature α-actinin+ cells in vivo with a conversion rate of approximately 1%. It is possible that the inflammatory environment generated after MI inhibited proper cardiac gene induction.2 Another possibility is that the 3 genes were not all sufficiently transduced into cells, due to the low infection efficiency in vivo. High doses of MyoD are required to induce skeletal muscle differentiation in cardiac granulation tissue,5 whereas Carey et al18 reported that a polycistronic vector encoding Yamanaka’s 4 factors by 2A system reprogrammed fibroblasts into iPSCs with only a single proviral copy insertion. In the present study, we used a polycistronic retrovirus to transduce all 3 genes into cells in vivo and generated more mature cardiomyocyte-like cells compared with introducing 3 separate vectors. However, we did not confirm the functional properties of the induced cardiomyocyte-like cells and improvement of global heart function due to the low cardiac induction efficiency in our study. In this sense, the present data are somewhat preliminary, and further investigation and optimizations are needed in future research. Nevertheless, this is the first study demonstrating that the polycistronic 2A system can be used for cellular reprogramming in vivo, which also suggests that this system could be useful for inducing other types of cells.
Although considerable challenges remain, our findings might introduce new regenerative strategies for repairing infarcted hearts.
We are grateful to members of the Fukuda laboratory, D. Srivastava and L. Qian, for critical discussions and reagents.
Sources of Funding
M.I. was supported by research grants from JST CREST, JSPS, Banyu Life Science Foundation International, The Uehara Memorial Foundation, and SENSHIN Medical Research Foundation.
In July 2012, the average time from submission to first decision for all original research papers submitted to Circulation Research was 11.2 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.112.271148/-/DC1.
Non-standard Abbreviations and Acronyms
- polycistronic vector expressing GMT
- cardiac fibroblasts
- green fluorescent protein
- Gata4, Hand2, Mef2c, and Tbx5
- Gata4, Mef2c, and Tbx5
- induced cardiomyocytes
- myosin heavy chain
- myocardial infarction
- MI-d3 CF
- cardiac fibroblasts from 3 days after myocardial infarction
- MI-d7 CF
- cardiac fibroblasts from 7 days after myocardial infarction
- Received April 11, 2012.
- Revision received August 9, 2012.
- Accepted August 10, 2012.
- © 2012 American Heart Association, Inc.
- Cleutjens JP,
- Blankesteijn WM,
- Daemen MJ,
- Smits JF.
- Anzai A,
- Anzai T,
- Nagai S
- Srivastava D,
- Ieda M.
- Jayawardena TM,
- Egemnazarov B,
- Finch EA,
- Zhang L,
- Payne JA,
- Pandya K,
- Zhang Z,
- Rosenberg P,
- Mirotsou M,
- Dzau VJ.
- Ly H,
- Kawase Y,
- Yoneyama R,
- Hajjar RJ.
- Carey BW,
- Markoulaki S,
- Hanna J,
- Saha K,
- Gao Q,
- Mitalipova M,
- Jaenisch R.
Novelty and Significance
What Is Known?
Gata4, Mef2c, and Tbx5 (GMT) can reprogram cardiac fibroblasts from intact hearts into cardiomyocyte-like cells in vitro.
Cellular reprogramming is a potentially useful strategy for regenerating damaged organs.
A polycistronic vector is a useful system to express multiple genes from the same promoter.
What New Information Does This Article Contribute?
GMT can convert fibroblasts in infarcted hearts into cardiomyocyte-like cells in vitro and vivo.
Retrovirus-infected cells are reduced in infarcted myocardium by the immune response.
Gene transfer of a polycistronic vector encoding GMT enhances cardiac differentiation in infarcted hearts.
After myocardial infarction, cardiac fibroblasts proliferate and synthesize extracellular matrix, leading to fibrosis and heart failure. Reprogramming the large population of endogenous cardiac fibroblasts into cardiomyocytes in situ could provide an ideal basis for regenerative therapy. We found that fibroblasts from infarcted hearts were converted into cardiomyocyte-like cells by GMT in vitro. Retrovirus-transduced cells were reduced in immunocompetent mouse hearts after myocardial infarction by the immune response. We thus used immunosuppressed mice in this study to show that direct gene transfer of GMT induced newly generated cardiomyocyte-like cells in infarcted myocardium and that a polycistronic retrovirus expressing GMT enhanced cardiac maturation. This approach may introduce new regenerative strategies for repairing injured hearts.