Proteomic characterisation of human LMNA-related congenital muscular dystrophy muscle cells

LMNA -related congenital muscular dystrophy (L-CMD) is caused by mutations in the LMNA gene, encoding lamin A/C. To further understand the molecular mechanisms of L-CMD , proteomic profiling using DIA mass spectrometry was conducted on immortalized myoblasts and myotubes from controls and L-CMD donors each harbouring a different LMNA mutation (R249W, del.32K and L380S). Compared to controls, 124 and 228 differentially abundant proteins were detected in L-CMD myoblasts and myotubes, respectively, and were associated with enriched canonical pathways including synaptogenesis and necroptosis in myoblasts, and Huntington’s disease and insulin secretion in myotubes. Abnormal nuclear morphology and reduced lamin A/C and emerin abundance was evident in all L-CMD cell lines compared to controls, while nucleoplasmic aggregation of lamin A/C was restricted to del.32K cells, and mislocalisation of emerin was restricted to R249W cells. Abnormal nuclear morphology indicates loss of nuclear lamina integrity as a common feature of L-CMD, likely rendering muscle cells vulnerable to mechanically induced stress, while differences between L-CMD cell lines in emerin and lamin A localisation suggests that some molecular alterations in L-CMD are mutation specific. Nonetheless, identifying common proteomic alterations and


Introduction
LMNA-related congenital muscular dystrophy (L-CMD, ORPHA:157973) is an extremely rare genetic condition, with an estimated incidence of <1/1,000,000 [1] caused by mutations in the LMNA gene that encodes the type V intermediate filament proteins, lamins A and C [2,3].L-CMD has been described as the most severe of the striated muscle laminopathies and is characterized by onset before the age of 2 years old [3,4].Distinct features include major muscle atrophy and weakness, mainly affecting the axial muscles, leading to a complete absence of or limited motor achievements [3,4].A hallmark of L-CMD is a "dropped head" due to weakness of muscles in the neck [2,[5][6][7], and other characteristics include the presence of multiple joint contractures, and life-threatening severe respiratory insufficiency, requiring mechanical ventilation [3,4].Cardiac arrhythmias have been identified in L-CMD patients, suggesting there is some cardiac involvement in this disease [3,4].Current treatment options are limited to physiotherapy, surgery to treat contractures, as well as managing the risk of respiratory and cardiac manifestations which are common causes of premature death.
Lamins A and C, along with B-type lamins, are the major components of the nuclear lamina (NL) [8][9][10].Lamin A/C are connected to other nuclear envelope (NE) proteins through their mutual association with the Linker of Nucleoskeleton and Cytoskeleton (LINC) complex, which functions to connect the NL to the cytoskeleton in mammalian cells [11].Lamin A/C is known to interact directly with LINC complex proteins Sad1 and UNC84 domain containing SUN proteins 1 and 2 [12], and associated protein, emerin [13], and is involved with a wide variety of cellular processes including the regulation of cell stability, cell motility, mechanosensing, gene regulation, chromosome organization, DNA damage repair, telomere protection, and cell differentiation, including myogenesis [14][15][16][17][18][19][20].Lamin A/C has been widely studied in many of these contexts, and in recent years, some progress has been made to elucidate the pathophysiological mechanisms of L-CMD, downstream of LMNA mutations.It was noted, for example, that fibroblasts from an individual with a severe L-CMD and lipodystrophy phenotype harbouring a heterozygous LMNA R388P mutation senesced prematurely, had abnormal cellular morphology and a small percentage of abnormally shaped nuclei [21].In C2C12 cells transfected with lamin A carrying the same mutation, mutant lamin A mostly accumulated within the nucleoplasm, with less lamin A being correctly localized at the nuclear periphery [22].Additionally, altered anchorage of the inner nuclear membrane protein and interaction partner of lamin A/C, emerin, and LAP2α, another nuclear envelope protein, was evident [22].In C2C12 cells in which the R249W LMNA mutation, a common cause of L-CMD, was introduced by CRISPR/Cas9 editing, mislocalised emerin was detected at the endoplasmic reticulum, nuclei showed a significant reduction in their circularity index, and there was evidence of DNA damage and impaired myogenic differentiation [23].Similarly, skin fibroblasts from L-CMD individuals carrying R249W and del.K32 LMNA mutations also exhibited mislocalised lamin A/C, which was almost exclusively found in the nucleoplasm [24].In the del.K32 myoblasts from a mouse model of L-CMD, a strong reduction of lamin A/C was noted, in addition to partially mislocalized emerin.The cells were found to have impaired differentiation, with highly misshapen myonuclei, that were elongated, enlarged, and situated in the middle of myotubes [24].In addition to emerin, lamin B1 and SUN2 were also mislocalized in del.K32 mouse myotubes, in contrast to the apparent correct localization of lamin B1 and SUN1 in the respective myoblasts, suggesting that the nuclear defects are exacerbated by differentiation [24].There is evidence too that L-CMD causing mutations interfere with mechanosignalling pathways in skeletal muscle, subsequently affecting muscle growth [25].
Human muscle stem cells carrying either the del.K32, R249W or L380S LMNA mutations were each found to have impaired myogenic fusion, due to disorganized cadherin/β catenin adhesion complexes, stretched myotubes and overloaded muscle fibres with aberrant regulation of the yes-associated protein (YAP) [25,26].Skeletal muscle from Lmna-CMD mice was also unable to hypertrophy in response to functional overload, due to defective accretion of activated satellite cells [25,26].
The wide range of LMNA mutations known to cause L-CMD complicates gene therapy development for the condition [27], and it may be too late for gene therapy to be fully effective by the time a diagnosis is made.Variable severity and symptom onset time with little obvious relationship to the mutation further complicates matters [3,4].Alternatively, therapies tailored towards conserved features of L-CMD might, in combination with gene therapy, offer maximum benefit to patients.This approach is gaining momentum for another inherited neuromuscular disease, spinal muscular atrophy, where gene replacement strategies have shown incomplete efficiency [28,29].While the studies described above have generated valuable insights into the cellular consequences of L-CMD, identification of targets for therapy development is hindered by limited insights into the molecular pathways downstream from LMNA to modulate pathogenesis.
Here, we have conducted targeted and unbiased proteomic analyses on immortalized myoblasts and myotubes from healthy controls compared to three individuals with L-CMD, each harbouring a different LMNA mutation (i.e., R249W, del.32K and L380S), with the aim of determining whether the L-CMD cells may share a core molecular signature.

