Cell Physiology

Implication of the satellite cell in dystrophic muscle fibrosis: a self-perpetuating mechanism of collagen overproduction

Catherine Alexakis, Terence Partridge, George Bou-Gharios


Because of its mechanical function, skeletal muscle is heavily influenced by the composition of its extracellular matrix (ECM). Fibrosis generated by chronic damage, such as occurs in muscular dystrophies, is thus particularly disastrous in this tissue. Here, we examined the interrelationship between the muscle satellite cell and the production of collagen type I, a major component of fibrotic ECM, by using both C2C12, a satellite cell-derived cell line, and primary muscle satellite cells. In C2C12 cells, we found that expression of collagen type I mRNA decreases substantially during skeletal muscle differentiation. On a single-cell level, collagen type I and myogenin became mutually exclusive after 3 days in differentiation medium, whereas addition of collagen markedly suppressed differentiation of C2C12 cells. Primary cultures of satellite cells associated with isolated single fibers of the young (4 wk old) mdx dystrophic mouse and of C57BL/10ScSn wild-type controls expressed collagen type I and type III mRNA and protein. This pattern persisted in wild-type mice at all ages. But, curiously, in older (18-mo-old) mdx mice, although the myogenic cells continued to express type III collagen, type I expression became restricted to nonmyogenic cells. These cells typically constituted part of a cellular sheet surrounding the old mdx fibers. This combination of features strongly suggests that the progression to fibrosis in dystrophic muscle involves changes in the mechanisms controlling matrix production, which generates positive feedback that results in a reprogramming of myoblasts to a profibrotic function.

  • collagen type I
  • myogenin
  • muscle single fibers
  • Duchenne muscular dystrophy

muscular dystrophies are inherited disorders characterized by muscle degeneration and associated progressive wasting and weakness (9). The most common, Duchenne muscular dystrophy (DMD), affects 1/3,500 newborn males. It is caused by mutations in the dystrophin gene encoding a large protein that is key to the assembly of a cell-surface complex connecting the muscle fiber cytoskeleton to the overlying basal lamina. In the absence of dystrophin, muscle fibers undergo repeated cycles of degeneration accompanied by regeneration that involves extensive proliferation of muscle precursor cells but also a chronic inflammatory response and cytokine production. This inflammation is associated with biosynthesis of extracellular matrix (ECM) molecules, most notably collagens type I and type III, whose excessive accumulation in the interstitial space was the basis of the original description of this disease as a myosclerosis. Although these various ECM components are commonly assumed to provide a mere passive space-filling substrate to maintain the structural integrity of the muscle, several lines of evidence suggest that this process may play an active pathogenic role in DMD, in particular the demonstration in vitro that collagen affects the growth and differentiation of myoblasts (21, 25). It is commonly held that fibrosis in dystrophic muscles is driven by the repeated bouts of muscle-fiber degeneration and ensuing inflammation dominated by macrophages and T lymphocytes (17). Certainly, depletion of T cells is reported to diminish pathology in mice (28, 29), and muscles of mdx nu/nu mice develop significantly lower levels of fibrosis than immunocompetent mdx mice (19, 20). This invasion is accompanied by the secretion of cytokines, such as TGF-β, which is thought to drive the increased fibrosis in vitro (23) and in DMD patients and mdx mice (2, 12, 36).

Here, we set out to investigate further the relationship between collagen expression and myogenesis. Initially we studied the expression of collagen type I by undifferentiated and differentiated C2C12 myoblasts, a cell line derived from a mouse skeletal satellite cell (32). We found that a large proportion of the undifferentiated C2C12 cells express collagen type I, that this was lost as differentiation proceeded, and, furthermore, that addition of type I collagen to C2C12 cultures inhibits their differentiation. These findings suggest a potential role for collagen type I in the regulation of skeletal muscle cell differentiation. Similarly, in primary cultures we found that proliferating myoblasts of young mdx and wild-type mice produce both type I and III collagens. Surprisingly, although this pattern of expression was maintained in old wild-type myoblasts, in old dystrophic muscle they continued to produce type III collagen but cease to synthesize type I collagen, which was found to be produced exclusively by nonmyogenic cells. We suggest that the fibrotic process in DMD may involve a major element of positive feedback whereby collagen type I directly inhibits muscle regeneration and favors increased collagen production. Such an active self-perpetuating mechanism within the disease has major implications for therapeutic strategies.



