Abstract
Purpose:
Skeletal muscle constitutes 30–40% of total body mass and is now considered an endocrine organ, given its secretion of a variety proteins, metabolites, and cytokines. We have previously shown that the absence of dystrophin in skeletal muscle contributes to lethal systemic stress pathology in the mdx mouse model of Duchenne muscular dystrophy through a mechanism that remains to be identified. Here we searched for secreted protein signaling factors, or myokines, released from dystrophin-deficient skeletal muscle that influence the organism-wide integrated stress response. We performed skeletal muscle extracellular fluid extraction and discovery proteomics for wild-type, mdx, and transgenic mdx mice rescued by expression of a dystrophin construct, all analyzed under basal conditions and following brief scruff restraint stress that causes inactivity in mdx mice.
Major Findings:
Our analysis demonstrated that skeletal muscle dystrophinopathy is associated with increased expression of numerous proteins in both intact mdx skeletal muscle and extracellular fluid compared to healthy mice. Brief scruff restraint revealed protein candidates with differential abundance in mdx extracellular fluid. Specifically, altered follistatin-like 1 protein and adiponectin secretion in response to scruff stress was shown to be dependent on skeletal muscle dystrophinopathy. The diverse signaling roles of follistatin-like 1 in the cardiovascular, musculoskeletal, and nervous system implicate it as a particularly intriguing myokine candidate regulating the mdx stress response.
Conclusions:
Our current study informs on the skeletal muscle secretory profile in mdx following a stressful stimulus and provides new leads to elucidate the mechanism by which mdx skeletal muscle orchestrates inter-organ stress signaling.
Background
Duchenne muscular dystrophy (DMD) is a progressive fatal neuromuscular disease characterized by loss-of-function mutations in the DMD gene that lead to a loss of dystrophin protein. Primary defects associated with the loss of dystrophin in skeletal, cardiac, and smooth muscle include progressive muscle degeneration and weakness, which are associated with chronic inflammation, muscle necrosis and fibrosis, respiratory distress, and cardiomyopathy. 1 DMD patients also suffer from secondary effects of dystrophinopathy in muscle and non-muscle tissues, with disease comorbidities including neurocognitive and behavioral challenges,2–4 heightened stress sensitivity, 5 scoliosis, 1 and metabolic dysfunction.4,6,7 Importantly, we have shown that the mdx mouse model of DMD exhibits dramatic stress pathology, with mild scruff stress provoking tonic immobility and transient hypotension, and more severe social defeat stress triggering sustained immobility, hypotension, and lethality.8,9 The MTBD/mdx mouse line transgenically restores dystrophin exclusively in the skeletal muscle (SkM) and rescues mdx stress pathology along with hallmark defects in mdx SkM morphology and force generation capacity.8,10 The pathological consequences of SkM dystrophinopathy highlight the centrality of SkM, which constitutes 30–40% of total body mass and serves as a primary hub for protein metabolism, 11 in regulating organismal health. While we have previously sought to identify how SkM metabolism influences systemic stress responses in mdx mice, 9 our current study investigates the role of SkM-secreted cytokines, peptides, and proteins – known as myokines – in regulating stress behavior in the context of healthy and dystrophin-deficient SkM.
Beyond its role in force production and locomotion, SkM is also an important endocrine organ. 12 The SkM secretome, comprised of myokines, nucleotides, and myometabolites released primarily during muscle contraction, has been shown to mediate inter-organ signaling crosstalk and participates in various organism-wide cellular and immune processes12–15. Given the systemic pathological consequences of SkM dystrophin deficiency,8,10 it is of great interest to identify altered patterns of myokine secretion in mdx SkM that regulate inter-organ crosstalk and central stress circuitry, as well as to observe how transgenic restoration of dystrophin in MTBD/mdx mice restores myokine and stress circuitry defects to protect against stress pathology. However, there are several experimental limitations to studying myokines within the SkM secretome, including the dynamic range problem of plasma or serum proteomics that masks detection of low-abundance molecules, 16 analogous signal dilution issues stemming from bulk tissue analysis, 17 and the need to unequivocally identify the tissue origin of an analyte within a broader circulating pool of molecules secreted from various organs. While exosome and extracellular vesicle isolation methods that allow for myokine detection have grown in popularity and sophistication over the years, they remain limited by reproducibility issues and the possible exclusion of myokines released through non-canonical secretion pathways.18,19
Recent work by the Chouchani and Spiegelman laboratories established a novel and simplified approach to isolate inter-myocyte extracellular fluid (ECF) from SkM for the proteomic characterization of myokines whose low abundance and extracellular localization preclude detection in a crude whole-tissue extract.17,20 Combining a SkM ECF and bulk SkM tissue proteomics analysis using the same muscle samples holds the unique advantage of being able to identify proteins that are actively secreted from SkM in response to an intervention. In this study, we utilized an ECF isolation and isobaric labeling discovery proteomics strategy with the goal of capturing stress-regulatory myokines at their site of secretion from quadriceps SkM in wild-type (WT), mdx, and MTBD/mdx mice at baseline and after mild scruff exposure that provokes mdx stress pathology.8,9 Herein, we demonstrate that mdx mice exhibit distinct ECF and bulk quadriceps muscle (QM) proteomic landscapes compared with WT or MTBD/mdx mice at baseline and after scruff stress. We show that mdx proteomic shifts are largely returned to WT levels in MTBD/mdx skeletal muscle, confirming a rescue of proteomic dysfunction in dystrophin-replete MTBD/mdx SkM and ECF that parallels a correction in stress pathology in the MTBD/mdx mouse. We mined our proteomics data for proteins with differential expression after scruff stress in the mdx ECF compartment compared to WT that are rescued in MTBD/mdx mice and identified follistatin-like 1 (FSTL1) protein as one interesting lead that matched the desired characteristics of altered expression in the mdx ECF compartment corresponding to stress. Elevated FSTL1 has been implicated in pathological adaptations to chronic stress exposure as well as inflammatory diseases while also demonstrating protective and regenerative cardiovascular capacities,21–23 suggesting its potential relevance to mdx stress pathology.
Methods
Mice
The mice used in this study were adult males (3–4 months of age). Animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Minnesota. C57BL6/J and C57BL10/J mice used as wild-type controls and C57BL/10ScSn-Dmdmdx/J (mdx) mice were bred in-house or purchased from The Jackson Laboratory. The previously described MTBD/mdx transgenic mouse line 10 expressing a human skeletal actin promotor-driven full-length dystrophin/utrophin chimera with microtubule-binding spectrin repeats 20–24 of dystrophin replaced by non-binding repeats 18–22 of utrophin was generated in our lab and bred in-house. Mice were group housed following standard specific pathogen free (SPF) procedures with ad libitum access to food and water on a 12-h light/dark cycle.
Scruff-induced inactivity
The scruff stress assay was implemented following an established protocol.8,9 Briefly, mouse locomotor activity was assessed using the SuperFlex Open Field activity monitoring system and associated AccuScan Fusion software (Version 3.4) by Omnitech Electronics, Inc. Following 30 min activity monitoring, mice were scruffed by the nape of the neck using the thumb and forefinger and placing the tail between the 5th digit and palm of the hand. Immediately following 30 s scruff, mice were released into the same activity box and locomotor activity monitoring resumed for 30 min, after which mice were anesthetized with tribromoethanol (Avertin) and euthanized via cervical dislocation. Quadriceps muscle was quickly dissected after euthanasia and processed for ECF isolation and proteomics analysis.
