Abstract
Although the role of satellite cells in muscle growth and repair is well recognized, understanding of the molecular events that accompany their activation and proliferation is limited. In this study, we used the single myofiber culture model for comparing the proliferative dynamics of satellite cells from growing (3-week-old), young adult (8- to 10-week-old), and old (9- to 11-month-old) rats. In these fiber cultures, the satellite cells are maintained in their in situ position underneath the fiber basement membrane. We first demonstrate that the cytoplasm of fiber-associated satellite cells can be monitored with an antibody against the extracellular signal regulated kinases 1 and 2 (ERK1 and ERK2), which belong to the mitogen-activated protein kinase (MAPK) superfamily. With this immunocytological marker, we show that the satellite cells from all three age groups first proliferate and express PCNA and MyoD, and subsequently, about 24 hr later, exit the PCNA+/MyoD+ state and become positive for myogenin. For all three age groups, fibroblast growth factor 2 (FGF2) enhances by about twofold the number of satellite cells that are capable of proliferation, as determined by monitoring the number of cells that transit from the MAPK+ phenotype to the PCNA+/MAPK+ or MyoD+/MAPK+ phenotype. Furthermore, contrary to the commonly accepted convention, we show that in the fiber cultures FGF2 does not suppress the subsequent transition of the proliferating cells into the myogenin+ compartment. Although myogenesis of satellite cells from growing, young adult, and old rats follows a similar program, two distinctive features were identified for satellite cells in fiber cultures from the old rats. First, a large number of MAPK+ cells do not appear to enter the MyoD-myogenin expression program. Second, the maximal number of proliferating satellite cells is attained a day later than in cultures from the young adults. This apparent “lag” in proliferation was not affected by hepatocyte growth factor (HGF), which has been implicated in accelerating the first round of satellite cell proliferation. HGF and FGF2 were equally efficient in promoting proliferation of satellite cells in fibers from old rats. Collectively, the investigation suggests that FGF plays a critical role in the recruitment of satellite cells into proliferation.
Keywords
Satellite cells, the myogenic precursors in postnatal and adult skeletal muscle, are situated between the basement membrane and the plasma membrane of myofibers in growing and mature muscle (Mauro 1961; Bischoff 1989; Yablonka-Reuveni 1995). At least some of these satellite cells are mitotically active in growing muscle, contributing myonuclei to the enlarging fibers (Moss and Leblond 1971). As muscle matures, the addition of myofiber nuclei ceases and the satellite cells become mitotically quiescent (Schultz et al. 1978). These quiescent myogenic precursors can become mitotically active in response to various muscle stresses, and their progeny can either fuse into preexisting fibers or form new myofibers (reviewed in Grounds and Yablonka-Reuveni 1993; Schultz and McCormick 1994). Satellite cell activation, proliferation, and formation of new fibers have been documented after massive muscle injury (reviewed in Carlson and Faulkner 1983; Grounds and Yablonka-Reuveni 1993). Regeneration in these instances involves obvious wound-related activities such as removal of necrotic debris by macrophages and revascularization. Overt muscle injury is not the only condition that leads to satellite cell proliferation. Recruitment of these precursors occurs in response to more subtle stresses such as stretch, exercise, and muscle hypertrophy (Appell et al. 1988; Snow 1990; Winchester et al. 1991; Schultz and McCormick 1994). Muscle denervation was also shown to lead to an increase in the number of satellite cells (Snow 1983; Weis 1994).
After various kinds of muscle trauma, satellite cells enter a program that involves the expression of the myogenic transcription factors MyoD and myogenin (Füchtbauer and Westphal 1992; Grounds et al. 1992; Koishi et al. 1995; Anderson et al. 1997). These factors are members of the basic-helix-loop-helix family of myogenic regulatory factors (MRFs) which are believed to be involved in the specification of the skeletal myogenic lineage during embryogenesis (reviewed in Yun and Wold 1996; Tajbakhsh and Giulio 1997). In early muscle development, MyoD is expressed first, followed by myogenin expression as myoblasts progress through differentiation. The MRFs are also detected in myogenic cultures (Wright et al. 1989; Hinterberger et al. 1991; Smith et al. 1993; Smith et al. 1994; Maley et al. 1994; Yablonka-Reuveni and Rivera 1994, 1997a; Cornelison and Wold 1997). This MRF expression by cells already committed to the muscle lineage probably reflects the role of MRFs in the transition from proliferation to differentiation (reviewed in Olson 1992, 1993; Weintraub 1993).
Because satellite cells are believed to be the only source of myogenic precursors in postnatal and adult muscle, myogenic cultures of cells isolated from juvenile or adult muscles are commonly presumed to be cultures of satellite cells. A number of growth factor families are involved in controlling proliferation and differentiation in such satellite cell cultures (Allen and Boxhorn 1989; Yablonka-Reuveni and Seifert 1993; Tatsumi et al. 1998). Similar conclusions regarding the role of growth factors have been reached in studies of myogenic cell lines derived from newborn and postnatal muscles (Clegg et al. 1987; Yablonka-Reuveni et al. 1990; Florini et al. 1995; Yablonka-Reuveni and Rivera 1997a; Anastasi et al. 1997). To further evaluate the molecular changes that accompany the transition of satellite cells from quiescence into the cell cycle, we have been using cultures of isolated rat myofibers (Yablonka-Reuveni and Rivera 1994). As first demonstrated by Bischoff (1986, 1989), these isolated myofibers retain their few satellite cells in the original site underneath the basement membrane, allowing analysis of satellite cells in their native position beside the myofiber without the complexity of the intact tissue. This association between the satellite cells and the basement/plasma membranes of the myofiber is potentially important for maintaining the satellite cells in a quiescent or proliferative state (Bischoff 1990a, b). In addition, myofibers close to the satellite cells can potentially serve as a paracrine source for various growth-regulating agents.
We reported previously on the use of indirect immunofluorescence to trace myogenesis of satellite cells in the isolated fiber model in analyzing fibers from “young adult” rats (8 to 12 weeks old). We demonstrated that the satellite cells follow a highly coordinated, multistep program of regulatory and structural protein expression (Yablonka-Reuveni and Rivera 1994). As the satellite cells enter the cell cycle, their nuclei become positive for MyoD and for proliferating cell nuclear antigen (PCNA) (Yablonka-Reuveni and Rivera 1994). PCNA is an auxiliary protein to DNA polymerase 8, whose levels correlate with DNA synthesis during the cell cycle, becoming maximal during the S-phase (Bravo et al. 1987; Baserga 1991). Indeed, in fiber cultures, the kinetics of the PCNA-positive nuclei correlated well with the kinetics of DNA-synthesizing satellite cells as determined by tracing [3H]-thymidine-labeled nuclei (Yablonka-Reuveni and Rivera 1997b). After approximately 24 hr in the PCNA+/MyoD+ compartment, the nuclei of the fiber-associated satellite cells become negative for PCNA and MyoD and positive for myogenin. Expression of sarcomeric myosin in satellite cell cytoplasms begins shortly after that of myogenin, but whereas the expression of myogenin is transient, lasting approximately 24 hr, myosin is expressed continuously (Yablonka-Reuveni and Rivera 1994). In the later study we were unable to view the satellite cell cytoplasm before the time that the cells become positive for differentiation-specific cytoplasmic proteins. Assigning the PCNA+, MyoD+, or myogenin+ nuclei to satellite cells undergoing myogenesis was based on the observation that such fiber-associated positive nuclei were uncommon (as expected for satellite cells). Moreover, myofiber nuclei were not expected to become positive for PCNA because it has been commonly accepted that under normal conditions myofiber nuclei do not enter proliferation.
The primary goal of the present study was to compare the proliferative dynamics of satellite cells from “growing” (3 weeks old), “young adult” (8-10 weeks old), and “old” (9-11 months old) rats. Whereas 3-week-old rats grow rapidly and their satellite cells are believed to be proliferative, older rats grow slowly and their satellite cells are believed to be quiescent. Experiments with primary cultures raised the possibility that quiescent satellite cells require, for their first round of cell proliferation, other growth factors then those needed during ongoing proliferation of satellite cells from growing animals (Allen et al. 1995). This possibility arose from the observation that the onset of proliferation in satellite cell cultures from older rats (9 months old) was delayed by about 18 hr compared to cultures from young rats (3-4 weeks old), regardless of the addition of FGF2 (Johnson and Allen 1993). The addition of hepatocyte growth factor (HGF) to cultures of satellite cells from the older rats accelerated the onset of cell replication (Allen et al. 1995). In vivo tracing of proliferating satellite cells after muscle injury demonstrated that cell replication commenced at 18–24 hr after injury in muscles of both young and old mice (McGeachie and Grounds 1995). Furthermore, no differences were found in the number of proliferating satellite cells when muscles from adult and aged quails were subjected to stretch overload (Carson and Alway 1996). Therefore, it is possible that the in vivo environment can provide growth signals that are lacking in cultures of tissue-dispersed satellite cells from the older rodents.