C57BL/10ScSn and dystrophic (mdx) mice aged 1 mo or 18 mo were obtained from the Biological Services Unit at the Medical Research Council Clinical Sciences Centre. Care and treatment of the animals was in compliance with the Animals Scientific Procedures Act, 1986.

Cell culture and single-fiber preparation.

Single fibers were isolated from collagenase-digested extensor digitorum longus (EDL) muscle of 1-mo-old and 18-mo-old C57BL/10ScSn and mdx mice, as described by Rosenblatt et al. (24). Briefly, the EDL muscle was carefully removed and incubated in a solution of 0.2% (wt/vol) type I collagenase (Sigma, Dorset, UK) in DMEM (Life Technologies, Paisley, UK) supplemented with 2 mM l-glutamine (Sigma) and 1% penicillin/streptomycin solution (Sigma) at 36°C for 90 min with agitation. The muscle was transferred to a serum-coated plastic Petri dish containing DMEM, and muscle fibers were separated from the whole muscle by repeated pipetting with a wide-mouth Pasteur pipette. Muscle fibers were sequentially washed in four changes of DMEM to remove any traces of interstitial and endothelial cells. Clean individual muscle fibers were placed into eight-well chamber slides (Life Technologies). Before plating, the wells were coated with 0.1% Matrigel (14) (Universal Biologicals, London, UK), a solubilized basement membrane to which muscle fibers and mononucleate cells adhere and on which differentiation of myogenic cells is enhanced (16). The fibers were incubated in DMEM supplemented with 10% horse serum and 0.5% chick embryo extract at 37°C, 5% CO2. In these conditions, the satellite cells migrate from the fibers and divide.

C2C12 cell culture conditions.

Cell-culture experiments were performed by using C2C12 mouse myoblasts (an immortalized mouse myoblast cell line; ATCC, Rockville, MD). Cells were grown in DMEM supplemented with 10% FCS and 1% penicillin/streptomycin under a humidified atmosphere of 5% CO2 in air and were maintained at low confluence on uncoated plastic to avoid any effect of the substratum. Cell differentiation was induced at 70% confluence by using DMEM supplemented with 2% FCS.


Synchronization was performed as follows: 2,000 cells/cm2 were plated. After 24 h in proliferation medium, cells were rinsed in PBS and shifted to DMEM without methionine supplemented with 1% FCS for 36 h. Methionine deprivation arrests C2C12 in a quiescent and nondifferentiated state. Quiescent myoblasts (G0) were then incubated for 15 h in either differentiation or proliferation medium in the presence or absence of bromodeoxyuridine (BrdU; 0.01 mM). Cells were then fixed and processed for immunofluorescence analysis by using antibodies against BrdU, myogenin, and collagen type I.

Immunofluorescent staining.

Cultures were fixed in 4% paraformaldehyde in PBS for 10 min at 37°C, permeabilized with 0.1% Triton X-100 in PBS at room temperature, and then incubated for 30 min with 20% nonimmune normal serum of the animals of the species in which the second-layer antibody was raised to block nonspecific antibody binding. Mouse anti-human MyoD1 (1:50, clone 5.8a; Dako) or rabbit anti-MyoD (1:50; Santa Cruz), mouse anti-myogenin (1:50, F5D; Dako), mouse anti-Pax7 (1:10; DSHB, Iowa City, IA), goat anti-human collagen type I UNLB (1:100; Cambridge Bioscience), and rabbit anti-human collagen type III (1:50; Chemicon) antibodies were applied at 4°C overnight in 6% horse serum in PBS. A two-step detection method was performed. Primary mouse antibodies were detected with Alexa Fluor 488-conjugated goat anti-mouse (1:200; Molecular Probes), primary rabbit antibodies were detected with TRITC-conjugated swine anti-rabbit (1:200; Dako), and primary goat antibodies were detected with Alexa Fluor 488-conjugated donkey anti-goat (1:200; Molecular Probes). Secondary antibodies were applied for 1 h in 6% horse serum in PBS, and washes were carried out once with PBS + 0.025% Tween 20 followed by two washes with 6% horse serum in PBS. Negative controls were carried out 1) by adding secondary antibodies only with 6% horse serum in PBS, 2) by adding primary antibodies without their corresponding secondary antibodies, and 3) by replacing primary antibodies with an equivalent concentration of their corresponding species IgG.

RT-PCR and Northern blotting.