Quadriceps tissue harvest and extracellular fluid isolation
We adapted a recent protocol from Mittenbuhler et al. 17 to isolate ECF from mouse quadriceps muscle. Quadriceps muscles from both hind limbs were dissected, transferred to the center of a 20μm nylon mesh (Millipore Sigma, NY2004700), then the mesh was folded into a conical shape and secured in a 1.5 mL centrifuge tube. The samples were centrifuged at 800xg for 10 min at 4°C, which produced 10–20μL ECF. Isolated ECF samples from the left and right hindlimb of the same mouse were pooled, snap frozen, and stored at −80°C. The remaining ECF-depleted quadriceps muscle tissue was snap frozen and stored at −80°C for downstream processing and mass spectrometry analysis.
Immunodepletion of extracellular fluid
Quadriceps ECF samples were immunodepleted to remove high-abundance plasma and serum proteins albumin and IgG using R&D Systems Proteome Purify 2 Mouse Serum Protein Immunodepletion Resin (R&D Systems, MIDR002020). Samples (10μL) were mixed with 1 mL of immunodepletion resin and rotated end-over-end at room temperature for 1hr, then resin was split between two SpinX filter tubes (R&D Systems, SPINX8160036) and centrifuged at 1,500xg for 2 min. Eluates were collected and protein concentration was determined using A280 absorbance measurements, then samples were concentrated using Vivaspin® 500 3 kDa centrifugal concentrator tubes (Sartorius, VS0192). Concentrated samples were snap frozen and kept at −80°C for downstream processing and mass spectrometry analysis.
Western blot analysis
Quad muscle was homogenized using mortar and pestle in liquid nitrogen. Crushed tissue was lysed with 1% sodium dodecyl sulfate in 1xPBS with added protease inhibitors (100 nM Aprotonin, 10 mg/mL E-64, 100μM Leupeptin, 1 mM PMSF, 1μg/mL Pepstatin, 0.79 mg/mL Benzamidine). Lysates were clarified by centrifugation, then protein content was assessed by A280 absorbance and samples were diluted to 2 mg/mL with PBS and β-mercaptoethanol for protein denaturation. Forty μg total protein was loaded on a 10% polyacrylamide gel for 20 min at 80 V, then 1hr at 120 V or until dye front ran off gel. Protein was transferred to a 0.45μm polyvinylidene fluoride membrane (Immobilon-FL IPFL00010, Millipore Sigma) overnight on ice at 20 V with methanol added to transfer buffer (25 mM Tris, 192 mM glycine, 10% methanol). Membranes were blocked with 5% bovine serum albumin (BSA) in TBST for 1 h before incubation with primary antibodies for 1.5 h at room temperature. Primary antibodies were diluted in 5% BSA in TBST: anti-FSTL1 (20182-1-AP, Proteintech) at a 1:5000 dilution and anti-α-tubulin (B512, Invitrogen) at a 1:1000 dilution. Secondary antibodies DyLight 680 and 800 (1:10,000, Cell Signaling) were incubated in the dark for 1 h at room temperature. Membranes were imaged and densitometry quantifications were made with a Licor Odyssey Infrared Imaging System and Image Studio Lite (5.x CLx/DLx) software.
ELISA analysis
A commercial FSTL1 ELISA kit (CSB-EL009025MO, CusaBio, Wuhan, China) was employed as an orthogonal approach to validate proteomics results. Quad muscle, ECF, and plasma were harvested from the same mice and used for ELISA analysis. Quad muscle and ECF were also used for WB and proteomics analyses. Aliquots of ECF-depleted quadriceps lysates processed for WB analysis were diluted 1:2 and used for ELISA quantification. Immunodepleted quadriceps ECF was diluted to 0.1 mg/mL and ELISA analysis was performed following the manufacturer instructions. Briefly, 100μL aliquots of diluted quadriceps lysates, quadriceps ECF, and plasma were incubated in antibody-coated wells for 2 h at 37°C along with a serial dilution of the provided FSTL1 standard to generate a standard curve. Technical duplicates were analyzed for quad lysates, while low sample amounts precluded technical replicate analysis for ECF and plasma. Biotin-antibody conjugation to antibody-immobilized samples was performed for 1 h at 37°C, followed by three washes and incubation with HRP-avidin for 1 h at 37°C. Samples were washed for five cycles, then TMB substrate was allowed to react with HRP for 15–30 min before quenching with stop solution. Colorimetric detection was performed with a SpectraMax iD3 Multi-Mode Microplate Reader (Avantor) at 450 nm within five minutes of adding stop solution. A standard curve was generated by plotting FSTL1 concentration on the x axis and blank-corrected absorbance units on the y axis, then a linear or polynomial regression fit was used for interpolation to determine analyte FSTL1 concentration.
Protein extraction, digestion, and peptide isobaric labeling
Frozen quadricep muscle tissue pieces (20–25 mg) were processed using a Percellys Cryolys Evolution bead beater (Bertin Technologies). Tissue samples were weighed in Percellys tissue homogenizing CKMix tubes (Bertin Technologies) and protein extraction buffer [7 M urea, 2 M thiourea, 0.4 M Tris pH 8.0, 20% (v/v) acetonitrile, 10 mM tris (2-carboxyethyl) phosphine (TCEP), 40 mM chloroacetamide, and 1μl/100ul buffer Pierce Universal Nuclease (Thermo Fisher Scientific)] was added at a ratio of 9ul lysis buffer per 1 mg tissue. A 150μl aliquot of each sample was transferred to a PCT tube with a 150μl cap for the Barocycler NEP2320 (Pressure Biosciences, Inc., South Easton, MA) and cycled between 35 kPSI for 20 s and 0 kPSI for 10 s for 60 cycles at 37°C. After barocycling, the samples were centrifuged at 15,000xg for 10 min. The samples were transferred to new 1.5 mL microfuge Eppendorf Protein LoBind tubes. Aliquots for each sample in the quadriceps muscle (QM) and ECF experiment were taken for protein concentration determination by Bradford assay.
A bridged pooled normalizing sample was made for ECF and muscle TMTpro 16plex (Tandem Mass Tag, Thermo Fisher Scientific, Waltham, MA) experiments. The pooled sample was composed of equal μg aliquots of each sample within the ECF or whole muscle experiment.
For the ECF TMTpro experiment, a 15μg aliquot of each sample and pooled sample was transferred to a new 1.5 mL Eppendorf Protein LoBind tube and brought to the same volume with PBS. A fourfold volume of extraction buffer was added to each sample and incubated at 37°C for 30 min. The extraction buffer volume was diluted fivefold with LC-MS grade water.
For the whole muscle TMTpro experiment, an 18μg aliquot of each sample and pooled sample was transferred to a new 1.5 mL Eppendorf Protein LoBind tube and brought to the same volume with extraction buffer. The samples were diluted fivefold with LC-MS grade water.
For all samples, trypsin (Promega, Madison, WI) was added in a 1:40 ratio of trypsin to total protein. Samples were incubated at 37°C overnight, then were acidified with 0.3% (v/v) formic acid. Samples were cleaned using a MCX Stage tip and eluates were vacuum dried. Samples were resuspended with 0.1 M triethylammonium bicarbonate, pH 8.5, to a final protein concentration of 1μg/μL.
For stable isotope labeling, a 14μg aliquot for each sample in the whole muscle TMTpro experiment and 12μg aliquot for each sample in the ECF TMTpro experiment was made and assigned a channel within a TMTpro 16plex. The samples were labeled with TMTpro 16plex isobaric label reagent in a 1:10 ratio of μg protein to μg TMTpro 16plex label according to the manufacturer's instructions. Isobaric tag-labeled samples within the same experimental screen were multiplexed together into a new 1.5 mL Eppendorf tube, then vacuum dried and cleaned with a 1 mL SepPak C18 solid phase extraction cartridge (Waters Corporation, Milford, MA). Each TMTpro 16plex sample was vacuum dried, resuspended in 20 mM ammonium formate, pH 10, 98% (v/v) water and 2% (v/v) acetonitrile and fractionated offline by high pH C18 reversed-phase chromatography as previously described. 24 After fractionation, concatenated peptide fractions were C18 Stage tipped 25 and eluates were dried in vacuo.
