EDMD-Causing Emerin Mutant Myogenic Progenitors Exhibit Impaired Differentiation Using Similar Mechanisms

Mutations in the gene encoding emerin (EMD) cause Emery–Dreifuss muscular dystrophy (EDMD1), an inherited disorder characterized by progressive skeletal muscle wasting, irregular heart rhythms and contractures of major tendons. The skeletal muscle defects seen in EDMD are caused by failure of muscle stem cells to differentiate and regenerate the damaged muscle. However, the underlying mechanisms remain poorly understood. Most EDMD1 patients harbor nonsense mutations and have no detectable emerin protein. There are three EDMD-causing emerin mutants (S54F, Q133H, and Δ95–99) that localize correctly to the nuclear envelope and are expressed at wildtype levels. We hypothesized these emerin mutants would share in the disruption of key molecular pathways involved in myogenic differentiation. We generated myogenic progenitors expressing wildtype emerin and each EDMD1-causing emerin mutation (S54F, Q133H, Δ95–99) in an emerin-null (EMD−/y) background. S54F, Q133H, and Δ95–99 failed to rescue EMD−/y myogenic differentiation, while wildtype emerin efficiently rescued differentiation. RNA sequencing was done to identify pathways and networks important for emerin regulation of myogenic differentiation. This analysis significantly reduced the number of pathways implicated in EDMD1 muscle pathogenesis.


Lentiviral Transduction
H2K myogenic progenitors expressing wildtype emerin (+EMD) and EDMD causing emerin mutations (S54F, Q133H, and ∆95- 99), an emerin mutation that does not cause the disease (M179), and a vector only control were generated using the following protocol. EMD −/y mouse myogenic progenitors (EMD −/y ) were seeded at a density of 1000 cells/well in 96-well plates coated with 0.01% gelatin. Cells were incubated at 33 • C and 10% CO 2 overnight in proliferation media and replaced with infection medium containing lentiviral particles (Genecopoeia, Rockville, MD, USA, #LPP-CS-G0746-Lv105,) at a multiplicity of infection of 350 and 8 µg/mL polybrene (Cyagen Biosciences, Santa Clara, CA, USA). Polybrene is a cationic polymer known to increase lentiviral transduction efficiency [39] by neutralizing the surface charge between the cell surface and the viral particles [40,41]. The infection medium was replaced with fresh growth media after 16-24 h. Cells were allowed to grow for 72 h post-transduction, then transferred to 12-well dishes containing growth media and puromycin (MilliporeSigma, Burlington, MA, USA, #P8833). EMD −/y cells transduced with control vector, S54F and ∆95-99 were selected using 15 µg/mL puromycin. EMD −/y cells transduced with Q133H and M179 vectors were selected using 10 µg/mL puromycin. EMD −/y cells transduced with wildtype emerin (+EMD) was selected using 6 µg/mL. Once the cells were 40-60% confluent, cells were subsequently moved to six-well dishes, followed by transfer to 10 cm dishes, and finally expanded to multiple 15 cm dishes. Cells were kept sparse to ensure cells don't undergo spontaneous differentiation or senescence. Media containing puromycin was replaced every 2-3 days.

Myogenic Differentiation Immunofluorescence Assay
Wildtype and EMD -/y cells were seeded at a density of 7800 cells per well into 96-well plates or 92,800 cells per well into 12-well plates coated with gelatin. For experiments in 12-well plates, +EMD, Q133H and ∆95-99 cells were plated at a density of 118,800 cells per well coated with gelatin. For 96-well plate experiments, +EMD, vector control, S54F and M179 were plated at a density of 10,000 cells per well coated with gelatin. Cells were incubated overnight in proliferation media. Differentiation was induced by replacing proliferation media with differentiation media (DMEM with high glucose, 5% horse serum, 1% penicillin/streptomycin) and incubating cells at 37 • C and 5% CO 2 .
EdU (5-ethynyl-2 -deoxyuridine, ThermoFisher Scientific, Waltham, MA, USA) was used as a proliferation marker. EdU is a thymidine analog which gets incorporated into newly synthesized DNA during S phase of the cell cycle. Ten micromolar EdU was incubated with cells for 2 h prior to fixing. Cells were fixed at 0, 24, 48, and 72 h post differentiation induction. After fixing, cells were washed three times and were processed according to the manufacturer's instructions (Click-IT EdU imaging Kit 647, ThermoFisher Scientific, Waltham, MA, USA). Cells were treated with 0.5% Triton-X100 in phosphate-buffered saline (PBS) for 20 min, washed twice with 3% bovine serum albumin in PBS, and treated with a Click-IT EdU reaction cocktail for 30 min. Cells were washed with PBS and then blocked for an hour with 3% BSA in PBS containing 0.1% Triton-X100.
