*3.7. Microarray Assessment of Gene Expression in TTRkd Cells and E*ff*ect of T4 on Gene Expression*

To explore TTR function in myoblast differentiation, TTRkd and TTRwt C2C12 cells were cultured with 2% FBS for two days. TTR/MYOG expression and myotube formation were decreased by TTRkd (Supplementary Figure S1A,B). Microarray analysis was performed with TTRwt and TTRkd cells. After applying two-fold cut-offs for down- and up-regulated genes, analysis of the effects of knocking down TTR on myoblasts revealed that, among the genes involved in sarcomere formation, specific genes are actively up- or down-regulated, and some novel genes that are not involved in sarcomere formation functioned at the onset of myogenesis. Among the identified genes, 29 and 7 genes were down- or up-regulated, respectively, by greater than two-fold in TTRkd cells (Table 1A,B; Supplementary Figure S1C,D). Many genes were previously reported to be involved in MSC maintenance (Heyl, Sox8), myogenesis (Fgf21, Ankrd2, Sox8, Asb2), proliferation (Ankrd2), myokine secretion (Fndc5), neuromuscular junction (Dok7), and Ca2<sup>+</sup> release of sarcoplasmic reticulum (Asph). Even though some of these genes have identified roles in myogenesis, many novel genes were also affected by TTRkd (R3hdml, Inpp4b, Igf2as, Btbd17, Sema6b and Ddc) (Table 1A; Supplementary Table S4A). However, the upregulated genes were mostly involved in the cell cycle, cell proliferation, and transcription regulation. Interestingly, there was little information indicating that those genes were related to muscle differentiation. Moreover, most of the up-regulated genes were novel genes, but their main functions have been studied in other tissues or organs (Table 1B; Supplementary Table S4B).


#### **Table 1.** Microarray analysis of TTR knockdown.

variant 2


**Table 1.** *Cont*.

TTRwt or TTRkd were cultured with 2% FBS for two days and microarray analysis was performed on TTRwt or TTRkd. (**A** and **B**) List of down- or up-regulated genes in TTRkd (2-fold≤). (**C** and **D**) Functional analysis by DAVID (2-fold≤). TTRwt indicates cells transfected with scrambled vector. Means ± SD (*n* = 3).

Down-regulated genes were analyzed at different myogenic times (0, 2, 4 and 6 days). Interestingly, most gene expressions were increased under myogenic conditions than that at the proliferating stage (Day 0). Similar to MYOG (the myogenic marker gene) the expression of 15 genes increased greatly during myogenic differentiation (Supplementary Figure S2). DAVID analysis was performed using the up- and down-regulated genes. More than half of the up-regulated genes were classified as transcription regulators (Table 1C), especially cell-cycle regulators. Although some of the down-regulated genes were identified as being involved in Ca2+-mediated signal transduction and were reported to regulate transcription, most down-regulated genes were classified as components or regulators of the sarcomere motor unit or the ATPase-related group, which are the main structural components of the sarcomere (Table 1D).

Even though genes were selected based on their high statistical significance among all differentially expressed genes, the genes were also cross-examined by performing real-time RT-PCR with TTRkd and comparing the results to those of TTRwt (Figure 7A). To investigate the effect of T4 on TR expression, cells were grown under serum-free conditions with added T4 and/or TR-specific antagonist 1-850 and examined both for morphological appearance and for changes in mRNA expression levels of certain genes. Myotube formation and mRNA levels of the myogenic marker genes (MYOD, MYOG and MYL2) were decreased by T4 + 1-850 treatment (Figure 7B). In contrast, T4 treatment increased myofibril diameter. The T4 treatment elevated most of the gene mRNA levels, whereas T4 + 1-850 treatment had opposite effects. However, T4 treatment reduced the suppressing effect of 1-850 on mRNA expression of Nmrk2 (40% rescue), Sox8 (40%), Myh1 (20%), and Myh8 (30%) (Figure 7C).

To determine whether the TTRkd effects were produced by TH and its specific receptor, the TR binding site was scanned in genome portions containing the 5 flanking region and the first intron of each gene. For precise analysis, two nuclear receptor scanning software programs, NHR-scan and NUBI-scan, were utilized. All binding site candidates were predicted by using the AGGTCA sequence arranged by the DR0, DR4, IR0, IR4, ER4, and ER6 patterns, as was used in the in silico thyroid hormone response elements (TRE) prediction models. Consequently, most of the genes contained more than one TRE at the 5 flanking region. However, some genes such as Fgf21 did not have a suitable TRE. In addition, Myh1 did not possess a TRE upstream of the first exon (Supplementary Figures S3 and S4). Altogether, these results showed that T4 transported to the cell interior activated TR to induce gene expression and modulated novel and major transcription regulating genes that markedly increased during myogenic differentiation in a TH-dependent manner.

