Inhibition of Gap Junctional Intercellular Communication Upregulates Pluripotency Gene Expression in Endogenous Pluripotent Muse Cells

Gap junctions (GJ) are suggested to support stem cell differentiation. The Muse cells that are applied in clinical trials are non-tumorigenic pluripotent-like endogenous stem cells, can be collected as stage-specific embryonic antigen 3 (SSEA-3+) positive cells from multiple tissues, and show triploblastic differentiation and self-renewability at a single cell level. They were reported to up-regulate pluripotency gene expression in suspension. We examined how GJ inhibition affected pluripotency gene expression in adherent cultured-Muse cells. Muse cells, mainly expressing gap junction alpha-1 protein (GJA1), reduced GJ intercellular communication from ~85% to 5–8% after 24 h incubation with 120 μM 18α-glycyrrhetinic acid, 400 nM 12-O-tetradecanoylphorbol-13-acetate, and 90 μM dichlorodiphenyltrichloroethane, as confirmed by a dye-transfer assay. Following inhibition, NANOG, OCT3/4, and SOX2 were up-regulated 2–4.5 times more; other pluripotency-related genes, such as KLF4, CBX7, and SPRY2 were elevated; lineage-specific differentiation-related genes were down-regulated in quantitative-PCR and RNA-sequencing. Connexin43-siRNA introduction also confirmed the up-regulation of NANOG, OCT3/4, and SOX2. YAP, a co-transcriptional factor in the Hippo signaling pathway that regulates pluripotency gene expression, co-localized with GJA1 (also known as Cx43) in the cell membrane and was translocated to the nucleus after GJ inhibition. Adherent culture is usually more suitable for the stable expansion of cells than is a suspension culture. GJ inhibition is suggested to be a simple method to up-regulate pluripotency in an adherent culture that involves a Cx43-YAP axis in pluripotent stem cells, such as Muse cells.


Introduction
The cells of multicellular organisms have many different ways to communicate with neighboring cells to maintain cellular and tissue homeostasis. Gap junction (GJ) channels are one of the representative systems that enables the exchange of small molecules, second messengers, and electrical signals to connect the cytoplasmic space of two adjacent cells. GJs consist of connexin (Cx) proteins that oligomerize into hexamers to form hemi-channels, namely connexons. One connexon is presented from each cell to form a GJ channel between the two cells [1,2].
GJ intercellular communication (GJIC) is also suggested to be important in regulating stem cell differentiation. For example, the ablation of GJIC led to the disruption of differentiation in embryoid bodies derived from mouse embryonic stem (ES) cells [3]. Moreover, human induced pluripotent stem cells (iPSCs) programming process is accompanied growth factor (bFGF), rather than the leukemia inhibitory factor (LIF), to maintain their proliferation and self-renewability [5].
Here, we show that GJIC inhibition in Muse cells with GJ inhibitors, such as dichlorodiphenyltrichloroethane (DDT), 12-O-tetradecanoylphorbol-13-acetate (TPA), and 18α-glycyrrhetinic acid (18α-GA), as well as the introduction of siRNA for GJA1, upregulates the expression of pluripotency-related markers in the adherent culture. An RNAseq analysis suggested that GJIC inhibition suppressed the expression of cell differentiationrelated pathways. Finally, our data suggested that there is involvement from the Cx43/Yesassociated protein (YAP) in the pluripotency gene up-regulation in GJ-inhibited Muse cells. YAP, a key transcriptional factor in the Hippo signaling pathway that plays an important role in cell proliferation, stem cell self-renewal, and tissue formation, is suggested to be anchored to connexins in the cell membrane and is released from the connexins and translocated to the nucleus under the presence of the GJ inhibitors. Since YAP is known to bind to the promoter regions of pluripotency genes, translocated YAP might have up-regulated the pluripotency genes.
Muse cells, as well as ES cells and iPSCs, were shown to up-regulate pluripotency gene expression when they were cultured in suspension [7,13,24,25]. Adherent culture is, however, more suitable for the stable expansion of cells on large scale than a suspension culture is. GJIC inhibition might be a simple method that enables both the stable cell expansion in adherent culture and the up-regulation of pluripotency gene expression.