Mass spectrometry data acquisition
ECF and QM proteomics experiments were performed at the University of Minnesota in collaboration with the Center for Metabolomics and Proteomics (CMSP) departmental core facility. Six experimental groups of mice (3–4 months old) were tested, including WT (C57BL/10J), mdx (C57BL/10 ScSn-Dmd mdx /J), and MTBD/mdx mice sacrificed at baseline or 30 min following a 30 s scruff stress exposure (n = 5 mice/group). Thirty total samples were collected and two TMTpro 16plex screens were run for each of the two biological compartments (ECF and QM) for a total of four screens, with a pooled normalization control sample included in each screen (Figure 1A). All data were collected on a Thermo Orbitrap EclipseTM mass spectrometer coupled to a DionexTM UltimateTM 3000 RSLCnano LC pump. Peptides from 20% (3μL) of each concatenated set of fractions were separated using a 199-min gradient at 0.315–0.350μL/min on a C18-AQ ReproSil-Pur column measuring 300 mm with an internal diameter of 100μm, 1.9μm resin size, and 120 Å pore size (Dr Maisch GmbH Ammerbuch, Germany). Buffer A consisted of water with 0.1% (v/v) formic acid and Buffer B consisted of acetonitrile with 0.1% (v/v) formic acid. High-field asymmetric-waveform ion mobility spectroscopy (FAIMS) was enabled during experimental acquisition with the following compensation voltage (CV) settings: −45 V, −60 V, and −75 V. Voltage was kept at 2.1 kV for positive ion mode and the ion transfer tube temperature was set to 275°C. At the MS1 stage, the mass spectrometer scanned masses in the range of 400–1400 m/z at a resolution of 120 K with an AGC target of 4.0E5 over a 50 ms maximal injection time. At the MS2 stage, ions were fragmented by high-energy collisional dissociation (HCD) with a collision energy of 38% at a detector resolution of 50 K with an AGC target of 1.25E5 (250% relative to default) over a 150 ms maximal injection time, and the Fourier transform first mass mode was fixed at 110 m/z.
Proteomics peptide spectrum matching and quantification
Raw MS files were processed by CMSP in Proteome Discoverer v3.0 (Thermo Fisher Scientific, Rockford, IL, USA). Peptide identification was performed by searching HCD MS/MS files against the UniProtKB/Swiss-Prot mus musculus database (UP000000589; accessed August 18, 2023) merged with a common lab contaminant database (https://github.com/HaoGroup-ProtContLib) with the Sequest HT search engine and a 1% false discovery rate (FDR) was set for peptide-to-spectrum matches using the Percolator algorithm in Proteome Discoverer v3.0. The following parameters were used for spectral processing: MS1 tolerance of 20 ppm, MS2 tolerance of 0.08 Da, trypsin (full) digestion with a maximum of two missed cleavage, minimum peptide length of 6 and maximum peptide length of 50, with 10 maximum peptides reported. Cysteine carbamidomethylation was set as a static modification, while TMTpro lysine and N-terminal modifications, asparagine and glutamine deamidation, methionine oxidation, pyro-glutamic acid, N-terminal acetylation, methionine-loss, and methionine loss with acetylation were set as dynamic modifications in Sequest. Only protein identifications with high FDR confidence (FDR<1%) and containing 2 or more peptides were accepted. Reporter ion quantification was conducted using the TMTpro 16plex Lot-YD372049 quantification method with a peak integration tolerance of 20 ppm and the most confident centroid method. Unique and razor peptides were used for quantification. All peptides were used for normalization and protein roll-up, and scaling was performed for inter-screen data normalization using a pooled average control sample. Hypothesis testing was performed using ANOVA (individual proteins) for pairwise ratios.
MS bioinformatics analysis
Scaled protein abundances were used to calculate pairwise fold changes based on the geometric means of all biological replicates from each sample group. Fold changes were calculated for pairwise comparisons between the following groups: mdx baseline/WT baseline, mdx baseline/MTBD-mdx baseline, MTBD-mdx baseline/WT baseline, mdx post-scruff/WT post-scruff, mdx post-scruff/MTBD-mdx post-scruff, MTBD-mdx post-scruff/WT post-scruff, WT post-scruff/WT baseline, mdx post-scruff/mdx baseline, and MTBD-mdx post-scruff/MTBD-mdx baseline. A two-way unpaired Student's t-test was used to calculate p-values for pairwise fold changes, and the Benjamini Hochberg method was used to control the false discovery rate (FDR). Corrected p-values were log-transformed and plotted against log-transformed fold change values to obtain volcano plots generated in R using the tidyverse package, and a minimum corrected p-value cutoff of 0.05 and minimum relative fold change cutoff of ±1 was applied to identify differentially expressed proteins (DEPs) in pairwise comparisons. Comparisons between all experimental groups was performed using two-way ANOVA statistical analysis with Tukey's multiple comparisons correction. Full protein quantification datasets generated in Proteome Discoverer and lists of DEPs were imported to R for data filtering and visualization using the gplots, VennDiagram, and dplyr packages. Venn diagrams were used to obtain lists of overlapping and non-overlapping DEPs between distinct two-group comparisons. DEPs with missing values for two or more biological replicates in an experimental group were excluded from further analysis.
Due to differences in ECF and QM biological matrix effects that impact chromatographic behavior, peptide ionization efficiency, and overall analytical sensitivity during LC-MS data acquisition, 26 we did not design our experiments to quantitatively compare the ECF and QM proteomes. Instead, we qualitatively sorted the ECF and QM proteome datasets to subset ECF-specific and QM-specific proteomes and identify secreted protein candidates that are uniquely present in the SkM ECF.
Functional enrichment analysis was performed using GOrilla, 27 g:Profiler, 28 NaviGO, 29 and SubcellulaRVis for visualization of Gene Ontology Cellular Compartment enrichment. 30 For DEP gene ontology analysis, the target set included the DEP list and the background set included all quantified proteins in the fraction-specific (ECF or QM) proteomics screen. For gene ontology analysis of proteins uniquely represented in ECF or QM proteomes, the compartment-specific dataset was used as the target set, and the Swiss-Prot/UniprotKB 31 Mus musculus reference proteome (dataset downloaded January 15, 2024) was used as the background set. Protein-protein interaction network functional enrichment was performed using STRING (version 12.0). The Swiss-Prot/UniprotKB Mus musculus proteome was filtered for proteins annotated as “secreted” (dataset downloaded December 29, 2024) and was cross-referenced against our proteomics datasets to obtain a subset of candidate secreted proteins. PCA plots were generated in R using the ggfortify package. Pairwise student's t-tests, two-way ANOVA statistical analysis, GO enrichment visualizations, and heat maps for DEPs of interest were performed in GraphPad Prism, version 10.2.
Results
Quadriceps ECF proteome is enriched in extracellular and secreted proteins distinct from bulk quadriceps muscle tissue
To identify novel secreted proteins originating from dystrophin-deficient SkM that may contribute to stress pathology in the mdx mouse model of DMD, we employed an ECF isolation technique recently reported by Mittenbuhler et al. 17 Quadriceps SkM was harvested from WT, mdx, and MTBD/mdx mice expressing transgenic dystrophin in SkM at baseline or 30 min following a 30-s scruff exposure. Quadriceps ECF and ECF-depleted QM from the six experimental groups were subjected to TMTpro 16plex proteomic analysis (Figure 1A).
