National Muscular Dystrophy Research Centre, Fitzroy, Victoria, Australia
Howard Florey Institute, Parkville, Victoria, Australia
Department of Medicine, The University of Melbourne, Fitzroy, Victoria, Australia
Centre for Neurology and Neuroscience Research, St Vincent's Hospital, Fitzroy, Victoria, Australia
The Bionic Ear Institute, East Melbourne, Victoria, Australia
National Muscular Dystrophy Research Centre, Department of Clinical Neurosciences, St. Vincent's Hospital, 35 Victoria Parade, Fitzroy, Victoria, 3065, Australia. Telephone: 61-3-9288-3341; Fax: 61-3-9288-3350
Bone marrow (BM)-derived cells (BMCs) have demonstrated a myogenic tissue remodeling capacity. However, because the myoremodeling is limited to approximately 1%–3% of recipient muscle fibers in vivo, there is disagreement regarding the clinical relevance of BM for therapeutic application in myodegenerative conditions. This study sought to determine whether rare selectable cell surface markers (in particular, c-Kit) could be used to identify a BMC population with enhanced myoremodeling capacity. Dystrophic mdx muscle remodeling has been achieved using BMCs sorted by expression of stem cell antigen-1 (Sca-1). The inference that Sca-1 is also a selectable marker associated with myoremodeling capacity by muscle-derived cells prompted this study of relative myoremodeling contributions from BMCs (compared with muscle cells) on the basis of expression or absence of Sca-1. We show that myoremodeling activity does not differ in cells sorted solely on the basis of Sca-1 from either muscle or BM. In addition, further fractionation of BM to a more mesenchymal-like cell population with lineage markers and CD45 subsequently revealed a stronger selectability of myoremodeling capacity with c-Kit/Sca-1 (p < .005) than with Sca-1 alone. These results suggest that c-Kit may provide a useful selectable marker that facilitates selection of cells with an augmented myoremodeling capacity derived from BM and possibly from other nonmuscle tissues. In turn, this may provide a new methodology for rapid isolation of myoremodeling capacities from muscle and nonmuscle tissues.
Disclosure of potential conflicts of interest is found at the end of this article.
Duchenne's Muscular Dystrophy (DMD) is an X-linked recessive (Xp21.1) disorder characterized by a lack of dystrophin protein in the membrane-associated cytoskeleton of muscle fibers [1, 2]. DMD involves progressive muscle weakness caused by ongoing degeneration of skeletal myofibers. The demonstrated capacity of transplanted normal myoblasts and muscle-derived stem cells (SCs) to form new myofibers and/or fuse with dystrophin-deficient muscle of the mdx mouse (a model of DMD) and restore dystrophin expression to host myofibers maintains a rationale for cellular therapy as a potential treatment for DMD [3, , , –7].
Several factors hinder the success of myoblast transplant therapy in DMD, particularly host immune rejection of donor myoblasts [8, –10]. Ultimately this may be circumvented by autologous SC transplantation subsequent to ex vivo gene correction strategies including chimeraplasty, exon skipping, and small fragment homologous replacement that are currently being developed to repair the mutation in the dystrophin (DMD) gene [11, , , , –16]. However, exhaustion of muscle-derived SCs in DMD and the low proliferative capacity of DMD myoblasts in culture (resulting from premature senescence) would make it difficult to isolate and expand sufficient numbers of cells for ex vivo gene correction and transplantation [17, , , –21]. An alternative cell source is, therefore, required for autologous remodeling of dystrophic muscle.
Bone marrow (BM)-derived cells (BMCs) have a demonstrated capacity to differentiate into mature cells of the heart, liver, kidney, lungs, gastrointestinal tract, skin, bone, skeletal muscle, cartilage, and brain in vivo [22, , , , , , , –30]. This suggests that there are reserves of SCs within the hemopoietic system that can remodel nonself tissue. The ability of BMCs to incorporate into myofibers has been established and suggests that they may be a renewable nonmuscle cell type that could be used to remodel dystrophic muscle [26, 31, , , , –36]. The contribution of whole bone marrow (WBM) to muscle remodeling is typically limited (< 5%–0.002%) [26, 31, 37, –39]. We propose that a rare subpopulation(s) of cells capable of remodeling muscle, but highly diluted when WBM is transplanted, exists in BM. Ideally, if the specific BMC population(s) with myoremodeling ability can be isolated and enriched, then SC transplantation may provide a clinically relevant level of muscle engraftment.