Reagents Preparation
18α-GA was purchased from Sigma (#G8503), and a 25 mM stock solution was prepared by dissolving in DMSO (Wako, Osaka, Japan) and was preserved at −30 • C. A fresh working solution of 120 µM was used for 24 h incubation in the main experiment.
TPA was purchased from AdipoGen Life Sciences (Liestal, Switzerland, #AG-CN2-0010), and a 500 µM stock solution was prepared by dissolving in DMSO and was preserved at −30 • C. A fresh working solution of 400 nM was used for 24 h incubation in the main experiment.
DDT was purchased from Tokyo Chemical Industry (TCI, Tokyo, Japan, #T0379), and a 250 mM stock solution was prepared by dissolving in DMSO and was preserved at −30 • C. A fresh working solution of 90 µM was used for 24 h incubation in the main experiment.
Etoposide was purchased from Sigma (#E1383), and a 50 mM stock solution was prepared by dissolving in DMSO and was preserved at −30 • C.
qPCR was performed using 7500 Fast Real-Time PCR System (Applied Biosciences, Waltham, MA, USA). β-actin was used as endogenous control. qPCR data analysis was performed, using (2 −∆∆CT ) relative quantification method in all calculations.
Signal intensity was measured using ImageJ software by selecting each individual cell and measuring the area, mean intensity, and integrated density. Corrected Total Cell Fluorescence (CTCF) was calculated using the following formula: CTCF = Integrated Density − (Area of selected cell × Mean fluorescence of background readings).

Apoptosis Assessment Assay
LIVE/DEAD™ Fixable Aqua Dead Cell Stain Kit (Thermo Fisher Scientific, # L34965) was used according to the manufacturer's instructions.

Cell Cycle Analysis
Samples were collected using trypsin into 1.5 mL microcentrifuge tubes, the cell pellet was washed with 1× PBS, and then the cells were fixed with 70% EtOH for 30 min at 4 • C. After washing, cells were treated with 100 µg/mL PureLink RNase A (Thermo Fisher Scientific, #12091021) for 15 min at 37 • C, followed by PI staining (Thermo Fisher Scientific, #P3566) diluted at 1:1000. Samples were analyzed, using CytoFLEX Flow Cytometer (Beckman Coulter, Brea, CA, USA). Cell cycle analysis was performed using Kaluza Analysis Software (Beckman Coulter).

Dye Transfer Assay
Dye transfer assay was performed as described previously [26]. Briefly, Muse cells were plated onto a 12-well plate at 80% confluency and were incubated overnight. The next day, the culture medium was changed to DDT, TPA, 18α-GA, or DMSO (control) contained medium. Five hundred nM DiIC (Fuji film) was also included in the culture medium described above, and cells were further incubated for 24 h. Muse cells were, afterward, co-cultured with 100 nM Calcein-AM (Dojindo, Kumamoto, Japan)treated MSC cells for 2.5 h. Finally, co-cultured cells were collected for FACS analysis, using CytoFLEX Flow Cytometer. DiIC-stained cells were detected by phycoerythrin (PE) filter, and Calcein-AM-stained cells were detected by fluorescein isothiocyanate (FITC) filter. Cells with active GJIC appeared as the FITC/PE-double positive population, due to the transfer of Calcein-AM dye into DiIC labeled cells through GJ channels, whereas cells with inactive GJIC were identified as cells positive only for PE.