To pursue the analysis of satellite cells in myofibers from the rats of different ages, we first needed to identify a cytoplasmic marker for satellite cells that allows, in combination with the nuclear expression of PCNA, MyoD, and myogenin, unequivocal identification of the cells. This was crucial for the present study; in various models of muscle injury and in different situations of muscle stress, the myofiber nuclei are also believed to express MyoD and/or myogenin, making the tracing of satellite cells on the basis of nuclear expression of the MRFs potentially problematic (Füchtbauer and Westphal 1992; Eppley et al. 1993; Weis 1994; Kami et al. 1995; Koishi et al. 1995; Anderson et al. 1998). For example, it is possible that myofiber nuclei that fused into the myofiber just shortly before fiber isolation (such as in the rapidly growing rats) can be positive for the MRFs. Likewise, it is possible that, in the environment of the aging rats, some of the myofiber nuclei may express the MRF proteins (Musarò et al. 1995). We were especially concerned with the source of the nuclei expressing MyoD and myogenin, because some researchers have considered changes in levels of MRF transcripts after stresses, including stretch, electrical stimulation, and denervation, to represent events occurring within the myofibers themselves (Buonanno et al. 1992; Jacobs-El et al. 1995).
We found that the cytoplasm of satellite cells undergoing myogenesis on isolated fibers can be monitored by immunostaining with an antibody against the extracelluar signal-regulated kinases 1 and 2 (ERK1 and ERK2). These ERKs are members of the mitogenactivated protein kinase (MAPK) superfamily, which is involved in the transmission of extracellular signals to their intracellular targets. The ERK1/ERK2 signal transduction pathway is typically initiated by binding of growth factors to their receptors, generating an activation cascade of specific protein kinases and leading to the activation of ERK1/ERK2. Activated ERK1/ERK2 can phosphorylate regulatory targets in the cytosol or can translocate to the nucleus and phosphorylate transcription factors. Although activated ERK1 and ERK2 are dually phosphorylated on specific tyrosine and threonine residues, at any given time the cell may also contain unphosphorylated ERK1/ERK2, along with ERKs that are singly phophorylated either at the tyrosine or the threonine residue (for reviews see Seger and Krebs 1995; Marshal 1996; Robinson and Cobb 1997).
The immunolabeling with the anti-MAPK antibody (which recognizes both the phosphorylated and nonphosphorylated forms of ERK1/2), combined with the immunolabeling of the nuclear proteins PCNA, MyoD, and myogenin, has provided a direct means for tracing satellite cells as they undergo proliferation and differentiation in fiber cultures, regardless of the age of the muscle from which the fibers are isolated. Using these cytological markers, we show in this study that FGF2 regulates proliferation of satellite cells from both young and old rats. The number of satellite cells entering proliferation and expressing MyoD, as well as the overall number of satellite cells subsequently transiting into the myogenin state, is enhanced in the presence of FGF2 regardless of the age of the muscle. This effect of FGF2 in the single fiber model differs from the commonly accepted convention that FGF suppresses MyoD expression and myogenic differentiation (Clegg et al. 1987; Vaidya et al. 1989; Hannon et al. 1996; Yoshida et al. 1996). We propose that the enhanced number of satellite cells undergoing myogenesis in the presence of FGF2 reflects a specific effect of the growth factor on recruiting satellite cells into active myogenesis in both young and old animals. Modulations in the levels of FGF, possibly induced by a variety of stresses, might be involved in the recruitment and proliferation of satellite cells in vivo.
Materials and Methods
Animals
Male rats (Sprague-Dawley) were used throughout the study and were purchased from B & K Universal (Kent, WA). Unless otherwise noted, the following three age groups were used: young or growing rats-3 weeks old (50-55 g, fast growers); young adult rats-8-10 weeks old (225-300 g, moderate growers); and old rats-9-11 months old (retired breeders; 600-750 g at 9 months, slow growers).
Isolation and Culture of Rat Muscle Fibers
Single muscle fibers with associated satellite cells were prepared from the flexor digitorum brevis (FDB) muscle of rat hind foot according to procedures previously described for 8-12-week-old rats (Bischoff 1986; Yablonka-Reuveni and Rivera 1994). Various modifications were included in the present study to facilitate fiber isolation from young and old rats. Typically, muscles from both hind feet of one to three young rats, or one to two young adult rats, or one to two old rats were used for each preparation (specific details are given in figure legends). For each rat, the outer connective tissue was removed and the two FDB muscles were immersed in a 5-ml solution of 0.2% collagenase type 1 (Sigma, St Louis, MO; resuspended in Eagle's minimal essential medium (MEM)) and incubated at 37.5C. Digestion times were typically 2.5 hr for muscles from young rats, 2.5-3 hr for muscles from young adult rats, and 5 hr (with collagenase change and brief teasing at 3 hr) for muscles from old rats. Fiber isolation from the old rats required a longer digestion because the increased extracellular matrix throughout the tissue. The collagenase-treated muscles were transferred into 3 ml of MEM containing 10% horse serum (Sigma) in a 60-mm culture dish and were further teased to separate out individual tendons with associated muscle fibers. The single tendons were gently teased to separate bundles of 30-100 myofibers and the stripped tendons were discarded. The resulting slurry of tissue was gently triturated for about 3 min with a 1-ml micropipet tip trimmed to a bore diameter of about 3 mm. The larger pieces of muscle bundles were transferred (with residual medium) to another 60-mm tissue culture plate containing 3 ml of 10% horse serum in MEM and triturated again as above. The residual larger muscle bundles were subjected to a final third round of trituration. The slurry of all three triturations was pooled and further triturated (50 X) with a trimmed, wide-mouth Pasteur pipet preflushed with horse serum. Fibers were then allowed to settle at 1 X g for 5-8 min at room temperature (RT) through 10 ml of MEM containing 10% horse serum in a 15-ml Sorvall polycarbonate tube. Fiber precipitation was repeated a total of three times to free the fibers from debris and connective tissue cells liberated by the digestion and teasing. Final fiber sediment in residual medium from the third fiber precipitation step (about 4.5 ml for material from two rats) was dispensed as 50-μl aliquots into 35-mm tissue culture plates coated with 0.12 ml of isotonic Vitrogen solution. Vitrogen 100 (about 2.9 mg/ml; Celtrix Laboratories, Palo Alto, CA) was made isotonic by the addition of 1 vol of 7 X DMEM to 6 vol of stock Vitrogen. To prevent fiber breakage during delivery to the tissue culture plates, the aliquots were dispensed using a micropipet tip trimmed to create a wide mouth. Fiber aliquots were dispensed to the center of the culture plates immediately after Vitrogen coating and the plates were gently swirled to facilitate even fiber spreading. Cultures were then preincubated for 20 min at 37.5C in humidified air containing 5% CO2 to allow formation of Vitrogen gel and adherence of fibers to the matrix. After this preincubation, cultures received 1.0 ml of basal medium (MEM containing 20% Controlled Process Serum Replacement (CPSR2; Sigma) and 1% horse serum (Sigma), and incubation continued under the same conditions as for the preincubation.
Medium (± additives when indicated) was replenished daily to allow adequate supply of the reagents. FGF2 (human recombinant, produced in yeast; kindly provided by Dr. S. Hauschka, University of Washington) was added to the medium of the fiber cultures at 2 ng/ml. Higher FGF2 concentrations in the range of 5-10 ng/ml had identical mitogenic effect as that obtained with 2 ng/ml FGF2. HGF (human recombinant; R&D Systems, Minneapolis, MN) was added at 10 ng/ml; this concentration was based on the analysis of the effect of 5 to 40 ng/ml HGF. Cytosine arabinoside (Sigma) was added to the medium at 10 μM unless otherwise noted.
Immunolabeling and Counterstaining of Nuclei
Single and double immunolabeling of fiber cultures were performed using indirect immunofluorescence as previously described (Yablonka-Reuveni and Rivera 1994). Cultures were rinsed three times with MEM at RT, fixed for 10 min at 4C with ice-cold 100% methanol, and air-dried at RT for 10-20 min. Cultures were then kept at 4C in sterile Trisbuffered saline containing normal goat serum (TBS-NGS; 0.05 M Tris, 0.15 M NaCl, 1% normal goat serum, pH 7.4) to block nonspecific antibody binding. After a minimum of overnight in TBS-NGS, cultures were rinsed with TBS containing Tween 20 (TBS-TW20; 0.05 M Tris, 0.15 M NaCl, 0.05% Tween 20, pH 7.4) and reacted with the primary antibodies. For reaction with single antibodies, cultures were incubated with the primary antibody (listed below) for 1 hr at RT followed by an overnight incubation at 4C. Cultures were then rinsed with TBS-TW20 and incubated at RT for 1-2 hr with secondary antibodies (from Organon-Technika Cappel; Downington, PA) diluted at 1:100 with TBS-NGS. A fluorescein-conjugated goat anti-mouse IgG was used for the mouse monoclonal primary antibodies. Fluorescein- or rhodamine-conjugated goat anti-rabbit IgG was used for the rabbit polyclonal primary antibodies. After exposure to the secondary antibody, cultures were rinsed again with TBS-TW20 and mounted in VECTASHIELD mounting medium (Vector Laboratories; Burlingame, CA).