Presently, it is still not known what characteristics of muscle-derived progenitors define a cell that can remodel muscle efficiently, although several have been investigated, including cell surface marker expression, in vitro adhesion properties, Hoechst efflux, or a combination of these characteristics [31, 35, –37, 40, , –43]. Fluorescence-activated cell sorting (FACS) is by far the most effective and most widely used method of isolating cells on the basis of cell surface marker expression and/or their ability to retain or efflux fluorescent DNA dyes such as Hoechst 33342. Recently, stem cell antigen-1 (Sca-1), a member of the Ly-6 multigene family commonly used for purification of murine hemopoietic cells [44, 45], has been reported as a marker for muscle-derived SCs [31, 46, –48]. Intravenously injected Sca-1-expressing muscle-derived SCs have also been found to migrate from the circulation, attach to capillaries of the muscles, and subsequently participate in myoremodeling in injured muscle [49, 50].
Although the use of Sca-1 cell surface antigen to identify cells with myoremodeling capacity has shown some variable success, it is still largely unclear whether Sca-1 expression is the most potent indicator of myoremodeling capacity in cells derived from BM or muscle itself [31, 37, 40, –42]. This study, therefore, initially investigated the myoremodeling capacity of BM and muscle sorted solely on the expression of Sca-1.
In 1999, Gussoni et al.  identified a subpopulation of BMCs able to incorporate into muscle, resulting in partial restoration of dystrophin expression in the dystrophic muscle of mdx mice. These so-called BM side-population (SP) cells were isolated on the basis of the expression of not only Sca-1, but also that of c-Kit (a receptor tyrosine kinase that is expressed on a number of cells including hemopoietic SCs [HSCs] and plays a role in HSC lodging to the BM niche) , as well as the expression of CD45 (a marker of all nucleated hemopoietic cells and their precursors) [31, 52, 53]. In recent years, researchers have identified phenotypically common, CD45−Sca-1+Kit−, multipotent progenitor cells from postnatal murine BM and muscle [54, 55]. They suggest that this population of cells may represent universal pluripotent SCs residing at different levels in multiple murine tissues. To date, myogenic differentiation of these cells has not been demonstrated in vivo. Furthermore, there is little clarity as to the true origin and identity of these cells. Nonetheless, these studies suggest that the BMCs with the optimal myoremodeling capacity may be elucidated by their cell surface marker expression profile, and CD45, Sca-1, and c-Kit may be important candidate markers for the isolation of these cells. This study was, therefore, extended to compare the relative myoremodeling capacity of various cellular components of WBM fractionated/enriched on the basis of their expression of cell surface antigens CD45, Sca-1, and c-Kit.
Materials and Methods
Wild-type (wt) male C57Bl/10 mice, 3–12 weeks of age, and male C57Bl/10 mdx mice, between 8 and 12 weeks of age, were obtained from Monash University Animal Services (Clayton, Victoria, Australia). All procedures were approved by the Animal Experimental Ethics Committee of St. Vincent's Hospital, Melbourne, Australia (protocol 01-11) and conformed to the guidelines for the care and use of experimental animals as described by the National Health and Medical Research Council of Australia.
BM Cell Suspension Extraction
Femurs, tibiae, and iliac crests were dissected from 20 male C57Bl/10 wt mice (8–12 weeks of age) and cleaned of all soft tissue, and BM cells were harvested by crushing the bones using a mortar and pestle. The cell suspension was filtered through a 40-μm cell strainer (Falcon, Becton Dickinson, Franklin Lakes, NJ, http://www.bd.com) to remove bone debris, washed twice, and resuspended in phosphate-buffered saline (PBS) supplemented with 2% newborn calf serum (PBS-2% Se). Approximately 1 × 108 WBM cells were obtained from each mouse. Low-density (LD) BMCs were isolated by discontinuous density gradient centrifugation using Nycoprep Animal (density 1.077 g/cm3; Nycomed Pharma, Oslo, Norway, http://www.nycomed.com/en) and collected into PBS-2% Se. Chromogenic donor cells (e.g., GTRosa-26) were not used in this study because of difficulties in breeding sufficient numbers of mice to generate the cell numbers needed for these experiments.
Negative Immunomagnetic Selection
LD BMCs were depleted of cells expressing mature hemopoietic cell lineage antigens by immunomagnetic selection using the MACS system (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com/en). The antibodies anti-B220, anti-CD4, anti-CD8, anti-Gr-1, anti-Mac-1, and anti-TER 119 were used as a cocktail, purified, or biotin-conjugated for MACS separation as previously described . The cells were washed and incubated with goat anti-rat microbeads for 15 minutes on ice. The cells were added to the column, run into the mesh, and left to magnetize for 5 minutes. The nonmagnetic fraction (mature cell lineage antigen-negative cells [Lin−]) was collected by eluting the cells through a 20-G needle with 100 ml of PBS-EDTA-0.5% bovine serum albumin.