RNA-Sequencing
Muse cells, plated onto a 12-well plate at 80% confluency, were treated with 90 µM DDT or DMSO control for 24 h and were collected after trypsinization for total RNA collection by using NucleoSpin RNA XS RNA Isolation Kit (Takara Bio), according to the manufacturer's instructions. Total RNA was, thereafter, subjected to RNA-sequencing by using Illumina HiSeq2500 (San Diego, CA, USA) at a rapid mode. RNA-seq data were mapped using tophat software, and the reads were aligned using hg19 genome database provided by Illumina. Cufflinks was used to detect gene expression levels in the mapped reads, and cuffdiff was used to compare expression levels between samples. Data were then analyzed using multiple data analysis tools, such as DEG analysis, GO analysis, and differential pathway regulation. The following software and tools were used for RNAseq data analysis: bowtie2, mapping, ver. cDNA library preparation, sequencing, and reads mapping were performed by the Division of Cell Proliferation at the Tohoku University Graduate School of Medicine.

Cx43 Knockdown
Cx43 siRNA was purchased from Dharmacon Inc. (Lafayette, CO, USA, #M-011042-01-0005), and a 20 µM stock solution was prepared by dissolving in distilled water (Nacalai tesque) following the manufacturer's instructions and was preserved at −30 • C. A fresh working solution of 2 nM was used in the main experiment.
Muse cells were plated at a density of 20,000 cells/cm 2 , incubated overnight, and then Cx43 siRNA was introduced into Muse cells by using Lipofectamine RNAiMAX (Invitrogen), following the manufacturer's protocol. Cells were incubated with the transfection medium for 24 h, washed, and then cultured with the culture medium (10% FBS in αMEM with 1 ng/mL bFGF) for 4 days. Total RNA was extracted, and qPCR analysis was performed.

Statistical Analysis
All statistical analyses were performed using Microsoft Excel, Graphpad Prism 9.3.1, or R programming (Version 3.5.1). Student's t-test or Mann-Whitney test was used to assess the significance between two groups, and one-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test or Dunnett's multiple comparisons test was used in the case of three or more groups. The data in this paper were evaluated as mean ± SEM. Statistical significance was determined as follows: ns, not significant *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001.

Connexin Gene Expression in Muse and Non-Muse Cells
The SSEA-3-positive Muse cells comprised approximately 5% of the hMSCs, consistent with previous reports [6,15,18]. Muse cells and cells other than Muse cells in hMSCs, namely non-Muse cells that are negative for SSEA-3, showed a limited differentiation capacity as they only differentiated into osteocytes, adipocytes, and chondrocytes, and do not basically express pluripotency genes [15]. They were separated and were subjected to qPCR to analyze the expression of the connexin proteins that form their GJs. Human cardiomyocyte cell line (AC-16) cells were used as a positive control. Muse and Non-Muse cells were shown to express GJA1, GJA5, GJC1, GJC2, GJD3, and GJD4 genes that encode Cx43, Cx40, Cx45, Cx47, Cx31.9, and Cx40.1 proteins, respectively. When compared to AC-16, both Muse and non-Muse cells had a higher expression of GJA5 (p < 0.05 for Muse cells and p < 0.01 for non-Muse cells), GJA1 (p < 0.05 for Muse cells), and GJD3 (p < 0.05 for Muse cells and p < 0.001 for non-Muse cells) ( Figure 1A). On the other hand, GJC1, GJC2, and GJD4 levels were lower than or at a similar level to those in AC-16 and in Muse and in non-Muse cells ( Figure 1A). When compared the expression level of each connexin in Muse cells, GJA1 was the highest, followed by GJC1, GJD3, and GJA5, with a statistical significance for GJA1 (p < 0.0001) ( Figure 1B). Non-Muse cells showed a similar trend ( Figure 1C).
We then selected the top two highly expressed connexins, GJA1 (Cx43) and GJC1 (Cx45), for Western blot analyses. Cx43 protein level was higher in Muse (p < 0.01) and non-Muse cells (p < 0.05) when compared to that in AC-16, with statistical significance, however, it did not show statistical significance between Muse cells and non-Muse cells ( Figure 1D,E and Supplementary Figure S1). The protein level of Cx45 was, however, higher in AC-16 when compared to that in the Muse and non-Muse cells, (both p < 0.0001) with statistical significance ( Figure 1D,E and Supplementary Figure S1). The protein levels of Cx45 did not differ between the Muse and non-Muse cells ( Figure 1D,F and Supplementary Figure S1). < 0.05 for Muse cells and p < 0.001 for non-Muse cells) ( Figure 1A). On the other hand, GJC1, GJC2, and GJD4 levels were lower than or at a similar level to those in AC-16 and in Muse and in non-Muse cells ( Figure 1A). When compared the expression level of each connexin in Muse cells, GJA1 was the highest, followed by GJC1, GJD3, and GJA5, with a statistical significance for GJA1 (p < 0.0001) ( Figure 1B). Non-Muse cells showed a similar trend ( Figure 1C).  Negative control stained only with 2nd antibody was displayed. Area inside white squares is shown in the high magnification panels. Bars: 50 µm, except for the high magnification fields: 10 µm. One-way ANOVA, followed by Tukey's multiple comparisons test, was used for statistical analysis in (A,E,F), and one-way ANOVA, followed by Dunnett's multiple comparisons test, was used in (B,C). Data are shown as the mean ± SEM. In immunocytochemistry, Cx43 was located nearby the cell membrane, as dotted signals in the Muse and non-Muse cells ( Figure 1G). Cx45 was faintly detected in the Muse and non-Muse cells (Supplementary Figure S2).
In summary, Cx43 was the major connexin expressed in the Muse and non-Muse cells, both in qPCR and Western blot analyses, and thus, we focused on Cx43 in the following experiments.

