Quantitative Proteomic Analysis Reveals Metabolic Alterations, Calcium Dysregulation, and Increased Expression of Extracellular Matrix Proteins in Laminin (cid:2) 2 Chain– deficient Muscle*

Congenital muscular dystrophy with laminin (cid:2) 2 chain deficiency (MDC1A) is one of the most severe forms of muscular disease and is characterized by severe muscle weakness and delayed motor milestones. The genetic ba-sis of MDC1A is well known, yet the secondary mechanisms ultimately leading to muscle degeneration and subsequent connective tissue infiltration are not fully under-stood. In order to obtain new insights into the molecular mechanisms underlying MDC1A, we performed a comparative proteomic analysis of affected muscles (diaphragm and gastrocnemius) from laminin (cid:2) 2 chain–deficient dy 3K /dy 3K mice, using multidimensional protein identification technology combined with tandem mass tags. Out of the approxi-mately 700 identified proteins, 113 and 101 proteins, respec-tively, were differentially expressed in the diseased gastrocnemius and diaphragm muscles compared with normal muscles. A large portion of these proteins are involved in different metabolic processes, bind calcium, or are expressed in the extracellular matrix. Our findings suggest that metabolic alterations and calcium dysregulation could be novel mechanisms that underlie MDC1A and might be targets that should be explored for therapy. Also, detailed knowledge of the composition of fibrotic tissue, rich in extracellular matrix proteins, in laminin (cid:2) 2 chain–deficient muscle might help in the design of future anti-fibrotic treatments. All MS data have been deposited in the ProteomeXchange with identifier PXD000978

Congenital muscular dystrophy with laminin ␣2 chain deficiency, also known as MDC1A, 1 is a severe muscle wasting disease for which there is no cure. MDC1A is caused by mutations in the LAMA2 gene that lead to complete or partial deficiency of laminin ␣2 chain (1)(2)(3). Although the primary defect in MDC1A is known, the secondary molecular mechanisms eventually leading to muscle degeneration are not fully understood. In normal muscle, laminin ␣2 chain binds to the cell surface receptors dystroglycan and integrin ␣7␤1, which both indirectly bind the cytoskeleton (4 -7). Both of these adhesion complexes are important for normal skeletal muscle function, and laminin ␣2 chain binding to dystroglycan contributes to the maintenance of sarcolemmal integrity and protects muscles from damage (8), whereas laminin ␣2 chain binding to integrin ␣7␤1 promotes myofiber survival (9,10). In MDC1A, laminin ␣2 chain is absent or severely reduced, and the expression of dystroglycan and ␣7␤1 is also dysregulated in MDC1A (9,11,12). Thus, the structural link is broken, and the yet to be determined downstream intracellular signaling pathways are also interrupted. Consequently, laminin ␣2 chaindeficient muscle fibers undergo degeneration-regeneration cycles, but rather quickly regeneration fails and muscle fibers die by apoptosis/necrosis followed by a major replacement of muscle tissue with connective tissue (3,7). In order to unravel novel secondary molecular mechanisms, which could indicate new therapeutic targets, we decided to evaluate the protein expression profile in laminin ␣2 chain-deficient dy 3K /dy 3K muscle. Several proteomic profiling studies of dystrophindeficient muscles (Duchenne muscular dystrophy) have been performed (13)(14)(15)(16)(17)(18)(19)(20), as well as some with dysferlin-deficient muscles (Limb-girdle muscular dystrophy type 2B, Miyoshi myopathy) (21,22). They all showed a great number of proteins that were differentially expressed in different dystrophic muscles and at different ages (13)(14)(15)(16)(17)(18)(19)(20)(21)(22). However, proteomic analyses of laminin ␣2 chain-deficient muscle have not yet been performed. We here used multidimensional protein identification technology with tandem mass tags (TMT), a powerful bated at 55°C for 1 h. Five microliters of 375 mM iodoacetamide was added, and samples were incubated for 30 min while protected from light. After this, samples were precipitated overnight with six volumes of prechilled (Ϫ20°C) acetone. Finally, protein pellets were resuspended in 100 l of 100 mM triethyl ammonium bicarbonate. At this point trypsin was added, and proteolytic digestion was allowed overnight at 37°C. Label reagents were prepared at room temperature, and 41 l of each reagent were used to label 100 g of protein.
