Transcriptional profiling and regulation of the extracellular matrix during muscle regeneration

Sean C. Goetsch1, Thomas J. Hawke1, Teresa D. Gallardo1, James A. Richardson2 and Daniel J. Garry1,3

1 Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas 75390-8573
2 Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas 75390-8573
3 Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, Texas 75390-8573


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
Muscle regeneration is a complex process requiring the coordinated interaction between the myogenic progenitor cells or satellite cells, growth factors, cytokines, inflammatory components, vascular components and the extracellular matrix (ECM). Previous studies have elegantly described the physiological modulation of the regenerative process in response to muscle injury, but the molecular response that characterizes stages of the repair process remains ill-defined. The recent completion of the Human and Mouse Genome Projects and the advent of technologies such as high-density oligonucleotide array analysis facilitate an expanded analysis of complex processes such as muscle regeneration. In the present study, we define cellular and molecular events that characterize stages of muscle injury and regeneration. Utilization of transcriptional profiling strategies revealed coordinated expression of growth factors [i.e., Tgfb1, Igf1, Egf, chemokine (C-C motif) ligand 6 and 7], the fetal myogenic program (Myod1, Myf5, Myf6), and the biomatrix (procollagen genes, Mmp3, Mmp9, biglycan, periostin) during muscle regeneration. Corroboration of the transcriptional profiling analysis included quantitative real-time RT-PCR and in situ hybridization analyses of selected candidate genes. In situ hybridization studies for periostin [osteoblast-specific factor 2 (fasciclin I-like)] and biglycan revealed that these genes are restricted to mesenchymal derivatives during embryogenesis and are significantly regulated during regeneration of the injured hindlimb skeletal muscle. We conclude that muscle regeneration is a complex process that requires the coordinated modulation of the inflammatory response, myogenic progenitor cells, growth factors, and ECM for complete restoration of muscle architecture.

microarray analysis; biglycan; periostin


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
ADULT MAMMALIAN SKELETAL MUSCLE is one of the few tissues that is capable of efficient, reproducible regeneration in response to an extensive injury (12, 24). The capacity for self-repair is due to a rare pool of undifferentiated myogenic precursor cells, termed satellite cells (or myogenic progenitor cells) because of their location at the periphery of mature skeletal myofibers (16, 23). Myogenic progenitor cells are arrested at an early stage of the myogenic program such that they do not express any of the members of the MyoD family of transcription factors (12, 16, 23, 24). Under unstressed conditions, these myogenic progenitor cells are quiescent. Following muscle injury or in response to increased work demand, myogenic progenitor cells are mobilized to proliferate, differentiate, and fuse into multinucleated myofibers in a manner that recapitulates the fundamental events of muscle development (12, 24). The myogenic progenitor cell population is a residual pool of self-renewing progenitor cells, capable of supporting additional rounds of regeneration and is reestablished after each discrete episode of muscle injury. However, this capacity for self-renewal is finite. The exhaustion of the myogenic progenitor cell pool is an important factor contributing to the clinical deterioration observed in patients with myopathies such as muscular dystrophy (5, 10, 12, 24). Furthermore, fibrosis and adipose tissue ultimately replace the myopathic, degenerating myofibers, resulting in further perturbation of skeletal muscle architecture and function (5, 10, 12, 24). Although a number of physiological studies have examined the responses of the myogenic progenitor cells to various stimuli, the molecular events that regulate the skeletal muscle repair response to an injury remain ill-defined.

Effective muscle regeneration requires a coordinated repair response involving the inflammatory system, capillary morphogenesis, and myogenic progenitor cell-mediated myogenesis to ultimately restore the architecture of the tissue without fibrosis (i.e., scar formation). This repair process is dependent on an orchestrated response between the inductive signals of cytokines, growth factors, and the extracellular matrix (ECM) (12, 16, 23, 24). Previous studies have established the critical role for ECM components during morphogenesis and tissue regeneration (15, 17, 20). Matrix degradation following injury initially releases growth factors that function as potent mitogenic signals for tissue-specific progenitor cells and the vascular endothelial cells (15, 17, 20). Additionally, at later stages of the regeneration phase, the ECM is capable of scavenging or inactivating these morphogens, resulting in the reestablishment of the tissue architecture. Although considerable progress has been made regarding our understanding of muscle development, the coordinated regenerative response of skeletal muscle is unclear. Applications using emerging technologies to examine the molecular response of muscle regeneration would enhance our understanding of this response and facilitate tissue engineering strategies that would have applications for patients with myopathies.

In the present study, we utilize cellular and molecular technologies to comprehensively examine skeletal muscle regeneration and further focus on the role of periostin (Osf2) and bigylcan (Bgn), which contribute to the ECM. The results of this study highlight the critical role of the ECM in skeletal muscle regeneration, which is a complex process involving thousands of genes, many of which are also expressed during the fetal program.


    EXPERIMENTAL PROCEDURES
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
Cardiotoxin-induced skeletal muscle injury.
Adult male C57BL/6 or mdx mice were bred according to the guidelines of the NIH and the Institutional Animal Care and Use Committee at UT Southwestern Medical Center. Gastrocnemius muscle injury was induced with the intramuscular delivery of 150 µl cardiotoxin (10 µM, Naja nigricollis; Calbiochem, La Jolla, CA) (7, 8). Uninjured (n = 3) and injured mice (n = 3) were killed, and the gastrocnemius muscles were harvested at 0.25, 0.5, 1, 2, 5, 7, or 14 days postinjury. Tissue was either snap frozen in liquid nitrogen and stored at -80°C or fixed overnight in 4% paraformaldehyde/DEPC-PBS following Avertin-induced anesthesia and transcardiac perfusion as described previously (68). Fixed tissue was paraffin embedded for rotary microtomy (68).