1768 proteins were quantified in the ECF proteomics screen and 2528 proteins quantified in the QM proteomics screen (Supplemental Data 1 and 2), with 1353 overlapping proteins between the ECF and QM compartments (Figure 1B), representing 76.5% of the ECF proteome and 53.5% of the QM proteome. The overlapping protein dataset, along with the remaining 415 (23.5%) proteins in the ECF-only proteome and 1175 (46.5%) QM-only proteins, were analyzed using gene ontology (GO) enrichment tools to validate the successful isolation of an ECF compartment with extracellular protein enrichment. Cellular compartment enrichment analysis and visualization 30 revealed that the QM only proteome was enriched for intracellular compartments, including cytosol, mitochondria, endoplasmic reticulum, ribosome, and peroxisome (Figure 1C and Table 1). A similar intracellular enrichment was observed in the subset of proteins identified in both ECF and QM matrices (Table 1). Conversely, the ECF only proteome was most enriched for extracellular followed by cytosolic proteins (Figure 1D and Table 1), suggesting that we successfully replicated previous efforts to capture a unique ECF proteome that is likely derived from the extracellular space between muscle cells.17,20 An orthogonal GO enrichment analysis using the g:GOst enrichment tool in g:Profiler 28 revealed a similar pattern of extracellular protein enrichment in the ECF only proteome (Figure 1E). GO enrichment comparison between compartments demonstrated enrichment of proteins associated with the extracellular space (GO:0005615) exclusively in the ECF only and overlapping proteome subsets and greater enrichment of cytoskeletal proteins (GO:0005856) in the overlapping and QM only proteomes than in the ECF only proteome (Figure 1F). Moreover, biological process GO terms associated with skeletal muscle were significantly overrepresented in the QM only and overlapping proteome subsets but not in the ECF only proteome (Figure 1G). We provided further confirmation of successful ECF compartment isolation by cross-comparing our compartment-specific and overlapping protein datasets with the same test set of common intracellular markers proteins used previously to validate ECF proteomics result. 17 None of the 28 intracellular proteins were found in the ECF-only proteome, while 19/28 intracellular proteins were detected in the overlapping ECF/QM proteome and 7/28 were found in the QM-only proteome (Figure 1H). Our results indicate that exclusion of proteins that overlap between the ECF and QM proteomes is effective in eliminating intracellular proteins and enriching the ECF-only dataset for extracellular and SkM-secreted proteins.
Subcellular compartment enrichment in ECF and QM proteomic screens from SubcellulaRVis.
Cellular compartment enrichment gene ontology analysis was performed using SubcellulaRVis 30 for proteins in the ECF and QM proteomic screens. A standard enrichment test was performed using the hypergeometric probability function, and p-values were adjusted using the false discovery rate correction. 30 P-values < 0.05 are bolded and represent significantly enriched pathways in the compartment corresponding only to the ECF fraction (ECF only proteome), only to the quadriceps tissue compartment (QM only proteome), or the overlapping ECF and QM compartment (ECF/QM overlapping proteome).
Dystrophin-deficient bulk SkM and ECF display many upregulated proteins and few downregulated proteins compared to wt or MTBD/mdx mice
Next, we compared protein expression differences between the three genotypes (WT, mdx, and MTBD/mdx) at baseline to examine dystrophin-dependent proteomic alterations in the quadriceps ECF and QM compartments. Volcano plots were generated to visualize differentially expressed proteins (DEPs) within two-group comparisons (Figure S1 and Figure 2). Compared to the WT ECF proteome at baseline, mdx ECF displayed 230 significantly upregulated proteins and 37 downregulated proteins (Figure S1A). In the QM compartment, there were 293 significantly upregulated proteins and 14 significantly downregulated proteins in mdx compared to WT quadriceps muscle (Figure S1B). Venn diagrams demonstrate similar basal proteomic alterations in mdx skeletal muscle compared with WT or MTBD/mdx, regardless of biological compartment (Figure S1C, D). Volcano plot comparisons between mdx and MTBD/mdx demonstrate 341 upregulated/23 downregulated mdx ECF proteins and 329 upregulated/12 downregulated mdx QM proteins (Figure S1E, F). One significantly upregulated protein and eight significantly downregulated proteins were observed between MTBD/mdx and WT ECF basal proteomes, and no significant differences between MTBD/mdx and WT were observed in the QM compartment (Figure S1G, H), indicating that transgenic expression of the full-length dystrophin/utrophin chimera in skeletal muscle efficiently restores a WT pattern of SkM protein expression and secretion.
Notably, cross-validation of our QM proteomics results against a published mdx global proteomics study with similar methodology 32 showed corresponding patterns of differential protein expression between WT and mdx skeletal muscle (Table 2 and Supplemental Data 3). Thus, the mdx QM proteome in our study exhibits a DMD proteomic signature that is consistent with previous studies,32,33 but is distinct from the ECF proteome, which yields its own unique pattern of mdx proteomic alterations compared with dystrophin-replete controls that may be used to identify novel DMD biomarkers.
Top mdx vs WT differentially expressed proteins (DEPs) in two independent studies.
Gene names of top basal differentially expressed proteins (DEPs) between mdx, WT, and MTBD/mdx SkM in the current study (Johnson et al.) compared with published proteomics data from Day et al. (Supplemental Data 1 in Day et al., 2022). 32 Gene names for downregulated proteins are indicated with a downward-facing arrow (↓), while gene names for upregulated proteins are indicated with an upward-facing arrow (↑). Bolded gene names represent proteins that are altered in common between studies. DEPs across both studies were found to be statistically significant with an adjusted p-value <0.05.
Our proteomics analysis also included WT, mdx, and MTBD/mdx ECF and QM samples collected 30 min after a 30 s exposure to scruff stress (Figure 1A; S2A) to identify myokines that may function as mediators of mdx scruff-induced phenotypes. Similar to the mdx proteomic landscape at baseline, an abundance of significantly upregulated proteins and relatively fewer downregulated proteins were observed between mdx and WT (213 upregulated/66 downregulated) or mdx and MTBD/mdx (194 upregulated/62 downregulated) after scruff exposure in the ECF compartment (Figure 2A, C). Principal component analysis (PCA) supports these findings and shows that 41.5% of variability (PC1) in the ECF proteomics dataset is attributable to the altered mdx proteome, while an additional 12.7% of variability (PC2) may be attributed to differences between the baseline and post-scruff mdx proteome (Figure S3A).
Corresponding patterns were observed in post-scruff QM pairwise comparisons, with 139 upregulated/12 downregulated proteins in mdx compared to WT quadriceps muscle (Figure 2B) and 142 upregulated/7 downregulated proteins in mdx compared to MTBD/mdx quadriceps muscle (Figure 2D). In contrast, the WT and MTBD/mdx proteomes closely resembled one another (Figure 2E, F). Meanwhile, there were no significant proteomic alterations between the baseline and post-scruff conditions in any genotype (Figure 2G, H; Figure S2B-E). The QM PCA plot demonstrates a subtle effect of scruff stress on the mdx QM proteome (Figure S3B), despite the failure of any protein to reach a significant fold-change value (adjusted FDR<0.05) between the mdx baseline and post-scruff proteomes (Figure 2H). While proteomic alterations between healthy and dystrophin-deficient SkM and ECF are similar under basal and post-scruff conditions, key differences were highlighted by PCA analysis that suggests the regulation of specific SkM proteins by stress exposure. These alterations are explored throughout the remainder of the manuscript.