Besides our four samples to be analyzed, a standard one, composed of equal fractions of the four others, was labeled. This standard sample was our reference, used to define the relative protein amounts for each analyzed sample. Our labeling design allowed a label swap, in order to avoid possible bias due to technical errors (Table I). The label reaction proceeded for 1 h at room temperature, and subsequently reactions were quenched with 8 l of 5% hydroxylamine for 15 min, mixed, and stored at Ϫ80°C. Pooled TMT-labeled samples were fractionated by strong cation exchange (Applied Biosystems, Foster City, CA) using 500 l of elution buffer (KH 2 PO 4 , 25 mM acetonitrile, pH 2.9) containing increasing concentrations of KCl (30,60,90,120,240, 300, 420, and 500 mM KCl) and collected as fractions 1-8, respectively. The fractions were cleaned on Ultra Microspin C18 columns (The Nest Group, Southborough, MA), dried, and resuspended in 30 l of 0.1% formic acid.
LC-MS/MS Analysis on LTQ-OrbitrapXL-Samples were analyzed at the Gothenburg Proteomics Core, Gothenburg University, following the protocol described below. Desalted and dried fractions were reconstituted into 0.1% formic acid and analyzed on an LTQ-OrbitrapXL (Thermo Fisher Scientific) interfaced with an in-houseconstructed nano-LC column. Two-microliter sample injections were made with an HTC-PAL autosampler (CTC Analytics AG, Zwingen, Switzerland) connected to an Agilent 1200 binary pump (Agilent Technologies, Santa Clara, CA). The peptides were trapped on a precolumn (40 ϫ 0.075 mm inner diameter) and separated on a reversed phase column (200 ϫ 0.050 mm). Both columns were packed inhouse with 3-m Reprosil-Pur C18-AQ particles. The flow through the analytical column was reduced by a split to ϳ100 nl/min, and the gradient was as follows; 0 -6 min, 0.1% formic acid; 6 -76 min, 7% to 35% acetonitrile, 0.1% formic acid; 76 -79 min, 40% to 80% acetonitrile, 0.1% formic acid. LTQ-OrbitrapXL settings were as follows: spray voltage, 1.4 kV; 1 microscan for MS1 scans at 60,000 resolution (m/z 400); full MS mass range, m/z 400 -2000. The LTQ-OrbitrapXL was operated in a data-dependent mode with one MS1 Fourier transform MS scan of precursor ions followed by collision-induced dissociation and high-energy collision dissociation MS2 scans of the three most abundant doubly, triply, and quadruply protonated ions in each Fourier transform MS scan. The settings for the MS2 were as follows: 1 microscan for high-energy collision dissociation MS2 at 7500 resolution (at m/z 400), mass range of m/z 100 -2000 with a collision energy of 50%, and 1 microscan for collision-induced dissociation MS2 with a collision energy of 30%. Dynamic exclusion of a precursor selected for MS2 was used for 120 s after one repeat, allowing most of the co-eluting precursors to be selected for MS2. All samples were analyzed a second time as described above, and also a third time using an exclusion list of all m/z within a 3-min retention window, already passing the identification criteria within the TMT set (eight fractions) in the database search.
