Gene expression profile of adhesion and extracellular matrix molecules during early stages of skeletal muscle regeneration

Abstract Skeletal muscle regeneration implies the coordination of myogenesis with the recruitment of myeloid cells and extracellular matrix (ECM) remodelling. Currently, there are no specific biomarkers to diagnose the severity and prognosis of muscle lesions. In order to investigate the gene expression profile of extracellular matrix and adhesion molecules, as premises of homo‐ or heterocellular cooperation and milestones for skeletal muscle regeneration, we performed a gene expression analysis for genes involved in cellular cooperation, migration and ECM remodelling in a mouse model of acute crush injury. The results obtained at two early time‐points post‐injury were compared to a GSE5413 data set from two other trauma models. Third day post‐injury, when inflammatory cells invaded, genes associated with cell‐matrix interactions and migration were up‐regulated. After day 5, as myoblast migration and differentiation started, genes for basement membrane constituents were found down‐regulated, whereas genes for ECM molecules, macrophage, myoblast adhesion, and migration receptors were up‐regulated. However, the profile and the induction time varied according to the experimental model, with only few genes being constantly up‐regulated. Gene up‐regulation was higher, delayed and more diverse following more severe trauma. Moreover, one of the most up‐regulated genes was periostin, suggestive for severe muscle damage and unfavourable architecture restoration.

| 10141 CEAFALAN Et AL. cells (SC)/myoblasts and various types of resident and migrated interstitial cells. 2 During the first few days after acute injury, the latter are represented mostly by several myeloid cell populations, which are recruited at different time-points of the regeneration process.
Mast cells and neutrophils are the first cells to respond. The second wave of cells attracted at the injury site are macrophages (MCs) whose numbers significantly increase 2 days postinjury. 3 After performing their phagocytic role, at around 4 days post-injury, the pro-inflammatory MCs suffer a phenotypic switch and become proliferative, anti-inflammatory MCs 4 that support myogenesis and myofibre growth. The phenotypic and transcriptional pattern induced by this transition coincides with changes in the expression level of developmentally regulated, myogenic genes. These genes govern different steps of the myogenic process, 5 from the activation and proliferation of myogenic precursors (MP) to their fusion, differentiation and growth. This sequence of the myogenic process can be easily demonstrated in vitro in the absence of the myeloid population. However, recruited myeloid cells release a large array of soluble molecules 6,7 and establish direct molecular contacts with resident myogenic cells. 8,9 This complex mechanism regulates the amplitude and the timing of each step along the myogenic process.
Extracellular matrix (ECM) remodelling is also a key process during inflammation, wound healing, and injury repair 10 and implies both protein synthesis and degradation. The ECM consists of various structural proteins-diverse collagen types, fibronectin, laminin-as well as non-structural proteins, such as matricellular proteins. The latter have cellular binding sites for cell-matrix interactions and diffusible growth factors for creating gradients as guiding cues for cell migration and signalling events during tissue regeneration. 11 The tight synchronization between interstitial and myogenic cells and ECM remodelling is an essential prerequisite for an efficient regeneration and functional recovery, while evading fibrosis. 4,12,13 To date, however, no studies have investigated ECM and adhesion molecule-associated gene expression changes during muscle regeneration, and there exist few data correlating gene profiles and types of injury.
The aims of our study were (a) to investigate the in vivo gene expression profile of ECM and adhesion molecules, as premises of homo-or heterocellular cooperation and milestones for skeletal muscle regeneration after acute crushing injury and (b) to identify specific genes related to different types of injury, by comparing the expression profile (for ECM and adhesion molecules) induced by our model (crushing) with previously reported injury models. 14 Thus, these gene expression profiles were measured at two different time-points after muscle injury, in the time-frame when inflammation and degeneration are peaking, and as it switches towards regeneration and remodelling. 15 Gene expression profile in injured muscles was compared both with the contralateral, non-injured muscles and uninjured controls. This double comparison allowed us to determine whether the crushing injury led to a systemic reaction which induced changes in gene expression on the contralateral, non-injured muscle.

| Experimental animals
All experiments were conducted on 12-week-old C57BL/6J male mice (The Jackson Laboratory # 000664), as they are the most commonly used strain for skeletal muscle injury models. 14,16 The animals included in this study were in perfect health. All mice were kept in individual standard cages at 21-24°C, with 40%-60% humidity and Histopathological evaluation was conducted on 9 (n = 9) animals, three mice at each time-point: 3rd, 5th and 14th day post-injury. For the gene expression analysis, a total of 19 (n = 19) male mice were used in this experiment. The animals were divided into three groups: 10 mice were used for generating the muscle injury model, 5 for each time-point, and 9 mice were used as non-injured, external controls.