Cells were then treated with emerin antibodies (Santa Cruz Biotechnology, Dallas, TX, USA, #FL-254,) for an hour, washed three times with PBS, and incubated with monoclonal antibodies against MyHC (Santa Cruz Biotechnology, Dallas, TX, USA, #MYH B-5,) for an hour. Cells were washed three times with PBS, incubated with secondary antibodies Alexa Fluor 488 goat anti-rabbit (1:200, ThermoFisher Scientific, Waltham, MA, USA) for one hour followed by Alexa Fluor 594 goat anti-mouse antibodies (1:200, ThermoFisher Scientific, Waltham, MA, USA) for one hour. Cells were washed twice with PBS and incubated with 0.2 µg/mL DAPI to label nuclei. Cells were washed once with PBS and imaged using the EVOS-FL Auto microscope (ThermoFisher Scientific, Waltham, MA, USA).
Multiple images were acquired at random (at least three) for each well using a 40× objective. A total of three wells were analyzed for each treatment for a given biological replicate. At least two biological replicates were done for each treatment. Each field was imaged in four different channels: green fluorescent protein channel for emerin positive nuclei, blue fluorescent channel for DAPI stained nuclei, Cy5 channel for Edu positive nuclei, and Texas Red channel for MyHC positive nuclei. Image analysis was done using the cell counter plugin in ImageJ. Cell cycle withdrawal was determined by dividing the total number of cells expressing EdU by the total number of nuclei in an image. Differentiation index (DI) was defined as the percentage of nuclei expressing the terminal differentiation MyHC. Myotube formation was determined by measuring the percentage of myotubes containing at least three nuclei and expressing MyHC.

Western Blots
Proliferating H2K myogenic progenitors in 15 cm dishes were harvested using SDS-PAGE sample buffer. A total of 50,000 cell equivalents for all cell lines except for +EMD and M179 expressing cells (100,000 cell equivalents) were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. Blocking of the membrane was done using 5% dried nonfat milk in PBS containing 0.1% Tween 20 (PBST) at room temperature for an hour. The membrane was then incubated with emerin antibodies (rabbit polyclonal, 1:2500, Proteintech, Rosemont, IL, USA) for an hour followed by three washes in PBST for five minutes. The membrane was then incubated for an hour at room temperature with goat anti-rabbit HRP antibodies (1:10,000, ThermoFisher Scientific), followed by three PBST washes for five minutes each. The blot was incubated with Amersham ECL western blotting detection reagents (GE healthcare, Chicago, IL, USA, #RPN2209) and imaged on the Bio-Rad Chemidoc equipment (Bio-Rad Laboratories, Hercules, CA, USA). The HRP enzyme was then inactivated with 0.1% sodium azide in PBST for 15 min followed by five washes (five minutes each) with PBS. The blot was incubated with mouse gamma-tubulin antibodies (1:10,000, ThermoFisher Scientific, Waltham, MA, USA) for an hour at room temperature followed by incubation with donkey anti-mouse HRP antibodies (1:10,000, ThermoFisher Scientific, Waltham, MA, USA) for an hour at room temperature. The blots were then incubated with Amersham ECL western blotting detection reagents (GE healthcare, Chicago, IL, USA, #RPN2209) and processed according to the manufacturer's instructions. ImageLab software (Bio-Rad Laboratories) was used to perform volume analysis.

RNAseq
For isolation of total RNA during proliferation, each H2K myogenic progenitor cell line was seeded at a density of 158,000 cells per 15 cm dish coated with 0.01% gelatin and incubated in proliferative media for two days. Two 15 cm dishes were combined and represent one technical replicate. Three technical replicates per sample were performed for each biological replicate (n = 2). Cells were trypsizined, washed three times with PBS, and total RNA was isolated using the mRNeasy Plus Kit (Qiagen, Germantown, MD, USA, #74136) per manufacturer's instructions.
Wildtype and EMD −/y H2K myogenic progenitors were seeded in six-well dishes coated with 0.01% gelatin at a density of 232,000 cells/well for isolation of total RNA during differentiation. 297,000 cells/well of vector control and emerin mutant cell lines were plated in six-well dishes coated with 0.01% gelatin. Cells were incubated overnight in proliferative media. In the morning, cell lines were switched to differentiation media. RNA was isolated after 0 h (to account for changes due to cell density), 24 h, 48 h, and 72 h of differentiation. RNA was isolated using the mRNeasy Plus Kit (Qiagen, Germantown, MD, USA, #74136) per manufacturer's instructions. Three technical replicates per sample were performed for each biological replicate (n = 2).