Cell culture
Immortalized myoblast cell lines were from three human control donors without neuromuscular disease, and from three individuals with L-CMD (Supplementary File 1, Table 1) [3,30].They were immortalized by transduction with human telomerase reverse transcriptase (hTERT) and cyclin-dependent kinase-4 (Cdk4-R24C mutant) containing retroviral vectors, at the Institut de Myologie, Paris, as described previously [31].Myoblasts

Cell proliferation assay
Control (C5d, C25, C41) and L-CMD (L380S, del.K32, R249W) myoblasts were grown in complete medium, as described above, seeded in T25 cell culture flasks at a density of 30,000 cells per flask (approximately 1% of the maximum confluency of a T25 flask).The cells were maintained in a humidified incubator (LEEC, Nottingham, UK) at 37°C and 5% CO2 for three days during their exponential growth phase.After this point, the media was removed from the flask and cells were trypsinized and counted.Each cell line was grown in three T25 flasks (to provide technical replicates), and each flask was counted four times.
Before the final cell count, each of the cell lines were less than 50% confluent, indicating that the myoblasts were still growing exponentially.A total of 12 measurements were performed for each cell line.Doubling time (DT) for each cell was calculated using the following equation: DT = ln(2) x t/(ln(C2) -(ln(C1)).Here, t represents the culture duration in hours, C2 is the number of cells at the end of the culture, and C1 is the number of seeded cells at the beginning of the experiment.

Sample preparation
Protein was extracted from myoblast and myotube samples from healthy control donors (n=3) and L-CMD patients harbouring mutations in LMNA (n=3) using 250μL of extraction buffer (8M urea (Cat no: U0631; Sigma Aldrich), 100mM ammonium bicarbonate, 2% sodium deoxycholate in sterile dH2O).Samples were sonicated at 5 microns for 10s and centrifuged at 13000 RPM for 5 min at 4°C to pellet insoluble material.A small aliquot of cell extract from each sample was used to determine protein concentration, whilst the remaining samples were stored at -80°C for downstream analysis.The protein concentration of each sample was determined using the Pierce™ BCA Protein Assay Kit.A minimum of 30μg of protein is needed to perform quantitative DIA MS analysis with a bespoke identification library therefore we ensured each sample contained more than the minimum protein concentration.
Each sample (50μg) was diluted with extraction buffer to obtain equal concentrations and then reduced with 5mM tris(2-carboxyethyl) phosphine (TCEP) at 30°C for 1 h followed by alkylation with 10mM iodoacetamide (IAA) in darkness at room temperature for 30 min.The reaction was quenched with 20mM DTT to deactivate any unreacted reagents.The samples were diluted to 1.5M urea and subsequently digested overnight with sequencing grade trypsin in a ratio of 1μg of protease to 50μg protein.
The peptides were subjected to cleanup using C18 columns, and the cleaned, digested sample was then dried and resuspended to 1μg/μl in loading buffer (100% water, 0.1% formic acid).Data independent acquisition was performed on individual samples (DIA-MS).
In addition, a pool of all the samples was prepared, and a portion subjected to nanoLC MS/MS analysis using data dependent acquisition (DDA-MS).The remnant of the pooled sample was then fractionated on high pH C18 Reverse Phase into 12 fractions before analysing the fractions individually in DDA mode.
The trap was then switched in-line with the analytical column (Easy-spray Pepmap RSLC C18 2μm, 50cm x75um ID).Peptides were eluted from the column using a linear solvent gradient using the following gradient: linear 4-40% of buffer B over 120 min, linear 40-60% of buffer B for 30 min, sharp increase to 95% buffer B within 0.1 min, isocratic 95% of buffer B for 15 min, sharp decrease to 2% buffer B within 0.1 min and isocratic 2% buffer B for 15 min.The mass spectrometer was operated in DDA positive ion mode with a cycle time of 1.5 s.The Orbitrap was selected as the MS1 detector at a resolution of 120000 with a scan range of from m/z 375 to 1500.Peptides with charge states 2 to 5 were selected for fragmentation in the ion trap using HCD as collision energy.
The raw data files were converted into mgf using MSconvert (ProteoWizard) and searched using Mascot with trypsin as the cleavage enzyme and carbamidomethylation as a fixed modification of cysteines, against the Swissprot database, restricted only to proteins from humans.Note that the iRT peptides were added to this database.The mass accuracy for the MS scan was set to 20 ppm and for the fragment ion mass to 0.6 Da.

Data independent acquisition (DIA) mode
For quantitative MS, sample (5 μg) was injected onto the same LCMS set up as above with the same gradient, however data acquisition was performed in data independent acquisition (DIA) mode.The DIA MS method alternates between a MS scan and a tMS2 scan containing 24 scan windows.The MS scan has the following parameters: the Orbitrap at 120000 resolution is selected as detector with a m/z range from 400 to 1000.The tMS2 scan uses HCD as activation energy with fragments detected in the Orbitrap at 30000 resolution.
The first 20 m/z windows are 20 mass units wide from 410 to 790 followed by a 30m/z window from 790-820, a 40m/z window from 820-860, a 50m/z window from 860-910 and a 60m/z window from 910-970.
All Mascot searches using the DIA data were exported as .datfile and assembled to a spectral library in Skyline by associating each peptide to its respective protein.After the quantitative spectra were imported, peaks were reintegrated using the mProphet peak scoring model [39] To identify differentially expressed proteins, the sum total area value for each protein identified by ≥2 peptides in the L-CMD (n=3) and control samples (n=3) was averaged, and only proteins with an average fold change of <0.8 and >1.25 with a p < 0.05 (as determined by a t-test) across the three L-CMD samples compared to controls were considered in further analysis.

