Effects of fibroblast growth factor 2 on muscle precursor cells from mouse limb and extraocular muscle

TL;DR

This study shows that FGF2 increases muscle cell proliferation but reduces fusion into myotubes in mouse limb and extraocular muscle cells.

Contribution

The study reveals a novel mechanism by which FGF2 affects myofiber formation in extraocular muscles.

Findings

FGF2 significantly increased proliferation of EECD34 and PAX7-positive myogenic precursor cells.

FGF2 reduced fusion into multinucleated myotubes and decreased myomereger expression in differentiating cells.

The findings suggest a potential mechanism for reduced myofiber size in Apert syndrome.

Abstract

Fibroblast growth factor 2 (FGF2) is known to play a role in skeletal muscle development and growth. We examined two populations of myogenic precursor cells for their responses to FGF2 in vitro using both extraocular and limb skeletal muscle. Fluorescence-activated cell sorting (FACS) was used to isolate two different populations of myogenic precursor cells, the EECD34 cells (positive for CD34, and negative for Sca1, CD31, and CD45) and PAX7-positive cells, from tibialis anterior and extraocular muscles of mice. These cells were cultured and treated with either proliferation or differentiation media in the absence or the presence of FGF2, followed by assays to determine the effects on proliferation and differentiation. EECD34 cells and PAX7-positive cells from both muscles responded to FGF2 with significantly increased proliferation. Both myogenic precursor cell populations from each muscle type showed increased percentage of desmin-positive mononucleated cells, but decreased rates of fusion into multinucleated myotubes in the presence of FGF2 in this in vitro system relative to control cells. FGF2 has pleiotropic effects on skeletal muscles. Contrary to the literature, FGF2 did not inhibit differentiation, but did appear to decrease fusion into multinucleated myofibers in vitro. Examination of immunostaining for myomerger in differentiating PAX7-positive cells in the presence or absence of FGF2 demonstrated a significant reduction of expression in the presence of elevated FGF2 levels. These results provide a potential mechanism for reduction in myofiber number and size in the extraocular muscles in individuals with Apert syndrome, where FGF receptor 2 mutations maintain the receptor in an activated state resulting in significantly reduced myofiber size.

Figures

FACS-isolated EECD34 cells from extraocular muscles (EOM) (A-C) and tibialis anterior (TA) (D-F) immunostained for EdU (green) and Hoechst dye (blue).Grown for 3 days in high serum (20%) media to promote proliferation. (A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Magnification bar is 75 µm. (G) Quantification of EECD34 cells derived from EOM and TA after 24 hours in FGF2-supplemented proliferation media. FACS-isolated EECD34 cells from both tissues were cultured until approximately 30% confluent and then treated with either 10 ng/ml or 40 ng/ml FGF2 in proliferation media for 24 hours. Proliferation was determined by examining EdU uptake for 1 hour immediately following the 24-hour treatment. Immunohistochemistry was used to visualize the EdU-positive cells (green) compared to non-proliferating cells (blue). These data represent 5 independent replicates for each treatment. N = 5-6 independent cell cultures, each isolated from different mice. * indicates significant difference with p < 0.001. indicates significant difference with p < 0.0001.

FACS-isolated EECD34 cells from EOM (A-C) and TA (D-F) immunostained for desmin (brown) and counterstained with hematoxylin (blue) grown in the low-serum medium to promote differentiation.(A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Bar is 100µm. Blue arrow: undifferentiated cells; Green arrow: desmin-positive singly nucleated cells; Pink arrow: desmin-positive multiply nucleated cells. Quantification of FACS-isolated EECD34 cells derived from the EOM or TA after being placed in differentiation media, treated with FGF2 at 30% confluency and stained two days later. (G) FACS-isolated EECD34 cells from EOM were cultured and treated with either 10 ng/ml or 40 ng/ml FGF2 and compared to cells grown in differentiation media only. Two days later, the cells were fixed, and differentiation was determined by examining expression of desmin. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent seven independent replicates. (H) FACS-isolated EECD34 cells from TA were cultured, treated, and assayed in the same manner as the EOM EECD34 cells. These data represent seven independent replicates. N = 4-5 independent cultures, each isolated from EOM and TA muscles from different mice. (I) Analysis of fusion index calculated as the number of nuclei with EECD34 cells that contain two or more nuclei as a percent of all nuclei in all cells counted. N = 4-5 independent cell cultures, each isolated from different mice. * indicates significant difference with p < 0.05, indicates significant difference with p < 0.01, * indicates significant difference of p < 0.001. indicates significant difference with p < 0.0001.

Quantification of the rate of proliferation of FACS-isolated PAX7-positive cells derived from EOM and TA treated with FGF2.FACS-isolated PAX7-positive cells were cultured until approximately 30% confluent and then treated with either 10 ng/ml or 40 ng/ml FGF2 in proliferation media for 24 hours. Proliferation was determined by examining EdU uptake for 1 hour immediately following the 24-hour treatment. Immunohistochemistry was used to visualize the EdU-positive nuclei (green) compared to all nuclei (blue). These data represent 3-4 independent replicates for each treatment. * indicates significant difference with p < 0.05, indicates significant difference with p < 0.01.

PAX7-positive cells isolated from EOM (A-C) and TA (D-F) immunostained for desmin (brown) and counterstained with hematoxylin (blue) grown in the low-serum medium to promote differentiation.(A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Green arrow: desmin-positive singly nucleated cells; Pink arrow: desmin-positive multiply nucleated cells. Bar is 100 µm. Quantification of FACS-isolated PAX7-positive cells derived from the EOM (G) or TA (H) after being placed in differentiation media and treated with FGF2 at 30% confluency, and stained two days later. (G) FACS-isolated PAX7-positive cells from EOM were cultured and treated with either 10 ng/ml or 40 ng/ml FGF2 grown in differentiation media only. Two days later, the cells were fixed, and differentiation was determined by examining expression of desmin. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent 4 independent replicates. (H) FACS-isolated PAX7-positive cells from TA were cultured, treated, and assayed in the same manner as the EOM Pax7-positive cells. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent 5 independent replicates. * indicates significant difference from control within the same treatment group. (I) Analysis of fusion index was calculated as the number of nuclei in cells that contain two or more nuclei as a percent of all nuclei in all cells counted. * indicates significant difference with p < 0.05. indicates significant difference with p < 0.01. * indicates significant difference with p < 0.001. indicates significant difference with p < 0.0001.