Purity of RNA was assessed using Nanodrop (ND-2000), and quantification was done using the Quant-IT RNA assay kit (ThermoFisher Scientific, Waltham, MA, USA, #Q10213). BGI Genomics performed library preparation and sequencing using BGISEQ-500 platform. BGI Genomics filtered reads using SOAPnuke and genome mapping was done to the mouse genome (mm10) using hierarchical indexing for spliced alignment of Transcripts-2 (HISAT2). Differentially expressed genes was detected using DEseq2. Transcripts were considered to be significantly differentially expressed if the q-value <0.05. The q-value is a p-value corrected for the False Discovery Rate (FDR) and represents the standard in the field for determining significance. All differentially expressed transcripts were also ≥2-fold increased or decreased. KEGG pathway enrichment analysis was done using phyper, a function of R.

Data Sharing Statement
RNA sequencing data is available through the NCBI Gene Expression Omnibus (Accession number GSE152226).
In total, 44.6% of EMD −/y cells ( Figure 3B; p < 0.02) and 40.2% of vector-alone cells (p < 0.02) expressed MyHC after 48 h of differentiation. In total, 61.3% of wildtype cells expressed MyHC. In total, 55.7% of +EMD cells expressed MyHC, demonstrating successful rescue of myoblast commitment. Expression of M179 and S54F in EMD −/y progenitors failed to rescue MyHC expression and the differentiation index, with 32.4% and 37.9% expressing MyHC, respectively ( Figure 3B; p < 0.02).  We next analyzed myotube formation, which was defined as the percentage of myotubes containing at least three nuclei and expressing MyHC. Myotube formation was impaired in EMD −/y and control cells, as only 19.2% of EMD −/y cells and 19.4% of control cells ( Figure 3A,C; p < 0.02) formed mature myotubes compared to 35.4% of wildtype cells. Wildtype emerin expression rescued myotube formation, as 35.8% of +EMD cells formed myotubes. Both M179 and S54F cells were significantly impaired in myotube formation, as only 19.5% of M179 cells and 21.5% of S54F cells fused to form mature myotubes ( Figure 3A,C; p < 0.02).

Identification of Pathways Shared Between all EDMD-Causing Emerin Mutants
RNA sequencing (RNAseq) was done on wildtype, EMD −/y , and each emerin mutant myogenic progenitor line throughout differentiation using Beijing Genomic Institute (BGI) Americas. Differentially expressed genes were identified by BGI by comparing all datasets to each other. The first comparison compared the changes in transcript expression at each day of differentiation for each EDMD-causing emerin mutant or EMD −/y myogenic progenitors to wildtype progenitors. Transcripts were considered to be significantly differentially expressed if the q-value <0.05. It was expected that the expression of a large number of differentially expressed transcripts would be seen during differentiation of each EDMD-causing mutant based on previous studies [34].
In total, 956 differentially expressed genes (456 down, 508 up) were identified in EMD −/y progenitors plated at high-density for differentiation (day 0, Figure 6A). S54F, ∆95-99, and Q133H myogenic progenitors differentially expressed 1468 genes (764 down, 702 up), 1536 genes (842 down, 694 up), and 1859 genes (1076 down, 783 up), respectively, when plated at high density (day 0, Figure 6A). There were 1338 genes (654 down, 784 up) differentially expressed in M179 progenitors upon plating at high density ( Figure 6A). In total, 874 differentially expressed genes (344 down, 530 up) were identified in EMD −/y progenitors one day after differentiation induction, which is when progenitors normally withdraw from the cell cycle to commit to differentiation (day 1, Figure 6A). S54F, ∆95-99, and Q133H myogenic progenitors differentially expressed 1570 genes (826 down, 744 up), 2859 genes (1666 down, 1193 up), and 3290 genes (2038 down, 1252 up), respectively (day 1, Figure 6A). There were 2188 (1011 down, 1177 up) differentially expressed genes in M179 progenitors at day 1. In total, 1286 differentially expressed genes (336 down, 950 up) were identified in EMD −/y progenitors two days after differentiation induction, which is when myoblasts normally commit to form myotubes (day 2, Figure 6A We predicted that comparing the differentially expressed genes in all the EDMD-causing emerin mutants and EMD −/y myogenic progenitors to wildtype progenitors would identify genes and pathways implicated in the impaired differentiation seen in EDMD. These analyses were focused first on differentiation day 1, as this timepoint is when wildtype progenitors exit the cell cycle to commit to becoming myotubes. In total, 433 genes were differentially expressed in each EDMD-causing emerin mutant cell lines and EMD −/y cells, compared to wildtype cells ( Figure 6B). KEGG pathway analysis was done to identify pathways shared amongst all differentiating EDMD-causing emerin mutants, but lacking in +EMD progenitors. Shared pathways were predicted to be important for the EDMD disease mechanism. Pathways enriched for these genes included Hippo signaling, ECM-receptor interactions, human papillomavirus infection, type I diabetes mellitus, proteoglycans in cancer, and amoebiasis ( Figure 6C).
