Muscle Regeneration by Bone Marrow-Derived Myogenic Progenitors

See allHide authors and affiliations

Science  06 Mar 1998:
Vol. 279, Issue 5356, pp. 1528-1530
DOI: 10.1126/science.279.5356.1528

This article has a correction. Please see:


Growth and repair of skeletal muscle are normally mediated by the satellite cells that surround muscle fibers. In regenerating muscle, however, the number of myogenic precursors exceeds that of resident satellite cells, implying migration or recruitment of undifferentiated progenitors from other sources. Transplantation of genetically marked bone marrow into immunodeficient mice revealed that marrow-derived cells migrate into areas of induced muscle degeneration, undergo myogenic differentiation, and participate in the regeneration of the damaged fibers. Genetically modified, marrow-derived myogenic progenitors could potentially be used to target therapeutic genes to muscle tissue, providing an alternative strategy for treatment of muscular dystrophies.

In postnatal life, growth and repair of skeletal muscle fibers are mediated by a resident population of mononuclear myogenic precursors, the satellite cells. These cells, which are located between the sarcolemma and the basal lamina of the muscle fiber, divide at a slow rate to sustain both self-renewal and growth of differentiated tissue (1). In response to muscle injury, or in individuals with chronic degenerative myopathies, satellite cells divide and fuse to repair or replace the damaged fibers. However, the self-renewal potential of adult satellite cells is limited, decreases with age (2), and can be exhausted by a chronic regenerative process such as that characteristic of severe muscular dystrophies, in which most muscle tissue is eventually lost and is replaced by connective tissue (3).

The number of resident satellite cells in adult muscle is much smaller than the number of committed myogenic precursors that populate the muscle tissue soon after an injury (4). Several explanations for this apparent paradox have been proposed, from migration of satellite cells from adjacent fibers, or even neighboring muscles, to recruitment to myogenesis of resident nonmyogenic cells such as fibroblasts or mesenchymal progenitors (5). Bone marrow (BM) stroma–derived mesenchymal cells, which serve as long-lasting precursors for bone, cartilage, and lung parenchyma in mice (6), can differentiate into contractile myotubes under certain conditions in vitro (7). However, recruitment to myogenesis of stroma-derived cells has not been observed in vivo (8).

To investigate whether BM cells can convert to myogenesis in response to physiological stimuli, we chemically induced muscle regeneration in the tibialis anterior (TA) of 10 immunodeficientscid/bg mice (9). Unfractionated BM cells (106 per muscle) obtained from the C57/MlacZtransgenic mouse line (10), in which a lacZ gene encoding a nuclear β-galactosidase (β-Gal) is under the control of the muscle-specific myosin light chain 3F promoter (the MLC3F-nlacZ transgene), were then injected into the damaged muscles. Expression of this transgene is restricted to cardiac and skeletal muscle in adult mice (11), although it can be activated in other cell types on induction of myogenic differentiation (12). Satellite cells were obtained from the same transgenic mice (13) and injected (5 × 105 cells per muscle) as a control in the contralateral leg of all recipient animals. TA muscles were examined at various times after injection (5 days to 5 weeks) for the presence of β-Gal–positive nuclei (14).

Whole-mount histochemical staining of a TA muscle 2 weeks after injection of total BM (Fig. 1A) revealed fibers containing β-Gal+ aligned nuclei similar to, although less numerous than, those observed in the contralateral leg injected with satellite cells (Fig. 1B). Transverse cryostat sections showed newly formed fibers with β-Gal+ centrally localized nuclei in four out of six mice at 2 to 5 weeks after injection of BM cells (Fig. 1C). No staining was observed in two mice analyzed after 5 days and one after 10 days. Conversely, centrally localized, β-Gal+ nuclei were observed in all muscles injected with satellite cells, as early as 5 days after injection (Fig.1D) (15).

Figure 1

Analysis of nuclear lacZexpression in whole-mount dissected fibers (A andB) or cryostat sections (C throughF) of regenerating TA muscles ofscid/bg mice. Mice were injected with unfractionated (A and C), adherent (E), or nonadherent (F) BM cells, or with control satellite cells (B and D), from C57/MlacZ transgenic mice. (A and B) Bright field; scale bars, 50 μm. (C through F) Nomarski optics; scale bars, 10 μm.

In a second series of experiments, BM from C57/MlacZmice was fractionated in vitro into adherent and nonadherent components (16). These components were then separately injected into the regenerating TA muscles of 15 scid/bg mice, and the mice were analyzed after 1, 2, and 6 weeks. β-Gal+ nuclei were observed in six mice injected with adherent cells and in three injected with the nonadherent fraction at 2 to 6 weeks after injection (Fig. 1, E and F). Activation of the MLC3F-nlacZ transgene in BM or blood cells was never observed in a nonmuscle environment in vitro or in vivo (for example, in inflammatory cells elicited by the intraperitoneal injection of thioglycollate). Thus, a population of cells within the BM entered a myogenic differentiation pathway when exposed to a regenerating muscle environment and actively participated in the formation of new muscle fibers.

To test whether myogenic progenitors could be physiologically recruited from BM and access a site of muscle regeneration from the peripheral circulation, we transplanted genetically marked BM cells from the C57/MlacZ line (H-2b) into 12 irradiated scid/bg mice (H-2d) (17). Five weeks after BM transplantation, muscle regeneration was induced in both TA muscles of nine surviving mice (18). All mice were examined 1, 2, and 3 weeks after induction of muscle damage for reconstitution of both the immune and nonimmune components of the hematopoietic system by analysis of the morphology and phenotype of BM, spleen, and peripheral blood cells. Flow cytometric analysis of peripheral blood nucleated cells revealed that all transplanted animals possessed a circulating lymphocyte population; such a population was virtually absent inscid/bg animals (Fig. 2A). CD4 and CD8 marker analysis confirmed that mature lymphocytes were present in a proportion (34.0 ± 2.9 and 15.2 ± 1.9%, respectively) similar to that of normal donors (Fig. 2B). Analysis of the H-2b (donor) haplotype in the total nucleated cell population of engrafted scid/bg mice revealed that all such cells were donor-derived (Fig. 2C). Virtually complete chimerism (80 to 90%) was also apparent in BM and spleen cells (19).