PAX7-positive cells isolated from EOM (A-C) and TA (D-F) treated with FGF2 and subsequently immunostained for myomerger expression using an antibody to the ESGP protein.(A-C) Cells derived from EOM. (D-F) Cells derived from TA. (A, D) Control cultures (B, E) Cultures to which 10 ng/ml FGF2 was added. (C, F) Cultures to which 40 ng/ml FGF2 was added. Positive cells are green. Nuclei were stained with Hoechst 33342 stain. Bar is 75 µm. (G) Total number of nuclei in myomerger-positive cells was calculated as a percent of total nuclei in 5 random fields across each culture. (H) Total number of myofibers positive for myomerger expression was calculated as a percent of total cell number in 5 random fields across each culture. These data represent 3 independent replicates for each treatment. * indicates significant difference with p < 0.05. indicates significant difference with p < 0.01. * indicates significant difference with p < 0.001.

Full text

Introduction

Skeletal muscle contains resident precursor cells that enable repair and/or regeneration after injury or disease [1,2]. The first of these cells to be described were the satellite cells, which largely sit between the sarcolemma of the muscle fibers and the basal lamina surrounding each individual myofiber [3]. These cells are responsible for the ability of skeletal muscle to regenerate in muscle disease or after injury due to their ability to become activated, proliferate, and repair or regenerate new muscle fibers [4]. A number of studies have investigated the factors that control the proliferation and differentiation of muscle precursor cells in limb skeletal muscle. A wide range of neurotrophic and growth factors have been tested for their ability to stimulate proliferation and/or differentiation [5,6]. For example, fibroblast growth factor 2 (FGF2) promoted proliferation of satellite cells in vitro while insulin-like growth factor 1 (IGF1) promoted differentiation in satellite cells isolated from adult limb skeletal muscle from a number of animal species [5]. These factors also play a role in muscle regeneration and repair after injury or in the presence of skeletal muscle disease [7,8]. FGF2 has been shown to be particularly important in the maintenance of adult muscles, as its expression is up-regulated in muscle hypertrophy [9] and its absence in adult mice results in gait disturbances and muscle dysfunction [10].

Recent studies demonstrated that the myogenic precursor cells in skeletal muscles are not the same in different muscles [11]. Even within the well-described PAX7-positive satellite cells, a well-accepted marker for quiescent myogenic precursor cells found in all skeletal muscles [12], there is biochemical and functional heterogeneity [13,14]. In the extraocular muscles, there are differences in the populations of myogenic precursor cells when compared to limb skeletal muscles. In an earlier study we calculated this by both the total number of nuclei per total myofiber number in tissue sections and as the number of PAX7-positive cells per muscle weight quantified via flow cytometry [15]. In previous studies, we found that the density of PAX7-positive satellite cells in extraocular muscles is significantly elevated in number compared to limb skeletal muscles [15]. We performed several studies demonstrating that these PAX7-positive cells are preferentially spared after injuries compared to those in limb skeletal muscle, such as after irradiation [16] and are preferentially retained in the mdx mouse model of muscular dystrophy [17–19]. We identified a second large population of myogenic precursor cells, the EECD34 cells, identified as CD34-positive but negative for Sca1, CD31, and CD45 [18]. CD34 is commonly used as a marker to identify and purify muscle precursor cells, although common to many stem cell types [20]. Its expression is linked to the capacity for proliferation during skeletal muscle regeneration [21]. In the extraocular muscles, we showed using FACS analysis that this EECD34 population was 70–80% PITX2-positive, and this was confirmed using cytospin analysis [22]. Additionally, this was confirmed with immunohistochemical staining, where there was minimal to no overlap between PAX7 and PITX2 [22]. In the absence of PITX2 expression in mice, the extraocular muscles lose many of their unique properties, including decreased expression of the extraocular muscle specific myosin heavy chain (MyHC) isoform (myh13) and loss of the en grappe endplates [23,24]. Based on isolation of these two populations using flow cytometry in our laboratory [22], only 3.65% of the PITX2-positive cells expressed PAX7. This population is highly elevated in the extraocular muscles, and preferentially returns after 18Gy irradiation in both a mouse model of muscular dystrophy and in control mice, but not in limb muscle [16]. This suggests that besides being elevated in number, these cells from extraocular muscles have properties that may be different than this same population in limb skeletal muscle.

In addition to the differences in overall myogenic precursor cell numbers in normal extraocular muscles, extraocular muscles display a significant level of myonuclear addition into existing myofibers throughout life, as demonstrated using bromodeoxyuridine labeling methods [25,26]. These data were confirmed in other laboratories using a PAX7-reporter mouse, which showed extensive myonuclear addition in the normal mouse adult extraocular muscles compared to limb skeletal muscles [27]. Collectively, these data suggest that the myogenic precursor cells in the extraocular muscles are likely to respond differently to neurotrophic and growth factors than similar cells isolated from limb skeletal muscle.

We isolated EECD34 cells, as a stand-in for PITX2-positive cells, and PAX7-positive cells from mouse extraocular muscles and limb skeletal muscles, specifically the tibialis anterior. EECD34 cells and PAX7-positive cells from mouse extraocular muscles and limb skeletal muscles were examined for their rates of proliferation and differentiation in vitro after the addition of fibroblast growth factor 2 (FGF2). These were compared to control cultures to determine if there were differences in their behavior that might shed light on how the extraocular muscles are able to maintain their relatively high numbers of these myogenic progenitor cells throughout life. In addition, examination of myomerger expression during PAX7-positive cell differentiation in the presence of increased FGF2 levels in vitro was examined to determine if there were changes in myomerger expression associated with myoblast fusion [28]. The response of the myogenic precursor cells to FGF2 might provide an explanation of their differential responses to short term compared to long term exogenously added FGF2 in rabbit extraocular muscles in vivo [29], where after short-term treatment the extraocular muscles showed increased muscle force generation, but with sustained treatment (3 months) muscle force generation was significantly decreased [29]. These results are interesting, as previous studies in limb skeletal muscle treated with FGF2 showed increased numbers of PAX7-positive cells, promoting regeneration [7,10,30,31]. Additionally, the results of these in vitro experiments may shed light on the phenotype of individuals with genetic mutations that involve FGFR expression, such as Apert syndrome [32], where we have shown significant reduction in overall muscle size in the extraocular muscles from individuals with this genetic disorder [33].