| Muscle injury model
The C57BL/6J mice were anaesthetized by intramuscular injection of 100 mg/kg Ketamine (Kepro BV) in PBS, in the anterior left leg prior to manipulations, before inducing the injury in the posterior left leg and again before being euthanized by cervical dislocation.
The level of anaesthesia was assessed by absence of reflexes. The muscle injury model was obtained by crushing the posterior left leg with an adjusted forceps, 1 cm away from the distal joint without fracturing the bone. The pressure was maintained for 2 min. Samples from the injured area of the crushed muscle and from the contralateral gastrocnemius were collected the 3rd and 5th day post-injury, after cervical dislocation while under anaesthesia, when response to stimuli was no longer detected. These time-points correspond to the peak of the inflammatory and degeneration stage dominated by inflammatory macrophage recruitment in the injured area, and the switch towards regeneration and remodelling after injury. 17 Samples from the 14th day post-injury were also collected for the histopathological assessment of the injury model.

| Histopathology
Histopathological evaluation was conducted on three mice at each time-point: 3rd, 5th and 14th day post-injury. Small fragments from the left gastrocnemius muscles were collected and fixed by immersion in 4% glutaraldehyde, post-fixed in buffered 1% OsO 4 with 1.5% K 4 Fe(CN) 6 (potassium ferrocyanide-reduced osmium), dehydrated in graded ethanol series and further processed for epoxy resin embedding (AGAR 100). One-micrometre-thick sections (semi-thin sections) were stained with 1% toluidine blue and examined by light microscopy for morphological analysis with Leica DM 600. Images were recorded using a Leica DFC7000 T camera.

| RNA isolation and gene expression analysis
Gene expression analysis was performed on total RNA isolated from five pair tissue samples from injured (I) and non-injured contralateral (N-I) muscles at 3 and 5 days post-injury. N-I was used as internal control. Another external control group (C) of 9 animals (without muscle injury) was enrolled in the study, and 3 pools of RNA were analysed for gene expression. Total RNA isolation was performed

| GEO data mining
A search of the NCBI Gene Expression Omnibus (GEO) was conducted in order to find data sets reporting gene expression changes in the skeletal muscle tissue of C57BL/6J mice following different methods of injury. Only GSE5413 was identified. 14 This data set reports gene expression in the skeletal muscles of uninjured and of injured mice at different time-points (6 hour, 1, 3 and 7 days), following eccentric contraction injury (CI) or freezing injury (FI), and using the Affymetrix Murine Genome U74A Version 2 Array. The candidate genes were analysed for differential expression using GEO2R in the following comparisons: 3 and 7 days (both CI and FI) vs non-injured, external controls.

| Statistical analysis
Gene expression analysis was conducted using the Statistical Package for Social Science (SPSS Version 17.0). Data normality was assessed using the Shapiro-Wilk test. Since data were normally distributed (P > .05), a paired t test was used to assess differences in gene expression levels between I and N-I (at 3 and 5 days). Comparisons of gene expression levels between I as well as N-I (at 3 and 5 days) and C were tested with an independent sample t test. Difference in gene expression was considered significant when P < .05 and fold regula-

Injured muscle vs contralateral (I vs N-I)
The paired analysis (I vs NI) revealed that at 3 days, 9 genes out of 84 were differentially expressed in injured muscle (4 up-regulated and 5 down-regulated).

| Non-injured muscle (N-I) and the injured muscle (I) vs external control group (C)
To exclude any systemic influence that might have induced a change in gene expression, we compared genes that were significantly regulated (presented in detail in Table 1)  Significant results are presented in Table 1.