Figure 2

Flow cytometric analysis of peripheral blood nucleated cells from a scid/bg mouse 8 weeks after transplantation with BM from C57/MlacZ mice (scid/bg + BMT). (A) Forward (FSC) and side (SSC) scatter plots. The gated lymphocyte (Ly) and total (T) cell populations are shown. (B) Lymphocyte population stained with fluorescein isothiocyanate (FITC)–conjugated antibodies to CD4 (x axis) and phycoerythrin (PE)–conjugated antibodies to CD8 (yaxis). (C) Total cell population stained with FITC-conjugated antibodies to H-2KbDb. Log-scale fluorescence is shown on both axes in (B), and on thex axis in (C). Analysis of cells from untransplantedscid/bg mice and immunocompetent, donor C57/MlacZ mice is shown in the left and right panels, respectively. Rel., relative.

Regeneration was analyzed histochemically in the TA muscles of all transplanted mice. Transverse cryostat sections showed regenerating fibers containing β-Gal+ nuclei in five of six reconstituted animals analyzed at 2 and 3 weeks after induction of muscle injury (Fig. 3, B through D) (20). Hoechst nuclear staining showed that β-Gal+ nuclei were present both in immature centrally nucleated fibers (Fig. 3, B and C) and in more mature peripherally nucleated fibers (Fig. 3D). Blue nuclei were observed in none of the three mice analyzed after 1 week, when TA muscles showed an early regeneration pattern characterized by marked infiltration of mononuclear cells and a majority of small newly formed myofibers (Fig.3A). In rare instances, β-Gal+ nuclei were apparent in mononuclear cells infiltrating the areas of muscle regeneration or in a peripheral position within a centrally nucleated fiber (Fig. 3, E and F).

Figure 3

Analysis of nuclear lacZexpression in cryostat sections of regenerating TA muscles fromscid/bg mice transplanted with BM from C57/MlacZ mice. Muscles were analyzed 1 week (A), 2 weeks (B), and 3 weeks (Cand D) after cardiotoxin injection. (E) A mononuclear cell with nuclear β-Gal staining (arrow) 2 weeks after cardiotoxin injection. (F) A peripheral β-Gal+ nucleus (arrow) within a centrally nucleated regenerating fiber. Nomarski optics; scale bars, 10 μm. (A) through (D) and (F) also show Hoechst fluorescent nuclear staining, revealing areas of early (centralized nuclei) or more advanced (peripheral nuclei) fiber regeneration.

Our data indicate the existence of BM-derived myogenic progenitors that can migrate into a degenerating muscle, participate in the regeneration process, and give rise to fully differentiated muscle fibers. These cells appear to be recruited by long-range, possibly inflammatory, signals originating from the degenerating tissue, and they appear to access the damaged muscle from the circulation, together with granulocytes and macrophages. The kinetics of differentiation of BM-derived progenitors differ from those for differentiation of committed adult myogenic precursors. Injected satellite cells fused into muscle fibers within 5 days, whereas β-Gal+ nuclei of BM origin were not detected in regenerating fibers before 2 weeks after induction of muscle damage. This observation, together with the observed clustering of β-Gal+ nuclei in regenerated fibers, may suggest that BM-derived progenitors undergo a longer, and possibly multistep, differentiation process, which may comprise migration, cell division, commitment to the myogenic lineage, and eventual terminal maturation and fusion.

The origin of the BM-derived myogenic cells, as well as their physiological role in the homeostasis of muscle tissue, are not known. It is possible that these cells originate from multipotent, mesenchymal stem cells in the BM stroma that have been shown, by similar transplantation experiments, to give rise to bone, cartilage, and connective tissue (6). Whether or not myogenic cells are derived from the same mesenchymal component, our experiments suggest that the BM could serve as a reservoir of progenitors for muscle tissue, and that, under conditions of extended damage, these progenitors might expand or maintain the pool of resident, more differentiated, muscle-forming precursors.

The existence of circulating myogenic progenitors has implications for cell or gene therapy for inherited muscle disorders. Efficient delivery to diseased muscles of genetically modified myoblasts, or even of viral vectors containing therapeutic genes, is one of the major hurdles currently limiting both ex vivo and in vivo approaches (21). Despite some anecdotal observations, it is generally accepted that satellite cells taken from skeletal muscle cannot colonize muscle tissue if delivered from the circulation (22). The availability of a cell population that could be engineered and then systemically delivered to a large number of muscles might aid in the development of a cell-mediated replacement therapy. In our experiments, BM-derived progenitors contributed only minimally to muscle regeneration. However, resident satellite cells are healthy inscid/bg mice and are unaffected by the low-dose radiation administered before BM transplantation. The situation might be substantially different in a dystrophic background characterized by chronic muscle degeneration, in which genetically corrected BM-derived cells could progressively replace the exhausted pool of satellite cells. The therapeutic potential of transplanted BM cells awaits further verification in such a model.

  • * To whom correspondence should be addressed. E-mail: cossu{at} (G.C.) or mavilio{at} (F.M.).


View Abstract

Navigate This Article