Methods

Mice

Two mouse genotypes were used for cell isolation experiments. All studies adhered to the ARVO Animal Statement as well as to the principles for use of animals in research set by the National Institutes of Health. For EECD34 cell isolation, we used C57BL/6 mice purchased from The Jackson Laboratory (Bar Harbor, ME). For PAX7-positive cell isolation, we used a PAX7-reporter mouse. Pax7-Cre^+/-^; tdTomato-stop^fl/-^ mice were generated by crossing tdTomato^fl/-^ mice (JAX strain #: 007909) with Pax7^cre/ERT2^ mice (JAX strain #: 017763). Resulting offspring were bred to produce Pax7-Cre^+/-^; tdTomato-stop^fl/fl^ mice for experimental use, and the colony was maintained with Pax7-Cre^+/-^; tdTomato-stop^fl/fl^ (male) and Pax7-Cre^-/-^; tdTomato-stop^fl/fl^ (female) breeding pairs. All mice were housed by Research Animal Resources at the University of Minnesota Twin Cities. All mice used for these experiments were euthanized by tank CO_2_ asphyxiation at 3 months of age. Reporter expression was turned on in Pax7-Cre^-/+^; tdTomato-stop^fl/fl^ mice via a once-daily/five-day injection series of tamoxifen. Tamoxifen was prepared at a concentration of 20 mg/ml in corn oil and injected intraperitoneally at a dose of 75 mg tamoxifen per kg body weight. A waiting period of two weeks after the fifth injection was required for full activation of tdTomato expression.

Isolation of myogenic precursor cells

PAX7-positive cells.

Extraocular muscle (EOM) and tibialis anterior (TA) muscles were collected and pooled from 3 mice into chilled Dulbecco’s Modified Eagle Medium (DMEM) (catalog # CX30313, Gibco, Billings, MT). EOM samples were placed into 1x phosphate buffered saline (PBS) containing collagenase B (catalog #11088831, Roche, Basel, Switzerland) at 10 mg/ml, dispase II (catalog #D4693-1G, Sigma-Aldrich, St. Louis, MO) at 2.4 units/ml, and CaCl_2_ at 2.5 mM. Prior to placement in the collagenase/dispase solution, TA samples were chopped quickly with a razor blade to aid in digestion. Samples were incubated at 37⁰C, 5% CO_2_, for 90 minutes, and briefly removed every 25 minutes for pipette dissociation. The resulting cell suspension was filtered through a 70 µm filter and pelleted at 1400 rpm for 10 minutes at 4^o^C. After pouring off the supernatant, the pellet was resuspended in PBS-based sorter buffer containing 95% v/v PBS (catalog # 10010023, Gibco), 2.5% v/v fetal bovine serum (FBS) (catalog #S11150, Sigma Aldrich) and 2.5% v/v HEPES buffer (catalog #15630080, Gibco). The resuspended cells were filtered through a 40 µm filter into 5 ml FACS tubes and centrifuged at 1400 rpm at 4^o^C for 5 minutes. Supernatant was discarded, and cells were resuspended in 2 ml of sorter buffer containing SYTOX Blue Dead Cell Stain (catalog #S11348, Invitrogen, Carlsbad, CA) at a 1:10,000 dilution. All work was performed in a sterile cell culture hood. Cells were kept on ice between steps.

EECD34 cells.

Samples used for EECD34 cell collection were prepared similarly to PAX7-positive cells with the following changes [22,34]. After filtering through a 70 µm filter, pelleting, and resuspending, the cells were incubated in the dark on ice for 45 minutes with fluorescent antibodies targeting CD34 (5 µl/100 µl sorter buffer, BV421 rat anti-mouse CD34, catalog #562608, BD Biosciences, Franklin Lakes, NJ), Sca1 (2 µl/100 µl sorter buffer, FITC rat anti-mouse Ly-6A/E, catalog #553335, BD Biosciences), CD45 (2 µl/100 µl sorter buffer, PE/Cyanine7 rat anti-mouse CD45, (catalog #103114, BioLegend, San Diego, CA), and CD31 (5 µl/100 µl sorter buffer, APC rat anti-mouse CD31, catalog #551262, BD Biosciences). Cells were filtered through a 40 µm filter, centrifuged, and resuspended in sorter buffer containing SYTOX Orange Nucleic Acid Stain (catalog #S11368, Invitrogen). All work was done in a sterile cell culture hood. Cells were kept on ice between steps.

FACS.

PAX7-positive and EECD34 cells were collected from the EOM and TA samples with a BD FACSAria II at the University of Minnesota Flow Cytometry Resource Center. For PAX7-positive cell sorting, tdTomato^+^/SYTOX Blue^-^ cells were collected. For EECD34 cell sorting, CD34^+^/Sca1^-^/CD45^-^/CD31^-^/SYTOX Orange^-^ cells were collected. All cells were collected into 2 ml of proliferation media containing 78.5% DMEM (Gibco), 10% FBS, 10% horse serum (catalog #12150, R&D Systems, Minneapolis, MN), 0.5% chick embryo extract (catalog #100356−958, VWR, Radnor PA), and 1% penicillin/streptomycin (catalog #CX30321, Gibco). EECD34 and PAX7-positive cell amounts as percentages of the total number of single cells were collected from FACS gating data (S1 Fig).

To validate that the EECD34 cells represent PITX2-expressing myogenic precursors, EECD34 cell cultures were immunostained for the expression of PITX2 at 24 and 72 hours in vitro. Cells were fixed in 4% paraformaldehyde at room temperature for 15 minutes, rinsed in PBS, and then incubated in blocking buffer containing 20% goat serum and 0.5% Triton X-100 (catalog #T9284, Sigma Aldrich) in PBS for 10 minutes at room temperature. The cell plates were then incubated with a primary antibody to PITX2 (1:1000; catalog number PA-1020–100, Capra Science, Ängelholm, Sweden) at room temperature for 30 minutes in the dark. After rinsing in PBS, cells were blocked for 15 minutes in 20% goat serum containing 0.2% BSA in antibody buffer, and then incubated at room temperature for 1 hour with goat anti-rabbit-AlexaFluor 488 in PBS (1:1000; Cat #A-11008; Invitrogen, Carlsbad, CA) at room temperature in the dark. Cells were rinsed in PBS before being cover-slipped with Vectashield Mounting Media containing DAPI (H-2000–2, Vector Laboratories, Burlingame, CA). Cells were imaged on a Leica DM4000B Fluorescent Microscope (Leica, Wetzlar, Germany). A minimum of 10 images at 20x magnification were quantified per treatment well. PITX2-positive cells were quantified as a percentage of all cells per field, and these were averaged for each of 3 independent replicates per time point using two mice per independent culture. Essentially all of the EECD34 cells in our in vitro system expressed PITX2 within the first 24 hours, and this expression was maintained in approximately 76% of the cells after 72 hours in vitro (S2 Fig).