QIAGEN Ingenuity Pathway Analysis (IPA®)
Dysregulated pathways were identified downstream of LMNA mutations in L-CMD patient myoblasts and myotubes (compared to controls) using QIAGEN Ingenuity Pathway Analysis (IPA®) software (QIAGEN; https://digitalinsights.qiagen.com/products-overview/discoveryinsights-portfolio/analysis-and-visualization/qiagen-ipa/).The fold change value for each protein along with the assigned p-value was inputted into IPA® in the form of an Excel spreadsheet.IPA® uses right-tailed Fisher's Exact Test to calculate the p-value determining the probability that each cellular and molecular function or canonical pathway assigned to that dataset is due to chance alone, and the final lists of functions and pathways were ranked accordingly to the resulting p-value.IPA® also produces a Z-score for enriched canonical pathways.The Z-score is a prediction of whether a pathway is activated or inhibited based on the direction of expression change in the input dataset.This is done by comparing the IPA® database, which predicts what to expect when an upstream regulator interacts with its downstream target, to the direction of differential gene/protein expression that was observed in the input dataset.Z-score of ≥2 represents the prediction of activation, while Z-score ≤ -2 represents the prediction of inhibition.

Statistical analysis
All statistical analysis were carried out using GraphPad Prism Version 9.0.0.for Windows (GraphPad Software, San Diego, CA, USA, www.graphpad.com).Western blot densitometry measurements were assessed using unpaired, two-tailed t-tests, with unequal variance to determine significant differences between L-CMD myoblast and myotube samples and healthy controls.To determine if changes in protein expression in L-CMD myoblasts and myotubes were significantly different to controls after MS analysis, multiple unpaired twotailed t-tests were performed assuming unequal variance applying no correction for multiple comparisons.Also following MS analysis, significant protein expression changes were determined between L-CMD myoblasts and L-CMD myotubes, as well as control myoblasts and control myotubes.For this, paired two-tailed t-tests with unequal variance were used.

Quantitative proteomic identification of differentially expressed protein profiles in L-CMD cells compared to healthy controls
To determine whether there is a molecular signature of L-CMD in myoblasts and myotubes that is conserved across each of the three different mutations (R249W, L380S, del.K32) a quantitative proteomic comparison was made with control cells (C5d, C25, C41) using Data Independent Acquisition MS This approach identified a total of 10,977 proteins in total with a Mascot significance threshold for of p<0.05 (Supplementary File 2).Following the subsequent filtering steps described in the methods section, 124 and 228 proteins met the criteria for differential abundance in the L-CMD myoblasts and myotubes, respectively, compared to control cells (Figure 1A & B).Of these, 85 and 171 proteins were downregulated in the L-CMD myoblasts and myotubes, respectively, compared to controls.
Two proteins, BCL7B (B-cell CLL/lymphoma 7 protein family member B) and transketolase, were commonly downregulated in both L-CMD myoblasts and myotubes compared to control cells.No commonly upregulated proteins were identified between the 39 and 57 proteins that were significantly increased in L-CMD myoblasts and myotubes, respectively, compared to controls (Figure 1A & B), though two proteins, DNA primase large subunit and F-box/LRR-repeat protein 3, were upregulated in L-CMD myoblasts but downregulated in L-CMD myotubes.Lamin A/C itself met the criteria for reduced abundance in the L-CMD myotubes compared to controls (ratio=0.450,p=0.038).Other LINC complex-associated proteins including lamin B1 and B2, emerin, FHL1, SUN1 and SUN2, and nesprin-1 and nesprin 2, were detected but did not meet the criteria for differential abundance.In the case of SUN2 and emerin, this was most likely due to the technical variability in the detection and measurement of individual peptides within each sample, as confirmed by inspection of the raw proteomics data.
To gain insights into whether the molecular response associated with myoblast differentiation was similar in the L-CMD cells compared to controls, the differentially expressed proteins identified in L-CMD myoblasts vs their respective myotubes were compared with those that were differentially expressed in the control myotubes vs their respective myoblasts.A total of 136 proteins were differentially expressed in control myoblasts compared to their myotubes, whilst 570 proteins were differentially expressed in L-CMD myoblasts compared to their myotubes.Of these, only 3 proteins were commonly upregulated and 28 commonly downregulated in L-CMD and control myoblasts compared to their respective myotubes (Figure 2A & B).The lack of overlap of differentially expressed proteins in the L-CMD and control myoblast vs myotube datasets, therefore implies that proteins that are involved at different points during muscle cell differentiation are dysregulated in L-CMD myoblasts and/or myotubes.

L-CMD myoblasts have a decreased proliferation rate and both myoblasts and myotubes exhibit nuclear morphology abnormalities
Next, we investigated the doubling time of each myoblast cell line, to determine the relevance of the finding described above that ten differentially abundant proteins in the L-CMD myoblast cell lines were associated with cell growth and proliferation canonical pathways.On average, the L-CMD myoblasts had a longer doubling time compared to the control cells during their exponential growth phase (36.24hrs vs 26.72 hrs, p = 0.023), indicating that the L-CMD cells had a decreased proliferation rate (Figure 3A).The quantitative proteomics analysis also identified that 32 proteins differentially abundant in the L-CMD myotubes compared to controls were associated with cell morphology.In culture, there were no obvious differences in the gross cellular morphology of L-CMD myoblasts compared to controls (Supplementary File 1, Figures 1-2).All L-CMD and control myoblasts appeared small and uniform in shape, except for the C5d control cell line which appeared to have slightly larger cells that were less uniform in shape, with some myoblasts appearing elongated.Across all cell lines, the myotubes also appeared similar in morphology when growing in culture (Supplementary File 1, Figures 1-2).Immunofluorescence microscopy analysis using DAPI nuclear staining, however, demonstrated that each of the L-CMD cell line myotubes exhibited disordered nuclei that did not all appear to fuse together properly and their placement within the myotube was more random and irregular compared to control myotube nuclei (Figure 3B).In contrast, control myotubes contained consistently fused nuclei that formed elongated, thick structures.In addition, some L-CMD myotubes were shorter, contained less nuclei or formed myotubes with "clumps" of nuclei which were not elongated.
To determine whether the L-CMD myoblasts used in this study also have nuclear morphology abnormalities, a classification method described previously by van Tienen et al.
(2018) was used to analyse immunofluorescence microscopy images from L-CMD myoblast and myotube cultures [43].Lamin A/C immunostaining in combination with DAPI allowed the nuclear envelope and nuclear defects to be visualized more clearly than with DAPI alone.A representative example of each type of abnormality that was identified is given in Figure 3C.Approximately 49% and 50% of myoblast nuclei were classified as abnormal in the L380S and del.K32 L-CMD myoblasts, while 47% and 54% of the nuclei in the myotubes for the respective cell lines were abnormally shaped (Figure 3D and E).The R249W L-CMD cells exhibited an increased amount of abnormally shaped nuclei in myotubes (95%) compared to myoblasts (54%), suggesting that nuclear defects in this L-CMD cell line are exacerbated during differentiation (Figure 3D and E).In comparison, only 4% of control myoblasts had nuclear abnormalities, which is comparable to the previous study by van Tienen et al., where 4.8% of healthy cell nuclei were abnormal, and only 2% of control myotube nuclei appeared to have abnormalities (Figure 3D and E).Nuclear shape abnormalities (NSA) were the most frequent abnormalities seen across all three L-CMD myoblasts and myotubes, followed by blebs.L-CMD L380S myoblasts also exhibited many nuclei with donut shapes, which was not observed in the other L-CMD cell lines.In comparison to the L-CMD nuclei, the control cell nuclei were spherical or slightly oval, as expected.