Total RNA was extracted by using Trizol reagent (Invitrogen). For RT-PCR analysis of C2C12, NIH/3T3 cells (used as a positive control), as well as nonpassaged cells of EDL single fibers of 1-mo-old and 18-mo-old control and mdx mice, 1 μg of total RNA was reverse transcribed in a 20 μl reaction by using SuperScript RNase H− reverse transcriptase (Life Technologies) primed with Oligo(dT)12–18. PCR was performed in a 25-μl PCR mixture containing 1× PCR buffer (Invitrogen), 0.1 mM of each dNTP (Invitrogen), 1 μM of specific primer, and 1 unit of Taq polymerase (Invitrogen). Primer sequences were as follows: α1(I)-procollagen, sense 5′-TCT CCA CTC TTC TAG TTC CT-3′, antisense 5′-TTG GGT CAT TTC CAC ATG C-3′, product length 269 bp; type III collagen, sense 5′-GGC TCC TGG TGA AGC GAG GAC G-3′, antisense 5′-CCT CTT GGA CCA CGT TTA CCC-3′, product length 522 bp, and hypoxanthine guanine phosphoribosyl transferase (HPRT), sense 5′-TCA GTC AAC GGG GGA CAT AA-3′, antisense 5′AGT CTG GGG ACG CAG CAA CT-3′, product length 503 bp. Amplification was performed with 35 cycles of 94°C for 1 min, 60°C for 1 min (α1(I)-procollagen) or 58°C for 1 min (type III collagen), and 72°C for 1 min. The PCR products were separated by agarose gel electrophoresis, and the amplified DNA fragments were visualized by ethidium bromide staining. For Northern blotting, 20 μg of C2C12 total RNA was denatured in glyoxal and fractionated on a 1.5% agarose gel in 10 mM sodium phosphate (pH 7). The gel was transferred onto Hybond-N+ nylon membrane (Amersham) and was hybridized according to the manufacturer's instructions with a α-[32P]ATP-labeled Col1α2 probe and HPRT. Levels of collagen type I mRNA and collagen type III mRNA were calculated relative to transcript for HPRT. PCR products and Northern blots were quantified by using ImageJ software (National Institutes of Health, Bethesda, MD; http://rsb.info.nih.gov/ij/).

Collagen type I dose response.

C2C12 were incubated for 2 days in proliferation medium, then switched to differentiation medium containing dilutions of collagen type I (C3511; Sigma) from 0.02 to 20 μg/ml. Myogenin expression was then assessed by immunocytochemistry as described above after 48 h. Myogenin-positive cells per field were counted.


Results are expressed as the mean ± SE of the number of cells in culture from mdx and C57BL/10ScSn mice. The one-way ANOVA test was performed, and a posttest using the Tukey test was used for all comparisons. Significance was defined as P < 0.05.


Collagen type I is expressed by undifferentiated C2C12 myoblasts.

Expression of collagen type I protein examined in undifferentiated C2C12 cultures by immunofluorescence staining (Fig. 1A) demonstrated that ∼50% of cells express collagen type I (total cell number 181.1 ± 8.3, collagen-positive cells 96 ± 5.5, 5 fields per well, 9 wells). Undifferentiated C2C12 cells also express collagen type I mRNA, as does a positive control NIH/3T3 fibroblast cell line (Fig. 1B). We then compared collagen type I immunoreactivity between undifferentiated and differentiated C2C12 cells (Fig. 1, C and D). Only single cells surrounding newly formed myotubes expressed collagen type I, whereas multinucleated desmin-positive myotubes showed no collagen type I staining. We further assessed the distribution of collagen type I during C2C12 differentiation by immunofluorescence staining using myogenin as a marker of differentiation. After 3 days in differentiation medium (2% FCS), we found that collagen type I and myogenin staining became mutually exclusive (Fig. 1E), with <0.2% of cells coexpressing collagen type I and myogenin (1/425 cells). This accorded with Northern blot analysis of collagen type I mRNA levels, revealing that collagen type I mRNA expression is significantly decreased by the seventh day of differentiation (Fig. 1F).

Fig. 1.