Chemical Inhibition of GJIC in Muse Cells
GJIC was inhibited either by the treatment of cells with 18α-GA, which reversibly inhibits GJ by altering the connexon particle packing in GJ plaques [27]; TPA that inhibits GJIC via a mitogen-activated protein kinase-extracellular receptor kinase 1/2 (MAPK-ERK1/2)-dependent mechanism; or DDT that inhibits GJ through a phosphatidylcholinespecific phospholipase C (PC-PLC)-dependent mechanism [28,29].
In order to set usable concentrations for 18α-GA (ranging from 60 to 120 µM), TPA (10 to 400 nM), and DDT (70 to 100 µM), hMSCs were used for the evaluation. Consequently, 24 h incubation with 120 µM 18α-GA, 400 nM TPA, and 90 µM DDT were shown to be less damaging to hMSCs than other concentrations (Supplementary Figure S3). We also confirmed that these concentrations were applicable to Muse cells with no apparent morphological changes ( Figure 2A).  Figure 2B). Thus, these three GJ inhibitor treatments did not largely induce apoptosis in the Muse cells.
We next analyzed the effect of the three GJ inhibitors on the cell cycle. Untreated Muse cells in the proportion of 72.3 ± 1.87% were in G0/G1 phase, while 7.3 ± 1.3% were in S phase, and 20.4 ± 0.6% were in G2/M phase. Compared to the untreated Muse cells, the Muse cells treated with 18α-GA, TPA, or DDT did not show significant differences ( Figure 2C).

Effect of GJ Inhibition on Pluripotency Gene Expression
qPCR exhibited up-regulation of the master pluripotency genes, such as NANOG, OCT3/4, and SOX2 in Muse cells after treatment with the GJ inhibitors. Compared to the untreated-Muse cells, the elevation of these factors was the most prominent in Muse cells treated with 90 µM DDT, where NANOG, OCT3/4, and SOX2 expression levels were 1.85 ± 0.2 (p < 0.001), 4.56 ± 1.26 (p < 0.001) and 3.92 ± 0.59 (p < 0.001) times higher, respectively ( Figure 3C).
When Muse cells were treated with 400 nM TPA, NANOG, OCT3/4, and SOX2 did not show statistically meaningful changes ( Figure 3C). Treatment of the Muse cells with 120 µM 18α-GA exhibited a 1.66 (p < 0.01) and a 2.3 (p < 0.05) times higher elevation of NANOG and SOX2, when compared to the untreated Muse cells, respectively, while OCT3/4 did not show significant changes ( Figure 3C).
Since the elevation of the pluripotency gene expression was the most prominent in DDT among the three GJ inhibitors, treatment with 90 µM DDT for 24 h (DDT-Muse cells) was used in the following experiments.