Compensatory changes in the MTBD/mdx qm proteome underscore high similarity with wt proteomic status and suggest altered regulation of mRNA processing
The transgenic MTBD/mdx mouse model expresses a full-length dystrophin/utrophin chimera with human skeletal actin (HSA) promoter-driven expression in SkM and dystrophin spectrin-like repeats (SLR) 20–24 in the microtubule (MT) binding domain replaced by utrophin SLR 18–22. 10 Despite this discrepancy between the WT and MTBD/mdx dystrophin sequences, only three DEPs were identified between the post-scruff WT and MTBD/mdx QM proteomes: peroxiredoxin-2 (PrxII), dipeptidyl peptidase 3, and utrophin (Figure S4A-C). PrxII expression was found to be significantly reduced in mdx QM compared to WT in accordance with previous data, 34 but was not restored in MTBD/mdx QM (Figure S4A). Dystrophin expression in MTBD/mdx QM was fully rescued compared to mdx QM and exhibited a non-significant elevated trend compared to WT QM (Figure S4D). Our group previously reported elevated dystrophin expression and unchanged utrophin expression compared to WT, as determined by western blot (WB). 10 Based on this discrepancy between our proteomic and previous WB data, we examined the peptide spectrum matches (PSMs) used for dystrophin and utrophin quantification in our proteomics analysis. We identified 14 PSMs within utrophin SLR 18–22, all of which displayed elevated abundance in MTBD/mdx QM compared to WT or mdx. Conversely, PSMs identified outside of the SLR 18–22 region in utrophin demonstrated similar or reduced abundance in MTBD/mdx compared to WT and mdx. Meanwhile, dystrophin PSM inspection revealed a depletion in MTBD/mdx PSM abundance (eight PSMs) within the dystrophin SLR 20–24 region, confirming that MTBD/mdx SkM expresses a transgenic dystrophin/utrophin chimera and that apparent elevated dystrophin and utrophin levels in MTBD/mdx QM are an artefact of omitting the MTBD/mdx transgene from the search database during initial data processing.
To confirm the restoration of near-WT protein expression levels in MTBD/mdx bulk muscle tissue, we used two-group comparison Venn diagram overlaps to identify baseline QM mdx DEPs with altered expression between mdx and WT but not between mdx and MTBD/mdx (intermediate MTBD/mdx restoration) and DEPs with altered expression between mdx and MTBD/mdx but not between mdx and WT (compensatory MTBD/mdx expression) (Figure S5). See Table 3 for a summary of inter-group DEP comparison terminology. Proteins with intermediate MTBD/mdx restoration are those whose levels are altered in mdx muscle and incompletely restored in MTBD/mdx muscle despite transgenic dystrophin expression. These intermediate DEPs are thus unlikely to be mediators of protection against mdx stress susceptibility in MTBD/mdx mice. DEP lists were filtered before heatmap analysis to exclude proteins that were not identified among all replicates within each sample group. Subtle differences were observed between WT and MTBD/mdx groups; however, most proteins altered in mdx QM were restored by dystrophin expression in MTBD/mdx muscle (Figure S5B, S5C). Proteins with altered MTBD/mdx expression compared to WT or mdx muscle (compensatory proteins) are associated with cellular processes including mRNA processing and striated muscle development (Figure S5D). Interestingly, a small subset of five proteins (bolded gene names) demonstrated similar abundances between WT and mdx baseline groups and reduced expression in the baseline MTBD/mdx QM compartment, including heat shock protein beta-1 (Hspb1; Figure S5E). Thus, our data confirm the near-complete rescue of the MTBD/mdx proteome to WT and highlight a handful of intriguing proteomic alterations between WT and MTBD/mdx that may regulate mRNA processing, MT network structure, and muscle cell integrity.
Inter-group DEP comparison terminology.
Summary table for relative expression levels of intermediate, rescued, and compensatory protein expression levels in MTBD/mdx extracellular fluid or quadriceps muscle.
Post-scruff ECF mdx DEPs reveal stress-dependent secreted protein alterations and suggest compensatory redox protein expression in MTBD/mdx SkM
A similar analysis was conducted to analyze pools of intermediate and compensatory expression changes in the MTBD/mdx ECF proteome compared to WT and mdx ECF following scruff exposure (Figure 3A). This analysis did not reveal major differences between the WT and MTBD/mdx post-scruff ECF proteomes (Figure 3B, C), but did draw out secreted protein candidates (bolded gene names in Figure 3B, C) that were reduced in mdx post-scruff ECF and may be related to mdx scruff pathology, including adiponectin (Adipoq; Figure 3D, E), calreticulin (Calr), and myeloid-derived growth factor (Mydgf). Additionally, comparison of mdx and healthy SkM ECF revealed diminished levels of antioxidant defense enzymes including glutathione S-transferase 1 (Gstm1), glutathione S-transferase 2 (Gstm2), and glutaredoxin (Glrx) in mdx ECF that were rescued back to WT levels in MTBD/mdx ECF (Figure 3C, F). Notably, glutathione S-transferase 1 (Figure 3G) abundance in the QM compartment was significantly elevated in MTBD/mdx SkM above WT and mdx levels, perhaps as an aspect of improved antioxidant function with restored SkM dystrophin expression.32,34 Altogether, these data do not suggest substantial differences between the WT and MTBD/mdx ECF proteome, but do reveal secreted and redox-regulatory protein factors that should be further investigated as contributors to mdx stress pathology and physiological stress susceptibility.
Compartment-specific proteomic enrichment analysis reveals stress-dependent myokine candidates in mdx SkM ECF
We previously showed that MTBD/mdx mice exhibit WT-level protection against scruff-induced hypotension and inactivity.8,9 Therefore, secreted myokine candidates that modulate scruff susceptibility in mdx mice are expected to be restored to WT-level expression in MTBD/mdx mice. Our experimental design allowed us to isolate the ECF fraction with greater enrichment in secreted proteins than intact muscle tissue. To identify secreted protein candidates with differential expression between WT, mdx, and MTBD/mdx, we filtered lists of DEPs by biological compartment: mdx DEPs in the ECF proteomics dataset also found in the QM dataset were excluded to obtain an ECF only dataset, and vice versa to obtain a QM only dataset. This filtering step produced an enrichment in extracellular and secreted proteins among mdx DEPs in the ECF compartment at baseline (Figure 4A) and after scruff exposure (Figure 4B). Filtering by the QM only proteome resulted in no significant enrichment of any compartment among mdx QM DEPs. Venn diagrams in Figure 5A and 5C were used to identify mdx baseline and post-scruff DEPs. We further filtered our DEP lists to exclude proteins that were not identified among all replicates within each sample group to obtain a robust set of secreted protein candidates relevant to mdx scruff pathology. After filtering, there were 12 significantly altered proteins in mdx ECF compared with WT or MTBD/mdx at baseline, while scruff exposure altered expression of five additional proteins in mdx ECF (Figure 5B). In the QM compartment, 25 DEPs were identified in mdx SkM compared to WT or MTBD/mdx at baseline, and one DEP was identified in mdx SkM compared to WT or MTBD/mdx after scruff stress (Figure 5D).
We overlaid our DEP datasets with the reviewed Swiss-Prot dataset containing secreted protein annotations (UniProtKB search term: cc_scl_term:SL-0243; downloaded January 2024) to identify secreted SkM DEP candidates, which are represented by the bolded terms in the ECF and QM DEP heatmaps (Figure 5B, D) and listed in Table 4. Three secreted proteins were identified among the DEPs in mdx QM at baseline (Figure 6A), while there were no DEPs annotated as secreted in the post-scruff mdx QM compartment. Eight secreted proteins were identified among the DEPs in mdx ECF at baseline (Figure 6B), a subset of which were also elevated in mdx post-scruff ECF to varying extents. A singular protein, follistatin-like 1 protein (Fstl1; FSTL1) was returned as a secreted protein that was detected solely in the ECF proteome with significant differential expression in mdx ECF after scruff exposure in two-group comparisons against WT or MTBD/mdx ECF (Figure 6C; Table 4). Two-way ANOVA statistical comparison of all six experimental groups indicated significantly elevated FSTL1 protein abundance in mdx ECF at baseline and post-scruff, with a non-significant trend (interaction term p-value = 0.06) toward further elevation in mdx post-scruff ECF (Figure 6C). Predictive linear regression analysis revealed a significant negative correlation between post-scruff activity levels and FSTL1 abundance in ECF (Figure 6D; R = -0.746; R2 = 0.558; p = 0.0014), suggesting that elevated FSTL1 may be a compelling predictive biomarker of stress susceptibility in mdx mice or that stress exposure leads to elevated FSLT1 secretion in SkM. GO enrichment analysis was performed for the small suite of secreted proteins identified as basally elevated in mdx ECF compared to WT or MTBD/mdx mice, which revealed significant enrichment in extracellular compartment localization, immune response, complement activation, protein activation cascade, and extracellular structure organization (Figure 6E).