Lamin A/C is reduced in expression in L-CMD myotubes compared to controls and is partially mislocalized in L-CMD del.K32 myoblasts
With lamin A/C having been identified by quantitative proteomics analysis as reduced in abundance in the L-CMD myotubes compared to controls (ratio=0.450,p=0.038), we next determined the relative levels of lamin A/C in myoblast and myotubes extracts using quantitative western blotting analysis.Quantitative western blotting analysis of myoblast cell extracts indicated slight, non-statistically significant reduction in lamin A, but not lamin C levels in the L-CMD cells compared to healthy controls (by 54.55%, p=0.244) (Figure 4A &E).The lamin A band exhibited a slightly faster electrophoretic mobility in the L-CMD myoblasts compared to control myoblasts (Figure 4A & C) that was not apparent in the myotube extracts (Figure 4B), where lamin A appeared to be of comparable molecular weight in the L-CMD and control samples.The lower molecular weight lamin A band in L-CMD myoblasts may indicate the presence of truncated lamin A in each of the L-CMD cell lines or a degraded form of mature lamin A, potentially due to lamin A being more unstable in the L-CMD myoblasts.It was also observed that there were higher molecular weight lamin A bands that were present in addition to the lamin A bands at the expected molecular weight (70kDa) in C25 and C41 control myoblast samples, as well as in all of the control myotube samples.This could indicate the presence of pre-lamin A, which would have a slightly higher molecular weight than mature lamin A [44], however further experiments are needed to confirm this.In myotube extracts, both lamin A and C were reduced in the L-CMD cells vs controls (by 98.68%, p=0.044, and by 63.29%, p=0.043, respectively) (Figure 4B & D).
With the observations of reduced levels of lamin A/C and differences in its molecular weight in L-CMD cells compared to controls, we next wished to determine whether lamin A/C is correctly localized in the L-CMD cells, which may help to gain insights into the cause and / or consequence of its reduced expression.In the L-CMD myoblasts harbouring the R249W and L380S mutations, lamin A/C was observed at the nuclear envelope, possibly with some diffuse lamin A/C immunoreactivity also in the nucleoplasm, to a greater extent than was evident in the control myoblasts (Figure 4F & G).Technically, this was unreliable to quantify, particularly in the R249W cell line, due to the abnormal nuclear morphologies causing wrinkling and invaginations around the nuclei.In the L-CMD del.K32 myoblasts, some lamin A/C was evident at the nuclear envelope, but most appeared to be present in nuclear aggregates within the nucleoplasm in all cells examined (out of 100 cells) (Figure 4G).This nuclear aggregation was not apparent in myotubes derived from the del.K32 cell line, and localization of lamin A/C in myotubes from the R249W and L380S cell lines also appeared to be relatively consistent with the distribution seen in healthy controls.To note, LMNA protein expression was not quantifiable using immunofluorescence microscopy and could not be compared to western blot analysis of total LMNA protein as samples were not imaged using consistent laser intensity.

Emerin is mislocalized to the cytoplasm in L-CMD R249W myoblasts and is reduced in expression in L-CMD myotubes compared to controls
The lamin A/C binding protein, emerin, was clearly localized at the nuclear envelope in all control myoblast and myotube cell lines examined, though some emerin was found to be mislocalized to the cytoplasm in approximately one third of the L-CMD myoblasts examined that harboured the R249W mutation (Figure 5A).On average across the three L-CMD myoblast cell lines, emerin was not significantly different to controls (p=0.097)(Figure 5B).
In L-CMD myotubes, emerin was, however, significantly reduced compared to controls (by 74.25%, p=0.040) (Figure 5C).An incidental finding, prompted by this observation, is that emerin expression generally appeared to be decreased on average by 81.37% in fully formed control myotubes, compared to their respective proliferating and differentiating myoblasts across six different cell lines.The most prominent reduction in emerin expression levels occurred between the second and third day (corresponding to timepoints 4 and 5) following differentiation initiation (timepoint 3) (Supplementary File 1, Figure 3).A similar pattern was noted in the L-CMD cell lines, albeit with emerin being less easily detectable at each timepoint examined (Supplementary File 1, Figure 3).

SUN2 was correctly localized at the nuclear envelope in L-CMD cells and reduced in expression in myotubes
SUN2 is a known interaction partner of lamin A at the INM (Crisp, et al., 2006), and is dependent on lamin A for its correct localization at the NE (Haque et al., 2010).SUN2 appeared to be correctly localised at the NE in the control and L-CMD myoblasts and myotubes, suggesting that the LMNA mutations harboured in the L-CMD cells do not affect SUN2 localization (Figure 6A).The intensity of SUN2 staining appeared reduced in L-CMD myoblasts and myotubes compared to control cells, when imaged using a consistent laser intensity (Figure 6A).When measured, it was found that in L-CMD myoblasts, SUN2 staining intensity was reduced by 32.23% compared to controls, but this reduction was not statistically significant (p= 0.439).In L-CMD myotubes, however, SUN2 staining was reduced by 85.35%, which was statistically significant (p= 0.010).Quantitative western blotting analysis confirmed this result, showing that SUN2 was not significantly different in expression in the L-CMD myoblasts compared to controls (p=0.401)(Figure 6B).In the myotube samples, SUN2 was significantly reduced in L-CMD myotubes compared to controls (by 52.21%, p=0.049) (Figure 6C).