Lin− cells were resuspended at 5.0 × 106 cells per 100 μl, and labeled with 1 μl of fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse Sca-1 and phycoerythrin (PE)-conjugated rat anti-mouse c-Kit. CD45+ cells were identified by labeling with biotinylated rat anti-mouse CD45 antibody and streptavidin-RED670 for 20 minutes on ice. The cells were then washed, and Lin−, CD45− cells within a predetermined rectilinear forward scatter (FSC) and side scatter (SCC) blast cell region  using a FACStarPlus cell sorter (Becton Dickinson), equipped with a 5-W argon ion laser (Innova 90; Coherent, Palo Alto, CA, http://www.cohr.com) running at 200 mW power, and a Spectra-Physics ultraviolet (UV) laser (Mountain View, CA, http://www.spectra-physics.com), running at 50 mW power. Green fluorescence (FITC) was collected through an FITC 530-nm filter, with a bandwidth of ±15 nm. Red fluorescence (PE) was collected through a 575DF26 filter, and RED670 fluorescence was collected through a long-pass RG655 filter.
Sorted cells were collected into serum-coated tubes at 4°C and reanalyzed to establish purity prior to storage at 4°C overnight in PBS-2% Se until injection into recipient mdx mice. These cells were counted and washed in PBS to remove all traces of serum before injection into mdx recipients. Immediately before injection, viable cell counts using trypan blue nuclear exclusion showed an average excess of 90% of viable cells in the injected population.
Isolation of Whole Muscle Cells from Fresh Muscle
Male C57Bl/10 wt mice, 3 weeks of age, were killed by cervical dislocation. Skeletal muscle from the hind limbs was removed, and visible nerves and fat were separated from the muscle. The muscle was finely minced and then digested at 37°C with 2 ml/g muscle of dispase (grade II, 2.4 U/ml; Roche Molecular Biomedicals, Indianapolis, http://www.roche-applied-science.com) and collagenase (class D, 1%; Roche Molecular Biomedicals) in 10 ml of Hams/F10 (Trace Biosystems, Sydney, Australia) supplemented with 2.5 mM CaCl2, 1× penicillin/streptomycin (Gibco Life Technologies, Gaithersburg, MD, http://www.invitrogen.com), and 25 ng/ml amphotericin. The muscle slurry was triturated every 20 minutes over 2 hours. After 2 hours, the muscle slurry was centrifuged at 840g for 10 minutes, and the supernatant was removed. Red blood cells were lysed with 17 mM Tris/144 mM NH4Cl, pH 7.2. The cells were washed twice in PBS by centrifugation before injection into mdx recipients.
Cell Transplantation into Dystrophic mdx Muscle
The tibialis anterior (TA) muscle of 8–12-week-old male mdx mice was exposed by surgical incision. Cells were injected into the TA in a 10-μl volume (refer to Table 1 for injected cell types and cell numbers). In the experiment involving i.m. injection of Sca-1-sorted components of muscle or BM, mice were injected 2 days before injection of cells, with 50 μl of 0.5% bupivacaine hydrochloride (BUP; Marcaine; Astra, NSW, Australia, http://www.astrazeneca.com), because this myotoxic agent has been shown to promote equal myoregenerative activities in mdx and wt rodent muscle [58, 59]. Right TAs served as internal contralateral saline controls and were likewise preconditioned with BUP 2 days before saline injection. The TA muscles were all harvested after 6 weeks and snap frozen in liquid nitrogen-cooled isopentane.
Table Table 1.. Percentage of each BMC fraction within WBM, actual number of cells injected, summary of myogenic engraftment by each BMC fraction, and their associated RMI
The TAs were sectioned and analyzed for dystrophin expression as described previously . Briefly, serial sections 8 μm thick were taken along the entire length of the muscles. Dystrophin expression was assessed with a 1:100 dilution of a polyclonal antibody raised in sheep against a 60-kDa dystrophin fusion protein (a gift from Prof. L. Kunkel, Children's Hospital and Harvard Medical School, Boston). The primary antibody was visualized with FITC-conjugated rabbit anti-sheep IgG (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com), diluted 1:100 in PBS. Where applicable, an area containing the maximum number of dystrophin+ myofibers for each injection was used to determine the Relative Myoremodeling Index (RMI; derivation of the RMI is discussed in the next section), to better facilitate comparison. The RMI was calculated from dystrophin+ myofiber numbers standardized for revertant fibers by subtracting the mean number of revertant fibers from the total number of dystrophin+ myofibers observed in each TA muscle. The revertant fiber numbers in each treatment groups was assessed by counting dystrophin+ fibers in TA muscles of littermates that were injected with saline (vehicle), but not cells. Statistical analysis between groups was analyzed by the Mann-Whitney U test (significance at p < .05) using Statistica software (StatSoft Inc., Tulsa, OK, http://www.statsoft.com). Images were viewed on an Olympus IX70 inverted fluorescence microscope (Olympus, Tokyo, http://www.olympus-global.com).