Cell culture and FGF2 treatment

After cells were sorted, they were pelleted and resuspended in proliferation media to a concentration of 32 cells/µl. Thus, for all in vitro experiments, 250 µl of these cells were plated on Matrigel Membrane Matrix (catalog #CB-40234A, Corning, NY) coated Lab-Tek 8 Well Chamber Slides (catalog #62407−335, Corning) and kept in an incubator at 37⁰ C and 5% CO_2_. Every 48 hours, proliferation media was aspirated and replaced with fresh proliferation media. For the proliferation assay, at ~30% confluency, cells were treated with recombinant mouse FGF2 (catalog #319-FB-025, Bio-Techne, Minneapolis, MN) at 10 ng/ml or 40 ng/ml in proliferation media. For the differentiation assay, when cells were at 30% confluency, roughly 4 days for EECD34 cells and 3 days for PAX7-positive cells, they were placed in low serum differentiation media supplemented with 10 ng/ml or 40 ng/ml of recombinant mouse FGF2 (Bio-Techne). The low serum media was aspirated and replaced with fresh media containing consistent amounts of FGF2 every 48 hours. The low serum media consisted of 94% v/v DMEM (Gibco), 5% v/v horse serum (catalog #S12150, R&D Systems, Minneapolis, MN), and 1% v/v penicillin-streptomycin (catalog #CX30321, Gibco). Control wells for both the proliferation and differentiation assays were supplemented with Dulbecco’s Phosphate-Buffered Saline (DPBS) only (catalog #CX30294, Gibco).

Cell proliferation assay

Still in their culture plates, the cells were incubated in proliferation media with 10 µM EdU for 1 hour at 37⁰C and 5% CO_2_ 24 hours following FGF2 treatment. Then, cells were fixed in 4% paraformaldehyde at room temperature for 15 minutes before being washed with 3% BSA w/v in PBS, and then permeabilized with 0.5% Triton X-100 (catalog #T9284, Sigma-Aldrich) in PBS for 10 minutes at room temperature. Cells were washed with 3% BSA w/v in PBS before being incubated with ClickiT EdU Cell Proliferation Kit reagents (catalog #C10337, Invitrogen) at room temperature for 30 minutes in the dark to label proliferating cells. After washing with 3% BSA w/v in PBS, cell nuclei were incubated with Hoechst 33342 (part of ClickiT kit, catalog #C10337, Invitrogen) in PBS at 1:2000 dilution for 15 minutes at room temperature in the dark. Cells were washed in 3% BSA w/v in PBS before being cover-slipped with H1000 Vectashield Mounting Media (catalog #NV9265087, Vector Lab.). Cells were imaged on a Leica DM4000B Fluorescent Microscope (Leica). A minimum of 10 images at 20x magnification were quantified per treatment well. Proliferating cell percentages from these images were averaged to determine the percent of cells in each category from the same biological replicate. These averages were used to calculate the percent of cells proliferating in a minimum of 4 independent replicates per treatment. Proliferating cell percentages were calculated for each image using the Bioquant Imaging System (Bioquant, Nashville, TN).

Cell differentiation assay

Once EOM cells reached ~80% confluency, cells in low-serum media treated with FGF2 were assayed for their differentiation status. This usually occurred 1–2 days after the change to differentiation media with FGF2 for PAX7-positive cells and 2 days after for EECD34 cells. Because TA cells generally grew slower than EOM and were not necessarily treated with FGF2 at the same time, TA cells were assayed on later days than EOM but on the same timeline as described here. For the assay, the media was aspirated, and wells were washed with PBS before being fixed in 4% paraformaldehyde in PBS. After washing in PBS, cultures were blocked in 10% horse serum (R&D Systems) for 30 minutes before use of avidin and biotin blocking solutions (catalog #SP2001, Vector Lab.) according to package instructions. Cells were incubated in desmin antibody (catalog #M0760, Dako, Troy, MI) at a 1:500 concentration for 1 hour at room temperature to label differentiating cells. The Vectastain Elite ABC-HRP kit (catalog #PK-4002, Vector Lab.) was used according to package instructions to enable colorimetric desmin staining when cells were consequently treated with diaminobenzidine (DAB). Slide chambers were removed, and slides were submerged in hematoxylin to visualize nuclei. The slides were coverslipped using Vectashield Antifade Mounting Media (Vector Lab.). Levels of cell differentiation were defined as follows: desmin-negative and thus non-differentiating with a single nucleus; desmin-positive and thus differentiating with a single nucleus; and desmin-positive differentiating with multiple nuclei. These populations were quantified using the Bioquant Imaging System (Bioquant). Approximately 20 images at 20x magnification were quantified per treatment well. These were averaged to determine the percent of cells in each category from the same biological replicate. These averages were used to calculate the percent of cells in each differentiation state in a minimum of 4 independent replicates per treatment.

Myomerger expression assay

An additional group of PAX7-positive cells in three independent replicates were grown for the differentiation assay as described in the preceding paragraph. At the same timepoint as for the differentiation assay, the media was aspirated, and wells were washed with PBS before being fixed in 4% paraformaldehyde in PBS. After washing in PBS, cultures were blocked in 20% donkey serum (catalog #10802–530, Rockland, Pottstown, PA) and 0.2% BSA (catalog #BP9706–100, Fisher, Waltham, MA) in PBS for 30 minutes. Without rinsing, the cells were incubated in an antibody to myomerger (5 µg/ml, ESGP; catalog # PA5–47639, Invitrogen) for 2 hours at room temperature. The slides were then rinsed in PBS, blocked again with the donkey serum blocking solution, and incubated in donkey anti-sheep IgG Alexafluor 488 for 1 hour at room temperature (1:500; catalog # 713545147, Jackson ImmunoResearch Lab., West Grove, PA). After rinsing in PBS, cell nuclei were incubated with Hoechst 33342 (part of ClickiT kit, Invitrogen) in PBS at 1:2000 dilution for 15 minutes at room temperature in the dark. Hoechst solution was removed, slide chambers were detached, and slides immersed in PBS. The slides were coverslipped using Vectashield Antifade Mounting Media (Vector Lab.). Staining was analyzed using a minimum of 5 fields randomly selected from across each culture well, and they were used to determine the number of nuclei in myomerger-positive cells out of the total number of nuclei per field, and averaged for each well.

Statistical analysis

All data are reported as means + /- SEM. All statistical analyses were performed in Prism 9 (GraphPad, La Jolla, CA). One-way ANOVAs, with multiple comparisons when necessary, were performed to determine significance. If the data were significantly different, post-hoc Tukey’s multiple comparison tests were performed for TA and EOM data to determine significance between treatment groups. Statistical significance was defined as P < 0.05.

Results

EECD34 cell assays

Proliferation.