Putative myokines with altered levels in the mdx skeletal muscle secretome.
Listed proteins have been annotated as secreted in the UniProtKB database and are differentially expressed in mdx ECF or QM compartments compared to WT or MTBD/mdx at baseline or after scruff stress exposure, suggesting potential altered myokine profiles in the mdx SkM secretome.
FSTL1 expression is elevated in mdx circulation and reduced in mdx ECF-depleted quadriceps muscle
Stress-induced alterations in mdx FSTL1 expression in quadriceps muscle and plasma were further validated with orthogonal techniques, including real-time quantitative PCR, WB, and enzyme-linked immunosorbent assay (ELISA). We demonstrated that Fstl1 mRNA levels were not significantly increased in mdx quadriceps muscle (Figure 7A; main effect p-value = 0.05), corresponding to similar FSTL1 protein levels between WT, mdx, and MTBD/mdx quadriceps muscle at baseline and after scruff exposure, as determined by WB (Figure 7B, C). Immunoblotting for FSTL1 protein revealed multiple bands at the expected FSTL1 molecular weight (35 kDa), as well as several bands between 37 and 50 kDa (Figure 7B, C). Three N-glycosylation sites and two O-glycosylation sites have been identified for FSTL1. 22 Published FSTL1 WB data in mouse adipose tissue and cell culture lysates reports a glycosylated FSTL1 band at 55 kDa, non-glycosylated FSTL1 band at 40 kDa, and a hypo-glycosylated FSTL1 band between 40 and 55kDa. 35 Therefore, multiple putative glycosylated FSTL1 bands were detected in our analysis between 37 and 50 kDa and were grouped together for comparison with non-glycosylated FSTL1 at 35 kDa. The FSTL1 glycosylation status was not significantly different between groups at baseline or after scruff exposure (Figure 7B, C).
In contrast, FSTL1 quantification with a commercial ELISA kit (CusaBio, Wuhan, China) demonstrated reduced FSTL1 levels in mdx quadriceps muscle at baseline and after scruff exposure, with a mild post-scruff reduction in SkM FSTL1 levels across groups (Figure 7D). ELISA quantification of FSTL1 in quadriceps-derived ECF samples analyzed in the ECF proteomics screen also demonstrated reduced mdx FSTL1 levels at baseline and post-scruff (Figure 7E). Circulating FSTL1 levels in mdx plasma were determined to be elevated by ELISA analysis, with a more dramatic elevation at baseline than after scruff exposure (Figure 7F). These immunoassay-based results contrast with one another and with the elevation in mdx FSTL1 levels reported by SkM ECF proteomics analysis (Figure 6C). The apparent discrepancy between FSTL1 detection methods may be related to the variable post-translational glycosylation modifications on FSTL1 and a higher avidity for certain glycosylated FSTL1 states over others. A strength of our discovery proteomics approach is the aversion of bias associated with antibody binding and interference effects, as well as the ability to quantify proteins indiscriminate of glycosylation status or other post-translational modification events.
Overlapping proteome analysis demonstrates secretion of contractile SkM proteins from mdx quadriceps muscle upon stress exposure
In addition to secreted proteins found exclusively in the extracellular compartment, our SkM ECF protein list contained a considerable number of proteins that are also found in intracellular compartments. 17 Therefore, we extended our ECF proteomics data analysis to include proteins with overlapping identification in the ECF and QM proteomes (Figures 8, S6). Venn diagram overlaps for two-group comparison DEP lists reveal 42 proteins with differential expression between mdx and WT or MTBD/mdx SkM ECF at baseline and 35 proteins identified uniquely in a post-scruff condition (Figure 8A). None of the post-scruff ECF DEPs were annotated as secreted in the UniProt database. Subcellular compartment enrichment analysis revealed cytosolic and mitochondrial enrichment in post-scruff ECF DEPs (Figure 8B), while baseline DEPs were enriched for the cytosol, extracellular space, and intracellular vesicles (Figure 8C). Further filtering was performed to retain 25 post-scruff ECF DEPs identified in all samples, whose abundances in the ECF and QM compartments are displayed by heatmap in Figure 8D. Of these, ten proteins were identified with unique elevation in mdx ECF after scruff exposure (Figure 8D).
Intriguingly, post-scruff protein expression changes in mdx ECF do not appear to be mirrored in mdx bulk SkM tissue, as demonstrated by the subtle pattern of expression changes for proteins in the QM compartment that were dramatically elevated or reduced in the ECF compartment after scruff exposure (Figure 8D). In contrast, proteins uniquely altered in mdx QM at baseline displayed a similar pattern of elevation after scruff (Figures S6A, B), and proteins with altered basal expression in mdx ECF were similarly altered after scruff and in the QM compartment (Figure S6C). These expression patterns suggest a unique subset of proteins with dramatically increased abundance in the mdx ECF compartment after scruff stress exposure that may be actively secreted from SkM to regulate the mdx central stress response. DEPs with significantly elevated post-scruff abundance exclusively in mdx ECF included myosin-binding protein C, fast-type (Mybpc2; fast MyBP-C), junctophilin-1 (Jph1; JP-1), junctophilin-2 (Jph2; JP-2), myomesin 2 (Myom2), and tropomyosin alpha-1 chain (Tpm1) (Figure 8D). GO enrichment analysis revealed that these post-scruff ECF DEPs are found in contractile fibers and junctional membrane complexes and serve as structural components of the cytoskeleton (Figure 1E). JP-1 expression in ECF and QM compartments is representative of the trend in elevated post-scruff ECF DEPs, with significantly increased JP-1 abundance in mdx post-scruff ECF and no differences in JP-1 abundance between any groups in QM (Figure 9A). Similar to FSTL1, JP-1 abundance in the ECF compartment significantly negatively correlates with post-scruff inactivity (Figure 9B; R = -0.736; R2 = 0.542; p = 0.0018), indicating that JP-1 may serve as an additional biomarker for mdx stress vulnerability. Our data introduce the possibility that contractile proteins are secreted from mdx SkM during scruff exposure, a mild form of stress that induces mdx hypoactivity.
Discussion
Our ECF proteomics investigation revealed distinct patterns of protein secretion from mdx SkM compared to WT and dystrophin-replete MTBD/mdx SkM, along with a subset of stress-dependent proteomic alterations that are most prominent in mdx SkM ECF. Some of the ECF mdx DEPs identified at baseline are associated with inflammation and intracellular calcium regulation, several of which have been previously identified as differentially expressed in mdx SkM, including protein S100-A4, 33 fibronectin,33,36 complement C1q subcomponent C, 37 and C4b-binding protein. 38 Mannose-binding protein A, complement C1q subcomponent subunit C, and complement factor I are all part of the complement system also involved in altered immune responses within dystrophin-deficient muscle. 39 Apolipoprotein M expression has not been reported as altered in DMD. However, fatty acid binding protein 3 levels are reduced in DMD SkM compared to Becker muscular dystrophy or healthy SkM, while apolipoprotein A-I levels are increased, 40 suggesting dysregulated lipid metabolism in mdx/DMD SkM and circulation.41,42 Our finding that fibronectin is over-secreted from mdx SkM agrees with an earlier characterization of the SkM secretome in cultured mdx myotubes 43 and extends these previous findings to include a novel panel of mdx myokines identified ex vivo in SkM tissue. FSTL1 was not detected in the quadriceps QM compartment, but was elevated in mdx ECF at baseline and after scruff exposure, with a trend toward further elevation in the post-scruff condition, and exhibited restored WT expression in MTBD/mdx ECF. Although volcano plot analysis did not highlight a significant elevation in FSTL1 between mdx post-scruff and baseline conditions, the fold-change thresholds and multiple hypothesis testing corrections applied to the dataset may have been too stringent to identify patterns that emerge from more focused investigation of smaller data subsets of high experimental relevance. Interestingly, FSTL1 levels moderately correlated with the magnitude of post-scruff inactivity in WT, mdx, and MTBD/mdx mice (R2 = 0.558), suggesting that FSTL1 may be a useful stress biomarker and possible signaling factor in the stress response pathway that modulates mdx scruff-induced inactivity.
