Heart repair by reprogramming non-myocytes with cardiac transcription factors

Reprogramming of adult fibroblasts into functional cardiac-like myocytes

A core set of evolutionarily conserved transcription factors (GATA4, Hand2, MEF2C, MesP1, Nkx2-5, and Tbx5) controls cardiac gene expression and heart development^14,15. Recently, GATA4, MEF2C, and Tbx5 were reported to be capable of converting neonatal fibroblasts to cardiomyocyte-like cells in vitro¹⁶. To search for an optimal combination of transcription factors capable of inducing cardiac lineage-reprogramming, we generated retroviruses to express each of the six core cardiac transcription factors (GATA4 (G), and Hand2 (H), MEF2C (M), MesP1 (Ms), Nkx2-5 (N), and Tbx5 (T)) in fibroblasts derived from adult mice bearing a cardiac-specific a-MHC-GFP transgene (Supplementary Figs. 1 and 2).

Potential cardiogenic activity of the above 6 factors (G, H, M, Ms, N, T) in adult tail-tip fibroblasts (TTFs) from α-MHC-GFP transgenic mice was quantified by analysis of GFP⁺ cells by flow cytometry 9 days after viral transduction. No GFP⁺ cells were observed in fibroblast cultures transduced with viral backbone or without infection. However, the 6 factors together generated a small population of cells (~7%) positive for GFP (Supplementary Fig. 3). After several rounds of withdrawing one factor, we identified multiple combinations, including GHMMsT, GHMT, HMMsT, GMT, HMT, and MT, that were capable of efficiently inducing α-MHC-GFP⁺ cells (Supplementary Fig. 3). The cooperativity amongst the different factors is consistent with their ability to synergistically activate cardiac gene expression and to activate each other’s expression^{14, 17–20}.

To determine whether the above factors could activate endogenous cardiac-specific genes in adult TTFs, we examined expression of cardiac troponin T (cTnT) and α-MHC-GFP by flow cytometry (Supplementary Fig. 4). Depending on the experiment, GHMT induced between 5.2% and 19.7% (average 9.2%) of cells to become positive for both α-MHC-GFP and cTnT. We refer to GHMT-transduced cells expressing cardiomyocyte markers as induced cardiac-like myocytes (iCLMs). By comparison, GMT induced ~2.9% (1.5% to 5.6%) of adult TTFs to adopt a cTnT⁺/α-MHC-GFP⁺ phenotype (Supplementary Fig. 4). The percentage of iCLMs generated from adult TTFs by GHMT (9.2%) was ~4-fold higher than previously reported for neonatal TTFs with GMT (2.5%)¹⁶. The percentage of cells expressing cTnT and α-MHC-GFP reached a peak at day 7 of transduction; the fraction of iCLMs declined thereafter due to overgrowth of non-reprogrammed fibroblasts (Supplementary Fig. 5).

Cardiac fibroblasts (CFs) are the most prevalent interstitial cell type in adult mammalian hearts. We transduced adult CFs with GHMT, GMT or empty viruses and analyzed expression of cardiac markers by flow cytometry one week later. GHMT induced 6.8% of CFs to become cTnT⁺/α-MHC-GFP⁺, compared with 1.4% double-positive cells with GMT (Supplementary Fig. 6a). GHMT induced both cTnT and cTnI (cardiac Troponin I) in 7.5% of cells (Supplementary Fig. 6b). Thus, GHMT represented the most optimal combination of factors for efficient initiation of cardiac gene expression in adult fibroblasts. α-MHC-GFP⁺ cells derived from adult TTFs and CFs by GHMT transduction showed strong immunostaining of the sarcomeric proteins α-actinin and cTnT (Fig. 1a, b, and Supplementary Fig. 7). More organized sarcomeres were observed in iCLMs at day 30 (Fig. 1b), compared to day 14 (Fig. 1a). In the presence of GHMT, three types of iCLMs, referred to as Types A, B and C, which appear to represent a spectrum of cardiac reprogramming, were induced from adult CFs and TTFs. Type A iCLMs only expressed α-MHC-GFP. Type B iCLMs expressed both α-MHC-GFP and α-actinin. Type C iCLMs not only expressed α-MHC-GFP and α-actinin, but also displayed sarcomere-like structures (Fig. 1a). GHMT induced ~15% of adult CFs and ~18% of adult TTFs to become α-MHC-GFP positive cells, and ~30% and ~10% of these became Type C iCLMs, respectively (Fig. 1c).

