Screening for maturation factors enhances conversion of mature cardiomyocte
To obtain mature CMs from mouse embryonic fibroblasts (MEFs) through direct cardiac reprogramming, we first optimized protocol by overexpressing three cardiac transcription factors (Mef2c, Gata4, and Tbx5; MGT) and screened various combinations of growth factors (24). MEFs were infected with MGT genes encoding retrovirus and transferred into iCM medium containing puromycin to select infected cells (Fig. 1a). After two weeks, we attempted to differentiate them into mature CMs by adding the test factors in maturation medium (StemPro-34) for an additional two weeks (22) (Fig. 1a).
Direct cardiac reprogramming is regulated by epigenetic modifications such as histone H3 Lysine residue position 4 (H3K4me3) and histone H3 Lysine 27 (H3K27me3). H3K27me3 silences fibroblast-specific genes and induces expression of cardiac-specific genes, whereas H3K4me3 activates cardiac-specific genes during direct mouse cardiac reprogramming (25). To examine remodeling of chromatin architecture, we analyzed levels of H3K27me3 and H3K4me3 by immunofluorescence staining at week 1. H3K27me3 was weakly detected, and H3K4me3 was strongly expressed in MGT compared to vehicle (Fig. 1b and 1c).
To investigate combination of maturation factors that promote functional iCMs, we used various small molecules, including FGF2, FGF4, FGF10, ascorbic acid (AA), VEGF, TGFb1, and 1-Thioglycerol (1-TG). We performed qRT-PCR for total CM marker (Tnnt2), atrial CM marker (Myl7), ventricular CM marker (Myl2), and t-tubule marker (Jph2) to determine whether maturation factors induced iCM maturation (Fig. 1d). We used FFV growth factors (FGF2, FGF10, and VEGF) as positive controls (22). Combination of FGF4 + AA (FA) and TGFb + AA (TbA) significantly increased levels of Tnnt2, Myl7, and Jph2 compared to MGT. Myl2 was higher in MGT + FA only than in MGT. The positive control, MGT + FFV, showed decreased levels of Myl7 compared to MGT. To enhance reprogramming efficiency, we explored best combination, FA, with/without small molecules such as TGFb1 and 1-TG (Fig. S1). Tnnt2, Myl7, and Jph2 levels were increased in MGT + FA, MGT + FATb, and MGT + FATG compared to MGT. Interestingly, Myl2 was significantly increased only in MGT + FA compared to that in MGT. Overall, we determined that combination of FGF4 and AA enhanced efficiency of direct cardiac reprogramming using MEFs.
Next, we investigated optimal timing for reprogramming MEFs into mature iCMs with FA by changing the timing of FA treatment to weeks 1–2, 2–4, and 1–4, sequentially, after MGT transfection. Expression of Tnnt2 showed a 4-fold increase after FA treatment between 2–4 weeks compared with MGT, although FA treatment for the other periods only slightly affected direct reprogramming (Fig. S2a). qRT-PCR analysis revealed that cardiac-specific maturation markers Mly7, Myl2, Cav3 and Jph2 were significantly increased after 2–4 weeks compared to MGT. These results suggest that FA enhances maturation of iCMs between–2–4 weeks of reprogramming and plays a crucial role in reprogramming MEFs into mature iCMs.
We also investigated proportion of iCMs in vehicle, MGT, and MGT + FA by flow cytometry at week 4. Flow cytometry analysis revealed that percentage of Tnnt2 + iCMs increased to approximately 8.3% in MGT and 17.4% in MGT + FA, compared to vehicle (2.8%) (Fig. 1e). Subsequently, we assessed expression of Tnnt2, Myl7, Myl2, and Jph2 using western blotting (Fig. 1f). Levels of Tnnt2, Myl7, Myl2, and Jph2 were higher in MGT than in vehicle and were further enhanced by FA treatment. Immunofluorescence staining was performed to examine morphology and cardiac structure using a Tnnt2 and a mature CM marker, Tnni3 (Fig. 1g and 1i). Tnnt2 and Tnni3 were expressed in MGT and MGT + FA but not in vehicle. Tnnt2 + and Tnni3 + iCMs in MGT + FA were larger and showed stronger expression than those in MGT. The percentage of Tnnt2 positive cells was approximately 30% in MGT + FA and 20% in MGT (Fig. 1h).