Reverse transcription, real-time, and semi-quantitative PCR.
Total RNA was isolated with TriPure isolation reagent (Roche, Basel, Switzerland) and reverse transcribed with SuperScript II (Invitrogen, Carlsbad, CA) to cDNA. Real-time PCR was performed with the iCycler iQ real-time PCR detection system (Bio-Rad Laboratories, Hercules, CA) and the QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA) (7, 8). The reaction was carried out in a final volume of 25 µl with 1 x SYBR Green PCR master mix, 0.3 µM each primer, and 20 ng cDNA for 50 cycles (20 s at 94°C to denature, 20 s at 60°C to anneal, and 30 s at 72°C to extend). The 75- to 100-bp extension was detected following each cycle and analyzed with the iCycler iQ software (Bio-Rad) for specificity, using melting curve, and threshold cycle (CT). Data are expressed with comparative CT method as an estimate of mRNA of target to reference control. Real-time PCR primers included the following: Ccnd1, forward 5' CGGATGAGAACAAGCAGACC 3' and reverse 5' GCAGGAGAGGAAGTTGTTGG 3'; Fn1, forward 5' TGTAGGAGAACAGTGGCAGAAA 3' and reverse 5' CAGGTCTACGGCAGTTGTCA 3'; Mmp3, forward 5' CCCAGGAAGATAGCTGAGGA 3' and reverse 5' CAACTGCGAAGATCCACTGA 3'; Myf5, forward 5' CTGTCTGGTCCCGAAAGAAC 3' and reverse 5' AAGCAATCCAAGCTGGACAC 3'; Myf6, forward 5' AATTCTTGAGGGTGCGGATT 3' and reverse 5' ATGGAAGAAAGGCGCTGAAG 3'; Pcna, forward 5' GCTTGGCAATGGGAACATTA 3' and reverse 5' CAGTGGAGTGGCTTTTGTGA 3'; and Rasa3, forward 5' ATTGATGGGGACCGTGAAAC 3' and reverse 5' GGGCCGTCATACACAGACTT 3'.

Semi-quantitative RT-PCR was performed with a final volume of 20 µl with variable cycle lengths as previously described for quantitative analyses. Conditions for these reactions include 20 s at 94°C to denature, 20 s at 60°C to anneal, and 30 s at 72°C to extend. Following the reaction, the amplicon (periostin, 92 bp; biglycan, 144 bp; 18S rRNA, 149 bp) was observed by using 1% agarose gel electrophoresis. Primer sequences are as follows: periostin, forward 5' CCCTTCCATTCTCATATA 3' and reverse 5' CCCTTCCATTCTCATATA 3'; biglycan, forward 5' CACCTGGACCACAACAAAA 3' and reverse 5' TCCGAATCTGATTGTGACCTA 3'; and 18S rRNA, forward 5' GGACCAGAGCGAAAGCATTTA 3' and reverse 5' TGCCAGAGTCTCGTTCGTTAT 3'.

Probe labeling and Affymetrix array hybridization.
Oligonucleotide array hybridizations were carried out according to the Affymetrix protocol summarized as follows. Total RNA was isolated from uninjured or cardiotoxin-injured gastrocnemius skeletal muscle at defined intervals and pooled from three animals as previously described. Five micrograms of total RNA was reverse transcribed using an oligonucleotide containing a T7 promoter with a poly-dT tail [T7(dT)24]. Ligase and polymerases were then added to produce the second strand of the single-stranded cDNA. Following precipitation, the double-stranded cDNA was converted to biotin-labeled cRNA by using the Enzo BioArray high-yield RNA transcript labeling kit (Enzo Biochem, New York, NY). The purified biotin-labeled cRNA was then fragmented by using Affymetrix fragmentation buffer for 35 min at 95°C. Labeled fragmented cRNA (15 µg) was then hybridized to the high-density oligonucleotide Murine Genome Array U74Av2 GeneChip (Affymetrix, Santa Clara, CA) (pooled RNA isolated from 3 animals/array). After 16 h of hybridization, the array was washed, stained, and scanned according to the manufacturer’s protocol.

Array analysis.
Array quality assessment and expression values were acquired with DNA-Chip Analyzer (dChip) model-based analysis of multiple arrays (n = 18 for modeling). This method utilizes invariant-set normalization and model-based expression indexes (MBEI) with standard error (SE) to measure accuracy. Comparative analysis was performed by using MBEI and SE to construct a 90% confidence interval (CI) of fold change (3, 4). Change was specified as significant if the lower 90% confidence bound of fold change was greater than or equal to 2 and the absolute difference of the group mean expression was greater than 100. Values were exported to GeneCluster for self-organizing map (SOM) clustering to determine common and unique expression profiles for sets of genes during muscle regeneration (27) (see the Supplemental Material, available at the Physiological Genomics web site).1

Gene ontology annotation as defined by the Gene Ontology Consortium (http://www.geneontology.org), either as biological process, molecular function, or cellular component, was performed with dChip.

Riboprobe synthesis.
IMAGE clones (403071, 425785) were sequence-verified and prepared for in vitro transcription following restriction enzyme digestion and gel isolation. Linearized template (500 ng) was transcribed by using either the T7 or SP6 RNA polymerase (Ambion, Austin, TX) with 7.0 µM [{alpha}-35S]UTP (1,000 Ci/mmol; Amersham, Piscataway, NJ) to produce the respective sense and antisense riboprobes (6, 25). The periostin riboprobe was 632 bp in length, and the biglycan riboprobe was 3.0 kb in length, for the sense and antisense probes (6, 25). The probes were purified with MicroSpin G-50 columns (Amersham), analyzed for integrity, and stored overnight at -80°C as previously described (25).

In situ hybridization.
In situ hybridization was performed as previously described (6, 25). Briefly, 5-µm paraffin sections of staged embryos or selected tissues (i.e., cardiotoxin-injured skeletal muscle or the mdx heart, diaphragm, and skeletal muscles) were cut, mounted on Vectabond-coated slides, dewaxed, permeabilized, acetylated, and hybridized at 70°C. Riboprobes were diluted in 50% formamide, 0.07 M NaCl, 20 mM Tris·HCl (pH 8.0), 5 nM EDTA (pH 8.0), 10 mM NaPO4 (pH 8.0), 10% dextran sulfate, 1x Denhardt’s, and 0.5 mg/ml tRNA (25). Following hybridization, the slides were rinsed with washes of increasing stringency, treated with RNase A (2 µg/ml, 30 min at 37°C), dehydrated, and dipped in K5 nuclear emulsion gel (Ilford). Autoradiographic exposure was undertaken for a 21-day period. In all cases, sections hybridized with the sense probe resulted in the absence of signal. Reference to all expression data presented in these studies pertains to the respective mRNA transcripts.