A more powerful analysis, which we have used previously [34], was to compare transcripts differentially expressed between wildtype progenitors and either EMD −/y or each EDMD-causing emerin mutant progenitors at important differentiation transitions. Differentially expressed genes between each day and the preceding day were determined for either differentiating +EMD, EMD −/y , or each EDMD-causing emerin mutant progenitor line. This analysis was anticipated to identify genes involved in the inability of EMD −/y and EDMD-mutant progenitors to properly differentiate. In total, 2196 transcripts were differentially expressed upon withdrawal of +EMD progenitors from the cell cycle (day 0 to day 1; Figure 7A). The transition to committed myocytes (day 1 to day 2) caused 2731 genes to be differentially expressed in +EMD cells. In total, 1293 transcripts were differentially expressed upon plating +EMD progenitors at high-density for differentiation (prolif. to day 0; Figure S6), and 1804 genes were differentially expressed during myotube formation and maturation (day 2 to day 3; Figure S6). All differentially expressed transcripts had q-value <0.05.
Differentially expressed genes during each transition were determined for EMD −/y myogenic progenitors and each EDMD-causing emerin mutant line. 2753 transcripts were differentially expressed upon the transition of EMD −/y myogenic progenitors from differentiation day 0 to day 1, which is when cell cycle withdrawal should occur. The transition to myocyte commitment caused 2440 transcripts to be differentially expressed in EMD −/y cells ( Figure 7A; day 1 to day 2). In total, 985 transcripts were differentially expressed upon plating EMD −/y myogenic progenitors at high-density ( Figure S6; prolif. to day 0). 694 genes were differentially expressed during myotube formation and maturation (day 2 to day 3, Figure S6). In total, 2204 transcripts were differentially expressed upon the transition of S54F myogenic progenitors from differentiation day 0 to day 1. The transition to myocyte commitment caused 2738 transcripts to be differentially expressed in S54F cells ( Figure 7A; day 1 to day 2). In total, 1046 transcripts were differentially expressed upon plating S54F myogenic progenitors at high-density ( Figure S6; prolif. to day 0). 1057 genes were differentially expressed during myotube formation and maturation (day 2 to day 3, Figure S6). In total, 1866 transcripts were differentially expressed upon the transition of ∆95-99 myogenic progenitors from differentiation day 0 to day 1. The transition to myocyte commitment caused 1897 transcripts to be differentially expressed in ∆95-99 cells ( Figure 7A; day 1 to day 2). In total, 1203 transcripts were differentially expressed upon plating ∆95-99 myogenic progenitors at high-density ( Figure S6; prolif. to day 0). In total, 1324 genes were differentially expressed during myotube formation and maturation (day 2 to day 3, Figure S6). In total, 1405 transcripts were differentially expressed upon the transition of Q133H myogenic progenitors from differentiation day 0 to day 1. The transition to myocyte commitment caused 1261 transcripts to be differentially expressed in Q133H cells ( Figure 7A; day 1 to day 2). In total, 930 transcripts were differentially expressed upon plating Q133H myogenic progenitors at high-density ( Figure S6; prolif. to day 0). 1179 genes were differentially expressed during myotube formation and maturation (day 2 to day 3, Figure S6). All differentially expressed transcripts had q-value <0.05.