FSTL1 is a glycoprotein in the secreted protein acid and rich in cysteine (SPARC) family and is an established myokine that has been implicated in exercise response,44–46 endothelial cell proliferation, 47 BMP and TGFβ signaling, cardiovascular disease, and a number of other disease mechanisms.22,48,49 FSTL1 exhibits proliferative and regenerative capacities, as well as pro-inflammatory and pro-tumorigenic functions depending on disease state, cell type-specific regulation of expression, post-translational glycosylation, and activation of diverse signaling pathways.21,22 Previously, Zhou et al. observed elevated Fstl1 mRNA levels in mdx mice that were further increased in DKO mice lacking both dystrophin and utrophin. 50 Elevated FSTL1 levels correlate with chronic unpredictable mild stress exposure in mice and genetic Fstl1 knockdown reversed depression- and anxiety-like behaviors as well as defects in synaptic plasticity in stress-exposed mice through microglial TLR4/MyD88/NF-κB signaling. 23 FSTL1 has also been shown to be upregulated as a pro-inflammatory signaling molecule in chronic immune disease conditions such as arthritis and Kawasaki disease.22,51 Pro-fibrotic signaling functions for FSTL1 52 may be particularly relevant to the elevated fibrotic SkM phenotype of mdx mice that are exposed to chronic daily scruff stress. 53 Thus, FSTL1 is actively secreted by skeletal muscle under numerous physiological states, promotes inter-organ crosstalk, particularly within the cardiovascular and central nervous system, and is known to be upregulated in chronic inflammatory conditions including DMD. Our results point to an exciting myokine candidate that may be implicated in the progression of dystrophinopathy and suggest that monitoring protein post-translational modifications may be an avenue for future investigation in probing dynamic changes that provoke the mdx stress response.
An unexpected result of our ECF proteomics investigation was the elevated abundance of several contractile proteins in mdx SkM ECF following scruff stress. Scruff handling represents a relatively mild physiological perturbation that is unlikely to cause mechanical injury to the animal's hindlimb muscle compartment, and the post-scruff hypoactivity observed in mdx mice suggests there is limited SkM contractile activity following scruff stress. Corroborating the lack of contractile injury is data demonstrating that SkM-type creatine kinase (CK-MM) levels were reduced, rather than elevated, in mdx bulk SkM or ECF, independent of scruff stress (main effect p-value<0.0001). Nonetheless, contractile fiber-associated proteins demonstrated elevated mdx post-scruff ECF abundance, including JP-1 (interaction p-value<0.0001), JP-2 (interaction p-value<0.0001), myomesin-1 (interaction p-value=0.0002), myomesin-2 (interaction p-value<0.0001), and tropomyosin alpha-1 chain (interaction p-value=0.002). Meanwhile, JP-1, JP-2, and tropomyosin alpha-1 chain did not show altered abundance in bulk SkM between genotype or treatment condition, while myomesin-1 and myomesin-2 each demonstrated a mild but significant reduction in mdx bulk SkM, independent of scruff (genotype main effect p-value<0.05).
Hallmarks of mdx pathology, including muscle force production deficits, basal MT lattice organization, post-exercise inactivity, and stress-induced pathology, are rescued in the MTBD/mdx model, in spite of the loss of a functional dystrophin MT binding domain.8,10 Despite a near-complete rescue of mdx pathology, MTBD/mdx mice do mimic the susceptibility of transgenic mdx mice expressing miniaturized dystrophin constructs to transverse MT loss following eccentric contraction in SkM, 54 a finding that is not readily explained by MT-related protein expression changes between MTBD/mdx SkM and transgenic mdx lines that do rescue ECC-induced MT disorganization. 54 As a corollary to our investigation of the mdx SkM secretome, we also sought to identify proteins in the basal QM proteome with compensatory or intermediate expression in MTBD/mdx mice compared to WT and mdx mice that may explain MTBD/mdx MT disorganization. Our data reinforce the broad molecular and phenotypic rescue in MTBD/mdx SkM and demonstrate that basal and post-scruff mdx SkM proteomic alterations are largely restored to WT status in MTBD/mdx SkM. Our finding that HSPB1 levels were reduced in MTBD/mdx SkM (main effect p-value = 0.004) aligned with our lab's previous MTBD/mdx SkM proteomics data 54 and led us to identify filamin-C (main effect p-value = 0.008) as an additional protein commonly reduced in MTBD/mdx SkM compared to MT-rescued control groups between our two proteomics studies. While filamin-C interacts with the sarcoglycan complex, 55 and β-sarcoglycan-null skeletal muscle presents with a disorganized subsarcolemmal MT lattice in the presence of dystrophin, 10 it is not immediately clear how the absence of filamin-C may be contributing to MT instability in dystrophic muscle. Thus, the few proteomic differences between WT and MTBD/mdx QM suggest that ECC-induced MT disorganization in MTBD/mdx mice is likely due to altered protein post-translation modifications and/or enzyme activities rather than alterations in protein expression.
Among MTBD/mdx QM proteins with altered expression compared to WT, we identified PrxII, HSPB1, ADP-ribosylhydrolase like 1 (Adprhl1), Rho GDP-dissociation inhibitor 2 (Arhgdib), tropomyosin 3, gamma (Tpm3), and sarco/endoplasmic reticulum Ca2+-ATPase 2 (SERCA2; Atp2a2). Our lab previously demonstrated that SkM PrxII expression is dispensable for MT lattice organization, 54 suggesting that the lack of PrxII rescue in MTBD/mdx SkM is unlikely to explain ECC-mediated loss of transverse MTs in MTBD/mdx mice. Tropomyosin 3 and SERCA2 regulate actin filament binding and Ca2+ dynamics during muscle contraction, respectively. Tropomyosin 3 levels were mildly but not significantly reduced in mdx and MTBD/mdx QM compared to WT (main effect p-value = 0.06), while SERCA2 levels were significantly reduced in MTBD/mdx QM compared to WT and mdx (main effect p-value = 0.01). Reduced tropomyosin 3 and SERCA2 expression in MTBD/mdx SkM may lead to altered excitation-contraction coupling; however, MTBD/mdx mice have fully rescued muscle function and do not display eccentric contraction-induced force drop compared to mdx mice. 10 ADP-ribosylhydrolase like 1 is a pseudoenzyme reported to be exclusively expressed in the heart for the regulation of actin cytoskeletal dynamics and Z-disc formation. 56 However, our proteomics data suggests that ADP-ribosylhydrolase like 1 is found in SkM and is reduced in MTBD/mdx SkM compared to WT and mdx (main effect p-value = 0.007), which may open the door to novel cytoskeletal and MT regulatory functions for ADP-ribosylhydrolase like 1 in SkM.
Additionally, we investigated compensatory or intermediate expression changes in MTBD/mdx ECF after scruff exposure that could rule out previously identified pathology-associated mdx proteomic alterations as contributing to mdx stress vulnerability. Following two-way ANOVA analysis, most of the MTBD/mdx DEPs identified by two-group comparisons were shown to demonstrate similar protein expression levels to WT mice, as exemplified by adiponectin and GSTM1.