Pluripotency Gene Expression after GJA1 Knockdown in Muse Cells
Cx43 siRNA was used to knock down the GJA1 gene expression and to observe the gene expression levels of NANOG, OCT3/4, and SOX2 in the Muse cells.
siRNA transfection did not induce remarkable morphological changes in the Muse cells (not shown). Effective GJA1 knockdown was confirmed by a Western blot analysis, which showed that the Cx43 protein levels were down-regulated to 10-20% for up to 5 days after transfection ( Figure 4A and Supplementary Figure S4).
siRNA transfection did not induce remarkable morphological changes in the Muse cells (not shown). Effective GJA1 knockdown was confirmed by a Western blot analysis, which showed that the Cx43 protein levels were down-regulated to 10-20% for up to 5 days after transfection ( Figure 4A and Supplementary Figure S4).

RNA-Sequencing
Three replicates of untreated-and DDT-Muse cells were subjected to RNA-sequencing ( Figure 5A). A differentially expressed genes (DEG) analysis revealed that 457 genes were up-regulated, and 751 genes were down-regulated in DDT-Muse cells, compared to the untreated-Muse cells (false discovery rate (FDR) < 0.05, fold change > 2) ( Figure 5B).
A GO pathway enrichment analysis was performed to identify the top enriched GO biological process terms in DDT treated-Muse cells. Top up-regulated terms included response to cytokine, cellular response to chemical stimulus, type I interferon signaling pathway, and response to organic substance. Down-regulated terms, however, included extracellular matrix organization, blood vessel development, circulatory system development, animal organ development, and response to growth factor ( Figure 5D).

Localization of NANOG, OCT3/4, and SOX2 before and after DDT Treatment in Muse Cells
The laser confocal microscopic observations of NANOG, OCT3/4, and SOX2 in untreated-Muse cells demonstrated that the signals for those three factors were detected both in the nucleus and cytoplasm, with a higher intensity in the nucleus than in the cytoplasm ( Figure 6A). In DDT-Muse cells, however, the absolute signal intensity in the cytoplasm was reduced, and instead, that which was found in the nucleus was elevated, as calculated using a cell fluorescence measurement system, ImageJ software. To evaluate the changes in the signal intensity in the nucleus and cytoplasm, the corrected total intensity ratio of nucleus/cytoplasm was measured and compared between the untreatedand DDT-Muse cells. The nucleus/cytoplasm ratio of NANOG increased 1.66 times in DDT-Muse cells (p < 0.001), compared with that in the untreated Muse cells ( Figure 6B). The nucleus/cytoplasm ratio of OCT3/4 and SOX2 slightly increased in DDT-Muse cells, compared with that of the untreated-Muse cells, but without statistical significance ( Figure 6B). labeled with NANOG, OCT3/4, or SOX2 (green). Nuclei were labeled with DAPI (blue). Bars: 100 µm; (B) The ratio of nucleus to cytoplasm signal intensity of NANOG, OCT3/4, and SOX2 in untreated and DDT-treated Muse cells (n = 20 each). Signal intensity was measured using ImageJ software. ***: p < 0.001, ns: no significant difference. Mann Whitney test was used for statistical analysis. Data are shown as the mean ± SEM.