To examine conversion of fibroblasts into structural-matured iCMs, we stained a cardiac sarcomere marker, Actn, and Jph2 in MGT and MGT + FA. Actn and Jph2 were co-expressed in MGT + FA, and their sarcomeres and T-tubules were regularly arranged compared to MGT (Fig. 1j). Percentage of Tnnt2/Jph2-double positive cells was approximately 1.3 times higher in MGT + FA than in MGT (data not shown). These results indicate that combination of FGF4 and AA induces formation of structurally mature iCMs from immature MEFs.
FA treated iCMs show elevated expression of ion channels genes
Cardiac functions, including rhythmic contraction, electrical activity, and ion channels, such as sodium, potassium, and Ca2 + channels, generate and regulate electrical impulses that drive heartbeat (26, 27). To evaluate effect of FA on generation of functionally mature iCMs from fibroblasts, we analyzed expression of L-type Ca2 + channel markers (ryanodine receptor 2; Ryr2), Ca2 + ATPase markers (Ca2+-ATPase; Atp2b1 and Phospholamban; Pln), plasma membrane Ca2 + pump (Atp2b1), and Na+-Ca2 + exchanger marker (Ncx1) using qRT-PCR (Fig. 2a and 2b). Expression levels of Ryr2, Atp2b1, and Pln were significantly higher in MGT + FA than in MGT (Fig. 2a). mRNA levels of Atp2b1 and Ncx1 also increased approximately 2-fold in MGT + FA compared to MGT (Fig. 2b). Furthermore, we employed Flou-4 AM staining to analyze intracellular Ca2 + flux in MGT and MGT + FA at week 4. A bright green fluorescence is observed in presence of high and low intracellular Ca2 + levels. To quantify Ca2 + level of MGT + FA compared to MGT, we determined fluorescence intensity; treatment with FA triggered maximal Ca2 + oscillation, lasting nearly 1 min, whereas rare Ca2 + oscillations were found in MGT (Fig. 2c).
In mature ventricular CMs, potassium channels play an important role in repolarization phase of action potentials (28). qRT-PCR analysis revealed that potassium channel markers Kcnj2, potassium inwardly rectifying channel; and Kcnh2, potassium voltage-gated channel were significantly upregulated in MGT + FA compared to MGT (Fig. 2d). These findings suggested that combination of FGF4 and AA enhanced cardiac function by promoting cardiac ion channel function during direct cardiac reprogramming of MEFs.
FA increases mitochondrial biogenesis, ROS production and activates mitochondrial Ca2+
In CMs, mitochondria perform major biological processes, including metabolic regulation, Ca2 + handling, and redox generation. To explore effect of FA treatment on mitochondrial biogenesis (29), we evaluated expression of mitochondrial biogenesis markers (Tfam1, Ppargc1a, Nfe2l1, and Nfe2l2) using qRT-PCR. Tfam1, Ppargc1a, Nfe2l1 and Nfe2l2 were elevated in MGT + FA compared to MGT. (Fig. 2e). Furthermore, mRNA levels of mitochondrial metabolism and ROS markers (Sirt1 and Sirt3) were significantly increased in MGT + FA than in MGT (Fig. 2g). We further examined protein levels of mitochondrial biogenesis markers by western blotting. Ppargc1a was more strongly detected in MGT + FA than in MGT but not Tfam (Fig. 2f). Ca2 + regulates mitochondrial function in energy production for cardiac activity (30, 31). Therefore, to evaluate function of cardiac contraction through Ca2 + regulation in mitochondria, we examined mitochondrial Ca2 + uniporter channel markers (Mcu, Micu1, and Micu2) using qRT-PCR. Mcu, Micu1, and Micu2 were upregulated in MGT + FA than MGT (Fig. 2h). In addition, we used Ca2 + transient markers (Fluo-4AM) and mitochondrial markers (MitoTracker) to evaluate mitochondrial structure and function in Ca2 + uptake. The Mitochondria in immature CMs are small and widely distributed in cytoplasm and are present in perinuclear region, whereas mature CMs exhibit larger mitochondria that are well-organized, primarily within intermyofibrillar or subsarcolemmal region (32, 33). Our results showed that treating iCMs with FA can increase mitochondrial content and arrange cristae. Furthermore, upon merging the two indicators (Fluo-4AM and MitoTracker), a distinct overlap appeared in yellow in MGT + FA, while there was no alignment between distribution of Fluo-4AM and MitoTracker in MGT. This result indicates that mitochondrial Ca2 + function was observed in mature cardiomyocytes (white arrow in Fig. 2i). These results demonstrate that overexpression of MGT transcription factor can induce differentiation of fibroblasts into CMs through cell conversion and that treatment with FA can induce high maturation and conversion rates in iCMs.