Microscopy and photomicrography.
Periostin and biglycan mRNA expression were visualized with a Leitz Laborlux-S microscope equipped with plan-apochromatic optics, standard bright-field condenser, and a Mears low-magnification dark-field condenser. Photomicrographs were obtained with an Optronics VI-470 CCD camera and Power Macintosh G3 with Scion Image 1.62 software.

Cell culture.
C2C12 myogenic cells (derived from satellite cells) were grown as monolayers in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 20% fetal bovine serum and antibiotics. Myotube differentiation was promoted by exposing 80% confluent myoblast cultures to differentiation medium [DMEM supplemented with 2% heat-inactivated horse serum, antibiotics, insulin, and transferrin as previously described (1, 9)]. Differentiation was assessed morphologically by the appearance of multinucleated myotubes. Monolayers of NIH 3T3 fibroblast cells were grown in DMEM supplemented with 10% fetal bovine serum. C2C12 myoblasts, myotubes, and NIH 3T3 fibroblasts were cultured in the presence and absence of supplemented transforming growth factor-ß1 (TGFB1, 5 ng/ml; Sigma, St. Louis, MO). At specified time periods, medium was removed from the respective cell populations, the monolayers of cells were rinsed, and RNA was harvested for RT-PCR analysis of gene expression.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
Cellular mechanisms of muscle regeneration.
Skeletal muscle regeneration is an efficient and reproducible process. In response to cardiotoxin delivery, ~75–90% of the muscle is destroyed in the mouse model (7, 8, 12). Following this chemical-induced injury, a well-orchestrated cellular response may be observed by using standard histological techniques. During the acute postinjury period (<=6 h), myofibers become hyalinized and vacuolated and their nuclei are pyknotic or lysed (Fig. 1, A and B). Interstitial edema, myonuclear dropout, and the presence of a neutrophilic (PMN) infiltrate further characterize this initial phase (Fig. 1A). Microvascular thromboses are evident (Fig. 1B), and myogenic progenitor cells assume an activated or primed state (not shown) (7, 8). Twelve hours following injury, myofibers lyse, resulting in the production of a protein-rich edema (Fig. 1C). The inflammatory response consists initially of PMNs (Fig. 1D), and at later stages macrophages that phagocytose necrotic myofibers (2–5 days following injury). This chemical-induced injury lacks a prominent lymphocytic response using standard histological techniques. Following activation, myogenic progenitor cells proliferate (days 2–4) and characterize the regenerative phase, forming crescent-like structures around necrotic or damaged myofibers (Fig. 1E). Newly regenerated myofibers are easily identified as small basophilic, centronucleated myofibers (Fig. 1, EG). A prominent remodeling of the ECM is clearly evident early during the repair process, which persists and matures throughout the repair period (Fig. 1, EG). Ultimately, the biomatrix organizes and matures corresponding to the completion of muscle regeneration resulting in preservation of the muscle architecture (Fig. 1, G and H). No evidence of residual scar formation or persistent fibrosis is observed with this extensive injury.



View larger version (125K):
[in this window]
[in a new window]
 
Fig. 1. Morphological analysis of muscle regeneration following cardiotoxin-induced injury. A: the acute period (6–12 h) following chemical-induced muscle injury is characterized by interstitial and myofiber edema. B: myonuclear dropout, myonecrosis, and vascular thrombosis (arrowheads) are evident within 12 h of injury. C and D: myofibers are lysed and a prominent PMN response is observed. PMNs (arrowhead) phagocytose myofibers and interstitial debris. E: satellite cells (myogenic precursor cells) are activated early following injury (<6 h) and proliferate extensively within 48 h of injury and form distinctive crescents (arrowhead) around injured myofibers. F: evidence of newly regenerated myofibers, which are characterized as small basophilic fibers that contain a centralized nucleus, is widely observed within 5 days of the injury. G and H: the biomatrix is organized (days 2–5) and matures (days 5–7) following injury. Reestablishment of the skeletal muscle architecture is observed by 14 days postinjury. Bar = 40 µm.

 
Molecular mechanisms of muscle injury and regeneration.
Transcriptional analysis was undertaken to define a signature of gene expression that corresponded to sequential stages of skeletal muscle injury and regeneration. Following cardiotoxin-induced muscle injury, mice were killed at defined periods (0 or uninjured, 0.25, 0.5, 1, 2, 5, 7, and 14 days; n = 3 at each time period). The gastrocnemius muscles were harvested, and RNA was isolated, labeled, and hybridized to the Affymetrix high-density oligonucleotide array. Using the MAS 5.0 expression report, the percent of probe sets present ranged from 40–47%, with the average signal intensity exceeding 200 and with the 3'-to-5' ratios less than three (0.9–2.26%). These results collectively verify the integrity of the RNA samples and the assay quality for these studies.

The analysis of the transcriptional profiling experiments revealed discrete molecular responses associated with sequential stages following cardiotoxin-induced muscle injury (Fig. 2). Although the molecular response involving the greatest number of genes occurred within 2 days of injury (1,225 genes dysregulated >=2 fold), the acute period following injury was associated with an induction of inflammatory and signal transduction genes (Fig. 2). Transcripts that encode nuclear proteins were increased within 0.5 days of injury followed by an induction of cell cycle and signal transduction genes at 2 days (Fig. 2). The expression profile of cell cycle genes corresponds to a marked increase in the proliferative capacity of the myogenic progenitor cell population at this time (i.e., Pcna expression, see Table 1 and clusters 10, 11, 15–17, and 21–23 in the Supplemental Material). Increased ECM/cell adhesion-associated gene expression observed at day 5 and thereafter (days 7 and 14) corresponds to the reorganization/maturation of the ECM and the differentiation of the myogenic progenitor cells (Fig. 1, CH) (Table 1 and Supplemental Material). This differentiation phase, corresponding to the maturation of both the ECM and the newly regenerated myofibers, ultimately results in regenerated skeletal muscle lacking any evidence of residual scarring or fibrosis.