Derivation of the RMI
The relative abundance of the various cell types used in this study differ significantly within the BM cellular compartment (Table 1). As a consequence, the absolute numbers of each cell type injected into recipient host muscles differed accordingly. A calibration factor that reflected the relationship between numbers of injected cells and cellular integration into the host tissue was derived to accommodate these differences in cell numbers injected into recipient mice. On the basis of the well-established Regeneration Efficiency Index [61, –63], which adequately describes the regenerative efficiency of cells injected into recipient muscle in previously published work, the numbers of cells used in some of the injections used here (i.e., > 1 × 106) raised the issue of an asymptotic “ceiling” effect regarding the incorporation of cells into the recipient muscle tissue matrix, previously shown by others to take effect in mdx muscle at between 5 × 105 and 1 × 106 cells . The logarithmic relationship between injected donor cell numbers and number of donor cells engrafted into the recipient muscle shows a linear relationship on a semi-log10 plot  that can be defined by:
where x = log10n
where y is the number of dystrophin+ myofibers, k (a constant which we have termed the “Regeneration Constant”) is the rate of increase in the number of dystrophin+ myofibers with an increase in injected cells on a semi-log10 plot, and x is the logarithm of n, the number of injected cells. This relationship holds true as long as the value of n is ≤ 106 (i.e., x ≤ 6), where 106 is the upper limit of the number of cells injected before a grafting plateau is reached .
If the relationship between injected and engrafted cells of the various cell types used in this study assumes equivalence, then the relationship can be simply described by a “Myoremodeling Index” for each cell type, defined as the number of dystrophin+ myofibers arising from the donor cells as a function of the number of donor cells injected:
where x1 and y1 represent log10(number of cell-type 1 injected) and the number of dystrophin+ myofibers resulting from engraftment of these cells, respectively. x2 and y2 represent similar parameters from grafted cell-type 2.
A RMI, which describes the relative capacities of two given cell types to engraft into muscle can thus be derived by:
The third equation therefore reflects the ratio of dystrophin+ fibers, y2 (generated by cell-type 2) as a function of the number of type 2 cells injected (x2) and number of injected type 1 cells required (x1) to generate the number of dystrophin+ fibers (y1) by the type 1 cells. As such, this transform defines the myoengraftment (and therefore myoremodeling) capacity of type 1 cells as a function of the engraftment capacity of the type 2 cells.
Allele-Specific Polymerase Chain Reaction Detection of wt DMD Gene and Transcript
Genomic DNA (gDNA) was isolated from the cryosectioned right TAs (QIAamp DNA Mini Kit, Qiagen, Valencia, CA, http://www1.quiagen.com). A 892-base pair, double-stranded DMD product (Amplicon A) was amplified from the gDNA, using antisense primer Dys In23 AS-02 (5′-CAGACAATCCAAGAAGGTATGAC-3′) and sense primer Dys In22 S-01 (5′-CACTATGATTAAATGCTTGATATTGAG-3′) (Supplemental Fig. 1). Reactions (50 μl) consisted of 100 ng of gDNA, 0.4 μM each primer, 0.2 μM each dNTP, 3.5 mM Mg(OAc)2, and 1.25 units of KlenTaq LA DNA polymerase in buffer supplied by the manufacturer (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). The reactions were subjected to 29 cycles of 92°C for 30 seconds, 62°C for 2 minutes, with an initial cycle of 92°C for 2 minutes and 65°C for 2 minutes on a Sprint PCR thermocyler (Eppendorf South Pacific, North Ryde, Australia).
Allele-specific polymerase chain reaction (AS-PCR) for the detection of wt DMD gene was performed with antisense primer Dys-wt AS-01 (3′ mismatch for the mdx nucleotide) and sense primer Dys In22 S-02, whereas wt DMD transcript was detected with antisense primer Dys wt AS-01 and sense primer C2917-S (supplemental online Fig. 1) as previously described [11, 60].