FA enhances transdifferentiation efficiency in adult mouse cardiac fibroblasts
Transfection of MGT into MEFs has been demonstrated to directly reprogram fibroblasts into iCMs without regression to a stem/progenitor state. According to other studies, efficiency of cardiac reprogramming in adult cardiac fibroblasts (MCFs) is lower than in MEFs (17, 34). Thus, we tested effect of FA on MCFs to determine whether it directly converts fibroblasts into CM-like cells. The schematic representation of experimental design is shown in Fig. 3a. To investigate chromatin architecture remodeling, we conducted immunofluorescence staining to examine levels of H3K27me3 and H3K4me3 at week 1. In MGT, we observed weak H3K27me3 expression and strong H3K4me3 expression compared to vehicle (Fig. 3b and 3c). We performed qRT-PCR to evaluate mRNA expression of cardiac-related markers, including Tnnt2, Myl7, Myl2 and Jph2. These results demonstrated that FA treatment upregulated Tnnt2, Myl7, and Jph2 compared to MGT but slightly affect Myl2 (Fig. 3d). Next, we confirmed that FA improved structural maturation of CMs by immunofluorescence staining with cardiac marker Tnnt2. MGT + FA demonstrated a higher quantity of MGT-infected cells stained with Tnnt2, unlike MGT, where such expression was not observed, as seen in vehicle (Fig. 3e). To determine whether iCMs exhibit functional characteristics typical of CMs, we analyzed Ryr2 and Kcnh2, along with genes Ppargc1a and Tfam, using qRT-PCR. mRNA expression of the genes significantly increased in MGT + FA but Kcnh2 was not significant (Fig. 3f). Finally, we used Fluo-4AM marker and mitochondrial marker MitoTracker to detect presence of mitochondrial Ca2+ (yellow; white arrow in Fig. 3h). We quantified expression of mitochondrial Ca2 + when it peaked maximally and found a significant increase in MGT + FA compared to MGT (Fig. 3g). These results indicate that combination of FA improves direct cardiac reprogramming efficiency and successfully converts fibroblasts into functional CMs, even in adult cardiac fibroblasts, known for their lower reprogramming efficiency.
FA enhances human fibroblasts to iCMs cardiac reprogramming efficiency
To investigate effects of FA on cardiac induction in human fibroblasts, we optimized a direct cardiac reprogramming protocol using transcription factors and FA. In humans, MGT, MESP1, and MYOCD (MGTMM) enhance efficiency of iCM reprogramming, promoting conversion of fibroblasts into iCMs (15, 16). We transduced Normal human dermal fibroblasts (NHDFs) with three lentiviruses encoding five-core cardiac-specific factors (MGTMM; Fig. 4a and Fig. S3a-c). We first investigated efficiency of MGT, MESP1, and MYOCD vectors by qRT-PCR in 293T cells (Fig. S3d). MEF2C, GATA4, and TBX5 were highly expressed in MGT vector-transfected 293T cells, MESP1 was significantly upregulated in MESP1 vector-transfected 293T cells, and MYOCD was upregulated in MYOCD vector-transfected 293T cells. Furthermore, we treated cells with FA at week 2 and analyzed them at week 4 (Fig. 4a). Distinct morphological changes were observed in each group, and levels of GATA4, MEF2C, TBX5, MESP1, and MYOCD were significantly increased in MGTMM and MGTMM + FA compared to vehicle (Fig. S4).