In studying the co-expression of various antigens (detected with monoclonal antibodies) along with MAPK (detected with a polyclonal antibody), the two primary antibodies were added to the cultures together and subsequently the two appropriate secondary antibodies were added together under the same conditions as for single antibody staining. For analysis of the co-expression of the various antigens (detected with monoclonal antibodies) along with MyoD (detected with a polyclonal antibody), the cultures were reacted first with the monoclonal antibody followed by a reaction with the secondary antibody, as discussed above. Cultures were then reacted overnight with both the specific monoclonal antibody and the polyclonal antibody against MyoD, followed by a reaction with both the fluorescein-conjugated goat anti-mouse IgG and the rhodamine-conjugated goat anti-rabbit IgG. The two cycles of reaction with the monoclonal antibody amplified the fluorescein signal, reducing eye fatigue when doubly stained nuclei were monitored in fiber cultures.
In many experiments, nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI; 1 μg/ml in PBS) to allow detection of all nuclei in the myofiber cultures. After removal of the secondary antibodies, the cultures were stained with DAPI for 15 min at RT. Cultures were then rinsed three times with TBS-TW20 and mounted in VECTASHIELD as above. DAPI-stained nuclei were observed with Hoechst filters.
Primary Antibodies
The following primary antibodies diluted in TBS-NGS were used to study fiber cultures.
Anti-PCNA (MAb 19F4). A mouse MAb against proliferating cell nuclear antigen (PCNA) was from Boehringer Mannheim (Indianapolis, IN). This antibody has been used by us in earlier studies to monitor proliferating rat satellite cells (Yablonka-Reuveni and Rivera 1994). We demonstrated similar kinetics of cell proliferation on isolated fibers in quantifying the number of PCNA+ nuclei via immunofluorescence or via autoradiography after [3H]-thymidine incorporation (Yablonka-Reuveni and Rivera 1997b).
Anti-MyoD (MAb 5.8A). A mouse MAb against murine MyoD (IgG fraction) was developed and kindly provided by Drs. P. Houghton and P. Dias (St. Jude Children's Research Hospital, Memphis) (Dias et al. 1992).
Anti-MyoD (Polyclonal Ab). A rabbit polyclonal Ab against rodent MyoD was prepared and provided by Dr. S. Alemá (Inst. of Cell Biology, CNR, Rome, Italy). We described in previous studies additional characterizations of this antibody (Yablonka-Reuveni and Rivera 1994; Anderson et al. 1998). Double immunofluorescence of C2 cells using the monoclonal and polyclonal antibodies against MyoD showed that the two antibodies have the same staining pattern (Yablonka-Reuveni and Rivera 1997a). Both the monoclonal and the polyclonal antibodies against MyoD did not react with myogenic cultures from MyoD(-/-) mice; these cultured myoblasts expressed other myogenic regulatory factor proteins, further indicating that both anti-MyoD antibodies are specific for MyoD (unpublished observations).
Anti-myogenin (MAb F5D). A mouse MAb against rodent myogenin was used in hybridoma supernatant form. The F5D hybridoma was developed and kindly provided by Dr. W. Wright (University of Texas). The utilization of this antimyogenin MAb to stain rat satellite cells on isolated fibers and mouse-derived myogenic cultures has been described in our previous publications (Yablonka-Reuveni and Rivera 1994, 1997a, b).
Anti-developmental Myosin (MAb F1584C10). A mouse MAb against developmental but not adult isoforms of sarcomeric myosin heavy chain (DEVmyosin) was developed and kindly provided by Drs. J.J. Leger and F. Pons (Faculty of Pharmacy, INSERM, Montpellier, France). This antibody recognizes developmental isoforms of myosin heavy chain in various species, including rodents. The antibody was originally prepared against fetal bovine myosin and was shown to recognize myosin heavy chain in embryonic and fetal but not adult human muscle (Marini et al. 1991). Using this antibody, we demonstrated that about 50% of the fibers in FDB muscle of 3-week-old rats are positive, but fibers positive for this antibody are rare in FDB muscle from 2- to 3-month-old rats.
Anti-MAPK (Polyclonal Ab). The antibody against mitogen activated protein kinase was made in a rabbit immunized with a peptide that represents residues 307-327 of the ERK gene product. The peptide sequence is conserved in both ERK1 and ERK2 and the antibody recognizes both ERKs (phosphorylated and nonphosphorylated forms) at the same sensitivity. The antibody was first produced by Drs. R. Seger and E. Krebs and is characterized in Gause et al. (1993) and in Seger et al. (1994).
Counting Positive Cells on Isolated Fibers
Fibers were monitored for the number of fiber-associated nuclei and/or cells positive for the different antibodies. In initial studies, the immunocytochemical observations were made with a Zeiss epifluorescence microscope using a X25 or X40 objective (depending on the intensity of the immunoreagent). More recent experiments were analyzed with a Nikon Optiphot 2 fluorescence microscope using a X20 or X40 objective. Two to three parallel 35-mm plates were used for each time point of an individual experiment. Positive cells were scored as the number of positives on each individual fiber, analyzing 30 fibers per plate. The total number of positive cells for 30 fibers was then averaged for the duplicate or triplicate plates. This value is eventually expressed per 10 fibers as shown in the figures. For some experiments (see Table 1), the data are also shown as the total number of positive cells per the 60 fibers analyzed. Although the data discussed under Results represent individual experiments, each experiment was repeated several times.
Immunoblot Analysis of Fiber Cultures with Anti-MAPK
Fiber cultures (two 35-mm plates per time point) were rinsed three times with MEM and received 100 μl of 2 X SDS sample buffer (4% SDS, 20% glycerol, 10% β-mercaptoethanol, 0.016% bromophenol blue, 50 mM Tris, pH 6.8). Cultures were then collected with a cell scraper, which led to the extraction of both the fibers and the Vitrogen into the SDS sample buffer. The culture extracts were then frozen at −20C until analysis by SDS-polyacrylamide gels. Samples were heated at 95-100C for 5 min just before the analysis. SDS gels, prepared with 12% polyacrylamide and 0.12% bisacrylamide, were run at 185 V until the dye front reached the bottom edge of the gel. Phosphorylation of MAPKs on the regulatory tyrosine and threonine residues changes the protein mobility in SDS-PAGE (Posada and Cooper 1992), and the lower concentration of bisacrylamide used in the present study allows the separation between the phosphorylated and nonphosphorylated forms of ERK1 and ERK2 (Seger et al. 1994). Gels were blotted for 1 hr at 100 V onto nitrocellulose paper. The blot was washed with the TBS-TW20 buffer (described above), blocked with 5% milk in TBS-TW20 (1 hr at RT followed by overnight blocking at 4C), and washed again with TBS-TW20. The blocked blot was then reacted for 2 hr with the rabbit polyclonal antibody against MAPK described above (diluted 1:10,000 in TBS-TW20). The MAPK immunosignal was finally detected using an alkaline phosphatase-conjugated secondary antibody (diluted 1:6000; 1-hr exposure) and an alkaline phosphatase developing buffer (Vector Laboratories).
Distribution of positive nuclei or cells detected by immunofluorescence with antibodies against PCNA, MyoD, myogenin, and MAPK on fibers from 3-week-old rats cultured in the presence of FGF2
Fibers were isolated from three 3-week old rats. All results shown are based on analysis of parallel plates within the same experiment. Results shown in Figure 6 are also derived from the same experiment summarized in the Table. Data of PCNA+ cells in Figure 6 and in the table are based on separate sets of plates, where as data of MAPK+ in the figure and the table are derived from the same analysis.
All cell counts reported are per 60 fibers. Numbers shown for the anti-MyoD/anti-PCNA and the anti-MyoD/anti-myogenin studies are derived from analyses of two plates monitoring 30 fibers per plate. Numbers for the anit-MAPK immunostaining are derived from analysis of three plates, monitoring 30 fibers per plate and averaging the total value for 60 fibers.
FGF2 was added at 2 ng/ml and medium (+FGF2) was replaced daily.
The value for “All MyoD+ cells” in this line is the average of the values obtained with anti-MyoD/anti-PCNA and anti-MyoD/anti-myogenin.
eBecause the temporal appearance of MyoD−/PCNA+ and MyoD−/myogenin+ cells differs markedly, it is assumed that these two phenotypes contribute independently to the total cells.
