Myogenic Transdifferentiation of Pig Fibroblasts Using the Porcine MYOD1 Inducible Vector

Background Myogenic transdifferentiation can be accomplished through ectopic MYOD1 expression, which is facilitated by various signaling pathways associated with myogenesis. In this study, we attempted to transdifferentiate pig embryonic broblasts (PEFs) myogenically into skeletal muscle through overexpression of the pig MYOD1 gene and modulation of the FGF, TGF-β, WNT, and cAMP signaling pathways. Results The MYOD1 overexpression vector was constructed based on comparative sequence analysis, demonstrating that pig MYOD1 has evolutionarily conserved domains across various species. Although forced MYOD1 expression through these vectors triggered the expression of endogenous muscle markers, transdifferentiated muscle cells from broblasts were not observed. Therefore, various signaling molecules, including FGF2, SB431542, CHIR99021, and forskolin, along with MYOD1 overexpression were applied to enhance the myogenic reprogramming. The modied conditions led to the derivation of myotubes and activation of muscle markers in PEFs, as determined by qPCR and immunostaining. Notably, a sarcomere-like structure was observed, indicating that terminally differentiated skeletal muscle could be obtained from transdifferentiated cells. Our


Background
To date, numerous studies have de ned various differentiation strategies for the differentiation of PSCs (pluripotent stem cells) into speci c cell types of ecto-, meso-, and endodermal lineages, which have advantages as in nite cell sources (1)(2)(3). Whereas these directed differentiation approaches recapitulate in vivo developmental processes through modulation of signaling pathways, they reportedly have defects in terms of time consumption and low e ciency because the small molecules used are also involved in differentiation into other cell types (4,5). For example, the derivation of striated myo bers, a form of mature skeletal muscle, from mouse ESCs (embryonic stem cells) and human iPSCs (induced pluripotent stem cells) reportedly takes 3 ~ 4 weeks, albeit with the use of an advanced protocol with a short experimental period (6,7). Therefore, direct reprogramming using the activation of the transcriptional program has been applied to facilitate the differentiation of PSCs into the desired cell type (5). However, problems derived from the use of PSCs still remain, including the potential risk of teratoma formation (8) and the requirement of a time-consuming and ne-tuning process for the derivation and maintenance of the pluripotent state. To overcome the above problems, a transdifferentiation approach is required, in which fully differentiated somatic cells, instead of PSCs, are induced to differentiate directly into target cell types via ectopic expression of transcription factors. Transdifferentiation toward the muscle lineage has been widely studied for a long time since the identi cation of the master transcription factor in myogenesis, Myod1 (9).
Forced Myod1 expression converts cell fates into muscle, which is responsible for Myod1 functioning as a transcriptional and epigenetic regulator, leading to the activation of muscle-speci c genes in a feedforward manner (10,11). Small molecules can be supplemented under culture conditions to Myod1/MYOD1-overexpressing somatic cells to improve the e cacy of transdifferentiation. The surface of myocytes and the basement membrane of myotubes secrete FGF2 ( broblast growth factor 2), which play roles in the activation of muscle stem cells and the expansion of myoblasts, inhibiting myogenic terminal differentiation (12,13). Because TGF-β signaling negatively functions in myogenesis by decreasing Myog activity, inhibition of TGF type I-β receptors, such as SB431542, reportedly enhances the elongation of myotubes through myoblast fusion (13). In fact, the expression of follistatin, which inhibits TGF-β signaling, is detected in the paraxial mesoderm in vivo (14). During in vivo myogenesis, the WNT activator secreted from the notochord, neural tube, and surrounding tissue is involved in a series of speci cations along with the presomitic mesoderm, somite, dermomyotome, and myotome, as determined by in vitro-directed differentiation in human and mouse PSCs (3,6,7). It demonstrates that WNT signaling induces the commitment of myogenic precursors and has been substantiated by previous research showing that the WNT activator CHIR99021 with SB431542 and FGF2 induces myogenic speci cation from human iPSCs (14). In a previous report, the adenylyl cyclase activator forskolin was identi ed as a myoblast proliferation-promoting factor through CREB-mediated WNT, leading to the upregulation of the Pax3, Myf5, and Myod1 genes (15). The combination of FGF2, a WNT activator, and forskolin stimulated skeletal muscle differentiation in human iPSCs and especially forskolin improved satellite cell expansion in mice (16).
In pigs, myogenic reprogramming could be used as a cell source for muscle disease models and the production of cultured meat. Although, in pig iPSCs, skeletal myotubes were reportedly observed through the activation of WNT signaling and ectopically expressed MYOD1 by overexpression vector and 5azacytidine (5-aza) (17), the aforementioned limitations derived from iPSCs remain. Moreover, it has been demonstrated that muscle stem cells lose the potential for proliferation and differentiation in long-term culture in vitro (18). As an alternative, broblasts can be obtained in high yield because of their large proportion in the body and are known to be the most effective cell type to accept expression of Myod1 because they are derived from the mesoderm, which is identical in origin to the muscle (19). Therefore, in this study, we attempted to transdifferentiate pig broblasts into skeletal muscle myogenically through overexpression of the pig MYOD1 gene and modulation of the FGF, TGF-β, WNT, and cAMP signaling pathways.