To evaluate effect of MGTMM overexpression on reprogramming, we examined expression levels of H3K4me3 using immunofluorescence staining on day 7 (Fig. 4b). H3K4me3 was more strongly expressed in MGTMM than in vehicle. Percentage of H3K4me3 area relative to total nuclei decreased in MGTMM compared to that in vehicle (Fig. 4c). Protein level of H3K4me3 was also higher in MGTMM than in vehicle (Fig. 4d).
We examined effects of FA combination on expression levels of TNNT2, MYL7 and MYL2 using qRT-PCR and western blot analyses (Fig. 4e and 4f). mRNA levels of TNNT2, MYL7, and MYL2 were significantly higher in MGTMM + FA than in MGTMM (Fig. 4e). Protein levels of TNNT2, MYL7, and MYL2 were also higher in MGTMM + FA than in MGTMM (Fig. 4f). Interestingly, nodal CM marker TBX18 revealed no significant differences between vehicle, MGTMM, and MGTMM + FA (Fig. S5a). Next, we investigated proportions of TNNT2 + iCMs in vehicle, MGTMM, and MGTMM + FA by immunofluorescence staining (Fig. 4g). Large TNNT2 + iCMs (5.6%) and long sarcomere structures (1.2 µm) were more frequently observed in MGTMM + FA compared to MGTMM (5.6% and 1.7 µm, respectively; Fig. 4h and 4i). We further assessed different cardiac markers (TNNI3, MYH6, and MYH7) and T-tubule markers (BIN1, CAV3, and JPH2) using qRT-PCR (Fig. S5b and S5c). TNNI3, MYH6, MYH7, BIN1, CAV3, and JHP2 mRNA expression was markedly increased in MGTMM + FA compared to MGTMM. A higher expression of CAV3 was also observed in MGTMM + FA than in MGTMM by western blotting (Fig. S5d). To investigate ultrastructural array between T-tubules and sarcomeres, JPH2 and TNNT2 were examined by immunofluorescence staining. Cross structures of T-tubules, Z-lines, and co-localization areas of JPH2 and TNNT2 were more frequently observed in MGTMM + FA than in MGTMM (Fig. S5e). These results demonstrate that FA treatment enhanced cardiac gene expression at reprogramming stage, indicating potential of FA to enhance efficiency of direct cardiac reprogramming of human fibroblasts.
FA improves iCM functions by enhancing Ca2+ channel and metabolism
We investigated whether FA improves cardiac function upon direct human cardiac reprogramming. To assess cardiac function, we analyzed expression levels of ion channels using qRT-PCR (Fig. 5a and 5b). Ca2 + channel markers (CACNA1C and RYR2) exhibited increased expression in MGTMM + FA compared to MGTMM (Fig. 5a). However, potassium channel markers (KCNH2 and KCNJ2) were not significantly different between two groups (Fig. 5b). Within CMs, mitochondria play a crucial role by producing over 90% of ATP required for cardiac function (35). Consequently, we investigated expression levels of mitochondrial biogenesis markers (TFAM, PPARGC1A, and SIRT1) by qRT-PCR (Fig. 5c). TFAM, PPARGC1A, and SIRT1 were upregulated in MGTMM + FA compared to MGTMM. Protein levels of TFAM and PPARGC1A were elevated in MGTMM compared to vehicle, and further elevated in MGTMM + FA (Fig. 5d). To quantify mitochondrial content in each group, we analyzed ratio of mitochondrial DNA (mtDNA) to nuclear DNA (nDNA) by qRT-PCR. We found a significant increase in mtDNA/nDNA ratio in MGTMM + FA compared to MGTMM (Fig. 5e).
Ca2 + are crucial for CM function, and dynamics of mitochondrial Ca2 + cycling and storage during excitation-contraction coupling (ECC) are strongly linked to cardiac physiology and pathophysiology (36–38). Mitochondrial Ca2 + uptake is crucial in regulating cellular metabolism and provides the necessary energy for contraction (39–42). To analyze relationship between mitochondria and Ca2 + handling, we investigated MitoTracker and Fluo-4 AM using immunofluorescence staining (Fig. 5f and 5g). Co-localized MitoTracker and Fluo-4 AM signals were increased in MGTMM + FA compared with MGTMM.