Results
Reactivity of Satellite Cells on Isolated Fibers from Young Adult Rats with the Anti-MAPK Antibody
Figure 1 shows fiber cultures from the FDB muscles of 8-week-old rats stained via double immunofluorescence with the polyclonal antibody against MAPK along with the monoclonal antibodies against PCNA (Figures 1A and 1A′), MyoD (Figures 1B and 1B′), or myogenin Figures 1C and 1C′). Cultures were also reacted with DAPI to trace the multiple nuclei within the myofiber and the nuclei of the fiber-associated satellite cells (Figures 1A“-1C”). In agreement with our earlier studies, the antibodies against PCNA, MyoD, or myogenin recognized a small number of nuclei which, by the DAPI counterstain, were indistinguishable from the rest of the myofiber nuclei. In contrast, MAPK immunostaining distinguished the cytoplasm of individual cells associated with the fibers and demonstrated that the nuclei stained with the antibodies against PCNA, MyoD, or myogenin are always co-localized to these MAPK-positive cells. This localization proves that the fiber-associated nuclei stained with PCNA, MyoD, or myogenin are within satellite cells (and are not myofiber nuclei). The MAPK immunostaining shown in Figure 1 represent results with fibers that were cultured for 2 or 3 days; these are the days on which the maximal number of PCNA+/MyoD+ or myogenin+ cells peaks (see Figure 2). Satellite cells in Time 0 cultures can also be recognized with the antibody against MAPK. However, at this early stage the satellite cell cytoplasm occupies just the periphery of the nucleus and shows a far lower staining intensity than at the later stages (not shown). For all time points, we cannot rule out the possibility that the cytoplasm of the myofiber itself is also positive for MAPK. Nevertheless, the staining of the satellite cells with the anti-MAPK is far more intense than that of the fiber cytoplasm, providing a means for tracing the cells.
Figure 2 demonstrates further quantification of the fiber-associated MAPK+ cells. Data in all three panels are based on experiments with fiber cultures prepared from 8-week-old rats and each panel represent an independent experiment. Similar results were obtained with cultures of fibers isolated from 6-12-week-old rats.
Figure 2A summarizes an experiment in which parallel plates were reacted via double immunofluorescence with the antibody against MAPK in combination with the anti-PCNA, the anti-myogenin, or the anti-DEVmyosin. The immunostaining with the different antibodies (excluding anti-DEVmyosin) is as in Figure 1. Micrographs depicting immunostaining of satellite cells with the antibody against DEVmyosin are included in Yablonka-Reuveni and Rivera (1994). Regardless of the antibody pair used, the total number of MAPK+ cells was similar (shown in Figure 2A as the average of the three determinations with the three antibody pairs). However, for each antibody pair, some of the MAPK+ cells were found to be negative for the second antigen tested. These cells probably represent cells that have not yet initiated or have already terminated the expression of the second marker examined. At the early time point examined (Day 1), the number of MAPK+ cells is higher than the number of PCNA+ cells. Because myogenin+ cells were not present at this Day 1 time point, the data suggest that the satellite cells can be recognized with the anti-MAPK antibody earlier than they can be recognized with the anti-PCNA (or anti-MyoD) antibody.
Micrographs of fiber cultures isolated from 8-week-old rats. Cultures were maintained in basal medium and reacted via double immunofluorescence with the following antibody combinations: the polyclonal antibody against MAPK and the monoclonal antibody against PCNA (
The decline in the number of MAPK+ cells at late time points probably reflects a reduction in the level of MAPK expressed by the cells as they differentiate. This is supported by the appearance of the DEVmyosin+ cells, which are negative for MAPK. The possibility that the decline in MAPK+ cells is due to fusion with the myofiber is not favored in view of the studies of Bischoff (1986, 1990a), which concluded that progeny of the satellite cells on isolated fibers do not fuse with their associated fibers as long as the fibers are intact.
A second independent experiment, shown in Figure 2B, summarizes a comparison of the accumulation of MyoD+/MAPK+ cells vs PCNA+/MAPK+ cells when cultures are maintained in the absence or presence of FGF2. Regardless of the presence of FGF2, all PCNA+ nuclei and all MyoD+ nuclei were within cell cytoplasm positive for MAPK. Although the overall kinetics of PCNA+/MAPK+ and MyoD+/MAPK+ shown in Figure 2B are similar, there are slight variations in the actual values of PCNA+/MAPK+ or MyoD+/PCNA+ cells, as noted by the error bars at the different time points. However, similar values were revealed for PCNA+ or MyoD+ cells when fiber cultures were doubly stained with the anti-PCNA and anti-MyoD antibodies. Almost all PCNA+ cells were also positive for MyoD and vice versa (Yablonka-Reuveni and Rivera 1994). Analysis of the total number of MAPK+ cells is not included for the experiment shown in Figure 2B, but the results are similar to those shown below in Figure 2C.
Temporal appearance of cells positive for PCNA, MyoD, myogenin, DEVmyosin, and MAPK in cultured fibers isolated from young adult (8-week-old) rats. Each panel represents a separate independent experiment. Basal medium (+FGF2 at 2 ng/ml when the growth factor was present) was added after the initial 20 min of fiber adherence to the Vitrogen and was changed every 24 hr. Plates were collected every 24 hr and were reacted via double immunofluorescence with the antibody against MAPK combined with the various monoclonal antibodies as indicated for each panel. Cells (or nuclei) were scored as the number of positives on each individual fiber, analyzing 30 fibers per plate. Total positives were then averaged for duplicate or triplicate plates and expressed per 10 fibers, as shown in the y-axis. (
Figure 2C summarizes the results from a third experiment in which we analyzed, via double antibody labeling, the effect of FGF2 on the total number of MAPK+ cells along with the effect on the PCNA+/MAPK+ cells. The results show that, at the time of culture establishment, the number of PCNA+/MAPK+ cells is markedly lower than the number of total MAPK+ cells. The increase in PCNA+/MAPK+ cells promoted by FGF2 correlates with an increase in the total number of MAPK+ cells. Regardless of the presence of FGF2, the increases in the number of PCNA+/MAPK+ cells or in total MAPK+ cells occur mainly after the first day in culture and peak by culture Day 2. As in Figure 2A, the total number of MAPK+ cells in Day 2 cultures is greater than the number of PCNA+/MAPK+ cells. This additional number of MAPK+ cells is primarily due to the emergence of myogenin+ cells (which are also MAPK+) by Day 2 (data not shown). Overall, the total number of MAPK+ cells is increased by 2.8-fold in cultures receiving FGF2 and only by 1.8-fold in control cultures. A similar relationship in the increase of the number of MAPK+ cells in the presence or absence of FGF2 is further demonstrated in Figure 3. The results suggest that the addition of FGF2 to the cultures allows more (or perhaps all) satellite cells present on the fibers to undergo cell division. Alternatively, FGF2 may accelerate the cell cycle time of the satellite cells which, without the addition of FGF2, proliferate at a far slower rate.
In the experiments shown in Figures 2C and 2A, there is a decline in the number of MAPK+ cells at later time points. This decline is more rapid in the experiment shown in Figure 2C than in the experiment shown in Figure 2A. The decline in the number of MAPK+ cells for Day 3 cultures (Figure 2C) and the plateau of MAPK+ cells for Days 2 and 3 (Figure 2A) have both been reproduced in many experiments. These slight variations in the timing of the decline in MAPK+ cells reflect how synchronous the cells are as they transit into the myogenin+ compartment.
Effect of the Mitotic Drug Cytosine Arabinoside on MAPK+ Cells
Culturing rat fibers with cytosine arabinoside led to the elimination of the fiber-associated satellite cells, as estimated by regular microscopic evaluation (Hinter-berger and Barald 1990). We adapted this drug treatment approach to determine whether all the MAPK+ cells are similarly sensitive to cytosine arabinoside, reasoning that if a fraction of the MAPK+ cells survive the drug treatment, these survivors might represent nonproliferating satellite cells. As suggested in the previous section, such nonproliferating cells might become proliferative when FGF2 is added. Data in Figure 3 show that the presence of cytosine arabinoside at 10 μM led to the elimination of MAPK+ cells or PCNA+ cells regardless of the absence (Figure 3A) or presence (Figure 3B) of FGF2. The sensitivity to the drug is already demonstrated by the first 24 hr in culture. Removal of the drug after 3 days in culture did not result in the reappearance of fiber-associated satellite cells, even after more than 2 weeks in culture. Although the number of fiber-associated satellite cells was reduced to a minimum, the fibers themselves remained intact and maintained their cross-striations. We further examined the effect of lower concentrations of cytosine arabinoside, reasoning that such lower concentrations may preferentially kill only the more rapidly dividing satellite cells, whereas MAPK+ cells, which proliferate only slowly (if at all), may survive. In the presence of 5 μM cytosine arabinoside, only 1-5% of cells survived to the second day of culture, regardless of whether cells were monitored by the anti-MAPK or anti-PCNA antibodies. Continuous maintenance of fiber cultures in 0.5 μM cytosine arabinoside allowed survival of 35-50% of the cells after the second or third culture, irrespective of the antibody used for monitoring the cells. Taken together, the results indicate that all the fiber-associated MAPK+ cells are similarly sensitive to the drug.