We examined the effect of FGF2 treatment of EECD34 cells derived from EOM and TA on proliferation rate using EdU labeling in vitro. Two-way ANOVA analysis showed that there was a significant main effect of both treatment (p < 0.0001) and tissue (p < 0.0001), as treatment with FGF2 significantly increased the percent of proliferating cells. Post hoc analyses showed increases from the control percentage of EOM-derived EECD34 cells from 41.4% to 60.6% for the 10 ng/ml treatment (p < 0.0001) and to 61.6% for the 40 ng/ml treatment (p < 0.0001) (Fig 1A-C,G). These represent percent increases of 46.6% and 49% respectively. Post hoc t-tests showed that the EECD34 control cells derived from EOM had a significantly larger percentage of proliferating cells, 41.5%, than those derived from TA, 24.8%, a 40.2% difference (p = 0.0001). For EECD34 cells derived from TA muscles, the 10 ng/ml treatment resulted in a significant increase in the percent proliferating cells compared to the TA control level, from 24.8% to 41.4%, a significant increase of 67.4% (p = 0.0001). The 40 ng/ml treatment similarly increased the percent of proliferating cells from the TA control level to 39.3%, a significant increase of 58.6% (p = 0.0006) (Fig 1G). Comparing the control cultures derived from EOM and TA cells, the EECD34 cells isolated from EOM had a significantly higher proliferation rate than those isolated from TA, a comparative increase 40.2% (p = 0.0001). In addition, proliferation rates in the EECD34 cells treated with 10 ng/ml FGF2 from TA showed significantly less proliferation, 31.7% less, compared to those derived from EOM (p = 0.0001). Proliferation rates for EECD34 cells treated with 40 ng/ml FGF2 from TA also displayed significantly reduced proliferation, 36.3% less, compared to EOM cells (p < 0.0001).

FACS-isolated EECD34 cells from extraocular muscles (EOM) (A-C) and tibialis anterior (TA) (D-F) immunostained for EdU (green) and Hoechst dye (blue).Grown for 3 days in high serum (20%) media to promote proliferation. (A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Magnification bar is 75 µm. (G) Quantification of EECD34 cells derived from EOM and TA after 24 hours in FGF2-supplemented proliferation media. FACS-isolated EECD34 cells from both tissues were cultured until approximately 30% confluent and then treated with either 10 ng/ml or 40 ng/ml FGF2 in proliferation media for 24 hours. Proliferation was determined by examining EdU uptake for 1 hour immediately following the 24-hour treatment. Immunohistochemistry was used to visualize the EdU-positive cells (green) compared to non-proliferating cells (blue). These data represent 5 independent replicates for each treatment. N = 5-6 independent cell cultures, each isolated from different mice. *** indicates significant difference with p < 0.001. **** indicates significant difference with p < 0.0001.

Differentiation.

EECD34 cells from EOM were grown in differentiation media and treated with either 10 ng/ml or 40 ng/ml FGF2, and ANOVA analysis showed significant differences (p = 0.0001) (Fig 2A-G). Post hoc multiple comparison tests showed a significant decrease of desmin-negative cells when treated with 40 ng/ml FGF2 compared to controls (70.28% decrease, p = 0.036), and a trend toward decreasing of desmin-negative cells when treated with 10 ng/ml FGF2 that was not significant (69.28% decrease, p = 0.056). There were significant increases in the number of desmin-positive cells with single nuclei after both 10 ng/ml and 40 ng/ml treatment (23.84% increase, p = 0.014; 25.5% increase, p = 0.008), respectively.

FACS-isolated EECD34 cells from EOM (A-C) and TA (D-F) immunostained for desmin (brown) and counterstained with hematoxylin (blue) grown in the low-serum medium to promote differentiation.(A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Bar is 100µm. Blue arrow: undifferentiated cells; Green arrow: desmin-positive singly nucleated cells; Pink arrow: desmin-positive multiply nucleated cells. Quantification of FACS-isolated EECD34 cells derived from the EOM or TA after being placed in differentiation media, treated with FGF2 at 30% confluency and stained two days later. (G) FACS-isolated EECD34 cells from EOM were cultured and treated with either 10 ng/ml or 40 ng/ml FGF2 and compared to cells grown in differentiation media only. Two days later, the cells were fixed, and differentiation was determined by examining expression of desmin. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent seven independent replicates. (H) FACS-isolated EECD34 cells from TA were cultured, treated, and assayed in the same manner as the EOM EECD34 cells. These data represent seven independent replicates. N = 4-5 independent cultures, each isolated from EOM and TA muscles from different mice. (I) Analysis of fusion index calculated as the number of nuclei with EECD34 cells that contain two or more nuclei as a percent of all nuclei in all cells counted. N = 4-5 independent cell cultures, each isolated from different mice. * indicates significant difference with p < 0.05, ** indicates significant difference with p < 0.01, *** indicates significant difference of p < 0.001. **** indicates significant difference with p < 0.0001.

EECD34 cells from TA were grown in differentiation media and treated with either 10 ng/ml or 40 ng/ml FGF2; ANOVA analysis indicated significant differences (p = 0.0001) (Fig 2A-G). Post hoc multiple comparison tests showed that compared to the untreated controls, there were no significant differences in the number of desmin-negative cells after either of the FGF2 treatments (p = 0.67; p = 0.58, respectively). There were significant increases in the number of desmin-positive cells with single nuclei after both 10 ng/ml and 40 ng/ml treatment (18.9% increase, p = 0.0015; 18.7% increase, p = 0.0007), respectively.

There was a significant decrease in cells with two or more nuclei in cultures from both EOM (Fig 2G) and TA (Fig 2H) after either 10 ng/ml or 40 ng/ml FGF2 (Fig 2G, H) compared to control. After 10 ng/ml treatment of EOM-derived and TA-derived cells, multinucleated cells decreased by 64.2% (p = 0.012) and 65.4% (p = 0.02) respectively. After 40 ng/ml FGF2, multinucleated cells significantly decreased by 55.75% (p = 0.02) in EOM cultures and 62.2% in TA cultures (p = 0.02).

A fusion index for the EECD34 cells was calculated based on number of nuclei within desmin-positive cells that contained two or more nuclei divided by the number of total nuclei in all cells (Fig 2I). There was a significant decrease in the fusion index of these differentiating cultures in the presence of 10 ng/ml or 40 ng/ml FGF2 compared to controls, decreasing by 65.4% (p = 0.012) and 62.2% (p = 0.02), respectively. No significant differences were seen when comparing identical treatment conditions between EOM and TA cells.

PAX7-positive cell assays

Proliferation.

