Matrisome proteomics reveals novel mediators of muscle remodeling with aerobic exercise training

Abstract

Skeletal muscle has a unique ability to remodel in response to stimuli such as contraction and aerobic exercise training. Phenotypic changes in muscle that occur with training such as a switch to a more oxidative fiber type, and increased capillary density contribute to the well-known health benefits of aerobic exercise. The muscle matrisome likely plays an important role in muscle remodeling with exercise. However, due to technical limitations in studying muscle ECM proteins, which are highly insoluble, little is known about the muscle matrisome and how it contributes to muscle remodeling. Here, we utilized two-fraction methodology to extract muscle proteins, combined with multiplexed tandem mass tag proteomic technology to identify 161 unique ECM proteins in mouse skeletal muscle. In addition, we demonstrate that aerobic exercise training induces remodeling of a significant proportion of the muscle matrisome. We performed follow-up experiments to validate exercise-regulated ECM targets in a separate cohort of mice using Western blotting and immunofluorescence imaging. Our data demonstrate that changes in several key ECM targets are strongly associated with muscle remodeling processes such as increased capillary density in mice. We also identify LOXL1 as a novel muscle ECM target associated with aerobic capacity in humans. In addition, publically available data and databases were used for in silico modeling to determine the likely cellular sources of exercise-induced ECM remodeling targets and identify ECM interaction networks. This work greatly enhances our understanding of ECM content and function in skeletal muscle and demonstrates an important role for ECM remodeling in the adaptive response to exercise. The raw MS data have been deposited to the ProteomeXchange with identifier PXD053003.

Keywords: Core matrisome; Exercise; Matrisome-associated; Proteomics; Remodeling; Skeletal muscle.

Fig. 1

Experimental workflow for the quantification of the skeletal muscle matrisome. 10 mg of pulverized gastrocnemius muscle was lysed with NaCl and Rapigest with mechanical disruption. The lysate was centrifuged to separate the supernatant (soluble fraction) from the pellet (insoluble fraction). The insoluble fraction was further digested and solubilized, and both fractions underwent reduction, alkylation and digestion for mass spectrometry. Peptides were labeled by tandem mass tags (TMT), fractionated, and analyzed by LC-MS3. Figure created with Biorender.com.

Fig. 2

Skeletal muscle matrisome quantification. (A) Matrisome proteins were annotated from the soluble (left) and insoluble (right) fractions of N=11 mouse gastrocnemius muscles. (B) ECM proteins from the core matrisome (collagens, glycoproteins, proteoglycans) and ECM associated (ECM regulators, ECM-affiliated, secreted factors) categories were identified in muscle by our ECM fractionation protocol.

Fig. 3

Regulation of the muscle matrisome by exercise training. (A) Exercise training induced significant remodeling of mouse gastrocnemius ECM in both the core matrisome (30/76; 40%) and ECM-associated proteins (13/85; 15%). (B) Heat map demonstrates log2 fold change of ECM proteins that are altered by exercise from soluble (left) and insoluble (right) fractions in Control (C1-C6) and Exercise-trained (E1-E5) mice. Only proteins with >1.2 fold-change and P-value <0.1 by moderated t-tests (sedentary vs. exercise-trained) are shown in the heat map.

Fig. 4

Western blot analysis of exercise-regulated ECM protein targets identified by MS. (A) A subset of core matrisome, and (B) matrisome-associated proteins were analyzed by Western blotting in the gastrocnemius muscle from sedentary (CON; N=12) and exercise-trained (ET; N=12) mice. These results validated exercise training-induced ECM remodeling of (C) FBN1, TINAGL1, LAMA4, and (D) SEMA3C, PLG, LOXL1. Mean, SD and individuals values are shown. P-value <0.05 by Mann-Whitney U Test was considered statistically significant.

Fig. 5

Validation and localization of exercise-regulated ECM proteins identified by MS. (A) Immunostaining for FN1 confirmed increased protein in the extracellular space of skeletal muscle with exercise training. (B) Training also increased LAMA4 staining in skeletal muscle, and co-localized with endothelial cells stained using griffonia lectin. (C) LAMA4 staining was significantly correlated with muscle capillary density and (D) oxidative fiber-type, suggesting a role in exercise-induced changes in muscle phenotype. Scale bar = 100µm for panel A, 50µm for panel B. P-value < 0.05 by Mann-Whitney U Test was considered statistically significant. Bar graphs demonstrate mean, SD, and individual values. Panels C and D show Spearman correlations and dotted lines represent 95% confidence intervals.

Fig. 6

Correlation between exercise-regulated ECM proteins and capillary density. Expression of (A) FBN1, (B) PLG, (C) TINAGL1, (D) LOXL1 (Mature; 32 kDa), (E) SEMA3C, and (F) FN1 are positively correlated with capillary density- a hallmark of exercise training-induced muscle remodeling. P-value < 0.05 for Spearman correlation was considered statistically significant. 95 % confidence intervals are shown with dotted lines.

Fig. 7

ECM genes expressed in diverse cell types in skeletal muscle. ECM gene expression in each cell type was analyzed from 3 published single-nuclei RNA sequencing datasets. Each color represents percentage of gene expression in corresponding cell types from Dos et al. (blue), Wen et al. (red) and Petrany et al. (green). Myh7, Type 1 myofibers; Myh4, Type 2B myofibers; Myh1 + 2, Type 2A/X/D myofibers, FAPS, fibroadipogenic progenitor cells. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 8

Exercise regulation of ECM interaction networks. Interactome analysis demonstrates the complexity of ECM remodeling by exercise training. ECM proteins that were upregulated (red) or downregulated (blue) by exercise, may in-turn regulate large or small networks of both ECM and non-ECM proteins. Exercise-regulated proteins from our dataset are colored according to the z-score from exercise vs. control; others are colored gray. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9

Identification of LOXL1 in human skeletal muscle. (A) Muscle lysates from N=22 human participants were analyzed by Western blotting using antibodies against LOXL1 and plasminogen (PLG). Stain-free gel images (BioRad) are shown as a loading control. All samples were run and analyzed on the same gel/membrane. Mature LOXL1 protein was positively correlated with (B) aerobic exercise capacity (VO₂peak), and (C) insulin sensitivity determined during an oral glucose tolerance test (SiOGTT). Mature LOXL1 was negatively correlated with (D) body mass index, and (E) glucose area under the curve during an oral glucose tolerance test, with higher values indicating lower glucose tolerance. P-value <0.05 for Spearman correlation was considered statistically significant.