Transcripts differentially expressed at each daily transition in differentiating EMD −/y progenitors or each EDMD-causing mutant progenitor line were then compared to wildtype progenitors at each transition. This identified transcripts differentially expressed only in differentiating wildtype or EMD −/y progenitors, or wildtype progenitors and each EDMD emerin mutant progenitor line, or any comparison between all the samples at each transition point ( Figure 7B). For example, transcripts that are differentially expressed during wildtype differentiation between day 0 and day 1 of differentiation were compared to the differentially expressed transcripts seen during the transition from differentiation day 0 to day 1 in EMD −/y cells, S54F cells, ∆95-99 cells, and Q133H cells. This comparison yields genes uniquely differentially expressed in wildtype cells, EMD −/y cells, S54F cells, ∆95-99 cells, and Q133H cells during the transition from day 0 to day 1 of differentiation ( Figure 7B). This approach also allows for comparisons between all possible combinations to identify differentially expressed genes shared between certain EDMD-causing emerin mutants or uniquely absent from certain EDMD mutant cell lines.  A large number of transcripts were differentially expressed between day 0 and day 2 of differentiation in wildtype, EMD −/y , and each EDMD-causing emerin mutant progenitor line ( Figure 7B,C), as expected [34]; this is when the differentiation program is initiated and commitment occurs. Of the 2753 transcripts differentially expressed during EMD −/y differentiation days 0 and 1, 2065 transcripts were also altered during wildtype differentiation, and 688 transcripts were uniquely altered in EMD EMD −/y cells ( Figure 7B; q-value <0.05). In total, 2440 transcripts showed altered expression in EMD −/y cells during the transition from day 1 to day 2 of myogenic differentiation ( Figure 7C; q-value < 0.05), with 1836 transcripts also altered during wildtype differentiation and 604 transcripts uniquely altered in EMD −/y cells. Of the 2204 transcripts differentially expressed during S54F differentiation days 0 and 1, 1542 transcripts were also altered during wildtype differentiation and 662 transcripts were uniquely altered in S54F cells ( Figure 7B; q-value <0.05). In total, 2738 transcripts showed altered expression in S54F cells during the transition from day 1 to day 2 of myogenic differentiation ( Figure 7C; q-value <0.05), with 1681 transcripts also altered during wildtype differentiation and 1057 transcripts uniquely altered in S54F cells. Of the 1866 transcripts differentially expressed during ∆95-99 differentiation days 0 and 1, 1152 transcripts were also altered during wildtype differentiation and 714 transcripts were uniquely altered in ∆95-99 cells ( Figure 7B; q-value < 0.05).
In total, 1897 transcripts showed altered expression in ∆95-99 cells during the transition from day 1 to day 2 of myogenic differentiation ( Figure 7C; q-value <0.05), with 1054 transcripts also altered during wildtype differentiation and 843 transcripts uniquely altered in ∆95-99 cells. Of the 1405 transcripts differentially expressed during Q133H differentiation days 0 and 1, 805 transcripts were also altered during wildtype differentiation and 600 transcripts were uniquely altered in Q133H cells ( Figure 7B; q-value <0.05). In total, 1261 transcripts showed altered expression in Q133H cells during the transition from day 1 to day 2 of myogenic differentiation ( Figure 7C; q-value <0.05), with 674 transcripts also altered during wildtype differentiation and 587 transcripts uniquely altered in Q133H cells. To identify genes implicated in the impaired differentiation of EDMD, transcripts differentially expressed between wildtype progenitors and each EDMD-causing emerin mutant progenitor line and EMD −/y cells were compared during these transitions. This identified 64 transcripts and 40 transcripts differentially expressed in each EDMD mutant and EMD −/y cells during transitions day 0 to day 1 and day 1 to day 2, respectively ( Figure 7B,C).
KEGG pathway analysis identified KEGG pathways enriched in the EDMD-causing emerin mutants, but absent from +EMD cells at each differentiation transition. These were compared to identify those pathways that were shared only amongst the differentiating EDMD-causing mutant myogenic progenitors ( Figure 7D). The shared pathways are predicted to be important for the EDMD disease mechanism. Pathways enriched among all EDMD mutants during differentiation transition day 0 to day 1 were ribosome, apelin signaling pathway, HIF-1 signaling, and metabolic pathways ( Figure 7D). The Rap1 signaling pathway and tight junction pathway were both enriched in two of the three EDMD emerin mutant lines, with Rap1 enriched in ∆95-99 and S54F, and tight junction enriched in S54F and Q133H. Pathways enriched amongst all EDMD mutants during differentiation transition day 1 to day 2 were the p53 signaling pathway and homologous recombination. The PI3K-Akt signaling pathway was enriched in ∆95-99 and S54F cell lines.

Discussion
In this study, we utilized lentiviral delivery to express EDMD-causing emerin mutants (S54F, ∆95-99, and Q133H) in an EMD −/y myogenic progenitor background. We next differentiated wildtype, EMD −/y , and +EMD progenitors to validate our experimental approach. The differentiation index and myotube formation were successfully rescued to wildtype levels in +EMD progenitors, while +EMD progenitors failed to rescue cell cycle withdrawal after 24 h of differentiation. This indicates emerin may not be important for regulating cell cycle exit, but rather regulate later differentiation transitions. This is similar to the effects we previously saw upon treatment with pharmacological activators of HDAC3 activity [30], histone acetylase inhibitors [38], and inhibitors of ERK activity [30]. Collectively, the results presented here, combined with our earlier studies [30,[34][35][36]38], strongly support emerin regulating transcriptional reprograming events early in differentiation to impact myoblast commitment and myotube formation through its functional interaction with HDAC3.