FGF2-treated PAX7-positive cells derived from EOM and TA were examined for its effect on proliferation rate (Fig 3). The two-way ANOVA analysis demonstrated a main effect for treatment (p < 0.0001) but not for tissue type. Post hoc analysis demonstrated that, as with the EECD34 cells, both 10 ng/ml and 40 ng/ml of FGF2 resulted in a significant increase in the proliferating EOM-derived cell populations. The 10 ng/ml treatment increased the percent proliferating cells from 30.7% to 52.2%, an increase of 70.1% (p = 0.001). The 40 ng/ml treatment similarly increased the percent proliferating cells to 47.7% compared to the controls, an increase of 55.4% (p = 0.0186). Post hoc analysis demonstrated that the 10 ng/ml treatment increased the percent proliferating cells from 34.0% to 47.3%, an increase of 39.1% (p = 0.04). Despite a 25.7% increase in cell number after treatment with 40 ng/ml FGF2, this was not statistically significant (p = 0.21). Post hoc tests confirmed that no significant differences were observed when comparing identical treatment conditions between EOM and TA samples.

Quantification of the rate of proliferation of FACS-isolated PAX7-positive cells derived from EOM and TA treated with FGF2.FACS-isolated PAX7-positive cells were cultured until approximately 30% confluent and then treated with either 10 ng/ml or 40 ng/ml FGF2 in proliferation media for 24 hours. Proliferation was determined by examining EdU uptake for 1 hour immediately following the 24-hour treatment. Immunohistochemistry was used to visualize the EdU-positive nuclei (green) compared to all nuclei (blue). These data represent 3-4 independent replicates for each treatment. * indicates significant difference with p < 0.05, ** indicates significant difference with p < 0.01.

Differentiation.

PAX7-positive cells from EOM were grown in differentiation media and treated with either 10 ng/ml or 40 ng/ml FGF2 (Fig 4A-C, G). ANOVA analysis showed a significant difference (p = 0.0001). Post-hoc multiple comparison tests showed that there was a significant increase in the number of desmin-positive cells with single nuclei at both 10 ng/ml and 40 ng/ml doses, increasing from 92.2% to 97.7% (p = 0.0005) and 98.1% (p = 0.0003), respectively. Similarly, there was a reduction in fusion compared to the control cells, with a decrease from the control at 7.7% down to 2.3% in cells treated with 10 ng/ml FGF2 (p = 0.0005) and 1.8% in cells treated with 40 ng/ml FGF2 (p = 0.0003). PAX7-positive cells derived from TA showed similar responses to those isolated from the EOM. ANOVA analysis showed a significant difference (p = 0.0001) (Fig 4D-F, H). Post-hoc multiple comparison tests showed that there was a significant increase in the number of desmin-positive cells with single nuclei at both 10 ng/ml and 40 ng/ml relative to the controls. The percent of these cells went from 82.3% in controls cells up to 92.9% in cells treated with 10 ng/ml FGF2 (p = 0.02) and 94.7% in cells treated with 40 ng/ml FGF2 (p = 0.014). Similarly, there was a reduction in fusion compared to the control cells, with a decrease from the control at 17.2% down to 5.9% in cells treated with 10 ng/ml FGF2 (p = 0.019) and 5.3% in cells treated with 40 ng/ml FGF2 (p = 0.014). As with the EECD34 cells, no significant differences were seen when comparing EOM and TA PAX7-positive cells.

PAX7-positive cells isolated from EOM (A-C) and TA (D-F) immunostained for desmin (brown) and counterstained with hematoxylin (blue) grown in the low-serum medium to promote differentiation.(A,D) Untreated control cells, (B,E) cells treated with 10 ng/ml FGF2, and (C,F) cells treated with 40 ng/ml FGF2. Green arrow: desmin-positive singly nucleated cells; Pink arrow: desmin-positive multiply nucleated cells. Bar is 100 µm. Quantification of FACS-isolated PAX7-positive cells derived from the EOM (G) or TA (H) after being placed in differentiation media and treated with FGF2 at 30% confluency, and stained two days later. (G) FACS-isolated PAX7-positive cells from EOM were cultured and treated with either 10 ng/ml or 40 ng/ml FGF2 grown in differentiation media only. Two days later, the cells were fixed, and differentiation was determined by examining expression of desmin. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent 4 independent replicates. (H) FACS-isolated PAX7-positive cells from TA were cultured, treated, and assayed in the same manner as the EOM Pax7-positive cells. Cells were categorized as desmin-negative, desmin-positive containing only one nucleus, or desmin-positive containing more than one nucleus. These data represent 5 independent replicates. * indicates significant difference from control within the same treatment group. (I) Analysis of fusion index was calculated as the number of nuclei in cells that contain two or more nuclei as a percent of all nuclei in all cells counted. * indicates significant difference with p < 0.05. ** indicates significant difference with p < 0.01. *** indicates significant difference with p < 0.001. **** indicates significant difference with p < 0.0001.

A fusion index for the PAX7 cells was calculated (Fig 4I). This analysis demonstrated that in the cultures of PAX7 cells derived from EOM there were significant decreases of the number of nuclei within multinucleated fibers in FGF2 treated cultures compared to control cultures, with decreases in 10 ng/ml FGF2 cultures of 93.7% (p = 0.0001) and 94.2% in cultures treated with 40 ng/ml FGF2 (p = 0.0001). Decreased numbers of nuclei within multinucleated fibers were also seen in cultures of PAX7 cells derived from TA, with decreases in cultures treated with 10 ng/ml FGF2 by 64.5% (p = 0.0001) and decreases in cultures treated with 40 ng/ml FGF2 by 62.8% (p = 0.0001). As seen with the EECD34 cells, the percent of nuclei within multinucleated fibers was significantly greater in control cultures of PAX7 cells derived from TA than EOM (34.5% increase, p = 0.0034). In the presence of 10 ng/ml FGF2, the EOM-derived cultures showed significantly lower numbers of nuclei within multinucleated cells compared to the TA derived cultures (88.5% decrease, p = 0.0008). Similarly, there was a decreased number of nuclei within multinucleated cells in the cultures treated with 40 ng/ml FGF2 derived from EOM compared to those from TA, with an 89.8% decrease (p = 0.005).

Myomerger analysis

Expression of myomerger was examined in PAX7 cells from both EOM and TA grown in differentiation media after treatment with either 10 ng/ml or 40 ng/ml FGF2 and compared to controls (Fig 5). ANOVA demonstrated that the results were significantly different (p < 0.0001), so post-hoc multiple comparison tests were performed. For cells from both EOM and TA, addition of 10 ng/ml FGF2 to the differentiating cells resulted in a significant decrease in numbers of cells expressing myomerger, an 85.9% decrease in EOM and a 61.9% decrease in TA (p = 0.0003) for both (Fig 5A-G). There was a similar decrease in the percent of cells expressing myomerger after addition of 40 ng/ml FGF2, with decreases of 90% in the cultured cells derived from EOM and 58.5% in those from TA (p = 0.0002 and p = 0.0006, respectively) (Fig 5A-G). None of the other pairwise comparisons were significantly different except the decrease in percent of cells expressing myomerger in EOM compared to TA after the addition of 40 ng/ml FGF2, which showed an 82% difference (p = 0.0367). There was also a significant decrease in percent of cells expressing myomerger after treatment by either 10 ng/ml or 40 ng/ml FGF2 compared to the total number of cells (Fig 5H). After the addition of 10 ng/ml FGF2 there were 89.7% fewer myomerger-positive cells in the EOM cultures and 66.6% fewer myomerger-positive cells in the TA cultures (p = 0.0028 and p = 0.0012, respectively). There was a similar decrease in percent of the cells in vitro expressing myomerger after addition of 40 ng/ml FGF2 in both EOM, with a 92.3% decrease, and TA, with a 62.5% decrease (p = 0.002 for both).