Database Search and TMT Quantification-MS raw data files from all eight strong cation exchange fractions per one TMT set and three MS runs were merged for relative quantification and identification using Proteome Discoverer version 1.3 (Thermo Fisher Scientific). The database search was performed with the Mascot search engine using the following criteria: Mus musculus in Swiss-Prot protein database from April 2012 (535,698 entries); MS peptide tolerance of 10 ppm; MS/MS tolerance of 0.5 Da; trypsin digestion allowing one missed  -terminal TMT6-plex label, lysine TMT6-plex label). The detected protein threshold in the software was set to a confidence using the 1% false discovery rate method, and identified proteins were grouped by shared sequence to minimize redundancy. For quantification, the ratios of TMT reporter ion intensities in MS/MS spectra (m/z 126.12, 127.13, 128.13, 129.14, 130.14) from raw datasets were used to calculate fold changes between samples via the relative ratio to the reference pool. Only peptides unique for a given protein were considered for relative quantitation, excluding those common to other isoforms or proteins of the same family. Only peptides with a score of Ͼ10 and below the Mascot significance threshold filter of p ϭ 0.05 were included. Single peptide identifications required a score equal to or above the Mascot identity threshold. Normalization of protein median was used, and the median of peptides was used for determining protein ratio; the resulting ratios were exported into Excel for manual data interpretation. The mass spectrometry proteomics data have been deposited to the Proteome-Xchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository (26) with the dataset identifier PXD000978. Please see the supplemental material for annotated spectra and Table I for the experimental setup. Data Analysis-To determine which proteins were differentially expressed, only proteins quantified in all three runs (see Table I for the experimental setup) were used; average values were compared, and Student's t test was performed to validate differences (p Ͻ 0.05) (supplemental Tables S1-S3). Still, compared proteins should present a variation coefficient of less than 25% and a protein ratio less than 1.25 or greater than 1.25. In order to avoid multiple hypothesis test error, q-values were estimated (27,28). Also, to ensure the reproducibility of our findings, the retrospective statistical power was calculated using the free statistical software package R, according to Ref. 29. Only results for ␤ Ն 0.8 were validated as true.
Western Blot Analysis-Muscles were collected and proteins were extracted as described above, but from a different animal cohort (n ϭ 4 for each genotype). Thirty micrograms of total protein were applied per well. SDS-PAGE was performed according to Laemmli's protocol using 12% pre-casted gels in a Mini Protean Tetracell Electrophoresis System (Bio-Rad), followed by electrophoretic transfer. PVDF membranes were blocked in 5% non-fat dry milk in TBS with 0.05% Tween-20 overnight at 4°C. Membranes were incubated with primary antibodies (rabbit anti-isocitrate dehydrogenase, 1:1000, NBP1-31599, Novus (Littleton, CO); mouse anti-SERCA1 ATPase, 1:4000, ab2819, Abcam; mouse anti-calsequestrin 1 and 2, 1:1000, ab3516, Abcam (Cambridge, UK); goat anti-annexin A1, 1:1000, NBP1-18842) for 1 h at room temperature and subsequently washed three times for 10 min in TBS with 5% Tween-20. Membranes were once again blocked in 5% non-fat dry milk for 1 h and then incubated with secondary antibodies (1 h at room temperature). After three 10-min washes with TBS with 5% Tween-20, membranes were incubated in ECL chemiluminiscence solution (Amersham Biosciences), exposed to Hyperfilm (Amersham Biosciences), and developed (Curix 60, AGFA, Mortsel, Belgium). We applied the Mann-Whitney U test to assess whether the difference in protein expression was statistically significant. Statistical significance was accepted for p Ͻ 0.05.

Shotgun Proteomic Analysis of Laminin ␣2 Chain-deficient
Gastrocnemius and Diaphragm-Two-dimensional polyacrylamide gel electrophoresis has mainly been used as the proteomic approach to analyze dystrophic muscle (13)(14)(15)(16)(17). To overcome several drawbacks associated with this technique, we decided to use a label-based proteomic approach combining TMT labels and the multidimensional protein identification technology method. In this kind of analysis, whole samples are digested with a protease prior to identification and quantification of proteins in the sample. In our case we used trypsin, and peptides were labeled with an isobaric tag (TMT) for quantification (24). Subsequently, samples were fractionated using biphasic capillary columns packed with strong cation exchange and reversed phase material to separate digested and TMT-labeled whole protein mixtures. Labeled peptides were thus separated based on their charge and hydrophobicity, and the identification and quantification of peptides in eluted fractions was achieved using a mass spectrometer (peptides were identified through database searching and quantified by evaluation of reporter ion intensities from labels) (31,32).