View larger version (53K):
[in this window]
[in a new window]
 
Fig. 2. Molecular response following cardiotoxin-induced muscle injury. Results of transcriptional profiling using high-density oligonucleotide array analysis of pooled RNA samples (n = 3 animals) following cardiotoxin-induced gastrocnemius muscle injury.

 

View this table:
[in this window]
[in a new window]
 
Table 1. Gene ontology: differential expression

 
Using an SOM algorithm, we identified clusters of regulated transcripts having similar expression profiles during the muscle repair process (see Supplemental Material). Transcripts encoding myogenic regulatory factors are expressed early following the injury and correspond to activation of the satellite cell population (i.e., Myod1 and Myf6, 3-fold upregulation at 0.25 days; cluster 18 in the Supplemental Material), and markers of muscle differentiation (sarcomeric components) at later time periods (myosin light chain, troponin I, troponin T3; see Table 1 and clusters 0–5 in Supplemental Material) further define the regeneration process. Cytokines such as the interleukins (1 and 10), Tnf, chemokine (C-C motif), ligand 2, 6, 7, and 9 (clusters 24–27 in the Supplemental Material), and Tgfb1 are significantly induced during the repair process and regulate the myogenic progenitor cell population as well as the organization and maturation of the ECM (Table 1). In the present study, transcripts such as biglycan, periostin, fibronectin, laminin, matrix metalloproteinases (Mmp 3, 8 and 9), fibril forming collagens (types I–III, and V), and network forming collagens (types IV and VIII), which are constituents of the ECM or involved with its remodeling, are significantly expressed during the repair process (Table 1 and clusters 0–5 in Supplemental Material). For example, one group of these transcripts which were significantly upregulated during the repair process was the metalloproteinases (Mmps). Mmps are important regulators for cellular proliferation, cell-cell interactions, cellular chemotaxis, and the release of matrix-bound cytokines and growth factors (20). The results of the profiling analysis establish and support the hypothesis that the ECM displays discrete, reproducible expression patterns in response to injury and is critical for effective regeneration. Furthermore, a number of transcripts that characterize this molecular response (induction of cytokines, growth factors, and ECM-related transcripts) during muscle regeneration (observed in the present study) have also been observed in differentiating C2C12 myoblasts, regenerating bone, and myopathic mouse models and provide further confirmation for the results observed in the present study (11, 18, 19).

In addition, the confirmation of the microarray results of selected candidate genes was undertaken by using quantitative real-time RT-PCR analysis (Fig. 3). As shown in Fig. 3, the real-time RT-PCR analysis confirmed the array analyses and further verified the stage-specific expression pattern for transcripts associated with the ECM remodeling (Mmp3 and Fn1), cell cycle regulation (Ccnd1 and Pcna), myogenesis (Myf6 and Myf5), and cellular signaling (Rasa3). These results, in combination with previously published data, provide further support for the molecular analysis undertaken in the present study (11, 18, 19).



View larger version (15K):
[in this window]
[in a new window]
 
Fig. 3. Confirmation of the microarray analysis using real-time RT-PCR. Log fold change of selected gene expression following cardiotoxin-induced injury (0.25, 0.5, 1, 2, 5, 7, and 14 days compared with uninjured) of the gastrocnemius muscle (n = 3 animals) by using microarray analysis (solid circles) and real-time RT-PCR analysis (open circles; performed in triplicate). Data are means ± SE for the real-time RT-PCR analyses.

 
ECM remodeling during embryogenesis and in response to injury.
Morphological and molecular mechanisms associated with muscle regeneration have been proposed to recapitulate those observed during embryogenesis. The ECM has multifunctional roles during embryogenesis and growth of various tissues (15, 17, 20). To further confirm and extend the microarray results, we evaluated the spatial temporal expression pattern of both periostin and biglycan using in situ hybridization techniques. These two transcripts were chosen for future study, as these were dynamically regulated during the muscle injury/regeneration process by microarray analysis.

Periostin was expressed early during murine embryogenesis. In the E9.5 embryo, periostin was restricted to the cardiac cushion, vascular outflow tract, amnion, vitelline artery, and vein (Fig. 4A). Persistent expression was observed in the cardiac cushion, outflow tract, peripheral nerves (but not dorsal root ganglia), head-fold mesenchyme, and the periaortic mesenchyme at E11.5 (Fig. 4B). The expression pattern broadens during the midgestational period (E13.5E15.5) to primarily include the mesenchyme as well as the skin, perichondrium, endocardium, perinephric stroma, choroid plexus, and diaphragm (Fig. 4, CG). These results support the conclusion that periostin is expressed principally in mesenchymal tissues.



View larger version (189K):
[in this window]
[in a new window]
 
Fig. 4. Expression of periostin during murine embryogenesis. A: using in situ hybridization, we observed periostin expression in the cardiac cushion (arrowhead), amnion, vitelline artery, and vein in the E9.5 embryo. B: parasagittal section of an E11.5 embryo revealing expression of periostin in the head-fold mesenchyme, cardiac cushion (arrowhead), and outflow tract, peripheral nerves, and the periaortic mesenchyme. C: periostin expression is present within the skin, choroid plexus, diaphragm, cardiac cushion, and mesenchyme in the E13.5 embryo. D: at E15.5 persistent periostin expression is observed within the diaphragm (E and G) (arrowhead), perichondrium (E and G) (open arrowhead), and perichondrium of the developing foot (F) (arrowhead). Scale bar for A = 200 µm; for B and G = 800 µm; for C and D = 2 mm; and for E and F = 800 µm.