PAX7-positive cells isolated from EOM (A-C) and TA (D-F) treated with FGF2 and subsequently immunostained for myomerger expression using an antibody to the ESGP protein.(A-C) Cells derived from EOM. (D-F) Cells derived from TA. (A, D) Control cultures (B, E) Cultures to which 10 ng/ml FGF2 was added. (C, F) Cultures to which 40 ng/ml FGF2 was added. Positive cells are green. Nuclei were stained with Hoechst 33342 stain. Bar is 75 µm. (G) Total number of nuclei in myomerger-positive cells was calculated as a percent of total nuclei in 5 random fields across each culture. (H) Total number of myofibers positive for myomerger expression was calculated as a percent of total cell number in 5 random fields across each culture. These data represent 3 independent replicates for each treatment. * indicates significant difference with p < 0.05. ** indicates significant difference with p < 0.01. *** indicates significant difference with p < 0.001.

Discussion

It is well established that the EOM have a complex and large pool of myogenic precursor cells. We specifically examined the effect of FGF2 on two populations: PAX7-positive cells, known to be the major quiescent myogenic precursor cell in all skeletal muscles [12], and population of myogenic precursor cells we call the EECD34 cells, which express CD34 but are negative for CD31, CD45, and SCA-1 [18]. These EECD34 cells are approximately 70% PITX2-positive in EOM and about 65% PITX2-positive in leg muscle [22] and represent a population of myogenic precursor cells particularly elevated in the extraocular muscles. Based on our flow cytometric cell isolations of PAX7 and PITX2-immunostained cells derived from normal genetically-unaltered mice followed by cytospin isolation, only 3.65% of PITX2-positive cells labeled directly ex vivo co-expressed PAX7 [22]. Thus, in our hands, the EECD34 cells and the PAX7 cells represent distinct populations. We assessed the responses of these two cell populations to FGF2 while growing in FGF2-supplemented culture media. The EECD34 cells isolated from adult mouse EOM muscle showed an increased rate of proliferation in the presence of FGF2 in vitro, similar to studies showing that FGF2 treatment of isolated limb-derived satellite cells promotes the proliferation in vitro [35,36]. It is also known that skeletal muscle satellite cells can express both FGF1 and FGF2 [37]. This supports experimental evidence that despite being isolated under the same parameters, defined by their expression of this set of specific precursor cell markers — CD34^+^/Sca1^-^/CD31^-^/CD45^-^ — they have some similar properties as the “same” cells derived from the TA, despite the comparatively unique properties of the EOM. This is further supported by our data showing a significant increase in the proportion of EOM EECD34 cells proliferating under control and FGF2-supplemented conditions compared to TA cells. It is interesting that no significant differences between EOM and TA PAX7-positive cells were seen, suggesting there is an inherent difference in EECD34 cells from different muscles.

Similarly, the PAX7-positive cells isolated from EOM and TA showed a significantly increased rate of proliferation in the presence of added FGF2 in vitro. This agrees with a large number of previous studies, where isolated satellite cells – which are largely PAX7-positive – responded to FGF2 in vitro with increased rates of proliferation [5,38]. Direct intramuscular injection of FGF2 also was shown to result in increased proliferation within the treated limb skeletal muscles, with short-term increases in Ki67-positive and PAX7-positive cells [39]. It is interesting to note a complexity in the response of extraocular muscle to FGF2. One week after injection of FGF2 directly into an extraocular muscle in adult rabbits there was a significant increase in the number of PAX7-positive cells within the treated muscle [29]. However, after sustained delivery for 3 months of FGF2 in adult rabbit extraocular muscle, there was a significant decrease in satellite cell numbers and significantly reduced myofiber cross-sectional areas [29]. In FGF2 knockout mice, significant gait disturbance and decreases in muscle size were present [10], supporting the hypothesis that FGF2 is critical in maintaining the pool of muscle precursor cells needed for maintenance of normal muscle function over time.

A number of recent studies have examined the myogenic precursor cell populations in the extraocular muscle with conflicting results. Studies have demonstrated, in agreement with our bromodeoxyuridine pulse-labeling studies [25,26,40–42], that there is a continual, albeit slow, process of replication of these cell populations in the extraocular muscles [27] and that these cells appear to have greater and more long-lasting proliferative potential throughout life [19,21,43,44]. These properties appear to be part due to the environment of the niche [45,46] and part due to properties intrinsic to the extraocular muscles [19,46,47]. A recent study using a number of transgenic mice described the PITX2-positive and PAX7-positive cell populations to be non-proliferative in the adult, while previous studies showing proliferation and myofiber fusion in many skeletal muscles [27,48], but the results agree that the extraocular muscles contain a higher content of PAX7-positive cells than limb skeletal muscles [49]. One potential explanation for this discrepancy could relate to transgenic mouse modifications in development that might alter the potential fates of the myogenic precursor cell populations in the adult mouse. Future studies will need to address this difference.

Using desmin as a marker for differentiation, FGF2 treatment of EECD34 and PAX7-positive cells in vitro resulted in a significant increase in the percent of differentiating cells, as identified by desmin expression, as well as a likely concomitant decrease in the number of desmin-negative undifferentiated cells. These results contrast with many previous studies showing that FGF2 significantly decreased differentiation in vitro and in vivo [5,7,36,50]. However, these sorts of pleiotropic effects are commonly seen with a variety of neurotrophic and growth factors. For example, insulin-like growth factor-1 can promote both the proliferation and differentiation of muscle precursor cells [51,52]. FGF2 also has been shown to have these pleiotropic effects [6,28].