**Figure 7.** Expression of down-regulated genes in TTR knock-down cells and effect of T4 treatment on down-regulated genes. (**A**) TTRwt or TTRkd were cultured with 2% FBS for two days. Down-regulated gene expression was assessed by real-time RT-PCR in TTRwt or TTRkd. (**B**) Cells were cultured with serum-free media supplemented with T4 or T4 + 1-850 and incubated for two days. Myotube formation and fusion index were observed by Giemsa staining and mRNA expression by real-time RT-PCR. (**C**) Cells were incubated without or with T4, T4 + 1-850 or 1-850 for two days. Expression of down-regulated genes without or with T4, T4 + 1-850 or 1-850 by real-time RT-PCR. Control indicates non-treated cells. Means ± SD (*n* = 3). \* *p* ≤ 0.05, \*\* *p* ≤ 0.001, \*\*\* *p* ≤ 0.0001.

#### *3.8. FNDC5 Expression During Myoblast Di*ff*erentiation*

To confirm the function of the genes that were down-regulated by TTRkd, myokine FNDC5 was selected. FNDC5 knockdown was performed followed by culture with 2% FBS for two days. Myotube formation and myogenic gene expression were decreased in FNDC5kd cells, whereas TTR and TRα expressions were increased at both the transcriptional and translational levels (Figure 8A). Next, cells were grown in serum-free media or supplemented with T4 for two days, and the FNDC5 mRNA level was analyzed in normal cells and exosomes from plasma and media of cultured cells (FNDC5kd and FNDC5wt). FNDC5 mRNA was evident in exosomes from culture media and plasma and decreased in FNDC5kd cells (Figure 8B). Additionally, decreased FNDC5 mRNA was observed in T4 + TTR treatment in MSCs exosomes (Figure 8B). Expression of FNDC5 was decreased in 26-week muscle compared with that in 16-week muscle (Figure 8C). These results show that FNDC5 positively regulates myoblast differentiation.

**Figure 8.** FNDC5 expression during myoblast differentiation. (**A**) FNDC5 knockdown was performed and cells were incubated with 2% FBS for two days. Myotube formation and fusion index were observed by Giemsa staining, mRNA expression using real-time RT-PCR and protein expression were observed by Western blot and immunocytochemistry. (**B**) Cells were cultured with only serum-free media for two days and exosomes were isolated from cultured media. FNDC5 mRNA level in normal cells, exosomes isolated from plasma, and media of cultured cells (FNDC5wt and FNDC5kd). MSCs were cultured with only serum-free media or supplemented with T4 for two days. FNDC5 mRNA level in exosomes from cell, media of cultured cells with T4 or T4 + TTR. (**C**) FNDC5 protein expression in 16 or 26-week muscle by immunohistochemistry and Western blot. FNDC5wt indicates cells transfected with scrambled vector. Means ± SD (*n* = 3). \* *p* ≤ 0.05, \*\* *p* ≤ 0.001, \*\*\* *p* ≤ 0.0001.

#### **4. Discussion**

Skeletal muscle accounts for nearly half of the body mass and represents the largest protein reservoir in the human body [31]. Although the importance of TH signaling in muscle physiology has been documented for several years, its precise mechanism in skeletal muscle during postnatal myogenesis remains unclear. Initially, we demonstrated the role of TTR in sustaining the cellular T4 level during myoblast proliferation and differentiation [6,29]. In this study, we give the first direct evidence of TTR secretion and uptake in C2C12 mouse myoblast cells. We also identify TTR mRNA in exosomes and its increased expression following T4 treatment, which may act as a mediator in this process. In addition, we studied the role of TTR in T4 transport into C2C12 cells and murine MSCs during the assessment of cell viability and differentiation. The appearance of TTR in cultured serum-free media from myoblasts strongly suggests that TTR synthesized by C2C12 cell is secreted. This suggestion was confirmed by TTR immunoneutralization using TTR antibody, which demonstrated a reduction in myotube formation and mRNA level of some myogenic marker genes (especially, MYOD and MYOG), and T4 uptake into cells, along with an increase in TTR retained in the cells. Further, it is important to emphasize the presence of T3 in exosomes, which indicates that T3 produced in cells is secreted out of the cell through exosomes. These results imply that muscle may not only utilize T4 but also act as a reservoir of T3 in order to distribute it to other tissues or to more distant sites.