A: expression of collagen type I (green) in C2C12 cell culture in proliferation medium. B: RT-PCR of collagen type I expression of C2C12 and NIH/3T3 (positive control) cells. C and D: coimmunofluorescence analysis of collagen type I (green) and desmin (red) in exponentially growing C2C12 (C) and in differentiated C2C12 (D) after 5 days in differentiation medium. E: coimmunofluorescence analysis of collagen type I (green) and myogenin (red) in differentiated C2C12 after 2 days in differentiation medium. F: 20 μg of total RNA isolated from exponentially growing C2C12 (undiff C2) and from C2C12 after 3 (d3), 5 (d5), and 7 (d7) days of differentiation were separated electrophoretically, blotted into nylon membranes, and hybridized with a 32P-labeled cDNA probe for α2(I)-collagen. The same membranes were hybridized with a HPRT probe. Northern blots (determinations of 3 different mRNA preparations) were quantified. *P < 0.05. Bar = 50 μm.

Quiescent myoblasts reentering proliferation specifically express collagen type I, whereas those undergoing differentiation do not.

We used methionine deprivation, which has been shown to block C2C12 myoblast proliferation without inducing differentiation, to examine collagen expression in synchronized myoblasts. On return to complete proliferation medium for 15 h, quiescent myoblasts reentering proliferation incorporate BrdU and express collagen type I (Fig. 2, A and B) but not the differentiation marker myogenin (Fig. 2, C and D). Conversely, methionine deprivation-induced quiescent myoblasts, when placed in low-serum differentiation medium, do not incorporate BrdU (Fig. 2, E and G) but express myogenin (Fig. 2H) and no longer stain for collagen type I (Fig. 2F).

Fig. 2.

C2C12 cells were cultured for 36 h in methionine-depleted medium + 1% FCS, then incubated for 15 h in either proliferation medium (AD) or differentiation medium (EH; see materials and methods) for 15 h in presence of 0.01 mM bromodeoxyuridine (BrdU). Cells were then fixed and processed for immunofluorescence analysis by using antibodies against BrdU (A, C, E, G) and collagen type I (B and F) or myogenin (D and H). Bar = 50 μm.

Collagen type I is expressed by most CD34-positive C2C12 cells.

CD34 is an established marker of hematopoietic stem cells and has also been shown to identify quiescent or “reserve” myoblasts and quiescent satellite cells (1). To investigate whether collagen type I is also a marker of undifferentiated myogenic cells, we costained C2C12 cells for collagen type I and CD34 following 48 h in differentiation medium. This procedure segregates the C2C12 cells into two populations, one of differentiating myoblasts and another of CD34-positive quiescent reserve cells. Following this treatment, we found that ∼70% of the collagen type I-positive C2C12 cells (Fig. 3A) are also CD34 positive (Fig. 3B).

Fig. 3.

Immunofluorescence analysis of collagen type I (A) and CD34 (B) in differentiated C2C12 after 2 days in differentiation medium. Bar = 50 μm.

Collagen type I added to the culture medium inhibits C2C12 differentiation.

We subsequently examined whether collagen type I expression by C2C12 cells might play an active role in preventing myogenic differentiation. C2C12 were cultured for 2 days in proliferation medium and then switched to differentiation medium containing various concentrations of collagen type I for a further 2 days. Comparison of the number of myogenin-immunopositive C2C12 cells after 48 h revealed a significant inverse relationship with a concentration of collagen type I >2 μg/ml (Fig. 4), with a twofold reduction in differentiated cells when 20 μg/ml of collagen type I was added to the medium.

Fig. 4.

C2C12 were incubated for 2 days in proliferation medium, then switched to differentiation medium containing dilutions of collagen type I (C3511; Sigma) from 0.002 to 20 μg/ml. Myogenin expression was assessed by counting myogenin-positive cells per field after 48 h in culture *P < 0.05 and **P < 0.01 compared with control (differentiation medium alone).

Collagen types I and III are expressed by mononucleated myogenic cells that emerge from young control and mdx isolated myofibers.

We then looked at collagen synthesis by myogenic cells emerging from isolated muscle fibers. This approach was more representative of the fibrosis that occurred in the muscle in vivo. Isolated single EDL muscle-fiber explants were cultured for 72 h on Matrigel. The fibers and the mononucleated cells that emerged from the fibers were fixed and immunostained for collagen type I and myoblast markers. To allow unequivocal identification of myogenic cells, we used a combination of the three nuclear-expressed markers Pax7, MyoD, and myogenin. This multimarker approach was crucial for the present study, since the previously used marker, desmin, is also expressed by some myofibroblast-like cells (32). Any cell positive for at least one of our three markers would be considered a myogenic cell. We found that all emigrating myogenic cells from 1-mo-old C57BL/10ScSn or from 1-mo-old mdx fibers express both collagen type I (Fig. 5, A and C) and collagen type III (Fig. 5, B and D).