Localization of Cx43 and YAP in Muse Cells
YAP is a transcription co-activator that acts through binding to TEA domain family member (TEAD) transcription factor [30]. Large tumor suppressor kinase (LATS) and the serine/threonine protein kinase (MST), two core members of the Hippo pathway, phosphorylate YAP and inhibit its transcriptional function by anchoring it to the cytoplasm [30]. It was also reported that YAP can be anchored to the cell membrane through binding to adherens junctions and tight junctions [30]. However, recently, several studies in mouse and rat astrocytes have shown that YAP can also be anchored to the GJ as well [31,32].
A laser confocal microscopic analysis of Cx43/YAP double staining showed that both signals were adjacent to each other, nearby the cell membrane, in untreated-Muse cells ( Figure 7A). However, after DDT treatment, the Cx43 signal was faintly detected in the cytoplasm and the YAP signal became more intense in the nucleus, compared to that in the untreated cells ( Figure 7A). A corrected signal intensity measurement demonstrated that the nuclear YAP intensity was higher in DDT-Muse cells rather than in untreated-Muse cells, with statistical significance (p < 0.01) ( Figure 7B).
A Western blot analysis exhibited the elevated YAP signal in the nuclei of DDT-Muse cells, compared to that in untreated-Muse cells, as shown by the quantification of protein bands using ImageJ software (p < 0.05). Histone H3, the nuclear loading control, was detected mainly in the nucleus, and α-tubulin, the cytoplasmic loading control, was detected in the cytoplasm ( Figure 7C, Supplementary Figures S5 and S6).