Quantification of PCNA+ nuclei and MAPK+ cells in cultured myofibers isolated from 10-week-old rats and maintained in the presence or absence of the mitotic drug cytosine arabinoside (cyt ara) without (
Immunoblot Analysis of the MAPK Isoforms Expressed by Satellite Cells
Immunoblot analyses of routine cell cultures with the anti-MAPK antibody used in the present study showed that the antibody recognizes ERK1 and ERK2 (44- and 42-kD molecular weight, respectively) in their inactive and active forms (Seger et al. 1994). A Western blot analysis of extracts of mouse C2 myoblasts and of cultured primary rat satellite cells confirmed that the anti-MAPK antibody recognizes the unphosphorylated (ERK1/ERK2) and phosphorylated (ERK1p/ERK2p) forms of the ERKs in these myogenic cultures as well (Figure 4). To further analyze the ERKs contributed by satellite cells in myofiber cultures, we compared extracts from control and cytosine arabinoside-treated fiber cultures. We reasoned that because the cytosine arabinoside eliminates the satellite cells, any reduction in specific ERK bands on the immunoblot of drug-treated cultures is likely to represent ERKs contributed by satellite cells. Fiber cultures that received cytosine arabinoside (±FGF2) on culture establishment and control cultures (±FGF2) were collected at daily intervals for immunoblot analysis with the anti-MAPK antibody. Results of this study are depicted in Figure 4. Figures 4A and 4B demonstrate that the nonphosphorylated forms of ERK1 and ERK2 are both prevalent in the fiber cultures between Days 1 and 5, regardless of the absence or presence of FGF2. The effect of cytosine arabinoside on the reduction in the ERK1 and ERK2 polypeptides become apparent by culture Day 2 (regardless of the absence or presence of FGF2), suggesting that both ERK1 and ERK2 contributed to the immunosignal demonstrated by satellite cells when cultured fibers were analyzed via immunofluorescence. It is noteworthy that the presence of the drug appears to lead to a more marked decline in the intensity of the ERK1 band compared to the ERK2 band.
Immunoblotting analysis of fiber cultures isolated from young adult (10-week-old) rats and reacted with the antibody against MAPK. For each time point, two culture plates were collected together with the Vitrogen substratum using 100 μl of 2 X SDS sample buffer. Equal volume aliquots of extracts from the different time points were separated by SDS-PAGE. Gels were prepared with 12% polyacrylamide and 0.12% bisacrylamide, which allowed separation of the phosphorylated and nonphosphorylated forms of ERK1 and ERK2. Reactivity with the anti-MAPK was visualized with an alkaline phosphatase-conjugated secondary antibody. Extracts of proliferating C2 cells (C2 in
In addition to the specific ERK polypeptides, fiber extracts show one more intense band when analyzed by immunoblotting (Figures 4A-4C). The migration position of this extra band, just above the 47-kD molecular weight marker, is marked with an arrowhead at the left of each panel. This is a nonspecific band contributed by the horse serum included in the fiber isolation medium and in the basal medium used for fiber cultures. This 47-kD protein cannot be washed away from the Vitrogen even after excessive rinses and is also present in extracts of sham fiber cultures that are devoid of fibers (data not shown). As shown in Figure 4D, extracts of fibers isolated and cultured without horse serum did not exhibit this band.
The immunoblot analysis in Figures 4A and 4B has shown a marked decline in the intensity of the ERK2p band between the initial 30 min in culture (i.e., 30 min in basal medium after the initial 20 min of fiber adherence to the Vitrogen) and the Day 1 time point. Figure 4C shows a more detailed analysis of the early time points after culture establishment. At time point 0 the prevalent forms of ERK1 and ERK2 are the shifted ones (ERK1p and ERK2p). ERK1p is drastically reduced by 30 min in culture, whereas ERK2p is drastically reduced in intensity between the 60-min and the 24-hr time points. These transitions are not influenced by the presence of FGF2. We further investigated the possibility that the horse serum included in the fiber preparation protocol (after the collagenase digestion) might provide the signals leading to the active ERK1 and ERK2 detected at the early time points. The horse serum, present routinely at 10% in the fiber isolation medium, was replaced with serum replacement (CPSR2, 10%) and the horse serum present in the culture medium was omitted. Figure 4D summarizes the immunoblot analysis of this horse serum-free protocol and shows that, in the absence of horse serum, ERK2p was present at high levels throughout the early time points tested, regardless of the presence or absence of FGF2. ERK2p declined to a lower level by 24 hr in culture. The pattern shown for the 5-, 10-, and 20-min time points was maintained for the 60- and 120-min time points as well (data not shown). Time points between 120 min and Day 1 were not analyzed. ERK1p was not detected in any of the extracts of horse serum-free fiber cultures. At present we do not have means for distinguishing whether the “collapses” in ERK1p and ERK2p seen by the first day in culture are specific to the satellite cells. Obviously, these transitions may reflect changes occurring in the myofiber itself on culturing.
Kinetics of Satellite Cell Myogenesis in Cultures of Fibers from Young Rats
Double immunostaining of fiber cultures from 3-week-old rats demonstrated that all nuclei positive for PCNA, MyoD, or myogenin are contained within MAPK+ cytoplasm of single cells (not shown; see Figure 1 for a similar staining). This indicates that, as in cultures from young adult rats, all positive nuclei in fibers from 3-week-old rats are within satellite cells and do not represent myofiber nuclei.
Figure 5 shows the results of two independent studies in which satellite cells in fibers from 3-week-old rats were monitored by their immunoreactivity with the antibodies against PCNA, MyoD, myogenin, and MAPK. Data in Figure 5 are based on analyses of fibers that were cultured in the presence of FGF2. Fibers cultured in the absence of FGF2 demonstrated only a small number of PCNA+ nuclei (Figure 6) and a similar small number of MyoD+, and subsequently myogenin+, nuclei (data not shown). In the experiment shown in Figure 5A, cultures were reacted with individual antibodies againt MAPK, PCNA, and MyoD. In the experiment shown in Figure 5B, cultures were reacted with the antibody against MAPK alone or with the polyclonal antibody against MyoD in combination with the anti-myogenin. The studies in Figure 5 demonstrate that satellite cells from 3-week-old rats follow a similar program as that seen with satellite cells from the young adult rats, first becoming positive for MyoD and PCNA and subsequently entering the myogenin+ state. Although the overall program of transition through proliferation and differentiation for satellite cells from 3-week-old rats is similar to that for satellite cells from the 8-10-week-old rats, the increase in the number of proliferating cells and the transition into the myogenin+ state is more rapid for satellite cells from the younger animals. Despite the presence of FGF2, the maximal number of proliferating satellite cells is lower for fibers from the 3-week-old rats compared to fibers from the young adult animals. Analysis of MAPK+ cells at the time of culture establishment is not shown because of the difficulty in differentiating between the cells and the myofiber nuclei in working with Time 0 fibers from the young rats. After about 12 hr in culture, satellite cell analysis by the antibody against MAPK becomes feasible.
Temporal appearance of cells or nuclei positive for MAPK, PCNA, MyoD, and myogenin in cultured fibers isolated from 3-week-old rats. The two panels represent separate experiments and for each experiment fibers were isolated in parallel from three rats and pooled together. Cultures were maintained in basal medium containing FGF2 (2 ng/ml), which was added at Time 0 and changed every 24 hr. (
Quantification of PCNA+ nuclei and MAPK+ cells in cultured myofibers isolated from 3-week-old rats and maintained in basal medium without (
Both panels in Figure 5 show a surplus of MAPK+ cells by Day 1, when the total number of MAPK+ cells is compared to the number of PCNA+ or MyoD+ cells. This surplus (as in the Day 1 fiber cultures from young adult rats) may indicate that the satellite cells become detectable by the anti-MAPK antibody before their de-tectability with the other antibodies. However, it is possible that some of the surplus cells represent additional satellite-like cells that do not enter the MyoD expression program. We further analyzed this issue in the study summarized in Table 1. Parallel plates were monitored by double antibody staining for the presence of MyoD±/PCNA± nuclei and MyoD±/myogenin± nuclei, or by single antibody staining for MAPK+ cells. At each time point, the total number of positive nuclei consisted of all MyoD+ cells (determined by averaging the number of MyoD+ cells from the aforementioned pairs), combined with the number of PCNA+ cells negative for MyoD, and the number of myogenin+ cells negative for MyoD. Indeed, the data in Table 1 show a surplus of MAPK+ cells in both Day 1 and Day 2 cultures (which amounts to 27 and 42 cells, respectively). This surplus may represent a subpopulation of cells that are incapable of entering the PCNA/MyoD/myogenin program under the culture conditions used (and may even be nonmyogenic). The discrepancy shown in Table 1 between the number of MAPK+ cells and the number of total positive nuclei in Days 3 and 4 is probably primarily due to the fact that most of the satellite cells are no longer traceable at late time points via the nuclear antigens studied.