Lentiviral vector production
Lentiviral vectors were produced as previously described (21). Brie y, HEK 293 LTV cells (Cell Biolabs, San Diego, CA, USA) were used as the packaging cell line, and ve plasmids were used for the production of lentiviral vectors: FUW-tetO-pMYOD1 and FUW-M2rtTA (the transfer plasmid); pLP1 and pLP2 (the packaging plasmids; Invitrogen, Waltham, MA, USA); and pLP/VSVG (the envelope plasmid; Invitrogen). These plasmids were transfected into HEK 293 LTV cells using the calcium phosphate precipitation method. Subsequently, the LTV culture supernatants were ltered and concentrated. The derived virus pellets were stored at -76 °C until use.
Genomic DNA (gDNA) extraction and polymerase chain reaction (PCR) Genomic DNA was extracted using the G-spin™ Total DNA Extraction Kit (iNtRON, Seongnam, Korea). Ampli cations were performed using the transgene-speci c primers listed in Tables 1 and 2 × PCR Master mix solution (iNtRON) containing 5 pmol of each primer set and 50 ng gDNA in a 10 µl reaction volume. PCRs were performed in a thermocycler under the following conditions: 94 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 30 sec, annealing for 30 sec (annealing temperatures depended on each primer set), and extension at 72 °C for 30 sec, with a nal extension at 72 °C for 7 min. The ampli ed PCR products were visualized using electrophoresis on a 1% agarose gel stained with ethidium bromide.  Similarity comparison between pig and the other species was con rmed on the score via the Protein BLAST program.

Immunocytochemistry (ICC) analysis
Before staining, all cell samples were preincubated for 5 min at 4 °C and xed with 4% paraformaldehyde for 15 min. After washing twice with Dulbecco's phosphate-buffered saline (DPBS; Welgene), the samples were treated for 1 h with 10% goat serum in DPBS to prevent nonspeci c binding. Serum-treated cells were incubated overnight at 4 °C with primary antibodies. The primary antibodies used were as follows:  Table 1 in a 10 µl reaction volume. Ampli cation and detection were conducted using the ABI 7300 Real-Time PCR system (Applied Biosystems) under the following conditions: one cycle of 50 °C for 2 min and 95 °C for 10 min, followed by 40 or 45 cycles of denaturation at 95 °C for 15 sec and annealing/extension for 1 min (annealing/extension temperatures depended on each primer set). The relative expression level was calculated by normalizing the threshold cycle (Ct) values of each gene to that of ACTB (beta-actin) via the Δ-Ct method (23).

Statistical analysis
The data from the qPCR analyses are presented as the mean ± standard error of the mean (SEM) and were analyzed using Prism 6 software (GraphPad Software; San Diego, CA, USA). The signi cance of differences was determined by one-way analyses of variance followed by Fisher's least signi cant difference test. Differences were considered signi cant at p < 0.05.