Microarray and qPCR analysis of gene expression patterns showed expression of a broad range of cardiac genes, indicative of a cardiac-like phenotype, and concomitant suppression of non-myocyte genes including Fsp1 (fibroblast-specific protein 1 or S100A4) in fibroblasts transduced with GHMT (Supplementary Figs. 8, 9). Following maintenance of GHMT-transduced adult CFs or TTFs in culture for more than 5 weeks, we observed spontaneous contractions in ~0.2% of Type C iCLMs. These beating iCLMs displayed calcium transients and action potentials (Supplementary Movies 1, 2 and 3, Fig. 1d and Supplementary Figs. 10, 11). iCLMs displayed a pattern of calcium transients most similar to neonatal ventricular cardiomyocytes (Fig. 1d). Transduction with inducible expression vectors showed that the cardiomyocyte-like phenotype was stable following termination of exogenous GHMT expression (Supplementary Fig. 12). Co-staining for cardiomyocyte markers and Myc-tagged GHMT suggests that induction of the cardiomyocyte-like phenotype by GHMT is cell-autonomous (Supplementary Fig. 13). Together, these results indicate that GHMT can activate cardiac gene expression in a sub-population of TTFs and CFs. The relatively small percentage of cells that adopts the cardiac-like phenotype perhaps indicates a precise stoichiometry of the cardiac factors required for phenotypic conversion, which is achieved in only a fraction of cells, or to differential susceptibility to reprogramming amongst heterogeneous cell types in the starting non-myocyte population. The efficiency of cellular reprogramming to iCLMs by GHMT is comparable to that of reprogramming of induced pluripotent stem (iPS) cells by pluripotency factors^7,21.

Reprogramming of Fsp1-tagged cells into iCLMs in vivo

Following MI or other forms of cardiac injury, cardiomyocytes undergo necrotic and apoptotic cell death and CFs are activated to produce collagen and other extracellular matrix components, causing fibrosis and impaired cardiac function²². Therefore, we sought to determine whether reprogramming CFs to a cardiomyocyte fate might blunt the decline in cardiac function post-MI. Adult mammalian cardiomyocytes do not divide and are therefore resistant to retroviral expression. We expressed the cardiogenic transcription factors in CFs and other dividing cells in vivo using a retrovirus expression system, which is specific for proliferating cells and directs expression preferentially in myofibroblasts of injured rodent hearts²³. We confirmed the specificity of retroviral infection for replicating non-cardiomyocytes by injecting concentrated GFP retroviruses into injured hearts after ligation of the left anterior descending coronary artery (LAD), which induces MI. GFP expression was clearly detected in the ischemic area. However, none of the GFP⁺ cells expressed the cardiac marker, cTnT, consistent with the specificity of retroviral infection for proliferating non-cardiomyocytes (Supplementary Fig. 14).

Fsp1 is expressed in non-cardiomyocytes such as fibroblasts and transitioning epithelia^{24, 26}. In mouse and human hearts, expression of Fsp1 primarily colocalizes with markers of CFs and increases following MI²⁶. Non-cardiomyocytes in mice carrying alleles of Fsp1-Cre and Rosa26-LacZ are specifically labeled with β-galactosidase (β-gal), providing a marker for fibroblast lineage tracing²⁴. To examine whether cardiac transcription factors were able to activate cardiac genes in non-cardiomyocytes in vivo, we performed LAD ligation on Fsp1-Cre/Rosa26-LacZ mice and injected concentrated retroviruses encoding GHMT or GFP into the border zone immediately following LAD ligation. We then analyzed β-gal activity in histological sections of hearts at various times thereafter. In uninjured hearts, less than one β-gal⁺ cardiomyocyte per section was observed. After injury, β-galexpression was readily detected in CFs throughout the infarct zone (Fig. 2a, b). In injured hearts infected with GFP viruses, 0.05±0.13% of cardiomyocytes in the injured area were β-gal⁺ (Fig. 2c), which may be due to low level ectopic activation or to a basal level of new cardiomyocyte formation from a stem cell pool, as reported previously²⁷. In contrast, abundant clusters of intensely stained β-gal⁺ cardiomyocytes were observed throughout the infarct and border zone of injured hearts infected with the GHMT retrovirus cocktail (Fig. 2b). We observed that 6.5±1.2% of cardiomyocytes in the injured area displayed β-gal activity (Fig. 2c). Generally, more β-gal⁺ cardiomyocytes were observed in the border zone adjacent to the infarct region, which may be due to intact vascular structures or higher viral infection in this region. Similar results were obtained upon injection with GHMMsT, whereas the inclusion of Nkx2-5 (GHMMsNT) diminished the efficacy of the other five factors (Supplementary Fig. 15), as seen in vitro. β-gal⁺ cardiomyocytes expressed cTnT and showed clear striations 3 weeks after viral transduction (Fig. 2b). Through quantification of β-gal⁺ cardiomyocytes in histological sections of infarcted hearts, we calculated that at least 10,000 new myocytes were generated in the injured area after 3 weeks of GHMT infection. This is likely an underestimate of the number of iCLMs generated in vivo since the reporter is only expressed in a subset of non-myocytes, such that unmarked cells could also be reprogrammed but go undetected.