We analyzed gastrocnemius and diaphragm muscles from 4-week-old laminin ␣2 chain-deficient dy 3K /dy 3K and agematched wild-type littermates. At this age, dy 3K /dy 3K muscles display typical pathological hallmarks such as small or rounded fibers (which indicate de-/regeneration), variation in muscle fiber diameter, centralized nuclei (indicative of regeneration), and the presence of inflammatory cells or adipose and fibrotic tissue (25,33). With the multidimensional protein identification technology approach, we identified around 700 proteins in gastrocnemius and diaphragm muscles (supplemetal Table S4). Among these, 113 and 101 proteins, respectively, were found to be differentially altered in their levels by at least a factor of Ϯ1.25 in dy 3K /dy 3K gastrocnemius and diaphragm muscle relative to corresponding wild-type muscles. In laminin ␣2 chain-deficient gastrocnemius muscle, there were 82 down-regulated and 31 up-regulated proteins, and in diaphragm muscle, 77 proteins were downregulated and 24 were up-regulated. Around 62% of the down-regulated proteins and ϳ67% of the up-regulated proteins in the two muscles were identical (Tables II-V). Overall, a vast majority of the differentially expressed proteins were intracellular proteins, but also several extracellular and plasma membrane proteins were identified. We used the PANTHER classification system to group the proteins with differential expression into different biological processes. Among the down-regulated proteins were proteins involved in metabolic processes, generation of precursor metabolites and energy, cellular processes, system processes, transport processes, and developmental processes. Among the upregulated proteins were proteins involved in cellular processes, metabolic processes, cellular component organiza- tion, cell communication, developmental processes, and system processes (Figs. 1A-1D). Oxidoreductases, transferases, cytoskeletal proteins, and calcium-binding proteins were among the most underrepresented proteins in dy 3K /dy 3K muscles, whereas receptors, calcium-binding proteins, cytoskeletal proteins, and extracellular matrix proteins were among the most overrepresented proteins (Tables II-V). Altered Metabolism in Laminin ␣2 Chain-deficient Muscle-A clear majority of the down-regulated proteins belonged to the class of metabolic processes/generation of precursor metabolites and energy, such as glycolysis, fatty acid ␤-oxidation, acyl CoA metabolic processes, pyrimidine and purin base metabolic processes, tricarboxylic acid cycle, oxidative phosphorylation, and respiratory electron transport chain (Tables II and IV). Each of these pathways is complex, involving numerous proteins. For example, the following proteins were down-regulated in laminin ␣2 chain-deficient muscles: fructose-biphosphate aldolase A, 6-phosphofrucokinase, L-lactate dehydrogenase, and glucose-6-phosphate isomerase (glycolysis-related); short-chain specific acyl-CoA dehydrogenase, medium-chain specific acyl-CoA dehydrogenase, and 2,4-dienoyl-CoA reductase (involved in fatty acid ␤-oxidation and acyl coA metabolic processes); adenylate kinase isoenzyme 1, phosphoglutamase-1, and various subunits of ATP synthase (pyrimidine and purin base metabolic processes); isocitrate dehydrogenase, succinyl-CoA ligase, malate dehydrogenase, and citrate synthase (tricarboxylic acid cycle-related); and various subunits of cytochrome c oxidase, electron transfer flavoprotein, and various subunits of ATP synthase (involved in oxidative phosphorylation and respiratory electron transport chain). Also, in a previous gene expression profiling study we noted that several down-regulated genes were assigned to metabolic themes (34). We next performed Western blot analysis to validate the proteomic findings, and indeed isocitrate dehydrogenase was significantly down-regulated in laminin ␣2 chain-deficient muscle (Fig. 2).