As mentioned above, fiber cultures from 3-week-old rats exhibited only a small number of PCNA+ nuclei (and a similarly small number of MyoD+ or myogenin+ nuclei) without the addition of FGF2 to the basal medium. Figure 6 summarizes typical results of such analyses on the effect of FGF2 in fiber cultures from 3-week-old rats, comparing the number of PCNA+ nuclei and MAPK+ cells. Figure 6 shows that the addition of FGF2 led to an increase in the number of PCNA+ nuclei and MAPK+ cells by Day 1 in culture (no effect was observed by culture Day 0.5; data not shown). On Day 2, the peak day for PCNA+ or MAPK+ nuclei, there is a discrepancy between the number of MAPK+ cells and PCNA+ nuclei. In the FGF2-treated cultures, a large portion of this discrepancy can be explained by the presence of myogenin+ cells (which are also MAPK+), but additional surplus MAPK+ cells are likely to be present too, as shown in Table 1. However, in the Day 2 control cultures, the number of myogenin+ cells is small and cannot account for most of the surplus MAPK+ cells. We therefore suggest that there is a larger number of surplus cells in control cultures and that the addition of FGF2 allows more of the surplus MAPK+ cells to enter the proliferative pathway (which is subsequently followed by myogenic differentiation).
We also analyzed the influence of cytosine arabinoside on the kinetics of PCNA+ nuclei and MAPK+ cells in fibers from the 3-week-old rats. Regardless of the absence or presence of FGF2, almost all of the PCNA+ nuclei were eliminated by Day 1 in culture. Some residual MAPK+ cells were still present by Day 1 in culture (about 20-25% of the cells) but were almost all eliminated by Day 2 in culture. Cultures maintained for 3 or 4 days in the basal medium (±FGF2) in the presence of cytosine arabinoside were practically devoid of PCNA+ nuclei or MAPK+ cells (data not shown). These findings suggest that all MAPK+ cells (PCNA+/MAPK+ or surplus MAPK+ cells) are sensitive to the anti-mitotic drug.
Monitoring Satellite Cells on Isolated Fibers from Old Rats
With the ability to co-localize nuclear antigens within MAPK+ cytoplasm of fiber-associated cells, we set out to investigate the dynamics of satellite cells on isolated fibers from the old (9-11-month-old) rats. Immunostaining of the fibers with the antibodies against MAPK and the nuclear proteins revealed more fiber-associated MAPK+ cells compared to fibers from the younger rats. Furthermore, a high proportion of these cells did not stain with the antibodies against the nuclear proteins MyoD, PCNA, and myogenin. Figure 7 shows two examples of fibers from the old rats stained with the antibodies against MAPK and MyoD (Figures 7/A′ and 7B/B′), demonstrating a larger number of cells together in single sites on the fibers compared to what we observed with the younger rats (Figure 1). The quantification of fiber-associated cells in cultures from the old rats is summarized in Figure 8 for isolated fibers maintained with or without FGF2. Figures 8A and 8B show the results of one experiment in which parallel plates were reacted via double immunofluorescence with the antibodies against PCNA and MAPK, or MyoD and MAPK. Figure 8C shows the results of a similar study with the anti-myogenin/anti-MAPK antibody combination. The analyses demonstrate that all PCNA+, MyoD+, and myogenin+ nuclei associated with fibers from old rats are always contained within MAPK+ cytoplasm, indicating that our method of monitoring satellite cell myogenesis is valid for the old rats as well. The kinetics of the PCNA+ cells are similar to those of the MyoD+ cells in either the presence or absence of FGF2, indicating that both PCNA and MyoD immunostaining mark the satellite cells (as shown above for fiber cultures from younger rats). Without the addition of FGF2, only a basal level of cells undergoes myogenesis, whereas in the presence of FGF2 a larger number of cells are detected. In addition, fibers from the old rats demonstrate a higher number of surplus MAPK+ cells than those seen in cultures from the younger ages. It is possible that some of the surplus MAPK+ cells reside outside of the myofiber basement membrane (especially in view of the elaborated connective tissue network between the myofibers in the older rats). However, similar to the MyoD+ and PCNA+ cells, these surplus cells also respond to FGF2.
Micrographs of fiber cultures isolated from 11-month-old rats. Left panels (
We further analyzed the effect of cytosine arabino-side in FGF2-treated fiber cultures from the old rats. Cytosine arabinoside in the range of 10-50 μM did not influence the total number of MAPK+ cells or the number of MyoD+ cells by the first day in culture. However, by the second culture day about 84% of the MyoD+ cells (and 40% of the total MAPK+ cells) were eliminated compared to FGF2-treated control cultures. By the third culture day, about 96% of the MyoD+ cells (and 81% of the total MAPK+ cells) were eliminated compared to the FGF2-treated controls. Therefore, some of the MAPK+/MyoD− cells are insensitive to cytosine arabinoside even by culture Day 3, whereas in the younger rats all MAPK+ cells were sensitive to the drug by this day. In addition, the effect of the drug on the elimination of MyoD+ satellite cells requires one additional day in culture from old rats compared to the younger rats.
Comparison of the Effects of FGF2 and HGF on Proliferation of Satellite Cells from Old Rats
The data in Figure 8 indicate that the satellite cells from the old animals spend a longer time in the proliferative compartment than satellite cells from younger animals. Furthermore, the delayed effect of cytosine arabinoside on the MyoD+ cells from the old rats suggests that the onset of satellite cell proliferation in fiber cultures from old rats lags behind the onset of proliferation in cultures from younger animals. Analyzing myogenic cultures from 9-month-old rats, Allen et al. (1995) suggested that HGF can induce satellite cell proliferation more rapidly than FGF2. Therefore, we tested whether the addition of HGF to cultured fibers isolated from old rats could potentially accelerate the progression of satellite cells through the proliferative compartment. Our findings, summarized in Figure 9, indicate that HGF and FGF2 support proliferation of satellite cells from old rats in a similar fashion during the first 3 days in culture. During later days in culture, the decline in PCNA+/MAPK+ cells (or total MAPK+ cells) is more rapid in cultures receiving HGF than in cultures receiving FGF2. Analysis of parallel cultures with antibodies against MyoD and MAPK has revealed a quantity of satellite cells similar to that seen with the anti-PCNA/anti-MAPK antibodies (data not shown). The addition of HGF and FGF2 together did not lead to an increase in the number of PCNA+ cells beyond that seen with each growth factor added individually. In the study discussed above, HGF was added to the cultures at 10 ng/ml. Lower and higher HGF concentrations (5 and 20 ng/ml, respectively) yielded a slightly lower number of PCNA+ cells compared to that seen with 10 ng/ml HGF (data not shown).
Temporal appearance of cells positive for PCNA, MyoD, myogenin, and MAPK on myofibers from 10-month-old rats. In all cases, positive cells were monitored via double antibody staining using the antibody against MAPK along with the monoclonal antibodies against PCNA (
Quantification via double immunofluorescence of PCNA+/MAPK+ (
Discussion
This study was undertaken to evaluate the proliferative potential of satellite cells from growing and old rats. Satellite cells are believed to be proliferative in growing rats, whereas in older animals satellite cells are believed to be quiescent. Therefore, it is possible that proliferation of satellite cells from younger and older animals are regulated differently by growth factors. We employed cultures of single myofibers, in which the original position of the satellite cells by the myofiber is preserved. This association between the satellite cells and the myofibers is potentially important because the myofibers might contribute some of the growth factors involved in controlling proliferation and differentiation of satellite cells. Furthermore, this association between the satellite cells and the myofibers provides a means for localizing the myogenic precursors even before they express myogen-specific traits.
We first demonstrated that satellite cells undergoing myogenesis in fiber cultures can be traced immunohistochemically by their cytoplasmic expression of ERK1 and ERK2, members of the MAPK superfamily involved in the transmission of extracellular signals to their intracellular targets. The antibody we used reacts with both the phosphorylated and nonphosphorylated forms of the two ERKs. Although phosphorylated, active ERKs are believed to be able to translocate to the nucleus, we were unable to detect ERK1/2-positive nuclei throughout the study. The immunohistochemi-cal staining with the anti-MAPK antibody was further employed for analyzing the proliferative dynamics of satellite cells from young (3-week-old), young adult (8-10-week-old), and old (9-11-month-old) animals. We demonstrated that satellite cells from all three age groups undergo a similar program, first becoming positive for PCNA and MyoD and later for myogenin. In addition, irrespective of the age of the muscle studied, FGF2 promotes the number of proliferating, MyoD+ satellite cells and does not suppress the transition of the cells into the myogenin+ state. Collectively, the present studies have led us to propose that skeletal muscle from both young and old animals contains satellite cells whose recruitment into active myogenesis, characterized by rapid proliferation and differentiation, is regulated by FGF2.