 
Periostin was absent in quiescent adult skeletal muscle. In response to cardiotoxin injury, low levels of periostin were localized to the ECM within 48 h of the injury (not shown). Robust expression was observed throughout the regenerating hindlimb skeletal muscle between 5 and 7 days following injury associated with fibroblasts, macrophages, and the ECM (Fig. 5, A and B). During this time period, low levels of signal were present within the myofibers (Fig. 5B). Expression was largely extinguished 14 days following injury, corresponding to reestablishment of the skeletal muscle architecture (Fig. 5C). The expression of periostin using in situ hybridization techniques further confirms the array analysis (see Table 1). Low level of periostin expression was observed in the adult wild-type heart (not shown). In contrast, focal but robust expression of periostin was observed within the interstitial fibroblasts and inflammatory cells of the myopathic diaphragm (Fig. 5D) and heart (Fig. 5E) of the dystrophin-deficient mdx mouse model. The dystrophin mutant mdx mouse model characteristically has persistent injury and repair cycles involving both myocardial and skeletal muscles and serves as transgenic mouse model that parallels the cardiotoxin injury model (described above).



View larger version (109K):
[in this window]
[in a new window]
 
Fig. 5. Expression of periostin during muscle regeneration and in heart. A: 5 days following injury, widespread periostin expression is observed throughout the regenerating hindlimb. B: higher magnification of regenerating skeletal muscle at 7 days following injury reveals that periostin is predominantly localized to the interstitium, with low levels associated with the myotubes. C: 14 days following injury, periostin expression is largely absent. Focal periostin expression is observed in the interstitium associated with focal regions of regenerating muscle in the mdx diaphragm (D) (arrowheads) and focal regions of inflammation in the mdx heart (E). Scale bars for A and C = 1 mm; for B, D, and E = 100 µm.

 
The spatial and temporal expression pattern for biglycan approximately parallels that of periostin, with several exceptions. Using in situ hybridization techniques, we observed biglycan expression in the endocardium of the developing heart and the vascular endothelium in the E8.5E9.5 mouse embryos (Fig. 6, A and B). In the E9.5 mouse embryo, biglycan expression was also observed in the mesenchyme of the head and branchial arches (Fig. 6B). Persistent expression of biglycan was observed in the vasculature (i.e., meningeal and cerebral vasculature), pericardium, endocardium, and the perichondria at E11.5 (Fig. 6C). At midgestational ages (E13.5E15.5) expression was observed in the encapsulating epithelial layers (i.e., pericardium, skin, hepatic capsule, etc.) with persistent expression in mesenchymal tissues (testis, kidney, and pancreas), the smooth muscle of the vasculature (i.e., aorta and pulmonary vasculature), and the choroid plexus (Fig. 6, DJ). Robust expression was observed within the tongue, diaphragm, cardiac cushion, proliferating chondrocytes, and the perichondrium (Fig. 6, DF and HJ). Low levels were observed throughout the myocardium of the developing heart (Fig. 6, DF).



View larger version (138K):
[in this window]
[in a new window]
 
Fig. 6. Biglycan expression during mouse embryogenesis. A: using in situ hybridization, we observed biglycan expression early during murine development (E8.5) in the vasculature, endocardium, and mesenchyme. B: at E9.5 persistent biglycan expression is observed in endocardium (arrowhead), pericardium, meningeal vasculature, and the mesenchyme (open arrowhead). C: E11.5 mouse embryo revealing persistent expression of biglycan in the endocardium, pericardium, cerebral vasculature, head mesenchyme, diaphragm, and the mesenchyme, which will contribute to the vertebrae (arrowhead). D: low magnification of an E13.5 embryo revealing robust biglycan expression in the endocardium, pericardium, cardiac cushion, smooth muscle of the aorta, diaphragm, skin (arrowhead), perichondrium, and mesenchyme. E: low magnification of an E15.5 embryo illustrating biglycan expression in mesenchyme of pancreas and testis, choroid plexus, serosal surfaces, endocardium, and pericardium. F: high magnification at E13.5 days gestation illustrating expression in the smooth muscle of the aorta (open arrowhead), cardiac cushion (arrowhead), and mesenchyme associated with the diaphragm (d). G: high magnification of the E13.5 revealing expression within the endothelium of the head vasculature (arrowhead) and the mesenchyme associated with the diaphragm (H) (d, diaphragm). I: high magnification of the E15.5 embryo revealing expression within the proliferating chondrocytes (arrowhead) and the mesenchyme of the testis (J). Scale bars for A and F = 200 µm; for B = 500 µm; for C = 800 µm; for D and E = 2 mm; for GJ = 100 µm.

 
Biglycan expression during muscle regeneration parallels that of periostin with biglycan absent in unperturbed skeletal muscle (Fig. 7A). Low signal intensity was associated with the ECM within 48 h of cardiotoxin-induced muscle injury (not shown). Robust expression of biglycan was observed 5–7 days (Fig. 7, B and C) during the regenerative process, and absence of signal was observed in the regenerated tissue at 14 days (Fig. 7D). Expression throughout the repair process appeared to be concentrated in the perimyofiber region associated with the ECM (i.e., fibroblasts). Biglycan expression was observed within the interstitium and coronary arteries of the adult heart (Fig. 7E) (corresponding to injured myocardium) and diaphragm (Fig. 7F) in the mdx mouse model.



View larger version (148K):
[in this window]
[in a new window]
 
Fig. 7. Expression of biglycan during muscle regeneration and in heart. A: absence of biglycan expression is observed at 12 h following cardiotoxin-induced muscle injury. B: 5 days following injury, widespread expression is observed throughout the regenerating skeletal muscle. C: higher magnification of regenerating skeletal muscle at 7 days following injury reveals that periostin is predominantly localized to the extracellular matrix (ECM) with low levels associated with the myotubes (asterisks). D: 14 days following injury, periostin expression is largely absent. E: higher magnification of the mdx adult mouse heart reveals robust biglycan expression observed in the coronary vessel (asterisk) endothelium (arrowhead) and the myocardial interstitium associated with focal regions of inflammation. F: focal regions of biglycan expression are observed in the interstitium (arrowhead) of the regenerating adult mouse mdx diaphragm. Scale bars for A, B, and D = 1 mm; for C, E, and F = 100 µm.