Impaired Calcium Homeostasis in Laminin ␣2 Chaindeficient Muscle-An increase in cytosolic calcium concentration has been connected to the pathomechanism of muscular dystrophy (35), but it is not yet clear whether impaired calcium homeostasis is pathogenic in MDC1A. In some muscular dystrophies (e.g. Duchenne muscular dystrophy), the muscle cell membrane is easily damaged, and this could lead to an influx of calcium in the cell; alternatively, enhanced calcium entry may occur through a class of stretch-activated channels. The increased calcium leads to necrosis and/or apoptosis of the muscle fiber through distinct mechanisms (36 -38). Also, the expression levels of proteins involved in the regulation of intracellular calcium concentration could be altered in muscular dystrophy (39). In MDC1A, there is little evidence of sarcolemmal disruption (40,41), but we noted that parvalbumin, a protein that acts as a soluble intracellular calcium buffer (and is present in high concentrations in fast-contracting skeletal muscles), was significantly downregulated in laminin ␣2 chain-deficient muscles (Table II and IV, first protein in the lists). Also, SERCA1 (which resides in the sarcoplasmic reticulum and transfers calcium from cytosol to sarcoplasmic reticulum lumen) was downregulated in laminin ␣2 chain-deficient muscles, as was calsequestrin-1 (which also resides in sarcoplasmic reticulum and helps hold calcium in sarcoplasmic reticulum) (Tables II and IV; Fig. 2). The expression of calsequestrin-2, in contrast, was up-regulated, at least in gastrocnemius muscle (Table III; Fig. 2).

Inflammation and Composition of Fibrotic Tissue in Laminin ␣2
Chain-deficient Muscle-We detected increased levels of annexins A1, A2, and A5 in laminin ␣2 chain-deficient muscle (Tables III and V). The annexins are phospholipid-and calcium-binding proteins involved in numerous physiological processes (42). Several members of the annexin family (e.g. annexins A1, A2, and A5) are differentially expressed in normal muscle relative to Duchenne muscular dystrophy, and the three annexins are highly expressed in macrophages and T cells in dystrophin-deficient muscle (18,43). The annexins could therefore have a role in the inflammatory response in muscular dystrophy, and MDC1A is indeed characterized by early-onset transient inflammation (44).
Furthermore, several of the up-regulated proteins are extracellular matrix proteins. Previous gene expression profiling experiments have demonstrated that up-regulation of genes encoding extracellular matrix components is more pronounced in congenital muscular dystrophies (including MDC1A) than in Duchenne muscular dystrophy (45). Yet the composition and properties of fibrotic lesions in laminin ␣2 chain-deficient chain muscle have not been extensively studied. The expression of periostin, galectin-1, collagen VI ␣2 chain, biglycan, lumican, and thrombospondin-4 was significantly increased in laminin ␣2 chain-deficient skeletal muscles (Tables III and V). Immunostaining of selected proteins (periostin and galectin-1) was performed to validate the proteomics data. Periostin expression was dramatically increased in laminin ␣2 chain-deficient muscle (Fig. 3), and galectin-1 was moderately augmented (Fig. 3).

DISCUSSION
Knowledge of the differences between normal and laminin ␣2 chain-deficient muscles is crucial for designing future treatment of MDC1A. Although several different approaches to combat muscular dystrophy in laminin ␣2 chain-deficient mouse models have been tested, and although these approaches include correction of both primary and secondary manifestations (11, 12, 25, 30, 41, 46 -55), the clinical appliance is still years away. Thus, we performed a comparative proteomic analysis of dy 3K /dy 3K muscles, and we report an extensive catalogue of differentially expressed proteins in dystrophic muscle. A majority of the differentially expressed proteins are involved in metabolic processes (of which a majority take place in the mitochondria), calcium handling, or extracellular remodeling, making them valuable candidates for potential drug targets for MDC1A.