Dystrophin is exclusively expressed in mature terminally myodifferentiated multinucleated fibers. Detection of wt DMD transcript arising from donor cell nuclei is, therefore, a good indicator of their myoremodeling capacity. The significance of wt DMD transcript expression (or absence) in mdx muscles bearing wt DMD loci was assessed by χ2 analysis (significance at p < .05; Statistica).
BMCs Contribute to Skeletal Myofibers
To assess the ability of BMCs to remodel muscle, recipient mdx mice each received i.m. injections of 3.0 × 106 WBM from C57Bl/10 wt donors. At 6 weeks post-transplant, their TAs were evaluated for the presence of dystrophin+ myofibers. Dystrophin expression was observed in its typical sarcolemmal distribution around the periphery of myofibers (Fig. 1A, 1B). The morphology of dystrophin+ myofibers was identical to the surrounding dystrophin− myofibers. In WBM-injected TAs, clusters of 2–5 myofibers of varying size and shape, scattered about the injection site, showed strong sarcolemmal expression of dystrophin (Fig. 1B). These dystrophin+ fibers were interspersed with typical mdx myofibers lacking this intense sarcolemmal staining. Dystrophin+ myofibers amounted to an average of 1.13% ± 0.13% of the total number of myofibers, whereas 0.74% ± 0.07% myofibers in saline-injected control TAs were found to express dystrophin (Fig. 1C). These revertant fibers arise from post-transcriptional splicing or second mutation events that bypass the mutation in the DMD gene and restore an open reading frame [12, 65, –67]. The increase in dystrophin+ myofibers in WBM-injected muscles compared to saline controls was slight but significant (p < .05), providing evidence that WBM does indeed participate in the myoremodeling process in mdx mice (Fig. 1C). This level of incorporation of WBM into skeletal muscle is consistent with what had been established by others [26, 39].
Sca-1+ Cells from BM or Muscle Sorted Solely on Sca-1 Expression Do Not Have a Myoremodeling Advantage Over the Sca-1− Subpopulations
To determine whether and to what extent BMCs capable of remodeling muscle are characterized by the expression of Sca-1, WBM was sorted exclusively on the basis of Sca-1 expression. Unfractionated WBM, Sca-1-positive (Sca +) and Sca-1-negative (Sca−) BMC populations (Fig. 2A, 2B) were injected into the TA muscle of mdx mice. If Sca-1 expression characterizes a subset of BMCs with optimal myoremodeling capacity, then the Sca-1 status of muscle-derived SCs may also coincide with their ability to remodel muscle. Therefore, whole muscle was similarly sorted on Sca-1 expression and unfractionated whole muscle, Sca+ and Sca− muscle fractions were transplanted into the TA muscle of mdx mice. The Sca-1 profiles of muscle and WBM are shown in Figure 2.
Whole muscle was significantly more efficient (p < .005) at regenerating dystrophic muscle than was WBM, which exhibited an RMI of 0.25 ± 0.07 (Fig. 3A, 3C, 3D). Although the Sca-sorted muscle fractions were more efficient at myoremodeling compared to the Sca-sorted BMC fractions (Sca+ muscle, 0.37 ± 0.09; Sca+ BM, 0.15 ± 0.03; p < .05; Sca− muscle, 0.28 ± 0.04; Sca− BM, 0.13 ± 0.01; p < .05), there was no significant difference in the myoremodeling capacity between the Sca+ and Sca− cell fractions within each group (Fig. 3A, 3E, 3F). Although the reason behind this is unclear, it is possible that the presence of Sca+ cells (e.g., Lin+, CD45+) other than the promyoremodeling cells may prevent the differentiation and/or fusion of myoremodeling cells in the Sca+ fraction. This observation was made in muscle and BM alike (Fig. 3A) and AS-PCR analysis showed detectable levels of wt DMD gene and transcript in TAs injected with whole muscle as well as the Sca+ and Sca− muscle fractions (data not shown). AS-PCR showed no preference for Sca+ over Sca− muscle fractions, and this was a similar trend observed in TAs injected with the Sca-sorted BM fractions. The absence of an AS-PCR product in saline controls confirms that the dystrophin+ myofibers are not revertants. These observations strongly suggest that, contrary to what was found in most literature to date, Sca− cells from muscle and BM were just as efficient at muscle remodeling as were their Sca+ counterparts.
Isolation and Characterization/Enrichment of BMC Populations within WBM
Because of the lack of distinction in myoremodeling capacity observed in i.m.-injected Sca+ and Sca− BMC fractions, WBM was sorted into various cellular fractions to further purify and enrich for cell populations that may be more efficient at myoremodeling.