Fig. 5.

Expression of myoblast markers (Pax7, MyoD, and myogenin) and collagen type I. One-month-old C57BL/10ScSn muscle fibers (A and B) and mdx muscle fibers (C and D) in culture immunostained for expression of myoblast markers (red) and collagen type I (green) (A and C) and for expression of myoblast markers (green) and collagen type III (red) (B and D). Eighteen-month-old C57BL/10ScSn muscle fibers (E and F) and mdx muscle fibers (G and H) in culture immunostained for expression of myoblast markers (red) collagen type I (green) (E and G) and for expression of myoblast markers (green) and collagen type III (red) (F and H). I: RT-PCR of collagen type I and collagen type III expression of cells migrating from 1-mo-old and 18-mo-old C57BL/10ScSn and mdx muscle single fibers from 72 h in culture to 184 h in culture. Only gels where collagen type I is expressed are shown, with HPRT as control housekeeping gene. Collagen type III is expressed at all time points. Levels of collagen type I mRNA and collagen type III mRNA were calculated relative to transcript for HPRT and noted. Bar = 50 μm.

Two distinct cell populations emigrate from old mdx fibers.

We found that cells emerging from 18-mo-old mdx fibers consist of two populations. One is defined as myogenic by being immunopositive for at least one of the three myoblast markers we used and does not coexpresses collagen type I. The second is not stained by any of the myoblast markers but is strongly immunopositive for collagen type I (Fig. 5G). Neither of these phenotypes was found in cells emigrating from 18-mo-old C57BL/10ScSn fibers; in all cells, myogenic markers and immunopositivity for collagen type I (Fig. 5E) were co-expressed. In contrast, collagen type III is expressed by all emigrating myogenic cells from 18-mo-old fibers of both C57BL/10ScSn and mdx muscles (Fig. 5, F and H).

To investigate the expression of collagen type I and collagen type III mRNA by cells emigrating from muscle fibers, total RNA from primary myogenic cultures of young control and mdx mice were analyzed by semiquantitative RT-PCR. We found an elevated expression of collagen type I mRNA in the cells emigrating from both 1-mo-old and 18-mo-old mdx fibers compared with the corresponding wild-type controls (Fig. 5I). Interestingly, collagen type I mRNA expression was detectable within 72 h in vitro in cells emigrating from 1-mo-old mdx fibers, whereas in wild-type controls collagen type I mRNA was only detected after 96 h in culture. No differences were detected in the time course of collagen type III expression from cells emigrating from young and old control and mdx fibers (Fig. 5I).

Accumulation in vitro of nonmyogenic cells from 18-mo-old isolated mdx muscle fibers.

Using combined expression of Pax7, MyoD, and myogenin, we assessed the myogenicity of the cells present in isolated myofiber cultures. We found that most cells emigrating from young control and young mdx fibers are positive for at least one of the myoblast markers we have used (Fig. 6A). In contrast, we found that only ∼60% of the cells emerging from 18-mo-old mdx fibers express a myogenic marker (446/792 cells). Cells were also immunostained for each myoblast marker individually and were counted (Fig. 6B).

Fig. 6.

A: myogenicity of cells from 72-h cultured single muscle fibers from extensor digitorum longus (EDL) muscles of 1-mo-old and 18-mo-old C57BL/10ScSn and mdx mice. Number of cells expressing myoblast markers (MyoD, Pax7, and/or myogenin) are expressed as a mean percentage of total number of cells present around fibers. B: MyoD and Pax7 expression by cells emerging from 18-mo-old mdx mice. MyoD and Pax7 are expressed as a mean percentage of total number of cells present around fibers. ***P < 0.001.

Presence of a “cellular sheet” surrounding old myofibers.