The genome undergoes dynamic reorganization during stem cell differentiation and development to regulate the coordinated temporal expression of differentiation genes [42][43][44]. It has been reported that active genes localize towards the nuclear interior whereas repressed genes preferentially localize at the nuclear periphery [43,45,46]. Activation, commitment, and differentiation of myogenic progenitors are controlled by various epigenetic mechanisms, which include histone modifications, RNA-associated silencing, and DNA methylation [42,44,47]. The importance of histone acetylation and deacetylation in the regulation of myogenesis has been established [48][49][50][51]. Although global histone acetylation decreases progressively during myogenesis, a subset of genes, such as MyoD target genes, show increased histone acetylation upon differentiation [48,52,53] concomitant with their expression. To date, class 1 and class 2 HDACs have been repeatedly reported to function in different phases of myogenesis. The results presented here, combined with our previous work [15,30,34,36,38], support an important role for emerin in activating HDAC3 activity to decrease histone acetylation during myogenic differentiation. We propose this interaction mediates the establishment, recruitment, and maintenance of repressed chromatin at the nuclear lamina [30], and that this reorganization of the genome at the nuclear lamina during differentiation induction is required for transcriptional reprogramming.
HDAC3 is the only reported binding partner that is altered in all four "special" EDMD-causing emerin mutations (Q133H, P183H, S54F, and ∆95-99) [15]. This highlights the direct binding and functional interaction between HDAC3 and emerin may be important for the EDMD disease mechanism. Our studies revealed EDMD-causing emerin mutants ∆95-99 and Q133H exhibit impaired differentiation. After 24 h of differentiation induction, ∆95-99 and Q133H cells fail to commit to differentiation; they fail to express MyHC, a terminal marker for differentiation; and fail to form mature myotubes. Notably, both of these mutants fail to bind HDAC3 [15]. EDMD-causing emerin mutant S54F also failed to rescue cell cycle withdrawal, the differentiation index, and myotube formation. Surprisingly M179, an emerin mutant not linked to EDMD, also failed to rescue cell cycle withdrawal, the differentiation index and myotube formation. The M179 mutant contains two alanine-substitutions at position 179 and 180 for leucine and serine. Although these mutants also disrupt binding to other partners [13,[15][16][17][18][19][20]37,54], only failure to bind HDAC3 is shared between all of the mutants tested here. We propose perturbation of this interaction in the EDMD-causing emerin mutants results in failure in genomic reorganization and failure to transcribe genes required for myoblast commitment and myotube formation. Supporting this hypothesis, inhibition of HDAC3 catalytic activity by RGFP966 blocked MyHC expression and myotube formation in both differentiating wildtype and EMD −/y myogenic progenitors [30]. It will be important to determine how emerin regulates genomic organization via HDAC3 regulation to affect transcriptional reprogramming. Understanding how specific molecular pathways are impacted by disruption of the functional interaction of emerin with HDAC3 will also be important for elucidating the mechanism(s) underlying the impaired differentiation seen in EDMD-causing mutants.
Emerin mutants M179, S54F, and Q133H all showed an increased number of cells cycling 24 h after differentiation induction compared to EMD −/y progenitors. ∆95-99 myogenic progenitors were similar to EMD −/y cells in their cell cycle exit. We predict this failure of M179, S54F, and Q133H progenitors to exit the cell cycle is caused by binding to germ cell-less (GCL). GCL is a transcriptional repressor that binds the DP3 subunit of the E2F-DP3 heterodimer to inactivate E2F-DP3-mediated transcription [55] and inhibit entry into S-phase. GCL binds directly to emerin via Regulator Binding Domains (RBD) 1 and RBD-2 in emerin [18]. We predict GCL and HDAC3 compete for binding emerin based upon their overlapping binding sites. There is precedent for this, as GCL and BAF compete for binding to emerin [18].
In our model, increased hyperphosphorylated Rb in EMD −/y mice is the primary driving force for passage through the restriction point in EMD −/y mice, as first described by Melcon and colleagues [27]. Here, increased hyperphosphorylated Rb caused increased E2F-mediated transcription that persisted after differentiation induction. This delay eventually resolved after a few days and myotubes eventually formed in vitro and in vivo. These events were predicted to be a compensatory mechanism allowing skeletal muscle formation and regeneration to occur in EMD −/y mice, which have surprisingly mild phenotypes [27,28]. In our model, emerin can bind HDAC3 and GCL in EMD −/y progenitors expressing wildtype emerin. GCL binding sites at the nuclear envelope would not be saturated because HDAC3 and GCL compete with each other for emerin binding. In this model there would be insufficient recruitment of GCL to alter the dynamics of E2F-DP-mediated transcription activation in the presence of high levels of Rb and hyperphosphorylated Rb in EMD −/y cells. ∆95-99 fails to bind GCL and HDAC3, so this mutant behaves just like the EMD −/y parental cell line. In M179-, S54F-, or Q133H-expressing EMD −/y progenitors, these mutants retain their ability to bind GCL but fail to bind HDAC3. Thus, emerin can bind more GCL, thereby releasing more GCL from E2F-DP to further increase the expression of E2F-DP-regulated genes that allow for passage through the restriction point.