Analysis of Myofibers from Young Adult Rats Suggests that FGF2 Promotes an Increase in the Number of Satellite Cells Entering Replication
Monitoring satellite cells on fibers from 8-10-week-old rats by their expression of MAPK has provided an overall measure of the number of fiber-associated cells. As the number of PCNA+ cells rises to a maximal level, the number of MAPK+ cells rises as well. Interestingly, the co-localization of PCNA and MAPK at the earlier days in culture (Time 0 and Day 1) demonstrated a far higher number of MAPK+ cells compared to PCNA+ cells. This indicates that satellite cells can be monitored by their cytoplasmic MAPK before cell replication. This number of initial MAPK+ cells increases minimally, if at all, between Days 0 and 1, regardless of the presence or absence of FGF2. Between culture Days 1 and 2 there is an increase in the total number of MAPK+ cells, representing an increase in the total number of satellite cells associated with the fibers. The majority of these MAPK+ cells are positive for PCNA but a smaller fraction of the cells have already transited to the myogenin+ state. Because this increase in MAPK+ cells between Days 1 and 2 is far more robust when FGF2 is added to the cultures, we considered the possibility that FGF2 promotes an increase in the number of satellite cells recruited from quiescence to proliferation. If this indeed is how FGF2 influences the satellite cells, then the proliferation of some satellite cells without the addition of FGF2 may indicate that a subpopulation of the satellite cells requires a lower mitogen level (FGF2 and/or other mitogens) to enter replication. This low-level mitogen(s) is perhaps contributed by the myofibers themselves, either naturally or because of the stress induced by the isolation/culturing procedure. However, at this stage we can not exclude an alternative recruitment mechanism in which the exogenously added FGF2 accelerates satellite cell proliferation, generating altogether more satellite cells. In this instance, satellite cells may proliferate more slowly without the addition of FGF2, generating a smaller number of satellite cells. In fact, the finding that the mitotic drug cystosine arabinoside rapidly eliminated all fiber-associated satellite cells provides support in favor of the second recruitment mechanism, i.e., the satellite cells can proliferate without the additional FGF2 but the addition of FGF2 facilitates more rapid proliferation of the cells.
The FGF-promoted increase in satellite cells was observed in the present fiber study by the second day in culture. It is noteworthy that, in another study of fibers from young adult rats, we showed that the effect of FGF2 on increasing the number of PCNA+ cells is already taking place by 36 hr in culture. At this 36-hr time point, the number of PCNA+ cells was about 75% of the maximal number seen at 48 hr in culture (Yablonka-Reuveni and Rivera 1997b).
The Transition of MAPK+ Cells to the Proliferative State in Fiber Cultures from 3-week-old Rats Is Highly Dependent on Exogenous FGF2
The quantification of fiber-associated cells in isolated fibers from 3-week-old rats revealed that, once the cells have been recruited into proliferation, myogenesis progresses in a similar fashion to that described for myofibers from young adult rats. In addition, as in the young adult rats, the FGF2-promoted increase in the number of PCNA+ cells in fibers from 3-week-old rats leads to a consequent increase in the number of cells entering the myogenin+ state. There are, however, distinct differences between the two age groups. First, the peak number of proliferating/differentiating satellite cells is lower in control or FGF-treated cultures from 3-week-old rats compared to cultures from young adults. Second, without the addition of FGF2, only a small number of MAPK+ cells are positive for PCNA in cultures from the 3-week rats (see Figure 5A, Day 2). On adding FGF2, the number of total MAPK+ cells increases along with a substantial increase in the number of PCNA+/MAPK+ cells and a decrease in the proportion of the surplus MAPK+ cells (negative for PCNA). The finding that satellite cells from 3-week-old rats demonstrate higher dependency on exogenous FGF2 for proliferation than satellite cells from 8-10-week-old rats is somewhat enigmatic, given the common convention that satellite cells from younger rats are proliferative, adding nuclei to the enlarging myofibers. The satellite cells in the 3-week-old fiber culture might be highly dependent on exogenous FGF2 because the culture system may lack endogenous FGF2 (and/or other FGFs). In contrast, FGF2 (or other FGFs) might be stored or produced in fibers from young adult rats, allowing some proliferation of the satellite cells without the addition of exogenous FGF2. Earlier studies have shown that FGF2 is present in the extracellular matrix surrounding the muscle fibers (DiMario et al. 1989) and can be released from focally injured myofibers (Clarke et al. 1993). Transcripts for different FGFs were also detected in myoblasts and myotubes both in vivo and in culture (Garrett and Anderson 1995; Hannon et al. 1996). Modulations in such endocrine/paracrine sources of FGFs during postnatal growth could potentially contribute to different requirements for FGF by satellite cells from growing and young adult rats.
It is possible that many of the PCNA−/MAPK+ cells detected in the absence of FGF2 in fibers from 3-week-old rats are slowly dividing cells that do not accumulate sufficient PCNA for tracing via immunocytochemistry (but are sensitive to the mitotic drug cytosine arabinoside). The addition of FGF2 may shift many of these cells from the slow-dividing to the rapid-dividing phenotype. Fibers isolated from young adult rats may also contain such surplus PCNA−/MAPK+ cells, although at lower numbers than the fibers from the young rats (see, e.g., the Day 2 time point in Figure 2A, where the discrepency between the total number of MAPK+ cells compared to the PCNA+ cells cannot be solely explained by the small number of myogenin+ cells present). The possible existence of rapid-dividing and slow-dividing satellite cells in growing rats has been previously proposed by Schultz (1996). Tracing proliferation of satellite cells in vivo, Schultz reported that about 20% of the satellite cells in 1-month-old rats divided far more slowly and suggested that these slower cells are true reserve cells which may generate, through asymmetric divisions, the rapidly dividing satellite cells. Schultz further proposed that the rapidly dividing cells go through minimal replication (perhaps just one round) and are readily available for fusion into the growing myofibers. In addition, Schultz (1996) proposed that the growth rate of the muscle could determine the rate at which these reserve cells produce the rapidly dividing satellite cells. We suggest that the PCNA−/MAPK+ (surplus) cells detected in isolated fibers from the 3-week-old rats during the early culture stages are the equivalent of the reserve cells in young rats proposed by Schultz. In the fiber culture, as satellite cells proliferate and express MyoD, they become committed to transition into the myogenin+ state and differentiation. Indeed, previous studies of myogenic lines suggested that MyoD, when present in its active form, leads to withdrawal of myoblasts from the cell cycle (Halevy et al. 1995). In contrast, the slowly dividing satellite cells might be able to escape MyoD expression (and subsequent rapid differentiation), maintaining a pool of stem cells ready to provide rapidly dividing progeny as demand dictates. FGF2 might represent one of several growth factors that can influence the transition of the stem cells into the rapidly-dividing, differentiation-destined satellite cells.
Recruitment, Proliferation, and Differentiation of Satellite Cells in Fiber Cultures from Old Rats
Our studies demonstrated that FGF2 influences myogenesis of satellite cells from old animals in a similar manner to that seen for the younger animals. In the presence of FGF2, there is an increase in the number of PCNA+ or MyoD+ cells by the second day in culture, reaching maximal numbers by culture Day 3 and reduced to nearly baseline level by culture Day 5. An increase in myogenin+ cells commences 24 hr after the increase in the PCNA+ or MyoD+ cells, indicating that the myogenic precursors rapidly transit from proliferation to differentiation. In the absence of FGF2, the number of cells that transit through the program is significantly lower. We also conclude that cultured myofibers from old rats support a higher number of proliferating MyoD+ satellite cells than cultured myofibers from the rapidly growing young rats. This finding suggests that the number of myogenic precursors increases with the age of the animal. This increase in the number of satellite cells from older animals was unexpected because muscle regeneration was found to be impaired in older animals (Sadeh 1988; Carlson and Faulkner 1996). Nevertheless, our findings fit well with the proposal that the impairment of muscle regeneration in old animals is due to the environment rather than to age-related changes in the satellite cells (Carson and Alway 1996; Carlson and Faulkner 1996). Our findings that the numbers of MyoD+ fiber-associated cells increase with aging is in contrast to the study of Bockhold et al. (1998), which reported on a decline in the number of myogenic cells with aging in the mouse. The starting material in both our rat study and the later mouse study was isolated myofibers. However, whereas in our study the satellite cells remained associated with the original myofibers, the study of Bockhold et al. (1998) used culture conditions that allow the satellite cells to emigrate from the fibers and proliferate in a serum-rich medium, and the analysis focused on cells that were no longer associated with the fiber. Further investigations are required to determine the basis for the discrepancy between the two reports. Differences in the species studied, the kind of the muscle used, and the technique of culturing need to be considered.