 
TGFB1 modulates biglycan and periostin expression.
C2C12 myoblasts (derived from the satellite cell population) and NIH 3T3 fibroblast cell lines were cultured in the presence and absence of TGFB1. The concentration and selection of TGFB1 was chosen based on the finding that this factor induced periostin expression in osteoblast cells (13). Using primers that span an intron and semi-quantitative RT-PCR, we found that both periostin and biglycan were expressed in C2C12 myoblasts and differentiated myotubes. Induction of periostin expression was observed in myoblasts exposed to TGFB1 (performed in triplicate) (Fig. 8). In contrast, only biglycan was expressed in NIH 3T3 fibroblasts, and expression was induced by TGFB1 exposure (Fig. 8). These in vitro studies further complement both the microarray results and the morphological analyses using in situ hybridization. These studies are consistent with the hypothesis that TGFB1 is an important regulator of the ECM.



View larger version (11K):
[in this window]
[in a new window]
 
Fig. 8. Expression of periostin and biglycan in C2C12 myoblasts and NIH 3T3 fibroblasts. Using semiquantitative RT-PCR analysis and primers which span an intron, we evaluated expression of periostin and biglycan in the presence (T+) and absence (T-) of TGFB1 (5 ng/ml) in selected cell populations. Periostin and biglycan were expressed in C2C12 myoblasts (GM, growth media) and myotubes (DM, differentiation media). Exposure of myoblasts to TGFB1 resulted in an induction of periostin with no significant changes observed in biglycan expression during C2C12 differentiation. In contrast, biglycan expression was induced by TGFB1 exposure in NIH 3T3 cells. Note that an absence of periostin expression was observed in NIH 3T3 cells. No signal was detected in the (-)RT lane. Ribosomal RNA (18S) transcript was used as a loading control.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
Skeletal muscle has a remarkable regenerative capacity in response to an extensive injury. Resident within adult skeletal muscle is a small population of myogenic precursor cells or satellite cells that are capable of multiple rounds of proliferation (estimated at 80–100 doublings) and are able to reestablish a quiescent pool of myogenic progenitor cells (i.e., self-renewal) following each discrete regenerative episode (12, 16, 23, 24). Although muscle regeneration is a highly efficient and reproducible process, it ultimately is exhausted, as observed in senescent skeletal muscle or in patients with muscular dystrophy (5, 10, 12). An enhanced understanding of the molecular program that is associated with injury and directs this regenerative process would promote therapies directed toward the treatment of debilitating myopathies.

In the present study, we pursued a comprehensive cellular and molecular analysis of the remodeling process following chemical-induced muscle injury. These results clearly establish the presence of phases associated with the inflammatory process, myogenic progenitor cell activation, and proliferation, myogenic differentiation, growth factor, or mitogenic stimuli and the contribution of the ECM. We further examined two candidate genes in detail to define the contribution of the biomatrix during muscle development and injury/regeneration.

Molecular mechanisms of muscle injury and regeneration.
Analysis of other vertebrate tissues that display increased regenerative capacity emphasizes common or shared molecular and cellular programs that characterize the injury-repair process and underscores the importance of the ECM in this process (15, 17, 20). Regeneration of hepatic parenchymal tissue following an extensive injury or partial hepatectomy highlights stages of the repair process, which occurs over a 2- to 3-wk period. The acute phase following an extensive hepatic injury is characterized by the induction of an immediate early gene program (Fos, Jun, JAK/STAT, NF-{kappa}B, etc.), a pronounced inflammatory response with the release of cytokines and mitogenic signals, which concludes with the activation and increased proliferative capacity of hepatic precursor cell populations (i.e., oval cells). Growth factors (Igf1, Igf2, Hgf, Fgf1, Tgfb1) in combination with cytokines (Il1, Il6, Tnf) coordinately promote cellular proliferation and the remodeling of the hepatic matrix composition which characterize the second stage of hepatic repair (17). Hepatic differentiation and matrix maturation resulting in preserved architecture characterize the final stages of the repair process. Notably, regeneration of hepatic tissue undergoes similar discrete phases of regeneration as observed in the present study associated with regenerating skeletal muscle.

Intense interest has focused on the regulatory mechanisms of the hepatic precursor cell population and the interaction of key signals (i.e., Hgf, Fgf, and Tgfb1) during these stages, but the signals that regulate the termination of the regenerative response are incompletely defined (15, 17, 20). One candidate factor that may initiate the termination signal for liver regeneration is Tgfb1, as it has been previously shown to inhibit hepatocyte proliferation and modulate the ECM of the regenerating liver. Although limited information is available regarding the termination signal for regeneration, it is highly probable that multiple factors (growth factors, cytokines, and ECM restoration) may deliver in aggregate a set of signals resulting in the termination of the repair process (17). A recent report further examined the molecular response during bone repair and complements the studies of hepatic regeneration and our studies of muscle regeneration. Specifically, transcriptional changes associated with bone formation and the ECM follow a similar expression profile observed in the present study of muscle regeneration, illustrating the stage-specific and essential role of the ECM during regeneration of different tissues. Although the molecular analysis utilized in the present study was performed on whole skeletal muscle tissue, the expression profiles of cell cycle, ECM, and muscle differentiation genes are consistent with array analysis performed on differentiating C2C12 myoblasts (18). The current study further focused on two ECM proteins that were robustly expressed during the muscle repair process.