In the heart, gap junctions composed of connexins ensure electrical and metabolic coupling between cardiomyocytes and coordinate their contractility²⁸. To determine whether reprogrammed cardiomyocytes could couple with surrounding endogenous myocytes through gap junctions, we performed immunostaining for Connexin 43 (Cx43), the major connexin in functional cardiomyocytes²⁹. Gap junctions were observed between β-gal⁺ and β-gal⁻ cardiomyocytes and between β-gal⁺ cardiomyocytes (Fig. 2d), suggesting coupling of reprogrammed cardiomyocytes with surrounding myocytes. Reprogrammed β-gal⁺ cardiomyocytes isolated from ventricular myocardium and identified by labeling with a fluorogenic, lipophilic β-gal substrate (C₁₂FDG) also displayed a pattern of contractility and calcium transients similar to normal β-gal⁻ ventricular cardiomyocytes in response to electrical pacing at 1 Hz (Fig. 2e), indicating their functionality.

Reprogramming of inducible Tcf21-tagged cells into iCLMs in vivo

To rule out the possibility that injury or viral infection might somehow activate the Fsp1 reporter in cardiomyocytes, we generated a strain of mice harboring an inducible MerCreMer expression cassette inserted by homologous recombination into the Tcf21 (capsulin/epicardin) locus (Tcf21^iCre), which is expressed specifically in non-cardiomyocytes in the heart (Supplementary Fig. 16)³⁰. Intercrossing of these mice with mice bearing the Cre-inducible R26R^tdTomato reporter showed specific expression of Tomato predominantly in cardiac fibroblasts with less expression in endothelial cells (Supplementary Fig. 16). However, we only observed one Tomato⁺ cardiomyocyte in over 40 histological sections of normal hearts. Tcf21^iCre/R26R^tdT mice were treated with tamoxifen for three days to mark Tcf21-expressing cells. Eight days after the last tamoxifen treatment, LAD ligation was performed and animals were injected with empty vector, GFP or GHMT retroviruses and analyzed 3–4 weeks later. Numerous Tomato⁺ cardiomyocytes were observed in GHMT-injected hearts compared to GFP-injected ones (Fig. 3a).

To quantify Tomato⁺ cardiomyocytes, we isolated cardiomyocytes from intact hearts using the Langendorff perfusion method. In GHMT-injected hearts, 2.4±1.5% of cardiomyocytes in the injured area were Tomato⁺, and these Tomato⁺ cardiomyocytes displayed clear striations (Fig. 3b and c). However, in empty vector-injected injured hearts, only 0.004±0.005% of cardiomyocytes were Tomato⁺ (Fig. 3c). Tcf21^iCre marks a smaller, more restricted fraction of non-myocytes than Fsp1-Cre. Hence the number of new myocytes marked with this reporter is lower than with Fsp1-Cre. These findings suggest that forced expression of GHMT in non-myocytes of the heart is sufficient to induce the expression of cardiac markers. The Tomato⁺ cardiomyocytes induced by GHMT in vivo were able to spontaneously contract in vitro (Supplementary Fig. 17 and Supplementary Movie 4). Recording of action potentials of reprogrammed Tomato⁺ cardiomyocytes and endogenous cardiomyocytes by whole-cell patch clamping suggested functionality of cardiomyocytes induced by GHMT in vivo (Fig. 3d).

To further test whether GHMT indeed promoted the formation of new cardiomyocytes following MI and to rule out the possibility that labeled cardiomyocytes obtained from Fsp1-Cre or Tcf21^iCre/Rosa marker mice injected with GHMT might arise from fusion of native cardiomyocytes with non-myocytes, we utilized mice with an inducible α-MHC-MerCreMer transgene and Rosa26-LacZ reporter. Gavage of these mice with tamoxifen for seven consecutive days resulted in labeling of 87.7±1.9% of cardiomyocytes in the left ventricular myocardium. Following LAD ligation and injection of GFP retroviruses, we observed a reduction in β-gal-labeled cardiomyocytes (83.2±4.6%) in the border zone adjacent to the infarct, suggesting replenishment of cardiomyocytes following injury, as described previously^{27, 31}. Injection of GHMT retroviruses further reduced the percentage of β-gal⁺ cardiomyocytes in the border zone (76.0±5.2%) (Supplementary Fig. 18). These findings suggest that GHMT promotes the formation of new cardiomyocytes from a non-α-MHC lineage in vivo following injury.

Enhancement of cardiac regeneration of injured hearts

We examined whether forced expression of GHMT in non-cardiomyocytes could lead to measurable improvement in function of ischemic hearts. Cardiac function following MI was assessed in a blinded fashion by fractional shortening (FS), ejection fraction (EF), and stroke volume using echocardiography and magnetic resonance imaging (MRI). Twenty-four hours after LAD ligation, FS and EF assessed by echocardiography of all mice decreased relative to sham-operated mice. Thereafter, cardiac function of GFP-injected mice continued to decline, reaching a stable value 2 weeks post-MI; FS ~13% and EF ~ 30%. In contrast, infection of injured myocardium with GHMT retroviruses blunted further worsening of cardiac function 3 weeks post-MI; FS ~26% and EF ~51% (Fig. 4a). By comparison, functional improvement was delayed and less complete with GMT, consistent with the reduced efficiency of this transcription factor combination in reprogramming in vitro.

To determine whether functional improvement was sustained, we assessed cardiac function at 6 weeks and 12 weeks by EF and stroke volume using cardiac MRI. EF of GFP-injected mice decreased to reach a stable value of ~28% 6 weeks post-MI (Fig. 4b). In contrast, infection of injured myocardium with GHMT blunted worsening of EF 6 weeks post-MI (~49%) with further significant improvement at 12 weeks post-MI (~57%) (Fig. 4b). This long-term effect on EF by GHMT was accompanied by significant increases in stroke volume at 12 weeks compared to 6 weeks (Fig. 4b). Individual mice in each group demonstrated similar functional changes in both cardiac parameters, indicating the reliability of cardiac MRI to assess cardiac function (Supplementary Fig. 19). These data suggest that expression of GHMT in non-cardiomyocytes in injured hearts can sustain cardiac function. Moreover, GHMT- and GHMMsT-infected hearts showed a pronounced reduction in fibrosis and increased muscle tissue, compared with GFP-infected hearts after MI (Fig. 5a, b and Supplementary Fig. 20).

Discussion

Our results demonstrate that GATA4, Hand2, MEF2C, and Tbx5 (GHMT) can reprogram CFs into functional cardiac-like myocytes in vivo and in vitro, confirming and extending prior studies¹⁶. Moreover, exogenous GHMT expression in non-cardiomyocytes of the heart post-MI also reduces fibrosis and improves cardiac function.

Improvement of cardiac function post-MI by different transcription factor combinations correlated with their ability to convert fibroblasts into iCLMs in vitro (Supplementary Fig. 21), suggesting that cardiac repair results, at least in part, from reprogramming of non-cardiomyocytes toward a cardiomyocyte fate. We feel it is important to emphasize, however, that the dramatic functional improvement of GHMT-injected hearts seems greater than might be predicted from the relative inefficiency of reprogramming of adult CFs to iCLMs in vitro. We speculate that the native milieu of the intact heart, containing extracellular matrix, growth factors, persistent contractility, surrounding contractile cells and other cell types, is more permissive than plastic tissue culture dishes for functional reprogramming. In this regard, reprogramming of pancreatic beta-cells has been observed in vivo but not in vitro³². It is also conceivable that other mechanisms, such as a blockade to the activation of CFs, enhanced survival of cardiomyocytes, facilitated differentiation of activated cardiac progenitors into cardiomyocytes, as described previously^{27, 31}, or improved angiogenesis contribute to the benefits observed upon expression of GHMT in the heart post-MI. Irrespective of these uncertainties, the strategy presented here, while still requiring optimization, provides a potential means of improving cardiac function in vivo, bypassing many of the obstacles associated with cellular transplantation.

Methods Summary

All experimental procedures with animals were approved by the Institutional Animal Care and Use Committee at UT Southwestern Medical Center.

Adult TTFs and CFs were isolated by an explanting method in which fibroblasts migrate from minced tissue and grow in fibroblast growth medium. Fibroblasts were transduced with a mixture of polybrene (Sigma; 6μg/ml) and freshly made retroviruses expressing transcription factors generated in Platinum E cells (Cell Biolabs). Twenty-four hours after viral transduction, the viral medium was changed to a cardiac induction medium. Medium was changed every two days. Expression of cardiac genes was analyzed by flow cytometry, microarray, and immunocytochemistry.

Adult mice, 8–10 weeks old, underwent either a sham operation or ligation of the LAD. Concentrated retroviruses (~10^8~9 pfu viruses) were injected into the border zone using a gastight 1710 syringe (Hamilton). Cardiac function was assessed using echocardiography and MRI. Hearts were harvested from euthanized animals for histological studies.

Isolation of cardiomyocytes, electrophysiological measurements, histology and immunohistochemistry were performed as described previously^{2, 33–37}.

Methods

Supplementary Material