We observed fragments of amorphous interstitial material intimately associated with the old mdx myofibers during the isolation process (Fig. 7A). This membrane-like structure was also observed around myofibers of mdx mice as young as 12 mo, whereas in C57BL/10ScSn mice it was seen only around very old myofibers (≥22 mo). This structure was not removed by the collagenase treatment used to produce isolated fibers and remains intimately associated with individual muscle fibers. DAPI staining showed the presence of cell nuclei within this membrane (Fig. 7B). Even in the absence of an appreciable surrounding membrane, most old mdx fibers demonstrated surface-adherent cells (Fig. 7, C and D). These are immunopositive for collagen type I and are negative for all of the myoblast markers we used (Fig. 7, E and F). Confocal microscopy of 18-mo-old mdx fibers revealed cells localized at the surface of the fiber, beneath the surrounding collagen-rich membrane (Fig. 7, H, I, and J). Fibers from old mdx mice often adopted a contorted shape in culture, and foci of contortion were commonly associated with adherent membranes.

Fig. 7.

Light micrograph of 18-mo-old EDL muscle single-fiber preparations (A and C) and DAPI (B and D) and collagen costainings (E) showing membrane seen surrounding old fibers. C and D are light micrograph and DAPI staining, respectively, of an 18-mo-old EDL muscle fiber showing nonsatellite cells associated with the fiber. H, I, and J are 3 different levels of a fiber by confocal microscopy showing cells comprised between basal lamina and membrane surrounding 18-mo-old mdx single muscle fiber. Bar = 50 μm.


Although the primary cause of DMD is a mutation in the dystrophin gene that lowers the resilience of muscle fibers to work-associated damage, the more severe clinical features of this disease are associated with secondary pathological processes. Most notable is the overwhelming accumulation of matrix components in the interstitial space, notably in the endomysium, isolating the fibers from one another within fascicles and constituting the original pathological hallmark that characterized the disease as a myosclerosis (8). This pathological feature is also seen in the main animal models of DMD, the Golden Retriever muscular dystrophy dog and the mdx mouse (see Ref. 5 for review), and is thought, in all three, to be instigated by the relentless muscle fiber degeneration. It has been postulated that this abnormal accumulation of connective tissue is, of itself, a major factor in the progressive failure of muscle regeneration and ultimately of muscle dysfunction (15).

The main finding of this study is that the myogenic C2C12 cell line, derived from a mouse skeletal satellite cell (32), and primary myoblasts emerging from young muscle fibers express collagen type I. This functional duality is perhaps surprising, especially in the pure myogenic satellite cells isolated from young myofibers, which have been shown to be the stem cells of postnatal skeletal muscle (6), able both to provide the material for robust regeneration of damaged muscle and to self-replace so as to repopulate the newly formed muscle with a further generation of fully competent satellite cells (18). Satellite cell involvement in muscle regeneration includes the need for cell adhesion, motility, spreading, and anchorage-dependent growth (34), for all of which collagens serve as a substratum, so it is of particular significance that collagens type I and III are secreted by these satellite/stem cells.

Investigation of the relative participation of C2C12 myoblasts in collagen expression and myogenic differentiation revealed an inverse relationship between the two activities. Thus we show that the majority of C2C12 that express collagen type I do not express differentiation markers, whereas differentiation-committed myogenin-positive cells do not express collagen type I at the protein level. This is substantiated by the observation of a significant decrease in levels of collagen type I mRNA over the course of differentiation. These findings are in agreement with other published data showing that type I procollagen mRNA decreases during the differentiation of adipocytes (33), a cell type with a close lineage relationship to muscle (27). It has also been found that stable expression of FGF-6 in C2C12 significantly reduces myotube formation by myogenic cells (7), consistent with the idea that the acquisition of a “fibroblastic” phenotype shifts the molecular program away from myogenesis. It is of interest here to draw a comparison with the paired-box transcription factor family member Pax7 (26). Quiescent satellite cells express Pax7 and on activation coexpress MyoD (11, 31, 35), a key transcription factor required for commitment to myogenic differentiation (30). After a bout of proliferation, most activated satellite cells cease to express Pax7 as they differentiate. In this study, we have documented a similar relationship that excludes simultaneous expression of collagen type I and myogenic differentiation genes. Differentiation of C2C12 myoblasts and myotubes are coupled with reduction in transcription of collagen type I mRNA. This negative relationship between expression of collagen type I and entry into myogenic differentiation is further demonstrated by the experimental release of C2C12 cells from mitosis block induced by methionine deprivation, which has been shown previously to resemble naturally occurring quiescence in that it does not lead to direct entry into myogenic differentiation (13). On such release of quiescent myoblasts in growth medium, they reentered proliferation, incorporated BrdU, and expressed collagen type I but not myogenin. In contrast, transfer of freshly released cells to differentiation medium resulted in no BrdU incorporation and in expression of myogenin but not collagen type I. Thus myoblasts make the choice to undergo differentiation or express collagen type I but do not do both.