Cosmo et al. reported that TH uptake by skeletal muscle can occur independently of monocarboxylate transporter 8 (Mct8). However, they found enhanced TH action, T3 content, and glucose metabolism in Mct8 knockout mice [32]. We speculate that TTR might maintain the TH content in Mct8 knockout mice and, hence, normal muscle metabolism and development. The binding affinity of TTR for T4 is high, hence, it serves as a primary distributor protein in muscle. We showed that TTR with T4 treatment significantly increased cell viability and differentiation compared to that of only T4 treated cells. This was consistent with our previous finding that TTR expression increases myoblast differentiation by increasing T4 transport into the cell [29]. Similar to what we observed in the C2C12 cell line, the T4 + TTR protein treated mouse MSCs also showed increased myotube formation with elevated T4 concentration. Additionally, a progressive increase in TTR expression was observed during differentiation (day 2) in primary MSC cultures. These data confirm that TTR promotes myogenesis by enhancing the transport of T4.

Kassem et al. showed that the availability of TTR in cerebrospinal fluid (CSF) was associated with enhanced T4 uptake into the choroid plexus and brain and this uptake was increased in the presence of TTR [33]. Accordingly, in the present study, low T4 concentration in media with its consequent high concentration in cells supplemented with T4 + TTR indicated that TTR enhanced the transport of T4 to the cell interior during myoblast viability. The enhanced cell uptake may be a simple consequence of the increased T4 level in serum, providing a concentration gradient that promotes TTR secretion and subsequent cell uptake. Furthermore, increased uptake of TH in cells treated with both T4 and TTR probably involves a T4 complex with TTR, as well as passive diffusion of T4, allowing for greater cell uptake than can be accomplished by diffusion only, which is consistent with observations in human ependymoma cells [34]. Although TTR has been reported to be the main component in maintaining high TH levels in CSF and brain [33,35], in this study we observed that TTR also sustained the TH concentration in skeletal muscle and, hence, promoted myogenesis.

TTR is one of three proteins required for T4 transport: TBG is the major transporter and albumin has the lowest affinity, acting as the third T4 binding protein in human plasma [25,36]. Consistent with this theory, we found that BSA reduced T4 transport to the cells, which was also increased with TTR treatment as it has high efficiency for TH. Additionally, BSA treatment decreased myotube formation and myogenic protein expression, while TTR and D2 expressions were increased at the translational level, which might reflect the drop in the T4 level in the cells. TH regulates several genes that are responsible for muscle development and homeostasis. Among those genes, MYOD, MYOG and contractility-determining proteins are transcriptionally regulated by TH and are important for regeneration and myogenesis [37]. MYOD expression regulated by TH is involved in the fast muscle fiber phenotype, with transcriptional stimulation of the myosin-1, myosin-2 and myosin-4 isoforms [38]. TH metabolizing enzyme D2 can activate TH by outer-ring deiodination and can influence local tissue TH levels [39]. Collectively, our findings suggest that TTR acts to maintain the TH level in myoblast cells.

Evidence of high fluorescence-labeled TTR protein levels in cells reveals that TTR was internalized into the myoblast cells. This supports previous results showing endocytosis of fluorescence-labeled TTR in ependymoma cells [34]. In other reports, 125I-TTR and digoxigenin labeled TTR were internalized by an endocytic process in rat yolk sac and β-cells, respectively [40,41]. Furthermore, increased uptake of T4 in TTR-overexpressing cells supplemented with T4 implies that an even distribution of T4 within the cell is not only dependent on the free fraction of T4 in serum but also on the T4 bound to TTR. The presence of T4 or T3 significantly enhanced TTR internalization in JEG-3 cells, with TTR entering the cells as a TTR-T4 complex [27]. In addition, Divino and Schussler [42] reported increased TTR internalization in HepG2 cells with increasing amounts of T4 and suggested that a T4-stimulated conformational alteration in TTR somehow enhanced the uptake of TTR.

Reduced T4 serum concentrations have been reported in old rats [43,44], though their serum T3 level remains more controversial [43]. We show that TTR and D2 expressions with T3 concentration have a correlation with muscle age. The reduced D2 activity is suggestive of impaired T4 conversion in 26-week muscle. Silvestri et al. observed reduced D1 activity in 26-month-old rats relative to that in young (6- and 12-month-old male) rats [44]. Interestingly, decreased TTR expression in 26-week muscle was consistent with the decreased TH transporter Mct8 protein level in liver of 24-month-old rats [44]. Furthermore, higher plasma T3 or T4 concentration in 26-week muscle could be associated with the reduced free T4 concentration in the 16-week muscle, probably due to a higher TBG expression, as described elsewhere [45]. Additionally, decreased T3 concentration in the 26-week muscle at the cellular level was consistent with the findings of Silvestri et al. [44] in which decreased T3 concentration was observed in 24-month-old rats. Nevertheless, T3 generation has been observed in 11-month-old rats relative to that in seven-month-old rats [46], indicating that the mechanisms of T3 production from T4 in old muscle remain poorly understood.

TH is the main endocrine regulator that acts by binding to TRs and imposing a signature type of gene expression [11]. TH primarily functions either via nuclear receptor-mediated stimulation that is T3 dependent or by switching off the gene transcription machinery [13]. In muscle, this signaling pathway is regulated by the THRA1 isoform of TR [47]. The heterodimer complex formed by the TR with RXR- binds to a TRE, leading to activation or suppression of gene transcription [13]. Accordingly, we showed that T4 treatment induced RXRγ expression. However, myotube formation and myogenic factors were decreased in RXRγ and TRα knockdown cells. Interestingly, RXRγ knockout mice are unable to increase their mass in response to high-fat feeding, suggesting a specific effect of RXRγ in skeletal muscle [48]. In muscle, the proteins whose expression are transcriptionally controlled by T3 are SERCA1a [12], SERCA2a [49], uncoupling protein 3 (UCP3) [50], GLUT4 [51], cytosolic malic enzyme (ME1) [52], muscle glycerol-3-phosphate dehydrogenase (mGPDH) [53], and myosin-7 [54]. Furthermore, we found that TTR and D2 expression were decreased in TRαkd cells, which explains the retarded transport of TH into the cell. The selective functions of TRs are controlled by local ligand availability [39,55] or by TH transport to the cell interior via Mct8 or other associated transporters [56]. The TH metabolizing enzymes D2 and D3, as well as transporters Mct8 and Mct10, are expressed in both rodent and human skeletal muscle [57,58].

The TTR-affected genes identified by TTRkd-based microarray analysis included important transcription factors or mediators that have the potential to control several other genes. For example, Rbm24 is reported to regulate MYOG expression [59] and mediate skeletal muscle-specific splicing events [60]. In contrast, Sox8, a negative regulator of myogenesis [61], has increased expression during myogenesis of C2C12 cells. In addition, Sox8 and Heyl genes are marker genes of MSCs [61,62]; however, the Heyl gene showed increased expression during myogenic differentiation. Interestingly, those opposing results were also observed for Nmrk2 [63]. Another research group reported that Ddc is not produced by myotubes [64], but in the present study, it was induced by suitable myogenic differentiation. Altogether, some genes that have been reported to be negatively correlated with myogenesis were markedly increased in expression during myogenic differentiation in this study.

Another interesting observation from the time-course expression study is that several novel genes that show increased expression during myogenesis responded to T4 as they did to TTR. However, Inpp4b and Asb2, genes that contain TREs in proximity to the transcription start site (TSS), did not show any change with T4 treatment. In the case of Inpp4b, TREs in the proximity of the promoter were only downstream of the TSS, and the first intron was approximately 130 kb. This characteristic indicates a rare aspect of the TTRkd-affected genes. Moreover, Fgf21 and Myh1, which do not seem to contain TREs, showed increased expression levels. The various TRE elements have only been predicted by a one-dimensional arrangement, moreover, a proper, precise, and complete nucleotide matrix for this one-dimensional arrangement is not present in public databases. Due to these limitations, many other researchers [65,66] have reported different nucleotide matrices for TREs and different reactivity of each.

Interaction and cooperation between TR and the mammalian insulator CCCTC-binding factor have already been reported [67,68]. An insulator can mediate multi-dimensional chromosomal changes [69]. In addition, based on the results of the TTRkd microarray analysis, the T4 affected sarcomere genes Myh1, Myh3 and Myh8 may be suitable candidates for TR-insulator mediated transcriptional regulation. In the case of the Myh1 gene, no TREs were present in its promoter region.

In contrast, the FNDC5 gene, downregulated by TTRkd, also showed a high expression level during myogenic differentiation and after T4 treatment. The FNDC5 gene encodes the irisin protein, which is considered as a circulating myokine. The most remarkable feature of the FNDC5/irisin protein is that it generates brown fat from white fat [70,71]. Recently, it has been shown that irisin injection stimulated muscle hypertrophy and increased regeneration in injured skeletal muscle [72]. Additionally, enhanced irisin levels have been found during myogenic differentiation and the additional irisin enhances the expression of p-Erk, which has a vital role in the protein synthesis pathway [73]. Thus, knockdown of the FNDC5 gene was undertaken. We showed that interruption of the FNDC5 gene produced a low level of myotube formation. In humans, FNDC5 protein is cleaved to provide detectable irisin levels in circulation. Additionally, increased irisin concentrations occur in response to exercise in humans [74]. Therefore, based on the pro-myogenic role of FNDC5 in the present study, we suggest that FNDC5 may be a potential curative target for the intrusion of muscle dystrophy. Thus, we conclude that one control pathway within TTR myogenesis is mediated by the protein FNDC5.

#### **5. Conclusions**

In conclusion, these results suggest that: (1) a portion of the extracellular T4 enters myoblasts or myocytes via MCT via passive diffusion and is converted to T3 by the D2 enzyme which, in turn, induces the expression of several genes including TTR; (2) synthesized TTR exocytoses the cell through exosomes; (3) TTR brings T4 inside the cells as a TTR-T4 complex through an endocytic mechanism; (4) intracellularly synthesized T3 can exocytose via exosomes (Figure 9A); and (5) TTR, through the action of T3 converted from T4, regulates gene expression of TTR intermediates, such as RXRγ and FNDC5 (irisin), which ultimately induces myogenesis (Figure 9B). In this study, we have shown that muscle cells use a much more active mechanism than previously thought to bring T4 into cells. Moreover, intracellularly-generated T3, besides being used in the target muscle cells, also moves out of the cell and affects adjacent cells as well as probably other tissues. Herein, we propose a novel mechanism for the uptake and release of T4 and T3 in myoblasts and for TTR to act as a sensor for intracellular T4 during myogenesis. However, this study has presented a most rudimentary picture of T4 and T3 transport into and out of muscle cells, and further studies will undoubtedly reveal more detailed mechanisms.

**Figure 9.** Hypothesis for the role of TTR with T4 during myoblast differentiation. (**A**) Hypothetical figure depicting role of TTR with T4 during myoblast differentiation. (1) T4 enters cells via Mct8 by passive diffusion and is converted to T3 by D2 enzyme, which in turn triggers the expression of several genes including TTR. (2) Synthesized TTR is exocytosed through exosomes, and (3) subsequently enters the cells as TTR-T4 complex via an endocytic mechanism. (4) T3 produced in the cells can exocytose via exosomes. (**B**) TTR positively regulates RXRγ and FNDC5 and triggers myogenic regulatory factors, hence promoting myogenesis. RXRγ and FNDC5 negatively regulate TTR while RXRγ and FNDC5 regulate each other.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4409/8/12/1565/s1, Figure S1: Microarray analysis of TTRkd cells, Figure S2: Time-course study of down-regulated genes during myoblast differentiation, Figure S3: Promoter of down-regulated genes was analyzed to predict TRE binding site, Figure S4: Promoter of down-regulated genes was analyzed to predict TRE binding site, Table S1: shRNA information, Table S2: Primer information, Table S3: Molecular weight of protein, Table S4: Functional analysis of up- or down-regulated genes affected by TTRkd.

**Author Contributions:** Conceptualization: E.J.L. and I.C.; formal analysis: Y.-W.K. and I.C.; funding acquisition: E.J.L. and I.C.; investigation: E.J.L. and D.C.; methodology: J.H.L., Y.-H.L. and S.-Y.P.; resources: S.J.P. and S.-Y.P.; writing—original draft: E.J.L., S.S. and I.C.; writing—review and editing: E.J.L., S.S., K.A. and M.H.B.

**Funding:** This research was supported by the National Research Foundation of Korea (NRF) funded by the Korean government (MSIP: grant no. NRF-2018R1A2B6001020) and a grant from the Next-Generation BioGreen 21 Program (project no. PJ01324701), Rural Development Administration, Republic of Korea.

**Conflicts of Interest:** The authors declare that they have no conflict of interest.