Interestingly, despite the increase in the percentage of desmin-positive cells, for both the EECD34 cells and the PAX7-positive cells isolated from extraocular muscles and leg muscle, there was a decrease in the desmin-positive multinucleated cells. This suggests that while differentiation significantly increased, there was a concomitant reduction in the ability of these differentiating cells to fuse. An earlier study showed that high levels of FGF2 were found in myogenin-positive cells, which indicate that they were differentiating but also singly-nucleated [53]. This lack of fusion is interesting in light of several studies suggesting this also occurs when FGF2 levels are increased within skeletal muscles. Using FGF2 embedded in microspheres, injection into injured leg muscle resulted in increased myogenic precursor cell proliferation but no change to overall muscle force generation [54]. The ability of FGF2 to increase functional muscle mass was tested using FGF2 collagen scaffold implants after volumetric muscle loss, and while it resulted in increased myoblast proliferation, it was unsuccessful in increasing myofiber density [55]. In another study, short-term treatment of skeletal muscle with FGF2 resulted in some improvement in number of regenerated myofibers in the mdx mouse muscular dystrophy model, but the effect was largely due to enhanced replication of satellite cells in the treated muscles [7,29]. Also, local injection of FGF2 into a previously denervated thyroarytenoid muscle within the larynx again resulted in increased numbers of proliferating cells as well as increased numbers of MyoD-positive cells within the treated muscles, suggesting that FGF2 had the potential to decrease atrophy in the denervated muscles [39]. Similar results were obtained in aging laryngeal muscles, with some evidence that differentiation was promoted by these treatments [39,56]. Long-term functional changes after these treatments were not assessed in these two studies. We previously showed significantly different effects of short-term and long-term treatment with FGF2 [29]. While direct injection of FGF2 into a superior rectus muscle of an adult rabbit resulted in increased numbers of PAX7-positive cells and increased muscle force, one-month and three-month sustained treatment with FGF2 resulted in a significant decrease in the numbers of PAX7-positive cells and a significant decrease in force generated by the treated superior rectus muscles. These results are interesting in that they provide a potential mechanism for the small extraocular muscles sizes seen in individuals with Apert syndrome, which is also supported by the present study [33]. In Apert syndrome, in the vast majority of cases there are one of two mutations in the gene for FGFR2 [57,58]. These gain-of-function mutations result in increased affinity for the FGF ligand, maintaining prolonged action even when FGF levels are limited [59]. Similarly, other craniosynostosis disorders that result in strabismus, such as Crouzon syndrome and Pfeiffer syndrome, have gain-of-function mutations in FGFR2 [60]. The effect on extraocular muscles and the resultant strabismus is not surprising, as human fetal extraocular muscles express FGFR2 [61]. Thus, this in vitro study serves as a potential mechanism for the effects of prolonged signaling of FGFR2.

To define further the altered fusion behavior of FGF2, treated cells were immunostained for myomerger expression, a protein required for fusion pore formation, a critical step in myoblast fusion [28,62]. We saw a significant decrease in the proportion of nuclei within myomerger-positive cells treated with FGF2 for both EOM and TA derived PAX7-positive cells. This confirms our observation from our differentiation experiments that FGF2 reduces cell fusion, and one potential mechanism is through the reduction in the expression of myomerger. Expression of myomerger, in conjunction with myomaker, is tightly regulated and limited to differentiating myoblasts, likely because they disrupt membrane integrity [63].

The apparent inhibition of fusion in the FGF2 treated cells was surprising, as FGF2 has not been linked to the control of myoblast fusion. This result supports our observation that after sustained treatment with FGF2, the extraocular muscles generated less force, and the myofibers were smaller than normal – presumably by further decreasing fusion of muscle precursor cells into existing fibers [29]. Interestingly, FGF2 is involved in controlling myoblast migration during development [64] and in vitro [65,66], which in turn could result in decreased fusion. Future studies will need to determine if this is the case. Interestingly, transforming growth factor-beta (TGFβ) also results in decreased myoblast fusion in limb skeletal muscle [67,68]. Expression levels of TGFβ appear to be quite low in normal extraocular muscles [69,70], but is expressed at elevated levels in strabismic muscles [70,71]. When we treated adult rabbit EOM with TGFβ, we found that the treated rabbit superior rectus muscles had significantly decreased mean myofiber cross-sectional areas from the already small fibers found in the extraocular muscles, and these muscles generated less force – similar to what was seen after extended FGF2 treatment [29,72]. These studies suggest that modulation of many different neurotrophic factors can result in similar outcomes when attempting to modify extraocular muscle structure and function.

With skeletal muscles, there are a large number of neurotrophic and growth factors whose expression overlap in normal adult, developing, and diseased or injured muscles. In a controlled tissue engineering experiment, after accelerating satellite cell activation and proliferation with FGF2, sequential addition of insulin-like growth factor 1 (IGF1) was a potent inducer of differentiation [73]. In the extraocular muscles, sequential expression of FGF2 and IGF-1 was observed in a time series shortly after an experimental recession in a rabbit model [71], again suggesting that multiple growth factors play a complex, interdependent role in maintaining normal function in the extraocular muscles and in modulating the responses within the muscles after strabismus treatment. Our previous studies showed that sustained treatment of rabbit superior rectus muscles with FGF2 resulted in muscles that were smaller and produced less force than controls. In individuals with overacting muscles causing strabismus, sustained FGF2 treatment has the potential to improve alignment by weakening the appropriate muscle.

Conclusions

In summary, similar to satellite cells isolated from leg skeletal muscle, the EECD34 myogenic precursor cells derived from EOM, representing a potentially more multipotent population [18], respond to FGF2 in vitro with increased proliferation. Contrary to many previous studies, both the EECD34 and PAX7 cell populations responded with increased levels of differentiation, as identified with desmin expression, in the presence of added FGF2. However, cell fusion was reduced significantly, suggesting that FGF2 influences the ability of these cells either to move or to recognize each other. Further studies are needed to understand this observation. These studies provide a potential mechanistic explanation for the observed decreases in muscle size and number seen in the extraocular muscles in individuals with Apert syndrome [33], which is characterized by prolonged binding of FGF and activation of the FGFR2 [59].

Supporting information

S1 FigMean percentages of EECD34 (A) and PAX7-positive cells (B) based on the total number of single cells counted during FACS isolation, compared between EOM and TA samples.Each data point represents one FACS sample tube. Significance determined by unpaired t-tests. * indicates significant difference with p < 0.05, ** indicates significant difference with p < 0.01, *** indicates significant difference with p < 0.001. **** indicates significant difference with p < 0.0001.(TIF)

S2 FigEECD34 cells isolated by FACS after 24 hours in vitro labeled with the nuclear stain DAPI (blue).(A) and immunostained for the expression of PITX2 (green) (B) and the merged image (C), and a second example stained for PITX2 and and co-stained with DAPI (D). (E) EECD34 cells after 72 hours in vitro labeled with DAPI and PITX2. Arrow indicates a PITX2-negative cell. Magnification bar indicates 50 µm. (F) Quantification of the percentage of EECD34 cells expressing PITX2 after 24 and 72 hours in vitro. Data represents averages from 3 independent cell isolations for each time point.(TIF)