Results
Pig MYOD1 overexpression vector construction Myod1/MYOD1 is reportedly identi ed as a master transcription factor in myogenesis, thereby inducing myogenic transdifferentiation in nonmuscle cells (9,(24)(25)(26)(27). Thus, comparative MYOD1 sequence analysis was performed among various species to assess whether pig MYOD1 contains functionally conserved sequences for myogenesis, as in the other species, before vector construction for ectopic pig MYOD1 (pMYOD1) expression. The pig MYOD1 whole protein is composed of 319 amino acids and has a similar size to MYOD1 protein in other species (Fig. 1A). Additionally, in all the analyzed species, including pig, the MYOD1 protein contains acidic domains, histidine and cysteine-rich (H/C) domains, basic-helix1-loop-helix2 (bHLH) domains, and helix3 domains, which are located at 3-56 a.a, 63-99 a.a, 102-167 a.a, and 246-259 a.a (in the case of mouse: 245-258 a.a), respectively, indicating there are identical locations and sizes among the species (28,29). Subsequently, the MYOD1 sequences from those species were aligned, and the similarity score was assessed at the amino acid level (Fig. 1B and Table 2). In the bHLH domain, Ala 114 -Thr 115 and Arg 111 reportedly endow the MYOD1 protein with myogenic activity (30,31). In fact, these residues (marked in red; Fig. 1B) were con rmed to be conserved in the basic region of all the analyzed species. The MYOD1 whole proteins of pig and ungulates (pighorse: 94.36%, pig-cow: 95.92%, and pig-sheep: 95.61%) were more similar than those of pig and human (93.12%) or rodent (89.03%). It is noteworthy that the bHLH domain showed a 100.00% similarity between pig and all other species. Altogether, porcine MYOD1, especially the bHLH domain, was identi ed as an evolutionarily conserved protein, which seems that its role is also conserved in myogenesis across species.
Based on the above analyses, doxycycline (DOX)-inducible pMYOD1 overexpression vectors were generated including the MYOD1 gene isolated from satellite cells in 3-day-old LYD biceps femoris. To verify the function of these constructed vectors, they were introduced into pig embryonic broblasts (PEFs) through lentiviral infection for stable transgene expression (4,27). First, the integration of the exogenous pMYOD1 gene was con rmed by PCR targeting the FUW-tetO-pMYOD1 sequence in gDNA of PEFs infected with a lentivirus carrying the vectors (pMYOD1-PEFs) ( Fig. 2A). At day 9, the gene of interest had been inserted stably into the genome of PEFs. Then, MYOD1 expression was con rmed at the protein level using immunostaining in pMYOD1-PEFs (Fig. 2B). The vectors were activated by the addition of DOX, leading to ectopic MYOD1 expression. Finally, the expression pattern of muscleassociated genes (Exo-MYOD1, Endo-MYOD1, PAX7, MYF5, and MYOG) was analyzed in pMYOD1-PEFs using qPCR (Fig. 2C). These genes have been characterized as myogenic lineage-speci c markers, such as skeletal muscle progenitor/myoblasts (PAX7, MYF5, and MYOD1) and myocytes (MYOD1 and MYOG) (3). Exo-MYOD1 overexpression by vector activation increased the expression of endogenous muscleassociated genes. During extended cell culture, the expression of these genes was stably maintained. In conclusion, we constructed a DOX-inducible pMYOD1 overexpression vector that triggered the expression of muscle markers in long-term culture, indicating its stable function.

Myogenic transdifferentiation of PEFs by overexpression of MYOD1
The expression of endogenous muscle markers, including Endo-MYOD1, was enhanced via ectopic MYOD1 expression, as shown in Fig. 2C. However, because of the mild changes in each gene, transdifferentiated muscle cells from broblasts were not observed. It has been shown that full transdifferentiation is achieved by genetic modulation along with suitable culture conditions for speci c cell types (32), suggesting that optimization of culture conditions is required. According to previous studies, a 2-step transdifferentiation protocol was employed with some modi cation to derive myogenic cells through MYOD1 overexpression (8,17,33). In the 'induction' step, the myogenic program was activated with the stimulation of transcription factors associated with myogenesis, thereby leading to the commitment into a myogenic lineage. Brie y, pMYOD1-PEFs were treated with various signaling molecules, such as FGF2, SB431542, CHIR99021, and forskolin, which have been shown to be involved in the regulation of myogenesis (8,14). Then, transdifferentiation was promoted through serum starvation in the 'differentiation' step.
To ensure e cient myogenic conversion, we investigated the transition point where exogenous MYOD1 leads to the peak expression of endogenous skeletal muscle-speci c genes. In addition, pMYOD1-PEFs were cultured in mitogen-rich media in which myogenic induction was sustained without entering the differentiation process (Fig. 3A). While the FSCHF group had a long cylindrical shape due to elongation of the cytoplasm, a typical broblastic and round shape was observed in the control and FGF groups, respectively (Fig. 3B). The alteration of the FSCHF group was maintained by the end of the culture period, resulting in a similar morphology to myoblasts, as previously reported (17,34). The myogenic genes were upregulated by forced MYOD1 expression (Fig. 3C). Compared to other groups, the relative gene expression of the FSCHF group was higher across all the genes and culture periods. These results suggested that the Exo-MYOD1 effect was enhanced by the combination of four signaling molecules rather than DOX-induced exogenous MYOD1 per se or additional FGF2. While the gene expression of the FGF group was gradually changed, that of the FSCHF group was signi cantly increased on day 6 and then decreased. The aforementioned expression patterns were observed in all the markers except PAX7, which was upregulated up to day 9 in both groups. Taken together, pMYOD1-PEFs treated with a cocktail of FGF2, SB431542, CHIR99021, and forskolin for 6 days were used for further experiments.
Based on the above observations, we established a myogenic transdifferentiation protocol in which the FSCHF medium for induction into a myogenic lineage was replaced with a low-serum medium for the initiation of differentiation on day 6 (Fig. 4A). The replaced culture condition was classi ed into two groups distinguished by the addition of DOX (+ DOX and -DOX) to assess whether consistent activation of the myogenic program could enhance transdifferentiation. Notably, multinucleated myotubes via fusion of myoblasts were observed on day 8 in both groups (Fig. 4B). According to the qPCR analyses performed with a sample from day 9, transcripts of the Exo-MYOD1, MYF5, MYOG, and Myosin heavy chain (MHC) genes were upregulated in the +/-DOX groups (Fig. 4C). Unlike that of the other genes, gene expression of Endo-MYOD1 was signi cantly decreased in the +/-DOX group. Across all of the genes, especially MYF5 and MHC, the + DOX group showed signi cantly higher expression levels than the -DOX group, demonstrating that continuous MYOD1 overexpression during the differentiation step facilitates myogenic transdifferentiation. The expression of MHC, a marker of late differentiation in myogenesis, was detected in the day 9 sample by immuno uorescence analysis (Fig. 4D). Interestingly, a sarcomerelike structure with a striated pattern was observed, as previously reported (35), indicating that the mature myotube could be assembled. Therefore, the established protocol using ectopic MYOD1 expression and signaling molecules associated with myogenesis, such as FGF2, a TGF-β inhibitor, a WNT activator, and a cAMP activator, enabled broblasts to be reprogrammed into skeletal muscle.

Discussion
Pig MYOD1 has an evolutionarily conserved bHLH domain that regulates myogenesis Previously, numerous attempts have been made to reprogram nonmuscle cells into skeletal muscle by modulating the expression of the transcription factor Myod1, which plays an important role in myogenesis (10,11). 5-azacytidine (5-aza), a DNA methyltransferase inhibitor, induces demethylation of the Myod1 locus in differentiated somatic cells and increases Myod1 transcripts, leading to transdifferentiation into the myogenic lineage (9,36). However, 5-aza does not target site-speci c demethylation; it induces genome-wide demethylation because 5-aza is incorporated into genomic DNA as a competitive analog of cytosine and then disintegrates DNA methyltransferase by trapping (37). In fact, when broblasts were treated with 5-aza, adipocytes and chondrocytes were also observed (38). Additionally, it has been substantiated that 5-aza exhibits cytotoxicity by causing DNA double-strand breaks and apoptosis (39). Thus, myogenic transdifferentiation using 5-aza is inappropriate in terms of safety and e ciency. As an alternative strategy, forced Myod1/MYOD1 expression without induction by 5aza has been used to activate the myogenic program in human and mouse broblasts (27,34,40). In pig, myotubes were differentiated from induced pluripotent stem cells (iPSCs) via a combination of a 5-aza treatment and ectopically expressed MYOD1 (17). Although contractile porcine myotubes with sarcomeres were obtained from iPSCs within eleven days, the entire process would be more timeconsuming because of the iPSC derivation period. Here, we established a direct transdifferentiation protocol with pig broblasts for the generation of skeletal muscle using porcine MYOD1 overexpression and myogenesis-associated signaling molecules, bypassing the induction of the pluripotent state.
Comparative analysis of the MYOD1 a.a sequence showed that the bHLH domain in MYOD1 was thoroughly conserved across various species. A previous study suggested that conserved sites are less permissive to evolutionary mutation due to their functional or structural importance (41). This nding indicates that the bHLH domain is an essential part of the MYOD1 protein, the function of which is associated with myogenesis, as in other species (34,42). The basic region in the bHLH domain of Myod1 recognizes and binds the E-box, which is a conserved DNA sequence that is frequently distributed throughout the genome rather than speci cally located in the regulatory region of myogenic genes (42).
The myogenic speci city of Myod1 is derived from Ala 114 -Thr 115 in the basic region (29,43). In fact, it has been veri ed that Myod1 binds to both canonical and noncanonical E-boxes (10). These myogenic codes direct Myod1 to bind to noncanonical E-boxes of Myog, one of the Myod1 target loci, and to interact with Pbx/Meis cofactors associated with myogenic genes (30). In particular, Ala 114 leads to an appropriate conformational change allowing myogenic activity by mediating the contact of Arg 111 and guanine in DNA (31). In addition to the basic region, the HLH motif of Myod1 is also required for myogenesis because the HLH motif dimerizes with other bHLH proteins, whose basic domains are involved in E-box binding (42). The other domains are also highly conserved and have been known to be functional Myod1 domains (11,44). For example, the acidic domain acts as a transcriptional activation domain (TAD) through additional DNA binding near E boxes, and both the H/C and helix3 domains are involved in chromatin remodeling to allow active transcription of sequential myogenic genes. Accordingly, pig MYOD1 is capable of muscle-speci c gene expression through conserved domains, especially the bHLH motif, preventing differentiation into nonmuscle cell fate.
Myogenic transdifferentiation is enhanced by activation of the FGF, WNT, and cAMP signaling pathways and inhibition of the TGF-β signaling pathway Doxycyline (DOX)-inducible pMYOD1 overexpression vectors were produced with the conserved pig MYOD1 sequence, as described above. Because differentiation is blocked when proliferation is enhanced by mitogen in serum, a low concentration of horse serum has been widely employed for myogenic differentiation in humans and mice (45)(46)(47). However, transdifferentiated muscle cells were not observed when the aforementioned conventional culture conditions were applied. This result was consistent with the qPCR data that showed mild upregulation of endogenous muscle genes (Fig. 2C). It was surmised that myogenic reprogramming in pigs requires more supportive culture conditions than in mice and humans. For these reasons, a mixture of small molecules, such as FGF2, a TGF-β inhibitor (SB431542), a WNT activator (CHIR99021), and a cAMP activator (forskolin), was selected to facilitate the conversion of cell fate into myogenic lineage, as applied in previous research (8). While activation of the FGF, WNT, and cAMP signaling pathways is required for speci cation into the myogenic lineage and proliferation of committed myoblasts, inhibition of the TGF-β signaling pathway is involved in myotube formation (3,13,15). These signaling molecules, except forskolin, are secreted during myogenesis in vivo, thus recapitulating the endogenous signaling pathway for muscle formation and regeneration. Forskolin has been reported to be involved in skeletal muscle differentiation from human iPSCs and to improve satellite cell expansion in mice (16). Supporting the function of forskolin, our preliminary study showed that the removal of forskolin resulted in the failure of myogenic transdifferentiation (data not shown).
With treatment of these four signaling molecules, endogenous myogenic genes were highly triggered by exogenous MYOD1 (Exo-MYOD1) (Fig. 3C). Consistent with these observations, it has been demonstrated that ectopically expressed MYOD1 upregulates endogenous MYOD1 (Endo-MYOD1), PAX7, MYF5, and MYOG, which reinforces the notion that MYOD1 is a key regulator in myogenesis. Exo-MYOD1 increased Endo-MYOD1 via an autoregulatory loop, as previously reported (26). Chromatin immunoprecipitation reportedly showed that mesoderm or myoblast markers, including Pax7/PAX7 and Myog/MYOG, were directly activated by Exo-Myod1, leading to an increase in Myf5 (27,48). Because MYF5 is upstream of MYOD1, it was reported that MYF5 was not expressed in MYOD1-overexpressing human iPSCs (33). However, when treated with signaling molecules, forced Myod1 expression converted mouse broblasts into Pax7and Myf5-positive myogenic progenitor cells (8). The temporal expression of myogenic genes has a hierarchy with stage-speci c markers: PAX7 in myogenic progenitor cells and MYF5 and MYOD1 in committed myogenic cells, followed by MYOG in the differentiation phase (3,49). Therefore, various cell types that belong to the myogenic lineage can be derived in mitogen-rich media. In the C2C12 myoblast cell line, mitogen-rich culture conditions stimulated myoblast growth without differentiation, whereas skeletal muscle was differentiated with the expression of differentiation-speci c genes under lowmitogen culture conditions (50). For the enrichment of terminal differentiated myotubes, the culture conditions were switched, and a medium containing 2% horse serum without signaling molecules on day 6 was used, in which endogenous muscle genes peaked by Exo-MYOD1.
As shown in the applied protocol presented in Fig. 4A, multinucleated and elongated myotubes were observed. These myotubes were formed through the fusion of mononucleated myoblasts exiting from the cell cycle, followed by reorganization of the cytoskeleton (3). Based on the higher expression of muscle genes in the + DOX group, continuous MYOD1 overexpression seems to enhance myogenic conversion in a feed-forward mechanism (42). Myotubes and myo ber express myosin heavy chain (MHC) (6), which is a downstream gene of Myod1 (26). MHC provides contractility to eukaryotic cells through lament assembly in the form of striated sarcomeres, such as skeletal and cardiac muscle (51). In fact, the expression of MHC was increased at the RNA level and a sarcomere-like structure was detected by immunostaining for MHC proteins in accordance with previous research (35). Altogether, our protocol enables the transdifferentiated muscle to undergo terminal differentiation and maturation into skeletal muscle.

Conclusions
In summary, fully differentiated somatic cells of pigs can be reprogrammed into mature skeletal muscle by the pig MYOD1 gene. Modulation of the FGF, TGF-β, WNT, and cAMP signaling pathways is required for the cell fate conversion into the myogenic lineage. The transdifferentiated muscle expressed skeletal muscle markers and had the structure of a striated sarcomere, implying that these matured myotubes possess contractile capacity. Given the role of pig as a signi cant livestock for supplying meat, the myogenic reprogramming of pig cells can be applied to increase agricultural yield and produce cultured meat. Additionally, pigs provide biomedical applications in preclinical studies for human disease because of their anatomical and physiological similarities with humans (52,53). For example, transdifferentiated pig muscle cells can offer a cell source for skeletal muscle disease modeling and drug screening in regenerative medicine. Finally, this study provides fundamental knowledge for developmental biology in revealing the genetic network and signaling pathways underlying myogenesis. The care and experimental use of pigs were approved by the Institutional Animal Care and Use Committee (IACUC) at Seoul National University (approval nos.: SNU-140328-2 for pig fetal broblast isolation).

Consent for publication
Not applicable.

Availability of data and material
Not applicable.

Competing interests
No potential con icts of interest relevant to this article are reported.