Low-density BM cells were first depleted of mature hemopoietic cells using antibodies directed against hemopoietic lineage antigens, and the lineage antigen negative (Lin−) fraction was further fractionated into CD45+ cells (enriched for HSCs) [40, 52, 53] and CD45− cells (enriched for mesenchymal/nonhemopoietic SCs) . The Lin− cell fraction comprised approximately 50% CD45+ and 50% CD45− cells. The nonhemopoietic CD45−Lin− fraction was then sorted into Sca−Kit−, Sca−Kit+, Sca+Kit− and Sca+Kit+ BMCs, each fraction representing approximately 85%, 12%, 1.5%, and 1.5%, respectively, of the entire CD45− BMC population. These cell fractions will be termed Sca−Kit−, Sca−Kit+, Sca+Kit− and Sca+Kit+ from this point forward. A schematic of the cellular fractions of BM that were sorted for and injected into mdx TAs is shown in Figure 4A, and a summary of the percentages of each cell fraction found within WBM is provided in Table 1. Sca+Kit− and Sca+Kit+ BMCs are very rare, with each fraction representing only 1 in 3,000 WBM (Table 1). Representative FACS dotplots are shown in Figure 4B–4D.
Myoremodeling Capacity in BMC Subpopulations That Express c-Kit
To investigate which BMC fraction possessed the highest myoremodeling capacity, we injected the various FAC-sorted BMCs into dystrophic mdx TAs and harvested them after 6 weeks. BMC contribution to skeletal muscle was assessed by dystrophin immunofluorescence.
As mentioned previously herein, the Sca+Kit− and Sca+Kit+ cells are extremely rare cell populations within WBM. All of the cells acquired in each FACS fraction were used for transplantation. As a result, fewer cells from the rarer subpopulations were transplanted into recipient mdx mice. To account for the disparate cell numbers injected, an RMI was utilized for comparison of the BMC fractions . The number of cells injected in each group and their RMI with respect to WBM are shown in Table 1. Comparison of the various subfractionated BMC populations demonstrated a variable RMI compared to WBM (Table 1; Fig. 5). The two BMC populations isolated on the basis of c-Kit expression (irrespective of Sca-1 expression), yielded a significantly higher RMI than did WBM (Sca−Kit+, p < .05; Sca+Kit+, p < .005; Table 1; Fig. 5B, 5D–5F), suggesting that the myoremodeling potential of BMCs copurified with the c-Kit surface antigen. This is further highlighted by BMC populations that expressed both Sca-1 and c-Kit, which displayed approximately two times more efficient myoremodeling activity than the Sca−Kit+ fraction which in turn, displayed three times the myoremodeling index of the Sca+Kit− fraction (Table 1; Figs. 5D–5F; p = not significant). Lin+ BMCs, largely comprising mature myelomonocytic cells [40, 42] did not contribute significantly to the myoremodeling process (Fig. 5F). The number of dystrophin+ myofibers observed in TAs injected with each of the sorted BMC fractions is summarized in Table 1.
AS-PCR detected wt DMD gene in all four BMC fractions sorted for Sca-1 and c-Kit (Fig. 6A). wt DMD transcript was consistently detectable in all TAs injected with Sca−Kit+ and Sca+Kit+ BMC fractions. However, this was not so in those injected with Sca−Kit− and Sca+Kit− BMC fractions (Fig. 6A), supporting the likelihood that cells expressing c-Kit are attributed with a higher myoremodeling capacity than cells not expressing c-Kit (χ2 = 8.00; p < .005; Fig. 6B). Absence of the 313-base pair wt locus in DMD transcript from recipient mdx mice 8 and 13 (both injected with CD45−Lin−, Kit− BMCs) indicates that, although engrafted into the recipient mdx muscle, the Kit− donor cells failed to express detectable levels of wt DMD transcript. Indeed, these genetic data correspond with dystrophin immunohistochemical data whereby 0 to negligible levels of dystrophin+ myofibers were detected in the injected muscles of mice 8 and 13 from the Kit− groups.
Using a systematic WBM FAC-sorting protocol incorporating lineage markers CD45, Sca-1, and c-Kit, this study showed that the Lin−CD45−Sca−Kit+ (Sca−Kit+) and Lin−CD45− Sca+Kit+ (Sca+Kit+) cell fractions demonstrated the highest myoremodeling capacity when injected into the dystrophic muscle of mdx mice. Three independent lines of evidence (DNA, RNA, and dystrophin protein) were used to generate statistically significant cross-referential data that our injected cell fractions functionally remodeled dystrophic mdx muscle. These results build on several studies, including our own earlier findings indicating that hemopoietic (CD45+) cell contribution to in vivo myogenesis is rare [69, , , –73] and that the main contributors to non-self tissue remodeling (including muscle remodeling) by BMCs were likely to be nonhemopoietic (CD45−Lin−) mesenchymal stem cells (MSCs) [36, 71, 74, , , , , , , –82]. Collectively, these observations give rise to the possibility that cells with efficient myoremodeling capacity may be enriched by FAC-sorting from the mesenchymal nonhemopoietic (CD45−Lin−) BMC fraction using “rare” stem cell surface antigen markers such as c-Kit.
A Lin−, CD45− cell surface antigen phenotype in adherent BM stromal cells has generally been accepted as defining a population rich in MSCs [68, 83, 84]. The detection of several cell surface antigens more usually associated with hemopoietic cells [36, 74, 77, 83, 85, 86], along with variation in MSC cell surface epitope expression between mouse strains , serves to occlude the “definitive” cell surface antigenic character of MSCs. This contrasts with HSCs, which are probably the most well characterized SCs [31, 87, , –90]. The sorting protocol adopted in this study ensured that an initially MSC-enriched fraction was obtained from WBM [68, 83].
After the initial observation of in vivo muscle remodeling by BMCs , investigators attempted to identify the specific population of BMCs that possessed myogenic potential . A significant myoremodeling potential was identified in FAC-sorted CD45+Lin−Sca+Kit+ HSCs (also termed “side population” [SP] cells). Subsequent identification of similar Sca+/Hoechst 33342Lo SP cells in muscle  generated interest in Sca-1 as a key identifier of myoremodeling potential in SCs from muscle [46, 47, 49, 91, –93] and nonmuscle tissues alike [40, –42]. After our previous study showed that myoremodeling potential is unlikely to originate from within the hemopoietic (CD45+) compartment of BM , our approach in this study was to subfractionate BMCs to Lin−, CD45− (mesenchymal) fractions and then into further subfractionated MSC populations using Sca-1 and c-Kit cell surface antigens. These cells' capacity to remodel muscle was scored by restoration of dystrophin expression at transcript and protein levels in mdx mouse muscle. An RMI was applied to account for the necessary injection of lower numbers of the rarer cell fractions and the logarithmic relationship between cell numbers injected and their contribution to restored dystrophin expression in recipient mdx muscle .
Muscle cells displayed a significantly greater myoremodeling potential compared to corresponding BMC fractions (Fig. 3A). On investigating the relative myoremodeling capacity of Sca+ and Sca− cell types, we observed no significant differences between transplanted Sca+ and Sca− muscle cells (Fig. 3A, 3E, 3F) or BMCs (Fig. 3A). This may result from the fact that the Sca+ population of BMCs contains within it CD45+Lin+ cells that may influence the myoremodeling capacity of promyoremodeling cells in the Sca+ fraction. Further subfractionation of the BMCs revealed that the Sca−Kit+ and Sca+Kit+ cells accounted for most of the myoremodeling activity of BMCs observed in this study (Figs. 5F, 6). This indicates an important role for c-Kit as a selectable marker for myoremodeling capacity in BMCs, particularly in conjunction with other selectable markers such as Sca-1.
Our results build on findings that cells with myoremodeling potential can be isolated using Sca-1 cell surface antigen [40, –42, 46, 47, 49, 91, –93]. In fact, our results implicate c-Kit as an additional marker that subfractionates Sca+ cells into a population with greater capacity to remodel dystrophic muscle than cells expressing Sca-1 in absence of the c-Kit antigen (Fig. 5F). In addition, our results show that cell selection using c-Kit defines myoremodeling cells from within the Sca− MSC fraction (Fig. 5F; Mann-Whitney p < .05; Fig. 6; χ2p < .005). In this regard, our observations are consistent with those of others demonstrating that multipotent BMCs capable of cardiac differentiation are not present in the Sca+ hemopoietic progenitor cell fraction  and that only BMC fractions inclusive of Kit+ precursors contribute to myofibers after i.m. injection into C57BL/6 mice .