Mature CMs are characterized by fatty acid metabolism, oxidative phosphorylation, and mitochondrial biogenesis. FA treatment increased mRNA level of CPT1B, CD36, and PPARA and protein level of CPT1B (Fig. 5h and 5i). These findings indicate that FA enhanced cardiac function by improving ion channel function, mitochondrial biogenesis, and metabolism during cardiac reprogramming.
Transcriptome analysis exhibits MGTMM + FA impact on cardiac reprogramming pathway
To understand genes and signaling pathways involved in maturation resulting from co-treatment with FGF4 and ascorbic acid, we analyzed transcriptome of vehicle, MGTMM, and MGTMM + FA. Among 25,737 genes, 358 genes in MGTMM compared to vehicle, 747 in MGTMM + FA compared to vehicle, and 757 in MGTMM + FA compared to MGTMM were upregulated (fold changes > 2; Fig. 6a). In contrast, 497 genes in MGTMM compared to vehicle, 921 in MGTMM + FA compared to vehicle, and 882 in MGTMM + FA compared to MGTMM were downregulated (fold changes > 2). Scatter plots highlighting upregulated cardiac-specific genes (TNNT2, CAV3, JPH2, CACNA1A, CACNA1C, and GJA1) are shown in Fig. 6b. Heatmap analysis further revealed upregulation of cardiac-specific genes in MGTMM + FA compared to MGTMM (Fig. 6c). To elucidate biological processes regulated by differentially expressed genes (DEG), we performed gene ontology (GO; Fig. 6d) and also analyzed Kyoto Encyclopedia of genes of genomes (KEGG) pathway databases to identify active signaling pathway involved in reprogramming processes (Fig. S6a). The most significantly regulated GO biological processes and KEGG pathway associated with these processes are shown in Table S1 and S2.
RNA-seq analysis reveals potential of JAK-STAT3 signaling in iCM maturation
To explore common signaling pathways involved in reprogramming induced by MGTMM and FA treatment, we performed KEGG pathway analysis and examined genes involved (Table S1 and S2). Results revealed that MGTMM upregulated MAPK, PI3K-AKT, VEGF, Ca2+, and p53 signaling pathway compared to vehicle. Furthermore, MGTMM + FA upregulated JAK-STAT, TGFB, and Toll-like receptor signaling pathways.
To validate these findings, we investigated ECM-receptor interaction and focal adhesions (Fig. S6b-d). COL1A1 and Laminin in ECMs, ITGAV, ITGA7, and ITGB3 in integrins were increased in MGTMM and MGTMM + FA compared to vehicle (Fig. S6b and S6c). However, FN1 was decreased in MGTMM and MGTMM + FA compared with vehicle. ROCK1, ROCK2, TLN1, and DES in focal adhesions were highly expressed in MGTMM and MGTMM + FA compared to vehicle, whereas VCL and TLN2 showed no significant differences between vehicle, MGTMM, and MGTMM + FA (Fig. S6d). These results suggest that MGTMM induces cardiac reprogramming by activating VEGF, ECM proteins, integrins, focal adhesions, and Ca2 + signaling.
To examine signaling pathways related to maturation of iCMs following FA treatment, we analyzed JAK-STAT and TGFB signaling pathways (Fig. 6f-j). TGFB1, TGFB2, and GDF5 were elevated in MGTMM + FA compared to MGTMM (Fig. 6f and 6h). Expression levels of BMP2 and BMP4 were not significantly different between groups. Therefore, we further analyzed downstream signaling of TGFB1 and 2, SMAD2 and SMAD3, and downstream signaling of GDF5, SMAD1/5 (Fig. 6g and 6i). The results revealed an increase in p-SMAD3 in MGTMM + FA compared to that in MGTMM, while p-SMAD1/5 showed no significant difference. Additionally, p-JAK2, p-STAT3, and p21 in JAK-STAT3 levels were increased in MGTMM + FA compared to MGTMM, whereas p-cJun showed no significant difference between groups (Fig. 6j). These findings indicate that FA contributes to maturation of iCMs by influencing JAK-STAT and TGFB signaling cascades.