F I G U R E 1
Light microscopy on toluidine blue-stained semi-thin sections of epoxy-embedded injured gastrocnemius muscle. Representative images from three different mice. A-C. 3 days post-injury oedema and some necrotic fibres are observed along with a massive inflammatory infiltrate in the interstitial spaces around damaged fibres. D-F. 5 days post-injury, regenerating myofibres (myotubes with central nuclei) and inflammatory infiltrate are observed. In some areas, muscle necrosis is still present. G-I. 14 days post-injury, inflammatory infiltrate and collagen deposition were still detected at the injury site. A, D, G longitudinal sections; B, E, H cross sections; boxed areas are presented at a higher magnification in C, F and I, respectively

| Gene ontology and pathway analyses
The results of the gene ontology and KEGG pathway are shown in Table 2a and b. Most relate to peptidase activity, modulators, as well as integrin-mediated cell-ECM interaction for migration and signalling.

| Gene expression comparison with eccentric contraction-induced muscle injury (CI) and freezeinduced muscle injury (FI) models
The GSE5413 data set reports the gene expression of 12 488 genes in C57BL/6J uninjured (control) and injured mice, using eccentric contraction (CI) and freezing injury (FI) models. In order to compare data from our model of crushing injury with those obtained by GSE5413, we had to compare the gene expression profile of I at 3 and 5 days with that of C (Table 1).
GEO data mining found that 10 transcripts out of the 84 we tested were significantly regulated at 3 days post-injury in the CI model. ITGAM, CD44 and TIMP1 were found up-regulated in both our model (I vs C) and CI model. Of note, ITGAM was up-regulated also in the paired analysis (Table 3 and Figure 3A). In the FI model, 12 transcripts out of the 84 we tested were significantly changed at 7 days after inflicting the injury. The ex-

| D ISCUSS I ON
The understanding of the cellular response and molecular composition of the microenvironment during muscle regeneration is mandatory for the development of clinical strategies to improve muscle function during aging, or after extensive trauma. In this study, we evaluated the gene expression profile of ECM and adhesion molecules in skeletal muscle regeneration after acute crush injury and by comparison with other previously reported injury models, such as CI and FI.
Our experimental procedure inflicted direct and severe muscle damage. The histopathology assessment showed collagen deposition and persistent inflammation even at 14 days post-injury, longer than in the case of small contusion injuries 18 and other previously reported injury models, like cardiotoxin injection, 19 CI or even FI. 14,18 In our model, inflammation gradually resolved and the architecture of the injured muscle was re-established only after day 21 post-injury (data not shown).  Bold fonts indicate genes differentially expressed both at 3 and 5 days.
Red fonts indicate the up-regulated genes.
Blue fonts indicated the down-regulated genes.  Besides facilitating myoblast migration and the angiogenic process, 22 in vitro studies showed that MMP-14 regulates myotubes formation by degrading interstitial ECM components, like fibronectin, that prevent cell fusion and laminin alpha 2, regulating the interaction of the mature myofibre with the BM. 23 In the time-frame TA B L E 3 Genes differentially expressed both in crush injury and CI at 3 days after injury Bold fonts indicate genes differentially expressed both at 3 and 5 days.

TA B L E 2 (a-b). Gene ontology and pathway analyses on the differentially expressed genes (both at 3 and 5 days post-injury) have been performed by GO Molecular Function 2018 (a) and KEGG 2019 Mouse (b) through Enrichr web server
Red fonts indicate the up-regulated genes.
Blue fonts indicated the down-regulated genes.  Bold fonts indicate genes differentially expressed both at 3 and 5 days.
Red fonts indicate the up-regulated genes.
Blue fonts indicated the down-regulated genes.
*Trend towards statistical significance. which suggested that the regeneration process was overwhelmed.
POSTN has been recently shown to promote fibroblast migration at the injury site and to favour scar formation. 33 POSTN was also previously reported to be transiently up-regulated around 4 days post-injury in a mouse model of cardiotoxin injury. The protein expression was first restricted to myoblasts and regenerating myofibres and then transferred to endomysial stromal cells, other than infiltrating myeloid cells. 19 One limitation of this comparative study is the potential difference in regeneration responses among the different muscles of the calf (in our study the gastrocnemius muscle). Multiple reports suggest that the expression pattern of many of the tested genes depends on the type of muscle and the phase of muscle regeneration. 25 Thus, a potential source of inaccuracy when comparing the data sets may be the different secondary time-point post-injury which in our case was at 5 days. However, in our view, this time-point is a more accurate window into the early stage of muscle regeneration.
Another source of imprecision could be the potential effect on gene expression levels of the ketamine anaesthesia that was not performed on the control group nor by the other studies. This was at least partially mitigated by performing the injection in a different leg to the one receiving muscle injury.

| CON CLUS ION
In conclusion, our study revealed remarkable changes in gene ex-

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All the new data generated or analysed during this study are included in this published article as File S1. The analysis reported in the File S3 refers to Data Set GSE5413 in the Gene Expression