We predict HDAC3 also plays a role in repressing transcription of cell cycle genes, but it is unclear whether it acts with Rb or GCL to repress E2F-regulated transcripts in cycling cells or if it acts to stably repress E2F-regulated chromatin in permanently arrested cells, like myotubes. Our data would suggest the latter. Thus, it will be important to test whether HDAC3 association with emerin plays a role in the compensatory mechanisms predicted to help stabilize repressed chromatin containing E2F-regulated genes, and if this is required for permanent cell cycle arrest. It will also be important to determine the dynamics of GCL and HDAC3 competition for emerin binding and how this affects passage through the restriction point. Whether expression of wildtype emerin in EMD −/y cells alters Rb phosphorylation dynamics will also need to be tested.
Emerin mutants Q133H and ∆95-99 had less MyHC expression after 48 h and formed fewer myotubes after 72 h of differentiation than EMD −/y cells. S54F and M179 were similar to EMD −/y cells in differentiation index and myotube formation. Both S54F and M179 bind normally to lamin A, whereas Q133H and ∆95-99 disrupt binding to lamin A [71]. We speculate that failure to bind lamin A in Q133H and ∆95-99 mutants contributes to their decreased differentiation. One possibility is loss of lamin A binding to emerin may impair nuclear structure [72,73], which may lead to increased senescence or death and subsequently less myotubes. Failure of lamin A to bind Q133H and ∆95-99 mutants may also disrupt stabilization of repressed chromatin at the nuclear periphery [74,75], which may affect the expression of genes important for myoblast commitment and myotube formation. Q133H and ∆95-99 also fail to self-assemble into curvilinear structures in vitro [26], suggesting this failure could contribute to a more severe phenotype either due to an unstable emerin conformation or due to failure to bind an unidentified protein that regulates myogenic differentiation. Another hypothesis would involve compensatory mechanisms via other inner nuclear membrane proteins such as Lap2β [27,29] and Lap1 [76]. Lap2β is similar to emerin in its LEM, transmembrane, and nucleoplasmic domains [1,71]. Both these proteins show some functional redundancy with emerin and their expression levels are upregulated in the absence of emerin [76,77]. It is possible that Lap1 and Lap2β are compensating for loss of some functions in EMD −/y cells that may involve binding to and repressing an unknown partner at the inner nuclear membrane. In this functional redundancy model, expression of the Q133H and ∆95-99 mutants in the EMD −/y cells specifically alters binding to this unknown partner and more of the partner is now active to negatively repress myogenic differentiation. Alternatively, release of this partner from emerin now allows it to bind to Lap2β or Lap1 and displaces other binding partners on these proteins that negatively impact myogenic differentiation.
The transcriptomic analysis of the EDMD-causing emerin mutants uncovered pathways implicated in myogenic differentiation. Members of the Hippo family play vital roles in skeletal muscle development [78], organ size control [79], and tissue regeneration [80] in mammalian cells. Two transcriptional cofactors within the Hippo pathway, Yap (Yap1) and Taz (Wwtr1), primarily regulate gene expression by binding Tead1-4 transcription factors. Together Yap, Taz, and Teads form the core of the Hippo signal transduction network, which includes the Hippo kinase cascade, comprising kinases Mst1 (Stk4), Mst2 (Stk3), Lats1, and Lats2 [80][81][82]. Yap and Taz were shown to regulate myoblast proliferation and terminal differentiation in regenerating myoblasts in vivo and in satellite cells derived from isolated mouse muscle fibers [83][84][85]. Increased Yap and Taz expression is associated with myogenic progenitor proliferation [83,85,86]. Sustained Yap upregulation is sufficient to inhibit the terminal differentiation program vital for myocyte fusion and formation of myofibers [85,86]. Interestingly, sustained Taz upregulation enhances terminal differentiation of myoblasts in vitro and in vivo [87][88][89]. Yap was recently shown to contribute to the pathogenesis in muscular dystrophy, as increased nuclear localization of Yap is seen in lamin-A-related congenital muscular dystrophy (L-CMD) [90]. Our analysis also revealed significant upregulation of Yap expression and activity along with upregulation of several other pathways' components in the EDMD-causing emerin mutants, suggesting there is increased Yap localization in the nucleus that promotes proliferation activity and inhibits the differentiation program [83]. Therefore, one can speculate that persistent activation of Yap in the EDMD-causing emerin mutants has negative effects on myogenic differentiation.