As in the younger rats, the addition of FGF2 to fibers from old rats allowed a larger number of satellite cells to undergo myogenesis compared to control cultures. However, the peak of proliferating cells is broadened to include culture Days 2, 3, and 4 (see Figures 8 and 9). It is interesting to note that the in vivo studies of McGeachie and Grounds (1995) also recognized a time delay in the peak number of proliferating satellite cells when young and old mice were compared after massive muscle injury. This delayed peak seen with satellite cells from old animals could be explained by a difference in the time when the appropriate growth-promoting agents become available. The effect of cytosine arabinoside in eliminating satellite cells was delayed by 1 day when satellite cells from the old animals were compared to satellite cells from the younger age groups. Therefore, the broadened proliferative peak and the delayed effect of the mitotic drug suggest that the onset of satellite cell proliferation is delayed in cultures from the old rats. This possibility of delayed proliferation is in agreement with earlier studies by Johnson and Allen (1993). These investigators showed that FGF2-induced proliferation is delayed in cultures of tissue-dispersed satellite cells from 9-month-old rats compared to cultures from young rats. Allen et al. (1995) also showed that the addition of HGF accelerates the onset of proliferation in satellite cell cultures from old rats, and further proposed that satellite cells from older muscle require HGF for their initial activation. However, in the present study, the addition of HGF to fiber cultures from old rats resulted in a schedule of satellite cell proliferation during the initial 3 days in culture similar to that seen with FGF2 (Figure 9). Therefore, our studies do not support the notion that HGF is more critical than FGF2 for the initial recruitment of satellite cells. We further noted that proliferation of satellite cells in fiber cultures from the old rats declines a day earlier when HGF is added compared to FGF2 (i.e., Day 4 vs Day 5, respectively). Perhaps both FGF2 and HGF are capable of recruiting satellite cells into proliferation, but of the two factors only FGF2 is capable of extending proliferation for one additional round. Alternatively, satellite cells from the old animals could contain a subpopulation that enters proliferation later and is recruited by FGF2 but not by HGF.
We detected many surplus cells which were PCNA−/MAPK+ or MyoD−/MAPK+ in fibers from the old rats. The kinetics of the total number of MAPK+ cells closely parallel the kinetics of the PCNA+/MAPK+ cells or MyoD+/MAPK+ cells in control, FGF2-treated, and HGF-treated cultures. This correlation suggests that the surplus MAPK+ cells in the old animals could be related to the satellite cells undergoing active myogenesis. Perhaps satellite cells from the old animals give rise to such surplus cells more frequently than do satellite cells from the younger animals. Myogenic and nonmyogenic phenotypes are routinely present in myogenic clones derived from a single progenitor (see discussion in Yablonka-Reuveni and Seifert 1993; Yablonka-Reuveni and Rivera 1997a). Obviously, we cannot rule out the possibility that some of these surplus cells, although resembling satellite cells by their position, are of a different phenotype altogether. Intrestingly, isolated fibers from aging mdx mice were also reported to give rise to myogenic and nonmyogenic cell phenotypes, as judged by immunostaining with an antibody against desmin and ability to form myotubes (Bockhold et al. 1998).
How Might FGF2 Operate in Influencing Satellite Cell Proliferation in the Isolated Fiber Model?
The results of the present study indicate that FGF2 can influence the number of satellite cells undergoing myogenesis in fiber cultures from growing, young adult, and old rats. For all three age groups, the increase in the number of satellite cells is due to the effect of FGF2 on enhancing the number of proliferating satellite cells, whereas the transition of the cells into the differentiative state is not suppressed by the growth factor. The kinetics of proliferation in the absence and presence of FGF2 suggest that for all three age groups FGF2 is likely to act during the first round of cell replication after fiber isolation. Although the results indicate that FGF2 is involved in generating more satellite cells, the specific mode of action of FGF2 is not resolved by the present study. FGF2 may recruit more satellite cells by accelerating the cell cycle of cells that have already been enabled by other factors to enter the mitogenic-myogenic program, and/or by potentiating the entry of slowly dividing (possibly almost quiescent) precursor cells into a regular progression through the cell cycle and MyoD expression. Additional experiments with delayed exposures to exogenous FGF2 (done with fibers from young adult rats) demonstrated that the addition of FGF2 as late as 15-18 hr after culture establishment results in a maximal number of fiber-associated PCNA+ satellite cells, which is similar to the number of cells seen when FGF2 is added at Time 0 (Yablonka-Reuveni and Rivera 1997b). This critical time for FGF2 addition precedes the onset of the increase in PCNA+ satellite cells and might represent an important step in the transition from G0 to G1, which requires FGF. Delaying the addition of FGF2 until the second culture day no longer yielded an increase in the number of PCNA+ cells above that seen in control cultures (not receiving FGF2). This has further suggested that FGF2 may play a role during the early phase of cell cycle entry.
In agreement with our proposal regarding the role of FGF2 in supporting recruitment of satellite cells, Floss et al. (1997) have recently suggested that FGF6 is critical for muscle regeneration in vivo. Their studies demonstrated that muscle regeneration is retarded in the FGF6(-/-) mouse and that this reduced regeneration correlated with a reduction in the number of MyoD+ myoblasts at the crush injury site (Floss et al. 1997). Floss and colleagues proposed that FGF6 may play a role in stimulating or activating satellite cells. FGF2 (employed in the present study) and FGF6 might both be involved in the specific recruitment of satellite cells because, as we have recently determined, both are present in the adult muscle and both promote activation/proliferation of satellite cells in isolated fibers in a similar fashion. FGF1 is also present in the muscle and can support recruitment/proliferation of satellite cells in fiber cultures in a similar fashion to FGF2 and FGF6 (unpublished work). Therefore, it is possible that different FGFs can be involved in the recruitment of satellite cells. Indeed, most FGF receptors can be activated by multiple FGFs (Green et al. 1996; Ornitz et al. 1996). We are studying the possibility that there is a strict specificity in the FGF receptors that are involved in satellite cell recruitment (rather than a specificity in the FGFs themselves).
Studies of routine myogenic cultures have reported that FGF2 (and other FGFs) can support continuous proliferation while delaying differentiation of myoblasts (Clegg et al. 1987; Hannon et al. 1996; Pizette et al. 1996; Yoshida et al. 1996). In contrast, in the fiber cultures, satellite cell differentiation is not suppressed despite the presence of FGF2. We examined the possibility that other members of the FGF family, or other growth factors including HGF, IGF, and PDGF, may facilitate continuous proliferation of satellite cells on isolated fibers when administrated to the cultures alone or with FGF2. None of the growth factors tested increased proliferation of satellite cells beyond what has been detected in the presence of FGF2 (unpublished work). We also observed this inability of FGF2 to block differentiation in routine cultures of rat satellite cells dissociated from the fibers after fiber isolation. Our investigation of the role of FGF2 during proliferation of the mouse-derived myogenic line C2 has raised the possibility that FGF2 enhances the proliferation of myoblasts before they enter the MyoD expression phase (Yablonka-Reuveni and Rivera 1997a). This active proliferation at the “pre-MyoD” stage might be a characteristic of the various frequently used myogenic lines. Such lines were selected on the basis of their ability to proliferate throughout multiple passages. However, in the fiber cultures the rapidity of MyoD expression by the satellite cells might limit the mitogenic effect of FGF to the initial step of recruitment from quiescence. It is further possible that the mechanism seen in the present study (rapid recruitment of satellite cells into the MyoD+ state by FGF2) might be critical for muscle maintenance during ongoing muscle growth and repair, and that the alternative mechanism, in which FGF enhances proliferation and delays differentiation, might operate when the muscle is subjected to major trauma and a large supply of myoblasts is required for repair. However, the studies of Grounds et al. (1992) demonstrated that satellite cells enter the MyoD-myogenin expression state within several hours after crush injury of mouse muscles, suggesting that the satellite cells follow a similar program in fiber cultures and in vivo after muscle injury. The study of Megeny et al. (1996), which demonstrated that muscle regeneration is compromised in the MyoD(-/-) mouse, provides further support for the hypothesis that MyoD expression is important for myogenesis of satellite cells in vivo.
Experiments are under way in our laboratory to determine which FGF receptors are involved in the recruitment of satellite cells into the proliferative PCNA+/MyoD+ state seen in the fiber model and whether the same (or additional) FGF receptors are operating in culture systems in which FGF extends proliferation and suppresses differentiation.
Footnotes
Acknowledgments
Supported in part by grants to ZY-R from the Cooperative State Research Service-US Department of Agriculture (agreements nos. 93-37206-9301 and 95-37206-2356) and the National Institutes of Health (AR39677 and AG13798).
We thank Priscilla Natanson, Stephanie Kästner, and Maria Elias for their important contributions to the study. We also thank many colleagues for providing us with valuable reagents: Dr S. Hauschka (FGF2), Dr S. Alemá (anti-MyoD, polyclonal), Drs P. Houghton and P. Dias (anti-MyoD, monoclonal), Dr W. Wright (anti-myogenin), and Drs L.L. Leger and F. Pons (anti-DEVmyosin).