Similar metabolic crises and mitochondrial abnormalities have been reported in gene and/or proteome expression profiling of different types of muscular dystrophy, including Duchenne muscular dystrophy and sarcoglycanopathy (56,57), and biochemical studies unveiled limitations in glycolytic and tricarboxylic acid cycle pathways and defective regulation of energy metabolism in mdx mice (58,59). Also, metabolic defects in ␤-sarcoglycan-deficient mice have been reported (60). Thus, the metabolic crisis and mitochondrial abnormalities most likely represent secondary effects of laminin ␣2 chain deficiency and of muscular dystrophy in general, but we cannot entirely rule out that these are unique features of MDC1A. It will be important to investigate whether metabolic alterations are present before the dystrophic pathology becomes visible and/or in myoblasts that generally are not affected by disease. For example, energy metabolism was shown to be affected in mdx myoblasts (61). Despite prevalent metabolic alterations in many muscular dystrophies, it is interesting to note that only few attempts to restore energy metabolism have been made (62), and future efforts should therefore be considered.
We further showed that the expression of several calciumbinding proteins was altered in laminin ␣2 chain-deficient muscles, indicating their role as well as a role for calcium in disease pathology. In line with the hypothesis that calcium dysregulation could be a primary event in the onset of muscular dystrophy, several treatment approaches aimed at normalizing the calcium homeostasis in dystrophic muscle have been evaluated in mice, and in particular in the mdx mouse model for Duchenne muscular dystrophy (63)(64)(65). Notably, it has been demonstrated that overexpression of SERCA1 is beneficial in the mdx mouse, as well as in a mouse model for sarcoglycanopathy (65). However, whether SERCA1 overexpression attenuates laminin ␣2 chain-deficient muscular dystrophy remains to be demonstrated. Ablation of cyclophilin D encoding gene (which prevents mitochondrial calcium overload) reduced muscular dystrophy in mdx and sarcoglycandeficient mice, as well as in laminin ␣2 chain-deficient animals, but whether pharmacological inhibition of cyclophilin diminishes pathology in laminin ␣2 chain-deficient animals (like in mdx and sarcoglycan-deficient mice) is not known (37). Thus, further approaches aimed at modulating calcium regulation in laminin ␣2 chain-deficient muscles should be tested.
Finally, it was recently demonstrated that anti-fibrotic therapy is beneficial in mouse models for MDC1A. Losartan and an analog derivative inhibited TGF-␤ signaling and reduced fibrosis and inflammation in skeletal muscle of two different mouse models (54,55). Yet, TGF-␤1 has a very broad spectrum of action. Therefore, it becomes important to elucidate the composition of the inflammatory/fibrotic tissue in laminin ␣2 chain-deficient muscle in more detail, as such exhaustive information might be helpful for the design of more specific anti-fibrotic treatments. For example, the expression of periostin, which has been suggested to function upstream but also downstream of TGF-␤ (66), was significantly increased in laminin ␣2 chain-deficient muscles. Thus, it is tempting to speculate that the removal of periostin would reduce the pathogenesis of laminin ␣2 chain-deficient muscular dystrophy by decreasing fibrosis, just like the deletion of periostin significantly improved sarcoglycan-deficient muscle structure and function (67).
In summary, we have here mapped alterations in the proteomic signature of laminin ␣2 chain-deficient mouse muscle. Although we used a powerful tandem mass spectrometry methodology with TMT isobaric labels, it also has its weaknesses, including the problem of ratio distortion, and we acknowledge that the expression ratio changes might have been underestimated (68,69). Nevertheless, we identified differences in around 100 proteins, and a major part of these proteins are involved in metabolic, calcium, and extracellular matrix homeostasis, indicating that these processes could be targets for MDC1A therapy.
Acknowledgments-We gratefully acknowledge SCIBLU Proteomics Resource Centre at Lund University (especially Karin Hansson) and the Proteomics Core Facility at Gothenburg University (in particular Carina Sihlbom) for excellent help with protein analysis. We also acknowledge the PRIDE team (Attilla Csordas) for the deposition of our data to the ProteomeXchange Consortium. The MS data have been deposited to the ProteomeXchange Consortium (http:// proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD000978. To access the data, please visit http://tinyurl.com/ojhrmot with username reviewer64343@ebi.ac.uk and password P07lsY9k.