Remodeling of the ECM during muscle injury and regeneration.
Periostin (also named "osteoblast-specific factor 2," or Osf2) was isolated using differential screening techniques and previously believed to be restricted to bone (26). Previous studies reported that periostin was a secreted protein that functioned in osteoblast recruitment, attachment, and spreading (13). Moreover, periostin has been shown to be regulated by Tgfb1 in osteoblast cells (13). More recently, embryological expression studies of periostin revealed that it was also expressed in the cardiac cushion (14). In the present study, we observed periostin to be robustly upregulated during the muscle repair process (>80-fold upregulation) using oligonucleotide array analysis.

Using in situ hybridization techniques, we observed periostin to be more broadly expressed in the head mesenchyme, periaortic mesenchyme, diaphragm, periosteum, skin, and cardiac cushion during mouse embryogenesis. Furthermore, periostin is robustly expressed in the remodeling ECM following muscle injury and in the regenerating diaphragm and heart of the mdx mouse model. In addition, we further establish that periostin is expressed in C2C12 myoblasts and regulated in this cell line by TGFB1. TGFB1 is a multifunctional, contextual cytokine that has previously been reported to function in modulating the tissue repair response through the regulation of the fibroblast cell population and matrix maturation as well as a regulator of cell migration and chemotaxis in regenerating tissues such as the skin (15). Further studies will focus on the complex interaction between growth factors and the biological role of periostin as important regulators of muscle regeneration.

Biglycan is an example of a secreted multifunctional leucine-rich proteoglycan that functions in matrix organization and the modulation of various growth factors (28, 29). Biglycan exhibits predominantly a pericellular location, and via its two glycosaminoglycan chains binds to {alpha}-dystroglycan located in the skeletal muscle plasmalemma (2). The implications of the interaction of biglycan with {alpha}-dystroglycan have not been elucidated, but one possibility is that it may serve to localize signaling molecules to the plasmalemma. For example, biglycan has been observed to bind growth factors or cytokines (i.e., Tgfb1, Tnf, etc.) and may modulate the presentation or the sequestration (i.e., inactivation) of these growth factors, thereby affecting muscle and/or matrix remodeling (28, 29). Additionally, biglycan may function in the matrix assembly through the binding of this proteoglycan to the collagen network (i.e., collagen VI and possibly collagen type I) (22, 30) or the elastin microfibril network (i.e., tropoelastin) (21). Furthermore, studies utilizing gene disruption strategies report that biglycan-deficient mice have perturbed bone remodeling as they manifest an osteoporosis-like phenotype characterized by decreased bone mass, suggesting that biglycan is a positive regulator of bone formation (31). Future studies using the biglycan-deficient mouse model will be of interest to further define a broader role for this proteoglycan during muscle regeneration and as a direct or indirect regulator of the myogenic progenitor cell population.

In the present study, we establish that biglycan is expressed in embryonic mesenchymal tissues early during mouse development and has an overlapping expression pattern with periostin in the embryo at midgestational ages. Similarly, biglycan is robustly upregulated during muscle regeneration and in the dystrophic mdx diaphragm and heart consistent with its role in regeneration and muscle remodeling. Using microarray analysis, we observed that Tgfb1 expression precedes that of biglycan and induces expression in the NIH 3T3 fibroblast cell line. These results further establish the dynamic role of the ECM during development and regeneration and underscores the coordinated regulation of myogenesis by growth factors and the ECM for efficient repair during the regenerative process.

In summary, the results of these studies underscore the molecular complexity associated with muscle regeneration and the essential role of the ECM proteins. Furthermore, these studies provide a foundation for future experiments that will enhance our understanding of injury/regenerative mechanisms and have applications for the treatment of chronic debilitating diseases.


    DISCLOSURES
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 
This work was supported by National Institutes of Health Grant AR-47850 and by grants from the Muscular Dystrophy Association, Donald W. Reynolds Foundation, and March of Dimes Association.


    ACKNOWLEDGMENTS
 
We thank J. Shelton, J. Stark, C. Pomjazl, and D. Sutcliffe for assistance with the histological and in situ hybridization analyses. We also thank Drs. R. S. Williams and R. Bassel-Duby for helpful discussions throughout the course of these studies.

Present address of T. J. Hawke: School of Kinesiology and Health Science, York University, Toronto, Ontario, Canada M3J 1P3.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: D. J. Garry, NB11.118A, 5323 Harry Hines Blvd., UT Southwestern Medical Center, Dallas, TX 75390-8573 (E-mail: daniel.garry{at}utsouthwestern.edu).

10.1152/physiolgenomics.00056.2003.

1 The Supplementary Material for this article (a figure and a table) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00056.2003/DC1. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 DISCLOSURES
 REFERENCES
 