Discussion
In the past couple of decades, quantitative proteomics has gained popularity as a useful method for identifying and quantifying all the proteins within a biological sample on a largescale in an unbiased manner.Proteomic studies have advanced our understanding of cellular signalling networks and have improved diagnosis and molecular understanding of disease mechanisms.Comparison of the proteome of L-CMD cells to controls, a total of 124 differentially expressed proteins were identified in the L-CMD myoblasts, and 228 proteins were identified in the L-CMD myotubes, which could be potentially linked to the development of the disease.It was noted that there were more dysregulated proteins in the L-CMD myotubes than the L-CMD myoblasts, compared to controls, indicating that processes are potentially more impaired in L-CMD myotubes.There was also found to be very little cross over between the dysregulated proteins that were identified in the L-CMD myoblasts and myotubes.This implies the processes and signalling pathways that are dysregulated in L-CMD myoblasts and myotubes differ.This was substantiated by the finding that the molecular and cellular pathways and the canonical pathways associated with the dysregulated proteins were different for the L-CMD myoblasts and myotubes.
Several particularly interesting canonical pathways were found to be associated with the dysregulated proteins that were identified in the L-CMD myoblasts and myotubes.A substantial number of the dysregulated proteins in myotubes were linked to carbohydrate metabolism.The two major energy sources for muscle contraction are glycogen and fatty acids, whose metabolic pathways converge into actyl-CoA for final oxidation via the Krebs cycle and the respiratory chain, therefore carbohydrate metabolism is important for muscle function [45].Metabolic dysfunction has been observed in Duchenne muscular dystrophy, and is characterised by reduced glycolytic and oxidative enzymes, decreased and abnormal mitochondria, decreased ATP, and increased oxidative stress [46][47][48][49][50][51].However, alterations in carbohydrate metabolism are often associated with neuromuscular disorders [52], therefore it may not be a direct consequence of LMNA mutations in L-CMD, and instead may be a result of muscle damage or abnormal function.Interestingly the insulin secretion signalling pathway, which is also a component of carbohydrate metabolism, was additionally found to be upregulated in L-CMD myotubes compared to controls.
The synaptogenesis signalling pathway was found to be elevated in L-CMD myoblasts compared to controls.Synaptogenesis is the formation of synapses between neurons in the nervous system.NMJ defects have been previously observed in models of autosomal dominant Emery-Dreifuss muscular dystrophy (AD-EDMD), which, similarly to L-CMD, is caused by LMNA mutations [53].Two AD-EDMD mouse models (Lmna H222P/H222P , Lmna -/-) were found to show innervation defects including misexpression of electrical activitydependent genes and altered epigenetic chromatin modifications, as well as aberrant NMJ architecture [53].Considering these results, it is plausible to infer that NMJ defects contribute to L-CMD pathophysiology, though confirmation would further investigation.
Huntington's disease signalling was upregulated in L-CMD myotubes compared to controls, and upon further examination of the dysregulated proteins involved in this pathway, two proteins that are involved in apoptosis were of particular interest; apoptotic proteaseactivating factor 1 (Apaf-1) and caspase-7.Apoptosis dysregulation has been identified as a feature of a number of neurodegenerative diseases, including Huntington's disease [54][55][56][57] In L-CMD myoblasts, it was found that the necroptosis pathway, another method of cell death, was significantly downregulated.Necroptosis has been found to be of central pathophysiological relevance in a variety of disease states including myocardial infarction and stroke [58,59], atherosclerosis [60], ischemia-reperfusion injury [61,62], pancreatitis [63,64], inflammatory bowel diseases [65], as well as neurological disorders [66,67].More recently, necroptosis has been implicated in neuromuscular diseases including Duchenne muscular dystrophy (DMD) and spinal muscular atrophy (SMA) [68,69].Considering the necroptosis pathway's emerging role in neuromuscular conditions as well as identification of its dysregulation in L-CMD myoblasts, it merits further verification.
We found that L-CMD myoblasts had a significantly decreased proliferation rate compared to controls.Furthermore, cellular growth and proliferation as well as cell development were enriched canonical terms associated with the differentially expressed proteins that were identified in L-CMD myoblasts using quantitative proteomics.This may suggest that in-vivo myoblasts fail to proliferate adequately, which could consequently hinder muscle regeneration and repair, as myoblast proliferation is a key step in this process.This would therefore impact L-CMD patients muscle integrity if it is subject to any damage or injury.For example, in models of other neuromuscular diseases, cell proliferation rate has previously been found to be altered compared to that of healthy controls.Satellite cells cultured from Duchenne muscular dystrophy patients (DMD) have been found to have an increased generation time, ceasing to proliferate beyond 100-1,000 cells, but still were capable of forming myotubes when differentiation was induced [70].In a study on spinal muscular atrophy (SMA), C2C12 myoblasts with differing levels of survival motor neuron protein (Smn) knockdown have reduced proliferation as well as fusion defects, correlating with Smn levels [71].These results suggest that increased doubling time of myoblasts could be related to neuromuscular disease pathology.The decreased proliferation rate observed in the L-CMD myoblasts could be a direct consequence of LMNA mutations, as cell proliferation rate has previously been shown to be altered when lamin A/C abundance is increased or decreased.The abundance of lamin A/C has been found to be downregulated in certain cancers, and depletion of lamin A/C abundance in healthy primary human fibroblasts leads to downregulation of the Rb family of tumor suppressors and a defect in cell proliferation [72].Additionally, in prostate tumor cell lines (LNCaP, DU145, and PC3), small hairpin RNA knockdown of lamin A/C decreased cell proliferation, whilst overexpression of lamin A/C stimulated cell growth [73].Despite these findings, here, we did not go on to observe a reduction in lamin A/C abundance in myoblasts, as would be expected if downregulation of lamin A/C decreases cell proliferation rate.However, another study on lamin A knockdown in lung carcinoma-derived A549 cells was also not found to perturb proliferation [74].
The nuclear morphology defects that were observed in the L-CMD cells are likely the consequences of a compromised NL in the cells.Abnormalities in nuclear structure are a well-known characteristic of LMNA mutations and have been previously observed in myoblasts from an Emery-Dreifuss muscular dystrophy (EDMD) patient with a confirmed LMNA mutation [75], myonuclei of muscle tissue from EDMD and Limb-Girdle (LGMD) muscular dystrophy patients carrying mutations in LMNA [76], as well as in skin fibroblasts from patients with a number of different laminopathies [77][78][79][80], and in myogenic cells derived from induced pluripotent stem cells (iPSCs) from patients with skeletal muscle laminopathies including L-CMD [81].How this is linked to disease pathology has already been well-discussed in the form of the structural hypothesis, with the idea that LMNA mutations lead to a weakened NL that results in the nucleus being unable to resist high mechanical strain in tissues exposed to tension, such as skeletal muscle [82].Of the different types of nuclear deformities observed, the L380S myoblasts exhibited many more donut shaped nuclei compared to the other L-CMD cell lines.It has previously been demonstrated that treatment of primary human skin fibroblasts with protein farnesyltransferase inhibitors (FTIs) resulted in a high frequency of cells with donut shaped nuclei [83], suggesting that the LMNA L380S mutation may affect farnesylation of lamin A/C.Across all L-CMD cell lines differentiated into myotubes, it was noticed that the nuclei within the myotubes were extremely disordered, which has previously been observed in del.K32 myotubes [84].Curiously, we also found that SUN2 was significantly reduced in abundance across L-CMD myotubes.It has been found that SUN1 and SUN2 variants disrupt myonuclear organization, as it has been found that myotubes from a patient carrying SUN1 mutations displayed defects in myonuclear organization [85].As well as this, events required for correct myonuclear arrangement, such as absence of recruitment of pericentrin, a centrosomal marker, to the nuclear envelope, and impaired microtubule nucleation [85].
Thus, perhaps reduction of SUN2 is contributing to nuclear disorganization in L-CMD myotubes.It would be of interest to upregulate SUN2 expression in L-CMD myotubes to determine whether this rescues myonuclear organization disruption, and whether this has any other effects on L-CMD myotube differentiation.Across all of the L-CMD cell lines, it was found that on average around half of the myoblast and myotube nuclei were irregularly shaped.For the L-CMD R249W cell line, though, almost all myotube nuclei exhibited abnormalities, meaning that the number of abnormal nuclei was almost doubled in R249W myotubes compared to myoblasts.This could suggest that nuclear abnormalities are exacerbated during differentiation in the R249W cells, which has not previously been described.Some emerin was also found to be mislocalized to the cytoplasm, perhaps specifically the ER, in R249W myoblasts, and not in any other L-CMD cells.This may suggest that the R249W mutation in particular may cause more severe nuclear defects.Emerin was found to be correctly localized in R249W myotubes, however, maybe not supporting the finding that nuclear defects are worsened in R249W myotubes.
Previously, emerin has been found to mislocalize and aggregate in foci in iPSC-derived myoblasts and the cytoplasm of C2C12 myoblasts harbouring R249W mutations [81,86].
Emerin has been identified as a well-characterised binding partner of lamin A/C, [13], and it has been found that lamin A is involved in tethering emerin to the NE [87].Consequently, mutations in LMNA may disrupt this interaction and cause emerin to become mislocalized.
Here, even when some emerin was mislocalized in the R249W myoblasts, there remained proportion of emerin that was correctly localized to the NE.This result suggests that lamin A-emerin interactions are not entirely lost and are merely compromised by the LMNA mutation.Loss of emerin from the NE could have implications in L-CMD disease development, as correct localization is critical for a protein's function [88].If a protein is in the wrong environment, it is unlikely to fold or assemble properly, and may also trigger secondary consequences such as protein degradation pathways, that could lead to further detrimental effects inside of the cell [88].Although emerin was only mislocalized in one L-CMD cell line, a reduction in emerin protein expression was observed across all L-CMD cell lines in myotubes compared to controls.It has previously been shown that mutant forms of emerin causing X-linked Emery-Dreifuss muscular dystrophy (EDMD) are mislocalised, and it has been proposed that an absence (caused by mutations that result in the deletion of EMD) or reduction of emerin at the NE may be responsible for the EDMD phenotype [89].It has also been found that COS-7 cells expressing mutant emerin (Del236-241, an EMD variant found in EDMD patients) that mainly localizes to the cytoplasm exhibit aberrant cell cycle length [90].Considering this has been linked to EDMD development, loss of emerin at the NE could also be contributing to L-CMD pathophysiology.This provides evidence that there may be some cross-over between the mechanisms behind EDMD and L-CMD disease pathways.This potentially can be expected as AD and AR-EDMD are also caused by LMNA mutations.
Whilst emerin was found to be mislocalized in the R249W myoblasts, the del.K32 myoblasts were the only cells to exhibit mislocalization of lamin A/C.In the del.K32 myoblasts, lamin A/C appeared to be aggregated in nuclear foci within the nucleoplasm.Increased nucleoplasmic aggregation of lamin A/C has been previously observed in L-CMD fibroblasts harbouring the same del.K32 mutation, as well as primary myoblasts derived from the Lmna dK32 mouse model of L-CMD [84].In this study, myotubes harbouring the del.K32 mutation were also found to have almost exclusively nucleoplasmic localization of lamin A/C, although this was not observed here [84].In the Bertrand et al. study, del.K32 myotubes were also found to have disrupted localization of other INM proteins such as emerin, lamin B1, SUN2 or nup153 (further information on where these proteins were mislocalized to is not provided), proving nuclear defects in cells harbouring the del.K32 LMNA mutation can also be severe [84].Absence of lamin A/C from the nuclear periphery and its accumulation in the nucleoplasm has been shown to be highly detrimental for myoblast differentiation in primary myoblasts derived from the Lmna dK32 mouse [91].This has been attributed to the inability of L-CMD post-mitotic myocytes to sequester musclespecific NE transmembrane proteins in the nuclear envelope that are required for chromatin remodelling [91].Depending on their level of phosphorylation, A-type lamins are able to assemble under the INM (where they are primarily localized), but they may also reside in the nucleoplasm [92].A-type lamins at the nuclear periphery are necessary for the nuclear sequestration of NE transmembrane proteins [93][94][95][96], and for the interaction with the cytoskeleton via the LINC complex [97].Other A-type lamin functions such as the regulation of gene transcription, DNA repair, and regulation of cell cycle and mechanotransduction also require lamin A/C to be correctly localized at the NL [98][99][100].
Based on the evidence outlined above, it appeared that cells harbouring the R249W or del.K32 mutations had different nuclear defects than the L380S L-CMD patient cells.This could be a consequence of the region of the LMNA gene that is affected by the mutations.The R249W mutation is located at the ERK1/2 binding domain in LMNA and causes an arginine residue to be substituted with a tryptophan residue.Arginine and tryptophan have very different properties.Arginine is amphipathic, whilst tryptophan in comparison is extremely hydrophobic in nature [101].Arginine is known to be common in binding sites, therefore, if arginine is substituted with another amino acid with different properties, this may affect interactions at these domains.The del.K32 mutation causes an in-frame deletion affecting a lysine residue.Similarly, to arginine, lysine is also commonly found in protein binding sites [101] .The del.K32 variant is located in the head region of lamin A/C, which is known to be involved in lamin assembly.Consequently, mutations in this region may particularly lead to nuclear assembly defects.The L380S mutation affects the HCD2 (highly conserved domain) region of LMNA, located at the end of the central rod domain, next to coil 2.
Lamin A/C were found to be significantly reduced across L-CMD myotubes compared to controls.This reduction in lamin A/C expression could be linked to the nucleoplasmic aggregation of lamin A/C observed in the del.K32 myoblasts, as this may indicate lamin A/C is unstable and could lead to increased degradation of the protein.In myoblasts and myotubes derived from Lmna ΔK32/ΔK32 mice, a mouse model of L-CMD, reduced levels of lamin A/C protein have also been observed, although LMNA mRNA levels were not reduced, suggesting a reduced translation efficiency or higher rate of degradation of mutant lamin A/C [91].Reduced lamin A/C levels could likely contribute to the L-CMD phenotype, as Lmna ΔK32/ΔK32 mice present with an even more severe phenotype than Lmna -/-mice (a lamin A/C deficient mouse model of AD-EDMD) [91].Whilst Lmna -/-mice present with growth retardation, skeletal and cardiac muscle involvement, hypoglycemia and sudden cardiac death, this phenotype was exacerbated in Lmna ΔK32/ΔK32 mice [91].The result presented here that lamin A/C were reduced across all myotubes differentiated from each of the L-CMD cell lines harbouring different LMNA mutations corroborates Bertrand et al.'s findings and suggests reduced lamin A/C may be conserved across different L-CMD-causing mutations.This finding may help to guide future research in L-CMD, as a therapeutic approach could be to upregulate lamin A/C production.Based on the finding that lamin A/C is reduced in Lmna ΔK32/ΔK32 mice, spliceosome-mediated RNA trans-splicing (SMaRT), an approach that targets RNA at the pre-mRNA level and converts endogenous mutated sequences into wild type ones has already been used to target the del.K32 L-CMD mutation [102].To achieve this, 5′-RNA pre-trans-splicing molecules containing the first five exons of Lmna and targeting intron 5 of Lmna pre-mRNA were developed and their efficacy at inducing transsplicing events on Lmna were tested and confirmed at the protein level in C2C12 myoblasts [102].This approach was then tested in-vivo in newborn mice using adeno-associated virus (AAV) delivery, although the efficacy of trans-splicing events were low [102].Despite this, these results provide the first evidence for reprogramming LMNA mRNA in vitro.

Conclusion
By conducting a quantitative comparison of the proteome of L-CMD myoblasts and myotubes and healthy control cells, a number of differentially expression proteins have been identified as well as potential dysregulation of a number of different molecular and cellular processes and signalling pathways.Amongst these, molecular and cellular processes common to the dysregulated proteins included cellular development, cell cycle, and cellular growth and repair.Upon this discovery, further examination of L-CMD myoblasts and myotubes revealed defects in cell proliferation, as well as nuclear morphology abnormalities.In addition to this, reduction of lamin A/C protein expression levels, and lamin A/C binding partners, emerin and SUN2 were observed across all L-CMD myotubes.In del.K32 myoblasts, lamin A/C was found to aggregate in the nucleoplasm, and in R249W myoblasts emerin was mislocalized.Proteomic analysis also revealed a number of canonical pathways that may be dysregulated including the insulin signalling pathway and Huntington's disease pathway in L-CMD myotubes, and the synaptogenesis and necroptosis signalling pathway in L-CMD myoblasts, which require verification and further investigation in future studies.The defects that were identified across the L-CMD cell lines may be associated with L-CMD pathophysiology, and might represent targets for the development of therapies.In future, it is important to study a greater number of cell lines harbouring the same mutations as the L-CMD cell lines used in this study, as well as different L-CMD causing mutations to determine whether the defects identified across the L-CMD cell lines are conserved features of L-CMD.Comparison of the proteomics data within this study with data from other published LMNA transcriptomic and proteomic datasets will also be useful to support the identification of potentially conserved changes to molecular pathways in L-CMD.were identified in L-CMD (R249W, L380S, del.K32) myoblasts and myotubes compared to controls (C5d, C25, C41).(A) Venn diagram illustrating the limited overlap of downregulated proteins that were identified in L-CMD myoblasts and L-CMD myotubes and showing no overlap between the upregulated proteins identified in L-CMD myoblasts and L-CMD myotubes.(B) Heatmap illustrating a total of 348 differentially expressed proteins were identified across the L-CMD myoblasts and L-CMD myotubes and the size of their fold-change, there were four commonalities across the L-CMD samples, with two proteins downregulated in L-CMD myoblasts and L-CMD myotubes, and two proteins upregulated in L-CMD myoblasts but downregulated in L-CMD myotubes.(C) Enriched molecular and cellular functions that were most common to the dysregulated proteins with the number of proteins associated with each term.(D) Enriched canonical pathways that were associated with the dysregulated proteins and the number of proteins associated with each term.A negative Z-score suggests inhibition, whilst a positive Z-score indicates potential activation.For terms where no activity pattern is available, this indicates that IPA® was not able to predict activity for a pathway.Representative images of control (C5d, C25, C41) and L-CMD (del.K32, R249W, L380S) myotubes stained with DAPI.On average, control myotubes had 16 nuclei per myotube, whilst L-CMD myotubes had a similar amount of nuclei, an average of 14 nuclei per myotube (ten myotubes were counted per cell line).Compared to the other L-CMD and control myotubes, however, the del.K32 L-CMD myotubes had a reduced number of nuclei per myotube.On average, the L-CMD myotubes contained 9 nuclei.(C) Abnormal L-CMD (L380S, del.K32, R249W) myoblast and myotube nuclei were characterised based on the defect type using lamin A/C immunostaining (yellow) and DAPI (blue).(a) Nuclear blebbing is defined as protrusions from the nucleus, sometimes these are seen to become detached from the nucleus itself, (b) honeycomb nuclei have multiple distinct holes that sometimes resemble a honeycomb, (c) nuclear shape abnormalities are nuclei that are shaped differently to control nuclei, elongated nuclei are an example of this, (d) donut shaped nuclei have a single hole through the centre, (e) correctly shaped control nuclei for comparison.(D) Graph depicting the percentage (%) of abnormal nuclei identified in L-CMD and control myoblasts.(E) Graph depicting the percentage (%) of abnormal nuclei identified in L-CMD and control myotubes.(A, C) In L-CMD myoblasts, lamin A appeared reduced compared to controls, although this finding was not statistically significant (p=0.244),whilst lamin C did not appear obviously reduced (p=0.805).(B) However, it was noticed that lamin A was of a lower molecular weight in L-CMD myoblasts compared to controls.(D, E) In L-CMD myotubes, lamin A and C were both reduced with statistical significance compared to controls (p=0.044,p=0.043).ns indicates not significant, Asterisk (*) denotes statistical significance.(F) Representative immunocytochemistry images showing lamin A/C (yellow) and DAPI (blue) staining in control myoblasts and myotubes (C5d, C25, C41).Lamin A/C was found to be correctly localized at the nuclear envelope across all control myoblasts and myotubes.(G) Representative immunocytochemistry images showing lamin A/C (yellow) and DAPI (blue) staining in L-CMD myoblasts and myotubes (L380S, del.K32, R249W).Lamin A/C was found in nucleoplasmic aggregates in del.K32 myoblasts, as indicated by white arrows.del.K32, L380S) myoblasts and myotubes.Emerin was found to be correctly localized at the nuclear envelope across all control myoblasts and myotubes, and all L-CMD myoblasts and myotubes except R249W L-CMD myoblasts where emerin was observed at the nuclear envelope and in the cytoplasm.(A) Emerin was not found to be significantly reduced in L-CMD myoblasts (L380S, R249W, del.K32) compared to age matched controls (C5d, C25, C41) (p=0.097).(C) Emerin was however significantly decreased L-CMD myotubes compared to controls (p=0.040).ns indicates not significant, Asterisk (*) denotes statistical significance.) and L-CMD (R249W, del.K32, L380S) myoblasts and myotubes.SUN2 was correctly localized at the nuclear envelope across all control myoblasts and myotubes, and all L-CMD myoblasts and myotubes.There did, however, appear to be reduced staining of SUN2 across L-CMD cells.(B) Expression SUN2 was not found to be reduced in L-CMD myoblasts compared to controls (p=0.401).(C) In myotubes, expression of SUN2 was significantly increased (p=0.049).ns indicates not significant, Asterisk (*) denotes statistical significance.