A further component of this switch between alternative fates is the finding that addition of collagen to cultures of C2C12 cells produced a dose-related drop in the proportion of cells entering myogenic differentiation. This has all the elements of a positive feedback loop whereby myogenic cells that produce collagen, rather than entering terminal satellite cell differentiation, act directly to bias potentially myogenic cells toward further production of collagen, thus directing the system as a whole toward fibrogenesis and away from myogenesis.

In differentiation medium, collagen type I was expressed in 70% of single CD34-positive C2C12 cells; these have been shown previously to become MyoD negative and not to fuse into myotubes, coming to resemble quiescent satellite cells (1). However, collagen type 1 is also expressed in proliferating cells and, in our study, is not expressed by satellite cells until the third or fourth day, suggesting that it is a feature of undifferentiated myogenic cells and may not be specifically associated with quiescence. It remains possible, however, that it is a factor predisposing the cells toward quiescence rather than terminal myogenic differentiation. A strong parallel is seen in the adipocyte, a cell type with a close lineage relationship to muscle (27), where a specific decrease in expression of the type I procollagen gene has been noted during differentiation (33).

Because C2C12 cells have been subjected to many years of selection for features that favor their propagation in tissue culture, we also examined collagen synthesis by satellite cells associated with freshly isolated muscle fibers, which are more relevant to fibrosis in vivo. The age-related increase in fibrosis is especially notable in muscles of the aging mdx as reported previously (19, 22). Preparations of young fibers, mdx or wild type, produced emerging cells that almost universally expressed myogenic markers and that, within a day or two, began to proliferate and to produce immunochemically identifiable collagen type I, thus establishing their strong comparability with C2C12 cells. Older fibers produced distinctly fewer such cells. In the case of wild-type fibers, these cells were again predominantly positive for both collagen and myogenic markers. Old mdx fibers, on the other hand, produced two distinct populations of cells: one devoid of any myogenic markers but strongly expressing collagen type I, the second expressing myogenic markers but no longer expressing collagen type I. This was accompanied by the presence around old dystrophic fibers of a sheet consisting of both amorphous and cellular material. This “cellular sheet” was seen around wild-type fibers only from much older animals (22-mo-old C57BL/10ScSn mice) but as early as 12 mo of age on mdx fibers. This rapid aging of the mdx myofiber may account for the small number of cells that emerge from an old dystrophic myofiber compared with young control and may explain the increase in nonmyogenic cells in aged mdx myofiber culture (4).

Intrigued by this apparent switch of phenotypes in old mdx myoblasts, we reinvestigated the single-fiber culture system for expression of collagen type III, which is the other major collagen component of the fibrotic tissue in dystrophic muscle. In satellite cells from both young and old fibers, wild-type or mdx, we found that the myogenic cells all stained positively for type III collagen. Thus in the older mdx myogenic cells, the loss of ability to synthesize type I collagen was specific and had become dissociated from the synthesis of type III collagen. This adds to the picture of a major dysregulation of gene expression in the satellite cell population of dystrophic muscle with continuation of expression of type III collagen but loss of type I expression exclusively in the older mdx muscle.

We had wished to address the remarkable discrepancy between the rapid onset of severe pathology, including conspicuous fibrotic change, in the diaphragm muscle of mdx mice and the much slower progression in the limb muscles. Unfortunately, however, we have not been able to isolate single fibers from the diaphragm with a consistency that would be required to make valid comparisons and so thus far have not been able to shed light on this important matter.

The formation of a sheet of collagen-producing cells around the myofibers provides a clear reflection of the fibrosis that characterizes this disease state. Even in normal muscle, the myofiber basal lamina becomes thicker and more rigid with age (3) and increased cross-linking of collagen molecules with age also renders the fibrils more resistant to degradation by collagenase (10). The dystrophic environment provides a range of inflammatory fibrogenic signals that may exacerbate this process. These, aided by the positive feedback mechanism we have identified here, would lead to the premature and exaggerated formation of this nondegraded connective tissue. Despite its prominent place in the clinical decline in DMD, this degradation of tissue structure and function is not addressed by any of the current gene or cell therapies. We propose that the fibrotic aspects of muscular dystrophy require much more attention if we are to achieve more than a stabilization of the disease state of DMD patients.


This study was supported by Aktion Benni & Co.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract