Cellular and Molecular Regulation of Skeletal Muscle Side Population Cells
http://www.100md.com
《干细胞学杂志》
a Departments of Internal Medicine,
b Cardiothoracic Surgery,
c Molecular Biology, and
d The Donald W. Reynolds Cardiovascular Research Center, University of Texas Southwestern Medical Center, Dallas, Texas, USA
Key Words. Side population cells ? Myoblast ? Foxk1 ? Skeletal muscle ? Bone marrow
Correspondence: Daniel J. Garry, M.D., Ph.D., NB11.118A, 5323 Harry Hines Boulevard, University of Texas Southwestern Medical Center, Dallas, Texas 75390-8573, USA. Telephone: 214-648-1654; Fax:214-648-1450; e-mail: daniel.garry@utsouthwestern.edu
ABSTRACT
Adult skeletal muscle is capable of self-repair in response to extreme training, injury, or myopathic diseases such as muscular dystrophy . The ability for self-repair or regeneration is attributable to a rare population of myogenic progenitor cells (MPCs), also referred to as satellite cells . The MPCs are capable of self-renewal and are arrested at an early stage of the myogenic program such that they do not express any of the myogenic basic helix-loop-helix proteins of the MyoD family . In response to an injury, the quiescent MPCs are activated, proliferate, and withdraw from the cell cycle to form differentiated myotubes, and ultimately restore the skeletal muscle architecture.
Foxk1 is a member of the forkhead/winged helix family of transcription factors and is expressed in the MPC population, where it functions as a regulator of the cell cycle . Mice lacking Foxk1 have a growth deficit and severely impaired muscle regeneration . This decrease in regenerative capacity is attributable, in part, to a decrease in MPC number, impaired MPC activation, and impaired MPC proliferation in Foxk1-deficient skeletal muscle . Studies undertaken in mice lacking both Foxk1 and p21 suggest that p21 is a downstream target for Foxk1, but other target genes for this forkhead/winged helix transcription factor are unknown .
Recently, an additional stem cell population referred to as side population (SP) cells has been identified in most adult tissues, including skeletal muscle . Using flow cytometry, these SP cells are isolated from adult tissues based on their ability to efflux Hoechst 33342 dye . This ability to efflux Hoechst dye is attributable to the ATP binding cassette (ABC) half-transporter, Abcg2, which has been shown to be the molecular determinant of the SP cell phenotype . SP cells have stem cell properties, because limited numbers of SP cells isolated from adult bone marrow were able to reconstitute the irradiated mdx (dystrophin mutant) mouse bone marrow. Later these cells were recruited from the bone marrow to contribute to differentiated myotubes during muscle regeneration . Previous studies suggest that SP cells may be the precursors of the MPC (satellite) population and require Pax7 for the specification of the satellite cell, because Pax7-deficient mice lack satellite cells but contain a complete repertoire of SP cells . Alternatively, SP cells may represent a second distinct progenitor or stem cell population that is resident in adult skeletal muscle. Although these elegant studies provide additional support for the role of skeletal muscle SP (SMSP) cells as a progenitor/stem cell population, the molecular regulation of this cell population is illdefined.
In the present study, we use morphological, molecular, and transcriptome analyses to enhance our understanding of the SP cell population isolated from wild-type (WT), injured, and Foxk1 mutant adult skeletal muscle. Comparison of these SP cell populations with embryonic stem cells and WT myoblasts additionally defines common and distinct molecular signatures for the respective SP cell populations. Collectively, this strategy will enhance our understanding of the SP cell population and impact therapeutic applications in the treatment of myopathic diseases such as muscular dystrophy.
MATERIALS AND METHODS
Abcg2 Is a Determinant of the SP Cell Phenotype
Using dual-wavelength flow cytometry, SP cells were isolated from several adult tissues based on the ability of the cell population to efflux Hoechst 33342 dye . Recent studies have demonstrated that the ability of SP cells to efflux Hoechst 33342 dye is dependent on the ABC half-transporter Abcg2 . Furthermore, the fungicide FTC has been reported to be a specific inhibitor of Abcg2, where it functions to inhibit the Abcg2-associated ATPase activity . Using these protocols, we isolated SP cells from adult murine skeletal muscle (Fig. 1A) and bone marrow (Fig. 1C). Routinely, the absolute number of SP cells obtained from bone marrow exceeded that obtained from adult skeletal muscle. Furthermore, the ability of these SP cell populations to efflux Hoechst dye was inhibited equally well by either the calcium channel blocker verapamil (data not shown) or after the addition of FTC (Figs. 1B, 1D). Preliminary studies undertaken in our laboratory reveal that overexpression of Abcg2 in C2C12 myoblasts confers the SP cell phenotype and that the ability to efflux the dye in the Abcg2-overex-pressing C2C12 myoblasts can be inhibited completely by FTC (C. Martin, D. Garry, personal observations).
Figure 1. Isolation of SP cell populations using flow cytometry. SP cell populations were isolated from adult skeletal muscle and adult bone marrow, respectively (A, C) using dual-wavelength fluorescence-activated cell sorter and Hoechst 33342 dye. Note that SP cells are located within the gated regions. Addition of fumitremorgin C to skeletal muscle (B) and bone marrow (D) cell preparations inhibited the ability of the SP cells to efflux the Hoechst 33342 dye. Abbreviations: MP, main population; SP, side population.
Morphological Characteristics and Differentiation Capacity of SMSP
SP cells were isolated from noninjured WT adult mouse hind-limb skeletal muscle using Hoechst 33342 dye efflux and FACS analysis. Sorted cells were stained with the Hema3 kit (Fischer Diagnostics, Middleton, VA) and examined at the light microscopic level. We observed that the SMSP cells were relatively small (6.6 ± 0.1 μm, n = 154) and had a high nuclear-to-cytoplasmic ratio (Fig. 2A). Ultrastructural examination of SMSP cells additionally confirmed that these small mononucleated cells had a high nuclear-to-cytoplasmic ratio, with few intracellular organelles (Fig. 2B). A high nuclear-to-cytoplasmic ratio is characteristic of quiescent stem cell populations, such as embryonic stem cells, and somatic stem cell populations, such as the MPCs (i.e., satellite cells) in adult skeletal muscle . SMSP are uniformly smaller (Fig. 2C) than the embryonic stem (ES) cells (Fig. 2D) that are propagated in culture.
Figure 2. Morphological analysis of SMSP cells. Microscopic analysis of wild-type SMSP cells using light microscopy (A) and electron microscopy (B) revealed a high nuclear-to-cytoplasmic ratio in SMSP cells. Comparison of SMSP cells (C) with ES cells (D) revealed morphological differences (i.e., size) between these two cell populations. Arrowheads denote the SMSP cells (C) and the ES cells (D). Side population and myogenic progenitor cells were sorted and cultured to evaluate their differentiation capacity. SMSP cells did not differentiate (E), whereas SMMP cells attached to the culture dish and differentiated after serum removal (F). In contrast, sorted SMSP (G) and SMMP (H) readily form multinucleated (stained with Hoechst dye) myotubes when cultured with myogenic progenitor differentiation media and on Matrigel-treated plates. SMSP cells are capable of differentiating into myotubes in vivo. Two days after cardiotoxin-induced skeletal muscle injury, 250 SMSP cells isolated from adult ROSA26 skeletal muscle were intramuscularly delivered into the injured skeletal muscle. Ten days later, the skeletal muscle was harvested and stained with X-gal. Note ?-galactosidase–positive fibers (blue) present within the regenerating skeletal muscle (I). Sorted SMSP cells were cytospun onto slides and were observed to express Foxk1 (arrowheads) using immunofluorescence techniques (J). Scale bars = 20 μm (A), 500 nm (B), and 20 μm (C–J). Abbreviations: ES, embryonic stem; SMMP, side population main population; SMSP, skeletal muscle side population.
Previous studies observed that isolated SMSP cells differentiate slowly (2 weeks compared with 1 week for differentiation of skeletal muscle MP ) into a mixture of myoblasts and fibroblasts or that isolated SMSP cells cultured in myoblast growth medium did not give rise to myogenic progenitors unless cultured in the presence of primary myoblasts . In the present study, we observed that sorted SMSP cells are viable but relatively quiescent, without any evidence of differentiation when cultured in F10 media supplemented with 20% FCS (Fig. 2E). In contrast, SMMP cells readily attach to the culture dish and differentiate even under high serum culture conditions (Fig. 2F). No cells exhibiting the morphology of the MP cells were detected in the isolated SP cell cultures. We further observed that sorted SMSP (Fig. 2G) and MP (Fig. 2H) cells cultured individually on Matrigel-pretreated plates in the presence of a MPC-differentiation media formed differentiated multinucleated myotubes but at different rates (3 days for SMMP compared with 7 days for SMSP; n = 3 separate experiments). It is possible that the Matrigel provides a surface similar to the extracellular matrix in skeletal muscle, allowing the SP cells to adhere and differentiate when exposed to this matrix .
Having established that the SMSP cells were capable of differentiation in vitro, we examined their ability to repopulate injured skeletal muscle. Two days after cardiotoxin-induced skeletal muscle injury, 250 SP cells (isolated from the skeletal muscle of the ROSA26 mouse model in which all cells constitutively express ?-galactosidase) were delivered intramuscularly into the regenerating skeletal muscle. After skeletal muscle regeneration (10 days after delivery of the cells), we observed evidence of ?-galactosidase–positive fibers (Fig. 2I). These results support the conclusion that limited SP cell numbers are capable of competing with endogenous progenitor cell populations in regenerating skeletal muscle to promote muscle repair. These results additionally establish that the SMSPs are capable of differentiation into skeletal muscle after the exposure to in vitro and in vivo environments.
We had previously established that Foxk1 was expressed in MPCs (satellite cells) using light microscopic and ultra-structural immunohistochemical techniques . Although the relationship between SMSP and MPC is ill defined, we undertook immunohistochemical techniques to examine Foxk1 expression in sorted SMSP cells. We observed that a subpopulation of the SMSP cells expressed Foxk1 (Fig. 2J). These results establish that Foxk1 is expressed in skeletal muscle progenitor cells (satellite or MPC and SP cells) .
Abcg2-Expressing Cells Are Present in Adult Skeletal Muscle
In addition to our results, previous studies support the conclusion that Abcg2 is a determinant of the SP cell phenotype . Using immunohistochemical techniques, we observed rare Abcg2-expressing cells in uninjured adult skeletal muscle (data not shown). After cardiotoxin-induced skeletal muscle injury (5 days after injury), we observed an increase in Abcg2-positive cells, which were in close approximation to vascular structures (Fig. 3A). To further corroborate our immunohistochemical results, we used RT-PCR techniques and primers spanning an intron and observed Abcg2 expression in the SP cells of skeletal muscle, the capillaries of the postnatal pupillary membrane and hyaloid vasculature, and the aorta (Fig. 3B). Abcg2 expression was absent in the MP population and isolated myofibers (n = 400 myofibers; performed in triplicate). Foxk1 was expressed in all of the samples, consistent with its expression in SP cells and MPCs. Furthermore, Ly6a (Sca-1), c-Kit, and Agpt2 are all expressed in SMSP cells, whereas the panendothelial marker PCAM (CD31) and the intermediate filament protein desmin (which is expressed in differentiated muscle) were absent in SMSP cells (Fig. 3B).
Figure 3. Abcg2 is expressed in SP cells. (A): Using immunohistochemical techniques, Abcg2-expressing cells (arrowheads) are located within injured wild-type muscle and closely associated with the vasculature (*). (B): RT polymerase chain reaction analysis reveals Abcg2 expression in the SP cell population isolated from skeletal muscle, the capillaries of the postnatal pupillary membrane (Cp), and the adult aorta (Ao). Note the absence of Abcg2 transcripts in SP myogenic progenitor and in isolated muscle fibers (SF). Foxk1 is expressed in skeletal muscle SP and MP cell populations, myofibers, capillaries of the postnatal pupillary membrane, and aorta, consistent with the expression of Foxk1 in SP and MP cell populations. Ly6a (Sca-1), cKit, and Agpt2 are all expressed in the SP cell populations, whereas the intermediate filament protein, desmin, is not expressed in SP cells but is expressed in differentiated muscle cells associated with isolated fibers and vascular structures (Cp and Ao). Pecam (CD31) is expressed in the Cp and Ao specimens. Gapd expression was used as a loading control. (C): Using Hoechst 33342 dye and dual-wavelength fluorescence-activated cell sorter analysis of the Tie2-GFP transgenic mouse heart, 84% (average of two preparations; 82% and 86%) of SMSP cells were GFP+. (D): This percentage was confirmed when the sorted skeletal muscle SP cells were examined using fluorescent microscopy (arrowheads mark GFP-expressing SP cells). Scale bar = 20 μm for (A) and 50 μm for (D). Abbreviations: MP, main population; RT, reverse transcriptase; SP, side population.
The angiopoietins are growth factors that bind and activate Tie2/Tek, which is a receptor tyrosine kinase. Tie2 signal transduction pathways are involved in cell survival and cell migration . To additionally examine the Abcg2 expression in vascular structures, we used the Tie2-GFP transgenic mouse model to determine the relationship between the SP cells and progenitor cell populations, because Tie2 is expressed in selected progenitor cell populations (hematopoietic progenitors) and endothelial cells throughout development and in the adult . Using FACS analysis, we observed that most (84%) of the sorted SMSP cells expressed GFP (Fig. 3C). These results were additionally confirmed using epifluorescent microscopy, because most of the sorted SP cells expressed GFP (Fig. 3D). In addition, these results are additionally supported by the expression of angiopoietin2 in the SMSP population (Fig. 3B). We conclude from these studies that most of the SMSP cells share signaling pathways with endothelial/hematopoietic precursor cell populations.
Decreased SP Cell Numbers in Foxk1-Null Skeletal Muscle That Increase after Muscle Injury
We had previously shown that mice lacking Foxk1 have a severe impairment in muscle regeneration attributable, in part, to decreased MPC numbers . Therefore, we examined the SMSP population in Foxk1-deficient skeletal muscle. Using flow cytometry, we have reproducibly isolated SMSP cells from WT skeletal muscle (Fig. 4A). Routinely, we obtain approximately 2,000 to 4,000 SP cells from the hindlimbs of eight adult C57Bl/6 mice. The ability of the SP cells to efflux Hoechst dye is inhibited with the fungicide FTC, which is an Abcg2 inhibitor (Fig. 4B). We further observed that Foxk1 adult mice had approximately half the number of SMSP (Fig. 4C) compared with the age- and gender-matched WT SMSP (Fig. 4A). After cardiotoxin-induced skeletal muscle injury in WT adult mice, we observed more than a 4.5-fold increase in SP cells 5 days after the injury (0.15 ± 0.01% at baseline, which increases to 0.72 ± 0.04%, n = 3; p < .05), consistent with the hypothesis that this cell population participates in muscle repair (Figs. 4D, 4F). Ten days after injury, the WT muscle architecture is largely restored and the SMSP approaches baseline levels (Fig. 4F). Foxk1 mutant skeletal muscle has decreased numbers of SMSP in unperturbed skeletal muscle, and this cell population increases after skeletal muscle injury (0.11 ± 0.01% at baseline, which increases to 0.39 ± 0.04%, n = 3; p < .05 at 5 days after injury). At all time periods after cardiotoxin-induced muscle injury, Foxk1 mutant skeletal muscle had fewer numbers of SMSP compared with WT controls (Fig. 4F). These results support the conclusion that although Foxk1 SMSP cells are decreased in number, they respond similarly to the signals and cues as WT control SP cells after muscle injury. These results may additionally explain the impaired regenerative capacity of the Foxk1 mutant skeletal muscle.
Figure 4. SP cells increase after injury but are decreased in Foxk1 skeletal muscle. (A): Representative FACS profile of skeletal muscle SP cells. Note the SP cells are located in the gated region and account for 0.21% of the total cell population. (B): Inhibition of the SP cell phenotype after the addition of the Abcg2 inhibitor, fumitremorgin C. (C): Fluorescence-activated cell sorter profile reveals fewer SP cells in the Foxk1 mutant muscle compared with WT skeletal muscle. (D): Increased SP cell numbers (compared with uninjured skeletal muscle in A) are observed 5 days after of WT skeletal muscle. (E): Increased SP cell numbers (compared with uninjured Foxk1 skeletal muscle) 5 days after cardiotoxin injury in Foxk1-null skeletal muscle. Note that the increase in SP cell numbers in injured Foxk1-null skeletal muscle is less than injured WT skeletal muscle. (F): Quantitation of the SP cell numbers in WT and Foxk1-injured skeletal muscle. Note that at each time period, WT skeletal muscle has increased numbers of SP cells (n = 3 at each time period; *p < .05). Data are presented as mean ± standard error of the mean. Abbreviations: MP, myogenic progenitor; SP, side population; WT, wild-type.
Foxk1–/–Adult Mice Have an Impaired Muscle Regenerative Capacity
Foxk1 mutant mice have severely impaired muscle regeneration because of decreased numbers of MPCs and SP cells. Ten days after cardiotoxin-induced muscle injury, WT skeletal muscle architecture is largely restored (Fig. 5A). In contrast, 10 days after injury, Foxk1 skeletal muscle is characterized by persistent myonecrosis and a hypercellular response with rare evidence of newly regenerated myofibers (Fig. 5B). We additionally evaluated the contribution of SP cells, which expressed the stem cell marker c-Kit (the receptor for stem cell factor) in unperturbed and injured WT and Foxk1-null skeletal mice. Using flow cytometry in combination with a FITC-conjugated c-Kit antibody, the SMSP from unperturbed WT and Foxk1 skeletal muscle expresses low levels of c-Kit, which is in agreement with previous observations. Five days after injury, c-Kit expression in the SMSP population from WT skeletal muscle increased more than eightfold (1.7 ± 0.4% c-Kit+ SP cells at baseline, increasing to 14.6 ± 0.8%, n = 3; p < .005; Fig. 5C) and approached baseline levels by 10 days after injury (2.0 ± 0.6%, n = 5). In contrast, the c-Kit expression in Foxk1-null SMSP cells was markedly delayed and did not increase until 10 days after injury (0.7 ± 0.04% c-Kit+ SP cells at baseline, increasing to 8.9 ± 2.1%, n = 3; p < .05; Fig. 5C). These results were further confirmed using immunohistochemical techniques for c-Kit expression in unperturbed and injured skeletal muscle. Low levels of c-Kit expression were observed in uninjured WT and Foxk1-null skeletal muscle (data not shown). After muscle injury, c-Kit–expressing cells increased abundantly by 5 days (Fig. 5D) and were largely absent by 10 days after injury in WT skeletal muscle (Fig. 5F). The temporal expression pattern for c-Kit expression differed in Foxk1-injured skeletal muscle, because c-Kit+ cells were absent at 5 days (Figs. 5E) but increased by 10 days after injury (Fig. 5G). At all time periods during the repair process, the c-Kit–expressing cells were less in Foxk1 skeletal muscle compared with WT regenerating skeletal muscle.
Figure 5. Morphological assessment of the skeletal muscle of WT and Foxk1-null mice after cardiotoxin injury. Hematoxylin and eosin–stained tissue sections of cardiotoxin-injured WT and Foxk1-null skeletal muscle. (A): Ten days after injury, WT skeletal muscle has many centronucleated myofibers (arrowheads) corresponding to complete regeneration. (B): In contrast, Foxk1-null skeletal muscle has severely impaired regeneration characterized by a hypercellular myonecrotic state with only rare centronucleated myofibers 10 days after injury. (C): Percentage of WT and Foxk1–/– skeletal muscle SP cells (days 0, 5, and 10 after injury) that are c-Kit+. Note an eightfold increase in c-Kit+ SP cells in the WT 5-day injured skeletal muscle (1.7 ± 0.4% c-Kit+ SP cells at baseline, increasing to 14.6 ± 0.8%, n = 3; *p < .005), whereas there is a significant delay (c-Kit+ cells peak at 10 days after injury to 8.9 ± 2.13%; n = 3 for each sample; *p < .05) in the Foxk1-injured skeletal muscle. (D): Immunohisto-chemical expression of c-Kit+ cells (arrowheads) in regenerating skeletal muscle reveals increased c-Kit expression 5 days after injury in WT skeletal muscle with a relative absence of cells present in Foxk1-null skeletal muscle (E). (F): c-Kit–expressing cells are absent in the 10-day regenerated WT skeletal muscle, whereas a small population of c-Kit+ cells (arrowheads) are observed in the Foxk1-null skeletal muscle 10 days after injury (G). Scale bars = 20 μm (A, B, D–G). Abbreviations: SP, side population; WT, wild-type.
SP Cells Have Distinct Transcriptional Signatures
A transcriptional analysis was undertaken to define a signature of gene expression that corresponded to WT and Foxk1–/–SP cell populations from unperturbed and injured skeletal muscle. Previous studies undertaken in our laboratory have rigorously characterized the isolation and amplification of RNA from limited ES cell numbers (100 to 1,000,000 cells), and we observed that these techniques were reproducible with minimal skewing of gene expression . The respective SP cell populations were sorted, and RNA was isolated, amplified (two rounds), labeled, and hybridized to the Affymetrix oligonucleotide array (U74v2A-Chip; performed in duplicate). Using the MAS 5.0 expression report, the percent of probe sets present ranged from 37% to 43%, with the average signal intensity exceeding 725 and 3' to 5' ratios less than 4.5%. These results verify the integrity of the RNA samples and the assay quality for these studies. The analysis was performed with Affymetrix MAS5.0 Software to obtain detection calls of "present" (p < .04) or "absent" ( p > .06). A 95% confidence interval for fold change was constructed using standard errors of expression values. To define a molecular signature for SP cell populations, gene expression was compared to Affymetrix array analysis of RNA isolated and amplified from 3-day-old murine neonatal myoblasts and ES cells (SM-1; passage 7). The respective cell numbers for each sample ranged between 3,000 and 5,000 cells. Gene expression for the respective cell populations was compared with a common denominator (i.e., irradiated STO fibroblast cells).
The analysis of the transcriptional profiling experiments revealed a discrete molecular program associated with the SMSP cell population. We confirmed the expression of Abcg2 in the SP cell populations using array technologies. Additionally, several transcripts expressed in endothelial cells (i.e., Vegf, Tie1, Vwf, Vcam1, Tie2/Tek, and Eng) were coexpressed in SMSP cells using Affymetrix array analysis. Several endothelial-restricted transcripts were absent in the SMSP cells, including the panendothelial marker CD31 (pecam1), the transmembrane ligand ephrinB2, and its receptor tyrosine kinase ephrinB4, which are molecular markers of embryonic arterial and venous endothelial cells, respectively (see online Tables 1 and 2) . Moreover, the expression of selected candidates was confirmed using RT-PCR analysis (Fig. 3B). Collectively, these results suggest either a common ancestry for these lineages (the precursor cell for SMSP and endothelial cells) or the use of shared signaling pathways between these cell populations.
Table 1. Selected candidate gene expression (fold change) of BMSP, SMSP, and ES cells compared with a common denominator (i.e., STO cells)
Table 2. Fold change of selected candidate transcript expression in the respective cell populations
As illustrated in the Venn diagram (Fig. 6 and Table 1), the SMSP:BMSP:ES shared transcriptional program consists of 78 transcripts and broadly includes genes related to chemokines, cell cycle regulatory genes, metabolism, RNA processing, oxidative stress, protein degradation, and Notch signaling (Table 1 and supplementary online Table 1). Other transcripts that are abundantly expressed in ES cells and absent in the SP cell populations include Esg1, Rex3, Oct4, Utf1, Sox2, and Nanog (Table 1 and supplementary online Table 1) . The SMSP:BMSP shared program consists of 467 transcripts (389 of these 467 transcripts are not shared with the ES cell population) related to nuclear proteins (Nmyc, Oct1, and Mef2A), channel proteins (e.g., chloride channel 3 and chloride channel 4), inflammation (colony stimulating factor 2, complement component 1, and cytokine inducible protein 2), cell-surface/adhesion proteins (fibrinogen-like protein 2, Icam, and Tyrobp), hematopoietic markers (Tie1 and Vwf), metabolism (Nd1, Pfkc, and lactotansferrin), stress (Hsp70 and thioredoxin reductase 1), and members of the Notch pathway (Table 1 and supplementary online Table 1). Transcripts expressed only in the SMSP (483 transcripts) and not the bone marrow SP cells include nuclear factors (Meox2, Foxc1, Sox17, and Sox18), endothelial transcripts (adrenomedullin and Ang2), and cell-surface proteins (fibrinogen-like protein 2 and heparin sulfate) (Table 1). These results suggest that the molecular programs of the respective SP cell populations are more similar to each other than to the ES cell population. Furthermore, although the SP cell populations share several housekeeping transcripts, they have distinctive molecular programs.
Figure 6. Transcriptional signatures of adult BMSP and SMSP are distinct from ES cells. Venn diagram of common and distinct molecular programs associated with BMSP, SMSP, and ES cells compared with a common denominator (i.e., STO cells) using Affymetrix array analysis. The number of transcripts significantly dysregulated are included in brackets. Abbreviations: BMSP, bone marrow side population; ES, embryonic stem; SMSP, skeletal muscle side population; WT, wild-type.
We additionally compared the molecular programs of Foxk1–/– SMSP, WT SMSP, and WT myoblasts (Fig. 7, Table 2). The Foxk1-null SMSP cell population shares several transcripts with the WT SMSP molecular program, and both of these programs (WT SMSP and Foxk1-null SMSP cells) differ considerably compared with the myoblast program (Fig. 7). The 269 transcripts that are distinctly expressed in the Foxk1 SMSP include MMPs (MMP8, MMP9, and complement factor H), schlaffen family members (schlafen 2 and schlafen 4), insulin growth factors (IGF2 and IGFBP5), and ectodermal neural cortex 1 (Table 2). Dysregulation of the schlaffen family members and insulin growth factors/binding proteins are consistent with the finding of perturbed cell-cycle regulation of the Foxk1 mutant SMSP cell population. Furthermore, the Foxk1-null SP cells have an induction of the hematopoietic molecular program (e.g., CD45, CD84, CD53, CD52, chemokine receptor 2, chemokine ligand 2, schlafen 4, and lipocalin 2), which may signify the recruitment of hematopoietic SP cells to skeletal muscle in the Foxk1-null mouse (online Table 3). In addition, we observed several transforming growth factor (TGF)-?–responsive genes to be dysregulated in the Foxk1-null SP cells (Eng, Pdgfa, Enam, Thbs1, Cyp1a1, Vcl, Map3k1, and Tgfbi). Future studies will examine the role of Foxk1 in the regulation of the TGF-? signaling pathway during embryogenesis and in the postnatal SP cell population. Both WT and Foxk1 mutant SMSP cells differ from myoblasts in their expression of Abcg2, endothelial markers (Ang2, endomucin, and endothelin 1), nuclear factors (Foxc1, Mef2A, and Meox2), muscle markers (MyoD, myogenin, myoglobin, troponin C, and skeletal -actin), leukemia inhibitory factor receptor, and members of the Notch pathway (Table 2 and supplementary online Table 1).
Figure 7. Molecular signatures of Foxk1–/– SMSP, WT SMSP, and WT myoblasts. Venn diagram of common and distinct molecular programs of Foxk1-null SMSP, WT SMSP, and WT myoblasts compared with a common denominator (STO cells) using Affymetrix array technology. The number of dysregulated transcripts associated with the respective cell populations is included in brackets. Abbreviations: SMSP, skeletal muscle side population; WT, wild-type.
Table 3. Fold changes of selected transcript expression in the respective cell populations
We observed that SMSP cells increase more than fourfold within 5 days after muscle injury. To additionally evaluate the increased SP cell population after muscle injury, we analyzed the transcriptional program of this cell population using array analysis (Fig. 8 and Table 3). We observed that 314 transcripts were shared between the cell populations (i.e., SMSP isolated from unperturbed versus 5-day injured skeletal muscle), which were categorized using gene ontology annotation. The largest categories included transcriptional regulators, metabolism, and intracellular signaling (data not shown). Selected transcripts shared between these SP cell populations included nuclear factors (Sox7, Sox17, and Sox18), endothelial markers (protein C receptor, Vwf, Tie1, thrombomodulin, and Icam2) interleukins (interleukin-6), and members of the Notch and Wnt pathways (Table 3). The molecular program of the SMSP cells (5 days after injury) express limited numbers of muscle markers (skeletal -actin, sarcoglycan epsilon, and myosin light chain alkali) and has decreased expression of proliferation markers (Pcna; Table 3). These results support the conclusion that SP cells participate in muscle repair.
Figure 8. Molecular signature of WT SMSP isolated from regenerating skeletal muscle. Venn diagram of common and distinct molecular programs associated with BMSP, SMSP (isolated from normal adult skeletal muscle), and SMSP isolated from WT skeletal muscle 5 days after injury compared with a common denominator (i.e., STO cells). Note the number of significantly dysregulated transcripts that are indicated in brackets. Abbreviations: BMSP, bone marrow side population; SMSP, skeletal muscle side population; WT, wild-type.
DISCUSSION
We thank Sean Goetsch and John Shelton for their technical assistance and Dennis Bellotto for assistance with electron microscopy. We thank Dr. Susan E. Bates (National Institutes of Health) for the anti-Abcg2 serum and Dr. Lee Greenberger (Wyeth Research) for the FTC. We also acknowledge Drs. Cindy Martin, Eric N. Olson, Rhonda Bassel-Duby, and Beverly Rothermel for helpful discussions throughout the course of these studies. We thank Dr. Margaret A. Goodell (Baylor College of Medicine) and Shannon McKinney-Freeman for technical assistance and discussions throughout this study. This work was supported in part by grants from the Muscular Dystrophy Association (to A.P.M. and D.J.G.), the NIH (AR47850), and the Donald W. Reynolds Foundation.
REFERENCES
Schultz E, McCormick KM. Skeletal muscle satellite cells. Rev Physiol Biochem Pharmacol 1994;123:213–257.
Hawke T, Garry DJ. Myogenic satellite cells: physiology to molecular biology. J Appl Physiol 2001;91:534–551.
Seale P, Rudnicki MA. A new look at the origin, function, and "stem-cell" status of muscle satellite cells. Dev Biol 2000;218:115–124.
Conboy IM, Rando TA. The regulation of notch signaling controls satellite cell activation and cell fate determination in postnatal myogenesis. Dev Cell 2002;3:397–409.
Garry DJ, Yang Q, Bassel-Duby R et al. Persistent expression of MNF identifies myogenic stem cells in postnatal muscle. Dev Biol 1997;188:280—294.
Hawke T, Jiang N, Garry DJ. Absence of p21 rescues myogenic progenitor cell proliferative and regenerative capacity in Foxk1 null mice. J Biol Chem 2003;278:4015–4020.
Garry DJ, Meeson AP, Elterman J et al. Myogenic stem cell function is impaired in mice lacking the forkhead/winged helix protein MNF. Proc Natl Acad Sci U S A 2000;97: 5416–5421.
Bunting KD. ABC transporters as phenotypic markers and functional regulators of stem cells. STEM CELLS 2002;20: 11–20.
Asakura A, Rudnicki MA. Side population cells from diverse adult tissues are capable of in vitro hematopoietic differentiation. Exp Hematol 2002;30:1339–1345.
Goodell MA, Brose K, Paradis G et al. Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo. J Exp Med 1996;183:1797–1806.
Goodell MA, Rosenzweig M, Kim H et al. Dye efflux studies suggest that hematopoietic stem cells expressing low or undetectable levels of CD34 antigen exist in multiple species. Nat Med 1997;3:1337–1345.
Zhou S, Schuetz J, Bunting K et al. The ABC transporter Bcrp1/Abcg2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype. Nat Med 2001;7:1028–1034.
Zhou S, Morris J, Barnes Y et al. Bcrp1 gene expression is required for normal numbers of side population stem cells in mice, and confers relative protection to mitoxantrone in hematopoietic cells in vivo. Proc Natl Acad Sci U S A. 2002;99:12339–12344.
Gussoni E, Soneoka Y, Strickland CD et al. Dystrophin expression in the mdx mouse restored by stem cell transplantation. Nature 1999;401:390–394.
Seale P, Sabourin LA, Girgis-Gabardo A et al. Pax7 is required for the specification of myogenic satellite cells. Cell 2000;102:777–786.
Motoike T, Loughna S, Perens E et al. Universal GFP reporter for the study of vascular development. Genesis 2000;28:75–81.
Jackson KI, Mi T, Goodell M. Hematopoietic potential of stem cells isolated from murine skeletal muscle. Proc Natl Acad Sci U S A 1999;96:14482–14486.
Meeson AP, Radford N, Shelton JM et al. Adaptive mechanisms that preserve cardiac function in mice without myoglobin. Circ Res 2001;88:713–720.
Gallardo T, Hammer R, Garry DJ. RNA amplification and transcriptional profiling for analysis of stem cell populations. Genesis 2003;37:57–63.
Kim M, Turnquist H, Jackson J et al. The multidrug resistance transporter Abcg2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is over expressed in hematopoietic stem cells. Clin Cancer Res 2002;8:22–28.
Scharenberg CW, Harkey M, Torok-Storb B. The Abcg2 transporter is an efficient Hoechst 33342 efflux pump and is preferentially expressed by immature human hematopoietic progenitors. Blood 2002;99:507–512.
Litman T, Drule T, Stein W et al. From MDR to MXR: new understanding of multidrug resistance systems, their properties and clinical significance. Cell Mol Life Sci 2001; 58:931–959.
Asakura A, Seale P, Girgis-Gabardo A et al. Myogenic specification of side population cells in skeletal muscle. J Cell Biol 2002;159:123–134.
Gaca MD, Zhou X, Issa R et al. Basement membrane-like matrix inhibits proliferation and collagen synthesis by activated rat hepatic stellate cells: evidence for matrix-dependent deactivation of stellate cells. Matrix Biol 2003;22: 229–239.
Ward NL, Dumont DL. The angiopoietins and Tie2/Tek: adding to the complexity of cardiovascular development. Stem Cell Dev Biol 2002;13:19–27.
Puri MC, Partanen J, Rossant J et al. Interaction of the Tek and Tie receptor tyrosine kinases during cardiovascular development. Development 1999;126:4569–4580.
Shin D, Garcia-Cardena G, Hayashi S et al. Expression of ephrin B2 identifies a stable genetic difference between arterial and venous vascular smooth muscle as well as endothelial cells, and marks subsets of microvessels at sites of adult neovascularization. Dev Biol 2001;15:139–150.
Gerety S, Anderson DJ. Cardiovascular ephrin B2 function is essential for embryonic angiogenesis. Development 2002;129:1397–1410.
Kim I, Ryu YS, Kwak HJ et al. EphB ligand, ephrin B2, suppresses the VEGF and angiopoietin 1 induced Ras/mitogen-activated protein kinase pathway in venous endothelial cells. FASEB J 2002;16:1126–1128.
Majka SM, Jackson K, Kienstra K et al. Distinct progenitor populations in skeletal muscle are bone marrow derived and exhibit different fates during vascular regeneration. J Clin Invest 2003;111:71–79.
Wulf GG, Luo KL, Jackson KA et al. Cells of the hepatic side population contribute to liver regeneration and can be replenished with bone marrow stem cells. Haematologica 2003;88:368–378.
Jiang Y, Jahagirdar B, Reinhardt RL et al. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature 2002;418:41–49.
Sutton J, Costa R, Klyg M et al. Genesis, a winged helix transcriptional repressor with expression restricted to embryonic stem cells. J Biol Chem 1996;271:23136–23133.
Ye H, Kelly T, Samandi U et al. Hepatocyte nuclear factor 3/forkhead homolog 11 is expressed in proliferating epithelial and mesenchymal cells of embryonic and adult tissue. Mol Cell Biol 1997;17:1626–1641.
Dou C, Lee J, Liu B et al. BF-1 interferes with transforming growth factor beta signaling by associating with Smad partners. Mol Cell Biol 2000;20:6201–6211.
Lai E, Clark KL, Burley S et al. Hepatocyte nuclear factor 3/forkhead or "winged helix" proteins: a family of transcription factors of diverse biological functions. Proc Natl Acad Sci U S A 1993;90:10421–10423.
Xuan S, Baptista CA, Balas G et al. Winged helix transcription factor BF-1 is essential for the development of the cerebral hemispheres. Neuron 1995;14:1141–1152.
Winnier GE, Hargett L, Hogan BL. The winged helix transcription factor MFH1 is required for proliferation and patterning of paraxial mesoderm in the mouse embryo. Genes Dev 1997;11:926–940.
Schwarz DA, Katayama CD, Hedrick S. Schlafen, a new family of growth regulatory genes that affect thymocyte development. Immunity 1998;9:657–668.
Rousse S, Montarras D, Pinset C et al. Up-regulation of Insulin-like growth factor binding protein-5 is independent of muscle cell differentiation, sensitive to rapamycin, but insensitive to wortmannin and LY294002. Endocrinology 1998;139:1487–1493.
Ramalho-Santos M, Yoon S, Matsuzaki Y et al. "Stemness": transcriptional profiling of embryonic and adult stem cells. Science 2002;298;597–600.
Ivanova NB, Dimos JT, Schaniel C et al. A stem cell molecular signature. Science 2002;298:601–604.
Benchaouir R, Rameau P, Decraene C et al. Evidence for a resident subset of cells with SP phenotype in the C2C12 myogenic line: a tool to explore muscle stem cell biology. Exper Cell Res 2004:294:254–268.(Annette P. Meesona, Thoma)
b Cardiothoracic Surgery,
c Molecular Biology, and
d The Donald W. Reynolds Cardiovascular Research Center, University of Texas Southwestern Medical Center, Dallas, Texas, USA
Key Words. Side population cells ? Myoblast ? Foxk1 ? Skeletal muscle ? Bone marrow
Correspondence: Daniel J. Garry, M.D., Ph.D., NB11.118A, 5323 Harry Hines Boulevard, University of Texas Southwestern Medical Center, Dallas, Texas 75390-8573, USA. Telephone: 214-648-1654; Fax:214-648-1450; e-mail: daniel.garry@utsouthwestern.edu
ABSTRACT
Adult skeletal muscle is capable of self-repair in response to extreme training, injury, or myopathic diseases such as muscular dystrophy . The ability for self-repair or regeneration is attributable to a rare population of myogenic progenitor cells (MPCs), also referred to as satellite cells . The MPCs are capable of self-renewal and are arrested at an early stage of the myogenic program such that they do not express any of the myogenic basic helix-loop-helix proteins of the MyoD family . In response to an injury, the quiescent MPCs are activated, proliferate, and withdraw from the cell cycle to form differentiated myotubes, and ultimately restore the skeletal muscle architecture.
Foxk1 is a member of the forkhead/winged helix family of transcription factors and is expressed in the MPC population, where it functions as a regulator of the cell cycle . Mice lacking Foxk1 have a growth deficit and severely impaired muscle regeneration . This decrease in regenerative capacity is attributable, in part, to a decrease in MPC number, impaired MPC activation, and impaired MPC proliferation in Foxk1-deficient skeletal muscle . Studies undertaken in mice lacking both Foxk1 and p21 suggest that p21 is a downstream target for Foxk1, but other target genes for this forkhead/winged helix transcription factor are unknown .
Recently, an additional stem cell population referred to as side population (SP) cells has been identified in most adult tissues, including skeletal muscle . Using flow cytometry, these SP cells are isolated from adult tissues based on their ability to efflux Hoechst 33342 dye . This ability to efflux Hoechst dye is attributable to the ATP binding cassette (ABC) half-transporter, Abcg2, which has been shown to be the molecular determinant of the SP cell phenotype . SP cells have stem cell properties, because limited numbers of SP cells isolated from adult bone marrow were able to reconstitute the irradiated mdx (dystrophin mutant) mouse bone marrow. Later these cells were recruited from the bone marrow to contribute to differentiated myotubes during muscle regeneration . Previous studies suggest that SP cells may be the precursors of the MPC (satellite) population and require Pax7 for the specification of the satellite cell, because Pax7-deficient mice lack satellite cells but contain a complete repertoire of SP cells . Alternatively, SP cells may represent a second distinct progenitor or stem cell population that is resident in adult skeletal muscle. Although these elegant studies provide additional support for the role of skeletal muscle SP (SMSP) cells as a progenitor/stem cell population, the molecular regulation of this cell population is illdefined.
In the present study, we use morphological, molecular, and transcriptome analyses to enhance our understanding of the SP cell population isolated from wild-type (WT), injured, and Foxk1 mutant adult skeletal muscle. Comparison of these SP cell populations with embryonic stem cells and WT myoblasts additionally defines common and distinct molecular signatures for the respective SP cell populations. Collectively, this strategy will enhance our understanding of the SP cell population and impact therapeutic applications in the treatment of myopathic diseases such as muscular dystrophy.
MATERIALS AND METHODS
Abcg2 Is a Determinant of the SP Cell Phenotype
Using dual-wavelength flow cytometry, SP cells were isolated from several adult tissues based on the ability of the cell population to efflux Hoechst 33342 dye . Recent studies have demonstrated that the ability of SP cells to efflux Hoechst 33342 dye is dependent on the ABC half-transporter Abcg2 . Furthermore, the fungicide FTC has been reported to be a specific inhibitor of Abcg2, where it functions to inhibit the Abcg2-associated ATPase activity . Using these protocols, we isolated SP cells from adult murine skeletal muscle (Fig. 1A) and bone marrow (Fig. 1C). Routinely, the absolute number of SP cells obtained from bone marrow exceeded that obtained from adult skeletal muscle. Furthermore, the ability of these SP cell populations to efflux Hoechst dye was inhibited equally well by either the calcium channel blocker verapamil (data not shown) or after the addition of FTC (Figs. 1B, 1D). Preliminary studies undertaken in our laboratory reveal that overexpression of Abcg2 in C2C12 myoblasts confers the SP cell phenotype and that the ability to efflux the dye in the Abcg2-overex-pressing C2C12 myoblasts can be inhibited completely by FTC (C. Martin, D. Garry, personal observations).
Figure 1. Isolation of SP cell populations using flow cytometry. SP cell populations were isolated from adult skeletal muscle and adult bone marrow, respectively (A, C) using dual-wavelength fluorescence-activated cell sorter and Hoechst 33342 dye. Note that SP cells are located within the gated regions. Addition of fumitremorgin C to skeletal muscle (B) and bone marrow (D) cell preparations inhibited the ability of the SP cells to efflux the Hoechst 33342 dye. Abbreviations: MP, main population; SP, side population.
Morphological Characteristics and Differentiation Capacity of SMSP
SP cells were isolated from noninjured WT adult mouse hind-limb skeletal muscle using Hoechst 33342 dye efflux and FACS analysis. Sorted cells were stained with the Hema3 kit (Fischer Diagnostics, Middleton, VA) and examined at the light microscopic level. We observed that the SMSP cells were relatively small (6.6 ± 0.1 μm, n = 154) and had a high nuclear-to-cytoplasmic ratio (Fig. 2A). Ultrastructural examination of SMSP cells additionally confirmed that these small mononucleated cells had a high nuclear-to-cytoplasmic ratio, with few intracellular organelles (Fig. 2B). A high nuclear-to-cytoplasmic ratio is characteristic of quiescent stem cell populations, such as embryonic stem cells, and somatic stem cell populations, such as the MPCs (i.e., satellite cells) in adult skeletal muscle . SMSP are uniformly smaller (Fig. 2C) than the embryonic stem (ES) cells (Fig. 2D) that are propagated in culture.
Figure 2. Morphological analysis of SMSP cells. Microscopic analysis of wild-type SMSP cells using light microscopy (A) and electron microscopy (B) revealed a high nuclear-to-cytoplasmic ratio in SMSP cells. Comparison of SMSP cells (C) with ES cells (D) revealed morphological differences (i.e., size) between these two cell populations. Arrowheads denote the SMSP cells (C) and the ES cells (D). Side population and myogenic progenitor cells were sorted and cultured to evaluate their differentiation capacity. SMSP cells did not differentiate (E), whereas SMMP cells attached to the culture dish and differentiated after serum removal (F). In contrast, sorted SMSP (G) and SMMP (H) readily form multinucleated (stained with Hoechst dye) myotubes when cultured with myogenic progenitor differentiation media and on Matrigel-treated plates. SMSP cells are capable of differentiating into myotubes in vivo. Two days after cardiotoxin-induced skeletal muscle injury, 250 SMSP cells isolated from adult ROSA26 skeletal muscle were intramuscularly delivered into the injured skeletal muscle. Ten days later, the skeletal muscle was harvested and stained with X-gal. Note ?-galactosidase–positive fibers (blue) present within the regenerating skeletal muscle (I). Sorted SMSP cells were cytospun onto slides and were observed to express Foxk1 (arrowheads) using immunofluorescence techniques (J). Scale bars = 20 μm (A), 500 nm (B), and 20 μm (C–J). Abbreviations: ES, embryonic stem; SMMP, side population main population; SMSP, skeletal muscle side population.
Previous studies observed that isolated SMSP cells differentiate slowly (2 weeks compared with 1 week for differentiation of skeletal muscle MP ) into a mixture of myoblasts and fibroblasts or that isolated SMSP cells cultured in myoblast growth medium did not give rise to myogenic progenitors unless cultured in the presence of primary myoblasts . In the present study, we observed that sorted SMSP cells are viable but relatively quiescent, without any evidence of differentiation when cultured in F10 media supplemented with 20% FCS (Fig. 2E). In contrast, SMMP cells readily attach to the culture dish and differentiate even under high serum culture conditions (Fig. 2F). No cells exhibiting the morphology of the MP cells were detected in the isolated SP cell cultures. We further observed that sorted SMSP (Fig. 2G) and MP (Fig. 2H) cells cultured individually on Matrigel-pretreated plates in the presence of a MPC-differentiation media formed differentiated multinucleated myotubes but at different rates (3 days for SMMP compared with 7 days for SMSP; n = 3 separate experiments). It is possible that the Matrigel provides a surface similar to the extracellular matrix in skeletal muscle, allowing the SP cells to adhere and differentiate when exposed to this matrix .
Having established that the SMSP cells were capable of differentiation in vitro, we examined their ability to repopulate injured skeletal muscle. Two days after cardiotoxin-induced skeletal muscle injury, 250 SP cells (isolated from the skeletal muscle of the ROSA26 mouse model in which all cells constitutively express ?-galactosidase) were delivered intramuscularly into the regenerating skeletal muscle. After skeletal muscle regeneration (10 days after delivery of the cells), we observed evidence of ?-galactosidase–positive fibers (Fig. 2I). These results support the conclusion that limited SP cell numbers are capable of competing with endogenous progenitor cell populations in regenerating skeletal muscle to promote muscle repair. These results additionally establish that the SMSPs are capable of differentiation into skeletal muscle after the exposure to in vitro and in vivo environments.
We had previously established that Foxk1 was expressed in MPCs (satellite cells) using light microscopic and ultra-structural immunohistochemical techniques . Although the relationship between SMSP and MPC is ill defined, we undertook immunohistochemical techniques to examine Foxk1 expression in sorted SMSP cells. We observed that a subpopulation of the SMSP cells expressed Foxk1 (Fig. 2J). These results establish that Foxk1 is expressed in skeletal muscle progenitor cells (satellite or MPC and SP cells) .
Abcg2-Expressing Cells Are Present in Adult Skeletal Muscle
In addition to our results, previous studies support the conclusion that Abcg2 is a determinant of the SP cell phenotype . Using immunohistochemical techniques, we observed rare Abcg2-expressing cells in uninjured adult skeletal muscle (data not shown). After cardiotoxin-induced skeletal muscle injury (5 days after injury), we observed an increase in Abcg2-positive cells, which were in close approximation to vascular structures (Fig. 3A). To further corroborate our immunohistochemical results, we used RT-PCR techniques and primers spanning an intron and observed Abcg2 expression in the SP cells of skeletal muscle, the capillaries of the postnatal pupillary membrane and hyaloid vasculature, and the aorta (Fig. 3B). Abcg2 expression was absent in the MP population and isolated myofibers (n = 400 myofibers; performed in triplicate). Foxk1 was expressed in all of the samples, consistent with its expression in SP cells and MPCs. Furthermore, Ly6a (Sca-1), c-Kit, and Agpt2 are all expressed in SMSP cells, whereas the panendothelial marker PCAM (CD31) and the intermediate filament protein desmin (which is expressed in differentiated muscle) were absent in SMSP cells (Fig. 3B).
Figure 3. Abcg2 is expressed in SP cells. (A): Using immunohistochemical techniques, Abcg2-expressing cells (arrowheads) are located within injured wild-type muscle and closely associated with the vasculature (*). (B): RT polymerase chain reaction analysis reveals Abcg2 expression in the SP cell population isolated from skeletal muscle, the capillaries of the postnatal pupillary membrane (Cp), and the adult aorta (Ao). Note the absence of Abcg2 transcripts in SP myogenic progenitor and in isolated muscle fibers (SF). Foxk1 is expressed in skeletal muscle SP and MP cell populations, myofibers, capillaries of the postnatal pupillary membrane, and aorta, consistent with the expression of Foxk1 in SP and MP cell populations. Ly6a (Sca-1), cKit, and Agpt2 are all expressed in the SP cell populations, whereas the intermediate filament protein, desmin, is not expressed in SP cells but is expressed in differentiated muscle cells associated with isolated fibers and vascular structures (Cp and Ao). Pecam (CD31) is expressed in the Cp and Ao specimens. Gapd expression was used as a loading control. (C): Using Hoechst 33342 dye and dual-wavelength fluorescence-activated cell sorter analysis of the Tie2-GFP transgenic mouse heart, 84% (average of two preparations; 82% and 86%) of SMSP cells were GFP+. (D): This percentage was confirmed when the sorted skeletal muscle SP cells were examined using fluorescent microscopy (arrowheads mark GFP-expressing SP cells). Scale bar = 20 μm for (A) and 50 μm for (D). Abbreviations: MP, main population; RT, reverse transcriptase; SP, side population.
The angiopoietins are growth factors that bind and activate Tie2/Tek, which is a receptor tyrosine kinase. Tie2 signal transduction pathways are involved in cell survival and cell migration . To additionally examine the Abcg2 expression in vascular structures, we used the Tie2-GFP transgenic mouse model to determine the relationship between the SP cells and progenitor cell populations, because Tie2 is expressed in selected progenitor cell populations (hematopoietic progenitors) and endothelial cells throughout development and in the adult . Using FACS analysis, we observed that most (84%) of the sorted SMSP cells expressed GFP (Fig. 3C). These results were additionally confirmed using epifluorescent microscopy, because most of the sorted SP cells expressed GFP (Fig. 3D). In addition, these results are additionally supported by the expression of angiopoietin2 in the SMSP population (Fig. 3B). We conclude from these studies that most of the SMSP cells share signaling pathways with endothelial/hematopoietic precursor cell populations.
Decreased SP Cell Numbers in Foxk1-Null Skeletal Muscle That Increase after Muscle Injury
We had previously shown that mice lacking Foxk1 have a severe impairment in muscle regeneration attributable, in part, to decreased MPC numbers . Therefore, we examined the SMSP population in Foxk1-deficient skeletal muscle. Using flow cytometry, we have reproducibly isolated SMSP cells from WT skeletal muscle (Fig. 4A). Routinely, we obtain approximately 2,000 to 4,000 SP cells from the hindlimbs of eight adult C57Bl/6 mice. The ability of the SP cells to efflux Hoechst dye is inhibited with the fungicide FTC, which is an Abcg2 inhibitor (Fig. 4B). We further observed that Foxk1 adult mice had approximately half the number of SMSP (Fig. 4C) compared with the age- and gender-matched WT SMSP (Fig. 4A). After cardiotoxin-induced skeletal muscle injury in WT adult mice, we observed more than a 4.5-fold increase in SP cells 5 days after the injury (0.15 ± 0.01% at baseline, which increases to 0.72 ± 0.04%, n = 3; p < .05), consistent with the hypothesis that this cell population participates in muscle repair (Figs. 4D, 4F). Ten days after injury, the WT muscle architecture is largely restored and the SMSP approaches baseline levels (Fig. 4F). Foxk1 mutant skeletal muscle has decreased numbers of SMSP in unperturbed skeletal muscle, and this cell population increases after skeletal muscle injury (0.11 ± 0.01% at baseline, which increases to 0.39 ± 0.04%, n = 3; p < .05 at 5 days after injury). At all time periods after cardiotoxin-induced muscle injury, Foxk1 mutant skeletal muscle had fewer numbers of SMSP compared with WT controls (Fig. 4F). These results support the conclusion that although Foxk1 SMSP cells are decreased in number, they respond similarly to the signals and cues as WT control SP cells after muscle injury. These results may additionally explain the impaired regenerative capacity of the Foxk1 mutant skeletal muscle.
Figure 4. SP cells increase after injury but are decreased in Foxk1 skeletal muscle. (A): Representative FACS profile of skeletal muscle SP cells. Note the SP cells are located in the gated region and account for 0.21% of the total cell population. (B): Inhibition of the SP cell phenotype after the addition of the Abcg2 inhibitor, fumitremorgin C. (C): Fluorescence-activated cell sorter profile reveals fewer SP cells in the Foxk1 mutant muscle compared with WT skeletal muscle. (D): Increased SP cell numbers (compared with uninjured skeletal muscle in A) are observed 5 days after of WT skeletal muscle. (E): Increased SP cell numbers (compared with uninjured Foxk1 skeletal muscle) 5 days after cardiotoxin injury in Foxk1-null skeletal muscle. Note that the increase in SP cell numbers in injured Foxk1-null skeletal muscle is less than injured WT skeletal muscle. (F): Quantitation of the SP cell numbers in WT and Foxk1-injured skeletal muscle. Note that at each time period, WT skeletal muscle has increased numbers of SP cells (n = 3 at each time period; *p < .05). Data are presented as mean ± standard error of the mean. Abbreviations: MP, myogenic progenitor; SP, side population; WT, wild-type.
Foxk1–/–Adult Mice Have an Impaired Muscle Regenerative Capacity
Foxk1 mutant mice have severely impaired muscle regeneration because of decreased numbers of MPCs and SP cells. Ten days after cardiotoxin-induced muscle injury, WT skeletal muscle architecture is largely restored (Fig. 5A). In contrast, 10 days after injury, Foxk1 skeletal muscle is characterized by persistent myonecrosis and a hypercellular response with rare evidence of newly regenerated myofibers (Fig. 5B). We additionally evaluated the contribution of SP cells, which expressed the stem cell marker c-Kit (the receptor for stem cell factor) in unperturbed and injured WT and Foxk1-null skeletal mice. Using flow cytometry in combination with a FITC-conjugated c-Kit antibody, the SMSP from unperturbed WT and Foxk1 skeletal muscle expresses low levels of c-Kit, which is in agreement with previous observations. Five days after injury, c-Kit expression in the SMSP population from WT skeletal muscle increased more than eightfold (1.7 ± 0.4% c-Kit+ SP cells at baseline, increasing to 14.6 ± 0.8%, n = 3; p < .005; Fig. 5C) and approached baseline levels by 10 days after injury (2.0 ± 0.6%, n = 5). In contrast, the c-Kit expression in Foxk1-null SMSP cells was markedly delayed and did not increase until 10 days after injury (0.7 ± 0.04% c-Kit+ SP cells at baseline, increasing to 8.9 ± 2.1%, n = 3; p < .05; Fig. 5C). These results were further confirmed using immunohistochemical techniques for c-Kit expression in unperturbed and injured skeletal muscle. Low levels of c-Kit expression were observed in uninjured WT and Foxk1-null skeletal muscle (data not shown). After muscle injury, c-Kit–expressing cells increased abundantly by 5 days (Fig. 5D) and were largely absent by 10 days after injury in WT skeletal muscle (Fig. 5F). The temporal expression pattern for c-Kit expression differed in Foxk1-injured skeletal muscle, because c-Kit+ cells were absent at 5 days (Figs. 5E) but increased by 10 days after injury (Fig. 5G). At all time periods during the repair process, the c-Kit–expressing cells were less in Foxk1 skeletal muscle compared with WT regenerating skeletal muscle.
Figure 5. Morphological assessment of the skeletal muscle of WT and Foxk1-null mice after cardiotoxin injury. Hematoxylin and eosin–stained tissue sections of cardiotoxin-injured WT and Foxk1-null skeletal muscle. (A): Ten days after injury, WT skeletal muscle has many centronucleated myofibers (arrowheads) corresponding to complete regeneration. (B): In contrast, Foxk1-null skeletal muscle has severely impaired regeneration characterized by a hypercellular myonecrotic state with only rare centronucleated myofibers 10 days after injury. (C): Percentage of WT and Foxk1–/– skeletal muscle SP cells (days 0, 5, and 10 after injury) that are c-Kit+. Note an eightfold increase in c-Kit+ SP cells in the WT 5-day injured skeletal muscle (1.7 ± 0.4% c-Kit+ SP cells at baseline, increasing to 14.6 ± 0.8%, n = 3; *p < .005), whereas there is a significant delay (c-Kit+ cells peak at 10 days after injury to 8.9 ± 2.13%; n = 3 for each sample; *p < .05) in the Foxk1-injured skeletal muscle. (D): Immunohisto-chemical expression of c-Kit+ cells (arrowheads) in regenerating skeletal muscle reveals increased c-Kit expression 5 days after injury in WT skeletal muscle with a relative absence of cells present in Foxk1-null skeletal muscle (E). (F): c-Kit–expressing cells are absent in the 10-day regenerated WT skeletal muscle, whereas a small population of c-Kit+ cells (arrowheads) are observed in the Foxk1-null skeletal muscle 10 days after injury (G). Scale bars = 20 μm (A, B, D–G). Abbreviations: SP, side population; WT, wild-type.
SP Cells Have Distinct Transcriptional Signatures
A transcriptional analysis was undertaken to define a signature of gene expression that corresponded to WT and Foxk1–/–SP cell populations from unperturbed and injured skeletal muscle. Previous studies undertaken in our laboratory have rigorously characterized the isolation and amplification of RNA from limited ES cell numbers (100 to 1,000,000 cells), and we observed that these techniques were reproducible with minimal skewing of gene expression . The respective SP cell populations were sorted, and RNA was isolated, amplified (two rounds), labeled, and hybridized to the Affymetrix oligonucleotide array (U74v2A-Chip; performed in duplicate). Using the MAS 5.0 expression report, the percent of probe sets present ranged from 37% to 43%, with the average signal intensity exceeding 725 and 3' to 5' ratios less than 4.5%. These results verify the integrity of the RNA samples and the assay quality for these studies. The analysis was performed with Affymetrix MAS5.0 Software to obtain detection calls of "present" (p < .04) or "absent" ( p > .06). A 95% confidence interval for fold change was constructed using standard errors of expression values. To define a molecular signature for SP cell populations, gene expression was compared to Affymetrix array analysis of RNA isolated and amplified from 3-day-old murine neonatal myoblasts and ES cells (SM-1; passage 7). The respective cell numbers for each sample ranged between 3,000 and 5,000 cells. Gene expression for the respective cell populations was compared with a common denominator (i.e., irradiated STO fibroblast cells).
The analysis of the transcriptional profiling experiments revealed a discrete molecular program associated with the SMSP cell population. We confirmed the expression of Abcg2 in the SP cell populations using array technologies. Additionally, several transcripts expressed in endothelial cells (i.e., Vegf, Tie1, Vwf, Vcam1, Tie2/Tek, and Eng) were coexpressed in SMSP cells using Affymetrix array analysis. Several endothelial-restricted transcripts were absent in the SMSP cells, including the panendothelial marker CD31 (pecam1), the transmembrane ligand ephrinB2, and its receptor tyrosine kinase ephrinB4, which are molecular markers of embryonic arterial and venous endothelial cells, respectively (see online Tables 1 and 2) . Moreover, the expression of selected candidates was confirmed using RT-PCR analysis (Fig. 3B). Collectively, these results suggest either a common ancestry for these lineages (the precursor cell for SMSP and endothelial cells) or the use of shared signaling pathways between these cell populations.
Table 1. Selected candidate gene expression (fold change) of BMSP, SMSP, and ES cells compared with a common denominator (i.e., STO cells)
Table 2. Fold change of selected candidate transcript expression in the respective cell populations
As illustrated in the Venn diagram (Fig. 6 and Table 1), the SMSP:BMSP:ES shared transcriptional program consists of 78 transcripts and broadly includes genes related to chemokines, cell cycle regulatory genes, metabolism, RNA processing, oxidative stress, protein degradation, and Notch signaling (Table 1 and supplementary online Table 1). Other transcripts that are abundantly expressed in ES cells and absent in the SP cell populations include Esg1, Rex3, Oct4, Utf1, Sox2, and Nanog (Table 1 and supplementary online Table 1) . The SMSP:BMSP shared program consists of 467 transcripts (389 of these 467 transcripts are not shared with the ES cell population) related to nuclear proteins (Nmyc, Oct1, and Mef2A), channel proteins (e.g., chloride channel 3 and chloride channel 4), inflammation (colony stimulating factor 2, complement component 1, and cytokine inducible protein 2), cell-surface/adhesion proteins (fibrinogen-like protein 2, Icam, and Tyrobp), hematopoietic markers (Tie1 and Vwf), metabolism (Nd1, Pfkc, and lactotansferrin), stress (Hsp70 and thioredoxin reductase 1), and members of the Notch pathway (Table 1 and supplementary online Table 1). Transcripts expressed only in the SMSP (483 transcripts) and not the bone marrow SP cells include nuclear factors (Meox2, Foxc1, Sox17, and Sox18), endothelial transcripts (adrenomedullin and Ang2), and cell-surface proteins (fibrinogen-like protein 2 and heparin sulfate) (Table 1). These results suggest that the molecular programs of the respective SP cell populations are more similar to each other than to the ES cell population. Furthermore, although the SP cell populations share several housekeeping transcripts, they have distinctive molecular programs.
Figure 6. Transcriptional signatures of adult BMSP and SMSP are distinct from ES cells. Venn diagram of common and distinct molecular programs associated with BMSP, SMSP, and ES cells compared with a common denominator (i.e., STO cells) using Affymetrix array analysis. The number of transcripts significantly dysregulated are included in brackets. Abbreviations: BMSP, bone marrow side population; ES, embryonic stem; SMSP, skeletal muscle side population; WT, wild-type.
We additionally compared the molecular programs of Foxk1–/– SMSP, WT SMSP, and WT myoblasts (Fig. 7, Table 2). The Foxk1-null SMSP cell population shares several transcripts with the WT SMSP molecular program, and both of these programs (WT SMSP and Foxk1-null SMSP cells) differ considerably compared with the myoblast program (Fig. 7). The 269 transcripts that are distinctly expressed in the Foxk1 SMSP include MMPs (MMP8, MMP9, and complement factor H), schlaffen family members (schlafen 2 and schlafen 4), insulin growth factors (IGF2 and IGFBP5), and ectodermal neural cortex 1 (Table 2). Dysregulation of the schlaffen family members and insulin growth factors/binding proteins are consistent with the finding of perturbed cell-cycle regulation of the Foxk1 mutant SMSP cell population. Furthermore, the Foxk1-null SP cells have an induction of the hematopoietic molecular program (e.g., CD45, CD84, CD53, CD52, chemokine receptor 2, chemokine ligand 2, schlafen 4, and lipocalin 2), which may signify the recruitment of hematopoietic SP cells to skeletal muscle in the Foxk1-null mouse (online Table 3). In addition, we observed several transforming growth factor (TGF)-?–responsive genes to be dysregulated in the Foxk1-null SP cells (Eng, Pdgfa, Enam, Thbs1, Cyp1a1, Vcl, Map3k1, and Tgfbi). Future studies will examine the role of Foxk1 in the regulation of the TGF-? signaling pathway during embryogenesis and in the postnatal SP cell population. Both WT and Foxk1 mutant SMSP cells differ from myoblasts in their expression of Abcg2, endothelial markers (Ang2, endomucin, and endothelin 1), nuclear factors (Foxc1, Mef2A, and Meox2), muscle markers (MyoD, myogenin, myoglobin, troponin C, and skeletal -actin), leukemia inhibitory factor receptor, and members of the Notch pathway (Table 2 and supplementary online Table 1).
Figure 7. Molecular signatures of Foxk1–/– SMSP, WT SMSP, and WT myoblasts. Venn diagram of common and distinct molecular programs of Foxk1-null SMSP, WT SMSP, and WT myoblasts compared with a common denominator (STO cells) using Affymetrix array technology. The number of dysregulated transcripts associated with the respective cell populations is included in brackets. Abbreviations: SMSP, skeletal muscle side population; WT, wild-type.
Table 3. Fold changes of selected transcript expression in the respective cell populations
We observed that SMSP cells increase more than fourfold within 5 days after muscle injury. To additionally evaluate the increased SP cell population after muscle injury, we analyzed the transcriptional program of this cell population using array analysis (Fig. 8 and Table 3). We observed that 314 transcripts were shared between the cell populations (i.e., SMSP isolated from unperturbed versus 5-day injured skeletal muscle), which were categorized using gene ontology annotation. The largest categories included transcriptional regulators, metabolism, and intracellular signaling (data not shown). Selected transcripts shared between these SP cell populations included nuclear factors (Sox7, Sox17, and Sox18), endothelial markers (protein C receptor, Vwf, Tie1, thrombomodulin, and Icam2) interleukins (interleukin-6), and members of the Notch and Wnt pathways (Table 3). The molecular program of the SMSP cells (5 days after injury) express limited numbers of muscle markers (skeletal -actin, sarcoglycan epsilon, and myosin light chain alkali) and has decreased expression of proliferation markers (Pcna; Table 3). These results support the conclusion that SP cells participate in muscle repair.
Figure 8. Molecular signature of WT SMSP isolated from regenerating skeletal muscle. Venn diagram of common and distinct molecular programs associated with BMSP, SMSP (isolated from normal adult skeletal muscle), and SMSP isolated from WT skeletal muscle 5 days after injury compared with a common denominator (i.e., STO cells). Note the number of significantly dysregulated transcripts that are indicated in brackets. Abbreviations: BMSP, bone marrow side population; SMSP, skeletal muscle side population; WT, wild-type.
DISCUSSION
We thank Sean Goetsch and John Shelton for their technical assistance and Dennis Bellotto for assistance with electron microscopy. We thank Dr. Susan E. Bates (National Institutes of Health) for the anti-Abcg2 serum and Dr. Lee Greenberger (Wyeth Research) for the FTC. We also acknowledge Drs. Cindy Martin, Eric N. Olson, Rhonda Bassel-Duby, and Beverly Rothermel for helpful discussions throughout the course of these studies. We thank Dr. Margaret A. Goodell (Baylor College of Medicine) and Shannon McKinney-Freeman for technical assistance and discussions throughout this study. This work was supported in part by grants from the Muscular Dystrophy Association (to A.P.M. and D.J.G.), the NIH (AR47850), and the Donald W. Reynolds Foundation.
REFERENCES
Schultz E, McCormick KM. Skeletal muscle satellite cells. Rev Physiol Biochem Pharmacol 1994;123:213–257.
Hawke T, Garry DJ. Myogenic satellite cells: physiology to molecular biology. J Appl Physiol 2001;91:534–551.
Seale P, Rudnicki MA. A new look at the origin, function, and "stem-cell" status of muscle satellite cells. Dev Biol 2000;218:115–124.
Conboy IM, Rando TA. The regulation of notch signaling controls satellite cell activation and cell fate determination in postnatal myogenesis. Dev Cell 2002;3:397–409.
Garry DJ, Yang Q, Bassel-Duby R et al. Persistent expression of MNF identifies myogenic stem cells in postnatal muscle. Dev Biol 1997;188:280—294.
Hawke T, Jiang N, Garry DJ. Absence of p21 rescues myogenic progenitor cell proliferative and regenerative capacity in Foxk1 null mice. J Biol Chem 2003;278:4015–4020.
Garry DJ, Meeson AP, Elterman J et al. Myogenic stem cell function is impaired in mice lacking the forkhead/winged helix protein MNF. Proc Natl Acad Sci U S A 2000;97: 5416–5421.
Bunting KD. ABC transporters as phenotypic markers and functional regulators of stem cells. STEM CELLS 2002;20: 11–20.
Asakura A, Rudnicki MA. Side population cells from diverse adult tissues are capable of in vitro hematopoietic differentiation. Exp Hematol 2002;30:1339–1345.
Goodell MA, Brose K, Paradis G et al. Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo. J Exp Med 1996;183:1797–1806.
Goodell MA, Rosenzweig M, Kim H et al. Dye efflux studies suggest that hematopoietic stem cells expressing low or undetectable levels of CD34 antigen exist in multiple species. Nat Med 1997;3:1337–1345.
Zhou S, Schuetz J, Bunting K et al. The ABC transporter Bcrp1/Abcg2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype. Nat Med 2001;7:1028–1034.
Zhou S, Morris J, Barnes Y et al. Bcrp1 gene expression is required for normal numbers of side population stem cells in mice, and confers relative protection to mitoxantrone in hematopoietic cells in vivo. Proc Natl Acad Sci U S A. 2002;99:12339–12344.
Gussoni E, Soneoka Y, Strickland CD et al. Dystrophin expression in the mdx mouse restored by stem cell transplantation. Nature 1999;401:390–394.
Seale P, Sabourin LA, Girgis-Gabardo A et al. Pax7 is required for the specification of myogenic satellite cells. Cell 2000;102:777–786.
Motoike T, Loughna S, Perens E et al. Universal GFP reporter for the study of vascular development. Genesis 2000;28:75–81.
Jackson KI, Mi T, Goodell M. Hematopoietic potential of stem cells isolated from murine skeletal muscle. Proc Natl Acad Sci U S A 1999;96:14482–14486.
Meeson AP, Radford N, Shelton JM et al. Adaptive mechanisms that preserve cardiac function in mice without myoglobin. Circ Res 2001;88:713–720.
Gallardo T, Hammer R, Garry DJ. RNA amplification and transcriptional profiling for analysis of stem cell populations. Genesis 2003;37:57–63.
Kim M, Turnquist H, Jackson J et al. The multidrug resistance transporter Abcg2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is over expressed in hematopoietic stem cells. Clin Cancer Res 2002;8:22–28.
Scharenberg CW, Harkey M, Torok-Storb B. The Abcg2 transporter is an efficient Hoechst 33342 efflux pump and is preferentially expressed by immature human hematopoietic progenitors. Blood 2002;99:507–512.
Litman T, Drule T, Stein W et al. From MDR to MXR: new understanding of multidrug resistance systems, their properties and clinical significance. Cell Mol Life Sci 2001; 58:931–959.
Asakura A, Seale P, Girgis-Gabardo A et al. Myogenic specification of side population cells in skeletal muscle. J Cell Biol 2002;159:123–134.
Gaca MD, Zhou X, Issa R et al. Basement membrane-like matrix inhibits proliferation and collagen synthesis by activated rat hepatic stellate cells: evidence for matrix-dependent deactivation of stellate cells. Matrix Biol 2003;22: 229–239.
Ward NL, Dumont DL. The angiopoietins and Tie2/Tek: adding to the complexity of cardiovascular development. Stem Cell Dev Biol 2002;13:19–27.
Puri MC, Partanen J, Rossant J et al. Interaction of the Tek and Tie receptor tyrosine kinases during cardiovascular development. Development 1999;126:4569–4580.
Shin D, Garcia-Cardena G, Hayashi S et al. Expression of ephrin B2 identifies a stable genetic difference between arterial and venous vascular smooth muscle as well as endothelial cells, and marks subsets of microvessels at sites of adult neovascularization. Dev Biol 2001;15:139–150.
Gerety S, Anderson DJ. Cardiovascular ephrin B2 function is essential for embryonic angiogenesis. Development 2002;129:1397–1410.
Kim I, Ryu YS, Kwak HJ et al. EphB ligand, ephrin B2, suppresses the VEGF and angiopoietin 1 induced Ras/mitogen-activated protein kinase pathway in venous endothelial cells. FASEB J 2002;16:1126–1128.
Majka SM, Jackson K, Kienstra K et al. Distinct progenitor populations in skeletal muscle are bone marrow derived and exhibit different fates during vascular regeneration. J Clin Invest 2003;111:71–79.
Wulf GG, Luo KL, Jackson KA et al. Cells of the hepatic side population contribute to liver regeneration and can be replenished with bone marrow stem cells. Haematologica 2003;88:368–378.
Jiang Y, Jahagirdar B, Reinhardt RL et al. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature 2002;418:41–49.
Sutton J, Costa R, Klyg M et al. Genesis, a winged helix transcriptional repressor with expression restricted to embryonic stem cells. J Biol Chem 1996;271:23136–23133.
Ye H, Kelly T, Samandi U et al. Hepatocyte nuclear factor 3/forkhead homolog 11 is expressed in proliferating epithelial and mesenchymal cells of embryonic and adult tissue. Mol Cell Biol 1997;17:1626–1641.
Dou C, Lee J, Liu B et al. BF-1 interferes with transforming growth factor beta signaling by associating with Smad partners. Mol Cell Biol 2000;20:6201–6211.
Lai E, Clark KL, Burley S et al. Hepatocyte nuclear factor 3/forkhead or "winged helix" proteins: a family of transcription factors of diverse biological functions. Proc Natl Acad Sci U S A 1993;90:10421–10423.
Xuan S, Baptista CA, Balas G et al. Winged helix transcription factor BF-1 is essential for the development of the cerebral hemispheres. Neuron 1995;14:1141–1152.
Winnier GE, Hargett L, Hogan BL. The winged helix transcription factor MFH1 is required for proliferation and patterning of paraxial mesoderm in the mouse embryo. Genes Dev 1997;11:926–940.
Schwarz DA, Katayama CD, Hedrick S. Schlafen, a new family of growth regulatory genes that affect thymocyte development. Immunity 1998;9:657–668.
Rousse S, Montarras D, Pinset C et al. Up-regulation of Insulin-like growth factor binding protein-5 is independent of muscle cell differentiation, sensitive to rapamycin, but insensitive to wortmannin and LY294002. Endocrinology 1998;139:1487–1493.
Ramalho-Santos M, Yoon S, Matsuzaki Y et al. "Stemness": transcriptional profiling of embryonic and adult stem cells. Science 2002;298;597–600.
Ivanova NB, Dimos JT, Schaniel C et al. A stem cell molecular signature. Science 2002;298:601–604.
Benchaouir R, Rameau P, Decraene C et al. Evidence for a resident subset of cells with SP phenotype in the C2C12 myogenic line: a tool to explore muscle stem cell biology. Exper Cell Res 2004:294:254–268.(Annette P. Meesona, Thoma)