Declaration of interest
The authors hereby confirm that they have nothing to declare in respect of the following manuscript: were cultured in skeletal muscle cell growth medium (Cat No: C-23060; PromoCell GmbH, Heidelberg, Germany) containing supplement mix (Cat No: C-39365; PromoCell) with 10% Fetal Bovine Serum (Cat No: 10270; Gibco; Thermo Fisher Scientific, Paisley, UK) and Penicillin-Streptomycin (Cat No: 15140122; Gibco; Thermo Fisher Scientific).Differentiation was induced once myoblasts had reached approximately 80-90% confluency by washing adherent myoblasts in serum free medium and then culturing in DMEM (Cat No: 31966-021; Gibco; Thermo Fisher Scientific) supplemented with Insulin-Transferrin-Selenium-Ethanolamine (ITS-X) (Cat No: 51500-056; Gibco; Thermo Fisher Scientific) and Penicillin-Streptomycin (Cat No: 15140122; Gibco; Thermo Fisher Scientific) without the addition of serum.After a further four days of cell culture, approximately 80% of the cells had fused into myotubes.

Figure 1 -
Figure 1-Quantitative proteomics analysis revealed 348 dysregulated proteins in L-CMD myoblasts and myotubes compared to healthy controls.Using SWATH-MS analysis, dysregulated proteins

Figure 2 -
Figure 2-Differentially expressed proteins identified in L-CMD myoblasts compared to myotubes, and control myoblasts compared to myotubes.There was limited overlap between the differentially expressed proteins that were identified in L-CMD myoblasts versus L-CMD myotubes, compared to

Figure 3 -
Figure 3-L-CMD myoblasts exhibited increased proliferation rate and nuclear deformities.(A) Cell proliferation rate is increased in L-CMD myoblasts and is expressed as doubling time in myoblasts from L-CMD patients (L380S, R249W, del.K32) compared to controls (C5d, C25, C41).Individual values for each cell line are presented, as well as the average for each group, with error bars indicating standard deviation from the mean.Asterisk (*) indicates significance (p=0.023).(B) Representative images of control (C5d, C25, C41) and L-CMD (del.K32, R249W, L380S) myotubes stained with DAPI.On average, control myotubes had 16 nuclei per myotube, whilst L-CMD

Figure 4 -
Figure 4-Lamin A/C is significantly reduced in L-CMD myotubes compared to controls and nucleoplasmic aggregation was found in del.K32 myoblasts only.Lamin A/C expression was compared in control (C5d, C25, C41) and L-CMD (R249W, del.K32, L380S) myoblasts (A,B,C) and

Figure 6 -
Figure 6-SUN2 is significantly reduced in expression in L-CMD myotubes compared to controls, and is correctly localized to the nuclear envelope in all L-CMD cell lines.(A) Representative immunocytochemistry images showing SUN2 (magenta) and DAPI (blue) staining in control (C5d,