JAK2-STAT3 signaling pathway modulates iCM maturation upon treatment of FA
Additionally, owing to the remarkable activation of JAK2-STAT3 signaling pathway in MGTMM + FA, which exhibited both structural and functional maturation compared to MGTMM, we further investigated the roles of JAK2-STAT3 signaling pathway in iCMs maturation. We treated MGTMM and MGTMM + FA cells with cryptotanshinone, a JAK2-STAT3 inhibitor, from week 2 to week 4 and performed qRT-PCR and western blot analyses (Fig. 7a). mRNA expression levels of FGFR1 and FGFR2 were significantly higher in MGTMM + FA and MGTMM + FA + inh than in MGTMM and MGTMM + inh. These results indicated that FGF signaling pathway was activated in FA-treated iCMs. To analyze effect of inhibitor, cryptotanshinone, we investigated JAK2-STAT3 signaling pathway using western blot analysis (Fig. 7b). Protein levels of p-JAK2, JAK2, p-STAT3, and STAT3 were lower in MGTMM + inh and MGTMM + FA + inh than in MGTMM and MGTMM + FA. These results demonstrated that cryptotansbinone inhibited JAK2-STAT3 signaling pathway.
JAK2-STAT3 signaling pathway promotes transcription of TGFB, collagen, and MYH7. Next, to investigate whether TGFB signaling was affected during iCM maturation, we analyzed markers of TGFB signaling pathway by qRT-PCR (Fig. 7c). TGFB1, TGFB2, TGFBR1, and TGFBR2 were significantly decreased in MGTMM + inh compared with MGTMM + FA, and there was no significant difference in MGTMM + Inh compared with MGTMM. TGFB2, TGFBR1, and TGFBR2 were decreased in MGTMM + FA + Inh compared to MGTMM + FA, but TGFB1 was not significantly different between MGTMM + FA and MGTMM + FA + Inh. We also observed enhanced phosphorylation of SMAD3 in MGTMM + FA compared to that in MGTMM and decreased phosphorylation in MGTMM + Inh and MGTMM + FA + Inh compared to that in MGTMM + FA (Fig. 7d). To examine regulation of collagen by JAK2-STAT3 inhibition, we performed western blotting for ECM markers (Laminin, COL1A1, and FN1) (Fig. 7e). Laminin and COL1A1 were decreased in MGTMM + FA + Inh compared to MGTMM + FA, whereas FN1 was not significantly different between MGTMM + FA and MGTMM + FA + Inh. These findings demonstrated that during direct cardiac reprogramming, JAK2-STAT3 signaling pathway regulates TGFB2, TGFBR1, TGFBR2, Laminin, and COL1A1.
To examine whether JAK2-STAT3 signaling pathway induces cardiac differentiation in FA-treated iCMs, we analyzed cardiac markers (MYH7, TNNT2, and MYL2) using qRT-PCR (Fig. 7f). MYH7, TNNT2, and MYL2 levels were lower in MGTMM + Inh mice than in MGTMM + FA mice. TNNT2 levels were not significantly different between MGTMM + FA and MGTMM + FA + Inh, whereas MYH7 and MYL2 were lower. Protein levels of TNNT2, MYL7, and MYL2 were also decreased in MGTMM + Inh and MGTMM + FA + Inh compared to those in MGTMM + FA (Fig. 7g). To investigate induction of cardiac maturation via JAK2-STAT3 signaling, we examined T-tubule markers (CAV3 and JPH2) and mitochondrial biogenesis markers (PPARGC1A and TFAM) using qRT-PCR (Fig. 7h and 7i). mRNA levels of CAV3, JPH2, PPARGC1A, and TFAM were significantly decreased in MGTMM + Inh and MGTMM + FA + Inh compared to MGTMM + FA; protein expression of JPH2 and PPARGC1A was also (Fig. 7j). Collectively, these results demonstrate that JAK2-STAT3 signaling pathway contributes to maturation of iCMs by regulating TGFB (TGFB2, TGFBR1, and TGFBR2) signaling pathway.