  1. Bassel-Duby R, Hernandez MD, Gonzalez MA, Krueger JK, and Williams RS. A 40-kilodalton protein binds specifically to an upstream sequence element essential for muscle-specific transcription of the human myoglobin promoter. Mol Cell Biol 12: 5024–5032, 1992.[Abstract]
  2. Bowe MA, Mendis DB, and Fallon JR. The small leucine-rich repeat proteoglycan biglycan binds to {alpha}-dystroglycan and is upregulated in dystrophic muscle. J Cell Biol 148: 801–810, 2000.[Abstract/Free Full Text]
  3. Cheng L and Wong WH. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci USA 98: 31–36, 2001.[Abstract/Free Full Text]
  4. Cheng L and Wong WH. Model-based analysis of oligonucleotide arrays: model validation, design issues and standard error application. Genome Biol 2: 1–11, 2001.
  5. Cossu G and Mavilio F. Myogenic stem cells for the therapy of primary myopathies: wishful thinking or therapeutic perspective? J Clin Invest 105: 1669–1674, 2000.[Free Full Text]
  6. Garry DJ, Bassel-Duby RS, Richardson JA, Grayson J, Neufer PD, and Williams RS. Postnatal development and plasticity of specialized muscle fiber characteristics in the hindlimb. Dev Genet 19: 146–156, 1996.[ISI][Medline]
  7. Garry DJ, Meeson A, Elterman J, Zhao Y, Yang P, Bassel-Duby R, and Williams RS. Myogenic stem cell function is impaired in mice lacking the forkhead/winged helix protein MNF. Proc Natl Acad Sci USA 97: 5416–5421, 2000.[Abstract/Free Full Text]
  8. Garry DJ, Yang Q, Bassel-Duby R, and Williams RS. Persistent expression of MNF identifies myogenic stem cells in postnatal muscles. Dev Biol 188: 280–294, 1997.[ISI][Medline]
  9. Grayson J, Williams RS, Yu YT, and Bassel-Duby R. Synergistic interactions between heterologous upstream activation elements and specific TATA sequences in a muscle-specific promoter. Mol Cell Biol 15: 1870–1878, 1995.[Abstract]
  10. Gussoni E, Blau HM, and Kunkel LM. The fate of individual myoblasts after transplantation into muscles of DMD patients. Nat Med 3: 970–977, 1997.[ISI][Medline]
  11. Hadjiargyrou M, Lombardo F, Zhao S, Ahrens W, Joo J, Ahn H, Jurman M, White DW, and Rubin CT. Transcriptional profiling of bone regeneration. Insight into the molecular complexity of wound repair. J Biol Chem 277: 30177–30182, 2002.[Abstract/Free Full Text]
  12. Hawke TJ and Garry DJ. Myogenic satellite cells: physiology to molecular biology. J Appl Physiol 91: 534–551, 2001.[Abstract/Free Full Text]
  13. Horiuchi K, Amizuka N, Takeshita S, Takamatsu H, Katsuura M, Ozawa H, Toyama Y, Bonewald LF, and Kudo A. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor ß. J Bone Miner Res 14: 1239–1249, 1999.[ISI][Medline]
  14. Kruzynska-Frejtag A, Machnicki M, Rogers R, Markwald RR, and Conway SJ. Periostin (an osteoblast-specific factor) is expressed within the embryonic mouse heart during valve formation. Mech Dev 103: 183–188, 2001.[ISI][Medline]
  15. Martin P. Wound healing: aiming for perfect skin regeneration. Science 276: 75–81, 1997.[Abstract/Free Full Text]
  16. Mauro A. Satellite cell of skeletal muscle fibers. J Biophys Biochem Cytol 9: 493–498, 1961.[Free Full Text]
  17. Michalopoulos GK and DeFrances MC. Liver regeneration. Science 276: 60–66, 1997.[Abstract/Free Full Text]
  18. Moran JL, Li Y, Hill AA, Mounts WM, and Miller CP. Gene expression changes during mouse skeletal myoblast differentiation revealed by transcriptional profiling. Physiol Genomics 10: 103–111, 2002. First published June 25, 2002; 10.1152/physiolgenomics.00011.2002.[Abstract/Free Full Text]
  19. Porter JD, Khanna S, Kaminski HJ, Rao JS, Merriam AP, Richmonds CR, Leahy P, Li J, Guo W, and Andrade FH. A chronic inflammatory response dominates the skeletal muscle molecular signature in dystrophin-deficient mdx mice. Hum Mol Genet 11: 263–272, 2002.[Abstract/Free Full Text]
  20. Reddi AH. Role of morphogenetic proteins in skeletal tissue engineering and regeneration. Nat Biotechnol 16: 247–252, 1998.[ISI][Medline]
  21. Reinboth B, Hanssen E, Cleary EG, and Gibson MA. Molecular interactions of biglycan and decorin with elastic fiber components: biglycan forms a ternary complex with tropoelastin and microfibril-associated glycoprotein 1. J Biol Chem 277: 3950–3957, 2002.[Abstract/Free Full Text]
  22. Schonherr E, Witsch-Prehm P, Harrach B, Robenek H, Rauterberg J, and Kresse H. Interaction of biglycan with type I collagen. J Biol Chem 270: 2776–2783, 1995.[Abstract/Free Full Text]
  23. Schultz E and McCormick KM. Skeletal muscle satellite cells. Rev Physiol Biochem Pharmacol 123: 213–257, 1994.[ISI][Medline]
  24. Seale P and Rudnicki MA. A new look at the origin, function, and "stem-cell" status of muscle satellite cells. Dev Biol 218: 115–124, 2000.[ISI][Medline]
  25. Shelton JM, Lee MH, Richardson JA, and Patel SB. Microsomal triglyceride transfer protein expression during mouse development. J Lipid Res 41: 532–537, 2000.[Abstract/Free Full Text]
  26. Takeshita S, Kikuno R, Tezuka K, and Amann E. Osteoblast-specific factor 2: cloning of a putative bone adhesion protein with homology with the insect protein fasciclin I. Biochem J 294: 271–278, 1993.[ISI][Medline]
  27. Tamayo P, Slonim D, Mesirov J, Zhu Q, Dmitrovsky E, Lander ES, and Golub TR. Interpreting gene expression with self-organizing maps: Methods and Application to hematopoietic differentiation. Proc Natl Acad Sci USA 96: 2907–2912, 1999.[Abstract/Free Full Text]
  28. Tufvesson E and Westergren-Thorsson G. Tumour necrosis factor-{alpha} interacts with biglycan and decorin. FEBS Lett 530: 124–128, 2002.[ISI][Medline]
  29. Ungefroren H and Krull NB. Transcriptional regulation of the human biglycan gene. J Biol Chem 271: 15787–15795, 1996.[Abstract/Free Full Text]
  30. Wiberg C, Hedbom E, Khairullina A, Lamande SR, Oldberg A, Timpl R, Morgelin M, and Heinegard D. Biglycan and decorin bind close to the n-terminal region of the collagen VI triple helix. J Biol Chem 276: 18947–18952, 2001.[Abstract/Free Full Text]
  31. Xu T, Bianco P, Fisher LW, Longenecker G, Smith E, Goldstein S, Bonadio J, Boskey A, Heegaard AM, Sommer B, Satomura K, Dominguez P, Zhao C, Kulkarni AB, Robey PG, and Young MF. Targeted disruption of the biglycan gene leads to an osteoporosis-like phenotype in mice. Nat Genet 20: 78–82, 1998.[ISI][Medline]