Discussion
Muse cells are endogenous pluripotent stem cells and are non-tumorigenic, but their expression levels of pluripotency master genes, such as Nanog, Oct3/4, and Sox2 in the adherent culture system, are lower than those in the suspension culture. When Muse cells are transferred to a suspension culture, the size of each cell becomes smaller and they form a sphere-shaped cluster. In this condition, the expression levels of Nanog, Oct3/4, and Sox2 are increased and become higher than those in an adherent culture [33]. ES cells and iPSCs were also shown to up-regulate the pluripotency genes in a suspension culture [7,13,24,25]. Thus, suspension culture is one of the simple approaches used to upregulate the pluripotency gene expression in these stem cells. On the other hand, methods to up-regulate pluripotency genes in adherent culture system have not been reported in Muse cells, except for the induction of iPSCs by introducing Yamanaka-four factors [15]. In this study, we propose that Cx43, a component of GJ, and YAP, a mechanosensory factor, play an important role in pluripotency regulation and that the disruption of Cx43-mediated GJIC up-regulates the master pluripotency genes, NANOG, OCT3/4, and SOX2, in an adherent culture system in Muse cells. The possible mechanism is described in Figure 8.
GJ channels mediate intercellular communication in multicellular organisms and play an important role in regulating cell growth and differentiation [34,35]. This important role can also be observed in cancer cells as the process of carcinogenesis is usually accompanied by the down-regulation of connexin expression, suggesting that the lack of communication between cells disrupts their homogeneity, allowing the appearance of abnormal cells with higher pluripotency and proliferation, namely cancer stem cells [34][35][36][37].
In this study, we found that GJA1 was the highest connexin subtype expressed in pluripotent-like Muse cells. GJA5, GJC1, GJC2, I, and GJD4 expression was also detected byqPCR, although to a far lesser extent than GJA1. Western blot and immunocytochemistry analyses supported these data. Therefore, we focused on GJA1 that encodes the Cx43 protein and analyzed its role in pluripotency regulation in Muse cells. Three GJ inhibitors, 18α-GA, TPA, and DDT, were chosen to block Cx43-mediated GJIC in Muse cells and the efficiency of GJIC inhibition after treatment with each inhibitor was confirmed by dye-transfer assay. regulate pluripotency genes in adherent culture system have not been reported in Muse cells, except for the induction of iPSCs by introducing Yamanaka-four factors [15]. In this study, we propose that Cx43, a component of GJ, and YAP, a mechanosensory factor, play an important role in pluripotency regulation and that the disruption of Cx43-mediated GJIC up-regulates the master pluripotency genes, NANOG, OCT3/4, and SOX2, in an adherent culture system in Muse cells. The possible mechanism is described in Figure 8.  [30]. YAP is also anchored to the cell membrane through binding to the gap junction, adherens junction, and tight junction [30]. Upon GJ disruption, YAP is translocated to the nucleus leading to the up-regulation of pluripotency factors. qPCR analysis of 18α-GA-, TPA-and DDT-treated-Muse cells showed that the expression levels of NANOG, OCT3/4, and SOX2 were increased-or showed a tendency of up-regulation-in all three kinds of inhibitors, when compared to the untreated-Muse cells. Although all of the three inhibitors effectively inhibited GJIC in Muse cells, their influence on pluripotency up-regulation was different from each other; DDT had the highest and most significant up-regulation of NANOG, OCT3/4, and SOX2; 18α-GA had the significant up-regulation of NANOG, and SOX2; TPA had a tendency of up-regulation but with no statistical significance in NANOG, OCT3/4, and SOX2. One of the reasons for these different responses might be due to the various mechanisms of GJ inhibition these inhibitors have. DDT causes actual GJ disruption by internalizing Cx43 into the cytoplasm and its degradation [28]. TPA, on the other hand, mainly inhibits GJIC through Cx43 hyperphosphorylation [38], while 18a-GA affects the membrane fluidity, disrupting the individual GJ units to assemble and form a GJ plaque [27]. The action of DDT for GJ inhibition might be more direct and efficient than TPA and 18a-GA.
GJA1 knockdown by siRNA transfection also showed the up-regulation of pluripotency genes in Muse cells, supporting the results of GJ inhibitor experiments.
RNA-seq data showed a difference in gene expression pattern between naïve-and DDT-treated Muse cells, with 457 up-regulated and 751 down-regulated genes in DDTtreated Muse cells. A pathway analysis showed an overall down-regulation of the cell differentiation-related pathways (e.g., blood vessel development, organ and tissue development). These data are consistent with previous reports suggesting that Cx43 is positively correlated with differentiation and negatively correlated with pluripotency gene expression in stem cells since GJIC plays a role in the differentiation of pluripotent stem cells and Cx43 levels are elevated during differentiation into many different cell types [39][40][41][42].
Hippo pathway plays an evolutionarily conserved role in organ size control by inhibiting cell proliferation, promoting apoptosis, and regulating the fates of stem/progenitor cells [43]. YAP, which is related to Hippo pathway activity, is a protein that acts as a co-transcriptional regulator by activating the transcription process of genes involved in cell proliferation and apoptotic gene suppression [44]. YAP is inhibited through Hippo pathway signaling and only exhibits its transcriptional effects after translocating into the nucleus upon the inhibition of Hippo pathway signaling [45].