Interestingly, adiponectin was shown to be elevated in mdx QM and depleted in mdx ECF, which to our knowledge has not been reported previously and suggests that reduced adiponectin observed in mdx and DMD biofluids57,58 may be due to impaired adiponectin secretion from dystrophin-deficient SkM. The last few years have shown an increased investment in the development of adiponectin-agonist therapies to treat the secondary pathological manifestations of dystrophinopathy in SkM, cardiac muscle, smooth muscle, and the brain59–65. Adiponectin secretion is regulated in part by PTMs of its high molecular weight oligomeric form to improve insulin sensitivity, glucose uptake, respiratory capacity, and anti-inflammatory macrophage phenotype switching.59,66 Recent work has highlighted impaired adiponectin secretion as a contributor to gut dysbiosis in mdx mice. 67 An improved mechanistic understanding of possible impairments in the adiponectin secretory pathway within mdx SkM will be important for the development of targeted therapies with high efficacy in improving DMD metabolic dysfunction along with muscle inflammation and fibrosis. Dystrophin repletion in MTBD/mdx SkM restores ECF adiponectin levels, exemplifying the multitude of proteomic alterations in mdx SkM and circulation that are rescued to WT-like status in MTBD/mdx mice.
Our mdx ECF and QM proteomics data supports existing literature demonstrating dysregulated glutathione (GSH) and oxidative stress-handling pathways in mdx mice and DMD patients68–70. Glutathione S-transferases (GSTs) are detoxification enzymes that bind a wide variety of endogenous and exogenous ligands in addition to their catalytic activity in conjugating electrophilic moieties to GSH and performing peroxidase and isomerase functions. 71 Our data demonstrated reduced GSTM1 and GSTM2 levels in mdx ECF and QM, while MTBD/mdx ECF displayed WT-like GSTM1 and GSTM2 levels in ECF and an additional increment in post-scruff QM levels of these enzymes beyond WT levels after scruff exposure. Glutaredoxin (GRX) enzymes perform cytoprotective functions and protect against protein oxidative stress by catalyzing the reduction of protein mixed disulfides with GSH. 72 GRX-1 levels were significantly reduced in mdx ECF compared to WT and MTBD/mdx (main effect p-value < 0.0001), while GRX-1 levels were also reduced in mdx QM and significantly restored in MTBD/mdx QM (main effect p-value = 0.0002). Superoxide dismutase (SOD) and catalase work in tandem to provide cellular antioxidant defense by SOD-mediated superoxide dismutation and catalase-mediated reduction of hydrogen peroxide formed by SOD. 73 We observed reduced SOD2 levels in mdx and QM compared to WT and MTBD/mdx (main effect p-value < 0.01). Catalase was reduced in MTBD/mdx QM compared to WT, with intermediate mdx expression (main effect p-value = 0.01); in contrast, ECF catalase levels were significantly reduced in mdx mice and increased in MTBD/mdx mice following scruff exposure (interaction p-value = 0.006), suggesting a potential compensatory secretion of SkM catalase in MTBD/mdx mice as an adaptive stress response. Altogether, our proteomics data supports a model in which SkM dystrophinopathy leads to defects in cellular antioxidant defense machinery and contributes to basally elevated oxidative stress. Additional perturbations in cellular redox homeostasis due to physiological stress exposure are met with poor oxidative stress handling in mdx mice and are rescued by SkM dystrophin expression in MTBD/mdx mice. Increased post-stress oxidative damage in mdx mice may lead to downstream proteomic alterations as well as stress pathology, suggesting that strategies to mitigate oxidative stress may be beneficial in reducing mdx stress susceptibility.
Conclusions
In conclusion, we characterized the mdx SkM secretome following scruff exposure to identify multiple putative secreted protein candidates whose secretion from SkM is regulated by stress and SkM dystrophinopathy. These protein candidates require further validation with orthogonal biochemical techniques, along with pharmacological and genetic approaches to assess their impact on mdx stress physiology. An important strength of our proteomics analysis is the inclusion of a phenotypically rescued group (MTBD/mdx) with targeted dystrophin restoration exclusively in the SkM compartment. This group enables us to identify proteomic changes that are specifically related to SkM dystrophinopathy and mdx stress pathology, since MTBD/mdx mice are also rescued for all mdx stress phenotypes.8,9 Additionally, the data from our current proteomics investigation can be further mined to identify novel secreted proteins that are regulated by dystrophin expression in SkM, with the inclusion of a comparison group that is phenotypically rescued and displays nearly WT-like proteomic expression patterns. Future biomarker identification and pre-clinical investigations of genetic and pharmacological DMD therapies should include MTBD/mdx mice or a similar phenotypically and genetically rescued group for comparison of biomarker specificity or treatment efficacy. Additional future investigations will assess the protein, metabolite, and nucleotide contents of SkM extracellular vesicles and exosomes to advance our understanding of the dystrophic SkM secretome during a physiological stress response.
Supplemental Material
sj-docx-1-jnd-10.1177_22143602251409302 - Supplemental material for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy
Supplemental material, sj-docx-1-jnd-10.1177_22143602251409302 for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy by Erynn E Johnson, Jacob Powers and James M Ervasti in Journal of Neuromuscular Diseases
Supplemental Material
sj-xlsx-7-jnd-10.1177_22143602251409302 - Supplemental material for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy
Supplemental material, sj-xlsx-7-jnd-10.1177_22143602251409302 for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy by Erynn E Johnson, Jacob Powers and James M Ervasti in Journal of Neuromuscular Diseases
Supplemental Material
sj-xlsx-8-jnd-10.1177_22143602251409302 - Supplemental material for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy
Supplemental material, sj-xlsx-8-jnd-10.1177_22143602251409302 for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy by Erynn E Johnson, Jacob Powers and James M Ervasti in Journal of Neuromuscular Diseases
Supplemental Material
sj-xlsx-9-jnd-10.1177_22143602251409302 - Supplemental material for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy
Supplemental material, sj-xlsx-9-jnd-10.1177_22143602251409302 for Identification of myokines associated with the pathological stress response in the mdx mouse model of Duchenne muscular dystrophy by Erynn E Johnson, Jacob Powers and James M Ervasti in Journal of Neuromuscular Diseases
Footnotes
List of abbreviations
Acknowledgements
The authors would like to thank Dr Kurt Prins at the University of Minnesota for his intellectual input and hypothesis generation for data pertaining to skeletal muscle transverse tubule morphology.
Ethical considerations
All animal experiments were approved by the Institutional Animal Care and Use Committee at the University of Minnesota (Protocol #2404-42023A).
Consent to participate
Not applicable
Consent for publication
Not applicable
Author contributions
EJ and JME conceived the study and designed experiments. EJ performed and analyzed proteomics and ELISA experiments. JP performed and EJ analyzed western blot experiments. EJ wrote the manuscript. EJ and JME made manuscript revisions. All authors read and approved the final manuscript.
Funding
This work was supported by the National Institutes of Health (NIH) Minnesota Muscle Training Grant [5T32AR007612-21, 5T32AR007612-22] to EJ, the NIH Functional Proteomics of Aging Training Grant [5T32AG029796-13, 5T32AG029796-14] to EJ, and by NIH grants [5R01AR042423-27] and [5R01AR049899-18] to JME. The Orbitrap Eclipse instrumentation platform used in this work was purchased through High-end Instrumentation Grant S10OD028717 from the NIH.
NIH Functional Proteomics of Aging Training Grant, National Institutes of Health (NIH) Minnesota Muscle Training Grant, National Institutes of Health, (grant numbers 5T32AG029796-13, 5T32AG029796-14, 5T32AR007612-21, 5T32AR007612-22, 5R01AR042423-30, 5R01AR049899-21).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability
The datasets supporting the conclusions of this article are included in the published article (Supplementary Data 1 and 2) and are publicly available in the ProteomeXchange Consortium via the Proteomics Identifications Database (PRIDE) repository, project accession PXD054593.
74
R script used to generate plots and filter data is publicly available at: 
.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