A number of factors potentially apply to our findings regarding Sca-1's contribution to the identification of cells with effective dystrophic mdx muscle myoremodeling capacity. Firstly, Sca-1 cell surface expression is dynamically modulated by the prevailing microenvironment  and transiently upregulated during myoblast cell-cycle withdrawal and differentiation [94, –96]. By this process, rare SCs isolated using Sca-1 alone may be overlooked depending on the cell cycle stage (and associated Sca-1 expression level) they are in at the time of isolation. These cells are likely to be recovered as Sca− cells. Secondly, Sca-1 is not only expressed on myogenic SCs [46, 47], but also on HSCs, B and T lymphocytes, mammary gland epithelium, cardiac SCs, endothelial cells, neural SCs, and MSCs derived from skin and adipose tissue in the mouse [97, , , , , , –104]. Thus, Sca+ cells isolated from any given tissue/organ, may contain Sca+ cells of different lineage(s) in addition to those of source tissue origin, which may in turn affect the net myoremodeling activity of the population.
Muscle progenitors are able to home to injured muscle when injected systemically [49, 50]. As well as being known to maintain/reconstitute SC pools, the c-Kit ligand (stem cell factor) also mediates HSCs and other progenitor/primitive cells' migration to target organs [105, , , , , , , , , –115]. It is plausible that BM-derived MSCs expressing c-Kit may utilize such a homing mechanism, enabling their recruitment to muscle tissues. That the Kit− BMC fraction contains cells with developmentally dormant myoremodeling capacity remains a possibility because of direct linear development of Kit+ from Kit− in cells [116, 117]. However, the myoremodeling potential demonstrated by the CD45−, Kit+ BMC fractions in this study (Fig. 6; χ2p < .005) suggests that the signaling pathways necessary to activate myoremodeling in the Kit− BMC fractions investigated here were at the very least less advanced than in the Kit+ cells. This result supports a role for c-Kit expression in the selection of CD45−Lin− cells from BM with effective myoremodeling capacity.
Myoremodeling in this study was achieved by all injected C57BL10/J wt BMCs with the exception of mature hemopoietic (Lin+) cells (Fig. 5F). From this perspective, our results promote c-Kit as a useful selectable FACS marker that complements existing cell isolation approaches for identifying cells with myoremodeling capacity from BM and possibly from other nonmuscle tissues. Nevertheless, it remains to be seen if these myogenic Sca−Kit+ or Sca+Kit+ BMCs can overcome the ceiling effect commonly observed with cell transplantation in dystrophic mdx muscle . From this viewpoint, dystrophic muscle remodeling levels achieved by either Sca−Kit+ and Sca+Kit+ BMCs are still too low for potential therapeutic application, and further work needs to clarify the specific issues that prevent efficient cell-mediated myoremodeling. Although some headway has been made in this regard using immunosuppression [118, –120], X-irradiation [121, –123] recipient muscle preconditioning with my toxins [124, 125], and hyaluronidase [125, 126], the generally small myoremodeling increments achieved via these avenues suggests as yet unidentified, but powerful antagonistic mechanisms limiting muscle remodeling by injection of myogenic and nonmyogenic cells alike.
Subfractionation of MSCs to identify cells enriched for myoremodeling potential from adult tissues is a promising step toward understanding and possibly developing the therapeutic potential of these cells in muscle cell replacement therapies. However, the existence of these cells as extremely low fractions of the BMC population gives rise to significant challenges for the development of their therapeutic potential. Further work is required to better understand and define the growth conditions by which these cells' myoremodeling characteristics are maximally retained  and how their myoregenerative performance and in vitro characteristic(s) compare to more conventional MSC populations. Likewise, cell cycle-related alterations in SC phenotype [128, 129] may necessitate cell synchronization before transplantation. Finally, the introduced cells need to be physiologically matched with their host environment in a manner that enables them to overcome physiological barriers to their engraftment. These barriers include host immune response [9, 130] and competing endogenous myoregenerative activity  in addition to physical barriers (e.g., extracellular matrix) within the muscle microenvironment.
Cells derived from BM displaying a MSC/stromal Lin−CD45−Sca−Kit+ and Lin−CD45−Sca+Kit+ cell marker profile were shown to possess a potent myoremodeling potential, defined by restoration of wt dystrophin protein and transcript expression in dystrophic mdx mouse muscle. Future research aimed at the optimization, expansion, and in vivo delivery of these cells will more clearly identify their potential application as therapeutic agents for neuromuscular disorders.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
We thank the Muscular Dystrophy Association (U.S.), National Health and Medical Research Council (Australia), Aktion Benni, and Muscular Dystrophy Australia for their support in funding the work communicated in this article; and Stephania Tombs, Kelly Steeper, and Judy Chin for their excellent technical support. This study was funded in part by the National Health and Medical Research Council of Australia, Muscular Dystrophy Association (Australia), Muscular Dystrophy Association (U.S.), and Aktion Benni and Co.