YAP has another important aspect. It is a primary sensor of the cell's physical nature, e.g., cell structure, shape, and polarity [43]. YAP activation reflects a cell's "social" behavior through cell adhesion and the mechanical signals that are perceived from the tissue architecture and the surrounding extracellular matrix [43]. Recent studies reported that YAP is physically associated with Cx43, and upon Cx43 disruption, YAP would translocate to the nucleus of mouse and rat astrocytes [31,32]. Similar to those observations, untreated Muse cells exhibited that YAP, co-localized with Cx43 in the cell membrane, dissociated from the cell membrane after DDT-treatment, as shown by immunocytochemistry imaging, and translocated to the nucleus, because the nucleic YAP signal intensity was significantly higher in the immunocytochemistry and Western blot analyses after DDT-treatment. Thus, it is possible that, when GJ connection is disrupted, YAP is translocated into the nucleus, leading to the activation of its co-transcriptional activity. Concomitantly, as mentioned above, an overall, down-regulation of cell differentiation-related pathways was observed.
It is reported that YAP and TEAD2 were highly expressed in self-renewing mouse ES cells and that YAP expression is up-regulated during iPSC reprogramming [44,45]. Indeed, TEAD2 associates directly with the promoter of OCT3/4, one of the master pluripotency genes [46]. Reversely, YAP gene silencing was reported to induce the loss of pluripotency in ES cells [47]. These reports indicate the important role of YAP in pluripotency regulation [46,48].
In this study, we focused on the role of Cx43 in Muse cells treated with these three different GJIC inhibitors that work by very different biochemical mechanisms to inhibit functional GJIC at the posttranscriptional level, as well as in GJA1 siRNA-transfected Muse cells. Other studies, using iPS cells (Cx43 knockout (GJA1 −/− )) iPSCs, generated using CRISPR-Cas9 gene ablation), maintained characteristics typical of iPSCs and successfully differentiated into cells of ectoderm, mesoderm, or endoderm lineages [49]. While these studies were well executed, the interpretation of these results is complicated and might not apply to normal adult organ-specific adult stem cells or normal Muse cells. These iPS cells' origin has been questioned, in that they might be derived from normal fibroblast stem cells rather than having been "reprogramed" [50,51].
These iPS cells have both the endogenous OCT4 gene, as well as the extra copy of the exogenous OCT4 gene. By definition, these iPS cells, when injected into an adult animal, form teratomas or the three primary germ lines. Whereas the organ-specific adult stem cells do not form teratomas. The normal human adult stem cells, i.e., human breast stem cells, do not express GJA1, nor do they have functional GJIC [52]. When these normal human breast stem cells are induced to differentiate into human breast epithelial cells, OCT4 is transcriptionally repressed, GJA1 is expressed, and functional GJIC is evident. If, however, the normal adult breast stem cells are transfected with the Large T antigen gene of SV40 virus, the cells do not express GJA1, maintain expression of OCT4, and do not differentiate. They stay in the stem cell state or remain "immortal". Clearly, further studies must be done to clarify the exact mechanism of how YAP interacts with the promoters of pluripotency related-genes in these cells, as well as the differences between these Muse cells and iPS cells in the future.
The expression of GJA1 is different between naïve and primed pluripotent states. Human primed pluripotent stem cells express higher GJA1 than they do in the naïve state [23], and naïve pluripotent stem cells are less affected by the pharmacological ablation of GJIC than primed pluripotent stem cells are [23]. Muse cells were suggested to be more similar to primed pluripotent stem cells than to naïve pluripotent stem cells for their proliferation and self-renewability and are ultimately dependent on the basic fibroblast growth factor (bFGF), rather than the leukemia inhibitory factor (LIF) [5]. The result of this study, namely the expression of GJA1 and the responsiveness of pluripotency gene expression by GJ inhibition, supports the assumption that Muse cells are similar to primed pluripotent stem cells.
While we demonstrated the involvement of a Cx43/YAP system in the up-regulation of pluripotency genes, NANOG, OCT3/4, and SOX2, when GJIC was inhibited, other systems might also be involved in this phenomenon. Furthermore, multiple steps/systems might intervene in the inhibition of GJ and pluripotency gene up-regulation. These interesting subjects need to be clarified in future studies.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cells11172701/s1, Figure S1: Western blot analysis of Cx43 and Cx45 in human cardiomyocyte, Muse, and non-Muse cells derived from hMSC.; Figure S2. Cx45 immunostaining in human cardiomyocyte, Muse, and non-Muse cells derived from hMSC; Figure S3. hMSC morphology after treatment with 18α-GA, TPA, and DDT for 24 h; Figure S4. Western blot analysis of Cx43 in GJA1-knockdown Muse cells; Figure S5. Western blot analysis of nuclear YAP in untreated-and DDT-Muse cells; Figure S6. Western blot analysis of cytoplasmic YAP in untreatedand DDT-Muse cells.