There is increasing evidence suggesting extracellular matrix (ECM) components are essential signaling mediators in the satellite cell niche, both for the maintenance of satellite cell identity and regulation of its activation [91,92]. Heparan sulfate proteoglycans syndecan-3 and syndecan-4 regulate satellite cell activation and satellite cell self-renewal is impaired in syndecan-4 knockout mice [93,94]. The ECM provides stimuli for muscle cell development, which is independent of the expression of muscle-specific transcription factors Myf-5 and MyoD [93,95,96]. During differentiation, heparan sulfate proteoglycans interact with multiple growth factors in the muscle basal lamina, including insulin-like growth factor (IGF), fibroblast growth factor (FGF), hepatocyte growth factor (HGF), and transforming growth factor beta (TGF-β), which are well established regulators of satellite cell proliferation and differentiation [92,97,98]. The dynamic expression of heparan sulfate proteoglycans during skeletal muscle regeneration and differentiation likely reflects different functions for these complex molecules [99,100]. EDMD-causing emerin mutants showed significantly decreased expression for collagen, laminin, von Willebrand factor, thrombospondin 1, reelin, and vitronectin, revealing alterations in the ECM-receptor pathway needed for regulation of cellular events vital for muscle growth, repair and to mediate cell-cell adhesion during skeletal muscle regeneration. Laminin is essential for key processes that occur during differentiation, where it enhances myoblast proliferation, migration, and alignment involved in myotube formation; absence of laminin impairs myogenic differentiation [101]. In EDMD, this would be predicted to cause aberrant myogenic regeneration due to the altered response of myogenic progenitors to the available growth factors needed for progenitor activation and myotube formation [102][103][104].
Apelin is locally produced by skeletal muscle fibers in response to exercise and acts locally to improve muscle metabolism and function [105]. Apelin is expressed, secreted, and responsive to exercise-activated signaling pathways in cultured human primary myotubes [105]. Apelin also stimulates energy expenditure by increasing vascular mass and mitochondrial biogenesis [106]. Several lines of evidence indicate mitochondrial activity and function are linked to myogenic differentiation [107][108][109][110][111]. Mitochondrial activity is increased when myoblasts differentiate into myotubes [112][113][114]. During myogenesis, there is a metabolic shift from glycolysis to oxidative phosphorylation as the major energy source, suggesting this metabolic shift may be a crucial event that regulates cell differentiation [115][116][117][118][119]. Our analysis revealed that EDMD-causing emerin mutants show alterations in apelin and associated metabolic pathways, suggesting that these pathways may be responsible for a metabolic shift that impairs myogenesis due to diminished energy utilization. The function of Hypoxia-inducible factor 1α (HIF1α) in myoblasts has been controversial. Two groups revealed hypoxia-induced HIF1α accumulation inhibited myoblast differentiation [120,121], whereas another study revealed that HIF1α knockdown inhibited myoblast differentiation in normoxic conditions [122]. A recent study reported that HIF1α inhibited ischemia-induced muscle regeneration by inhibiting Wnt signaling [121]. Collectively, the function of HIFs in muscle stem cells in vivo is not well understood. Whether Apelin or HIF-1 signaling are important in the impaired myogenic differentiation seen in EDMD remains to be determined.

Conclusions
These studies show, for the first time, that emerin mutations found in EDMD patients cause impaired myogenic differentiation. The impaired differentiation is likely caused by loss of HDAC3 binding, since these mutants share in their failure to bind HDAC3. The use of these individual mutants allowed for a robust approach to identifying pathways implicated in the impaired differentiation in EDMD. Using this approach, we narrowed the implicated pathways to four to six pathways. Future studies to independently test each of these pathways will be needed to confirm their involvement in the impaired differentiation seen in EDMD.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4409/9/6/1463/s1, Figure S1: Cell cycle withdrawal 48 h after differentiation induction of M179 and S54F mutant myogenic progenitors, Figure S2: Cell cycle withdrawal 72 hours after differentiation induction of M179 and S54F myogenic progenitors, Figure S3: Cell cycle withdrawal 48 h after differentiation induction of EDMD mutant progenitor lines Q133H and ∆95-99, Figure S4: Cell cycle withdrawal 72 h after differentiation induction of EDMD mutant progenitor lines Q133H and ∆95-99, Figure S5: Significant changes in the transcriptome are seen during differentiation of EDMD mutant myogenic progenitors, Figure S6: Significant changes in the transcriptome are seen at each daily transition during differentiation of EDMD mutant myogenic progenitors.