• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Biol. Author manuscript; available in PMC Nov 15, 2008.
Published in final edited form as:
PMCID: PMC2233612

Myogenic reprogramming of retina-derived cells following their spontaneous fusion with myotubes


Satellite cells are recognized as the main source for myoblasts in postnatal muscle. The possible participation of other cell types in myofiber maintenance remains a subject of debate. Here, we investigated the potential of vascular preparations from mouse retina to undergo myogenesis when cultured alone or with differentiated primary myogenic cultures. The choice of retina, an organ richly supplied with capillary network and anatomically separated from skeletal muscles, ensures that the vasculature preparation is devoid of satellite cells. We demonstrate that retinal-derived cells spontaneously fuse with preexisting myotubes and contribute additional myonuclei, some of which initiate expression of muscle-specific genes after fusion. Myogenic differentiation of retinal cells prior to their fusion with preexisting myotubes was not detected. Although originating from vasculature preparations, nuclei undergoing myogenic reprogramming were contributed by cells that were neither endothelial nor blood borne. Our results suggest smooth muscle/pericytes as the possible source, and that myogenic reprogramming depends on the muscle specific transcription factor MyoD. Our studies provide insights into a novel avenue for myofiber maintenance, relying on nuclei of non-myogenic origin that undergo fusion and subsequent myogenic conversion within host myofibers. This process may support ongoing myofiber maintenance throughout life.

Keywords: Retina, vasculature, endothelial cells, pericytes, skeletal muscle, myogenesis, satellite cells, Sca-1, MyoD, Myf5


Skeletal muscle is composed of myofibers, multinucleated syncytia, established during embryogenesis by fusion of myoblasts. Myoblasts continue to fuse with the growing myofibers during postnatal development until the myofibers reach their mature size. Nuclei within myofibers are postmitotic and normally do not contribute new myonuclei. Injury of adult muscle results in reinitiation of myogenesis to provide myoblasts that fuse with existing myofibers or form new myofibers, depending on the degree of the damage. Myoblasts in postnatal muscle are classically considered to be derived from satellite cells, myogenic progenitors residing beneath the myofiber basal lamina (Zammit et al., 2006). Satellite cells contribute proliferating and differentiating myoblasts and also can self-renew, thus, fulfilling the definition of tissue-specific stem cells (Collins et al., 2005).

Satellite cells and their progeny are characterized by temporal expression of transcription factors associated with different stages of myogenic progression. Quiescent satellite cells commonly express the paired-homeobox transcription factor Pax7, while their proliferating progeny co-express Pax7 and the myogenic determination factor MyoD. The induction of the muscle-specific transcription factor myogenin along with a concomitant decline in Pax7, mark the transition of satellite cell progeny into the differentiation phase and is followed by the expression of muscle structural genes such as sarcomeric myosin and fusion into myotubes (Halevy et al., 2004; Shefer et al., 2006; Zammit et al., 2006). The muscle-regulatory gene Myf5 is also transcriptionally active in satellite cells and their proliferating progeny, and this activity declines following formation of myotubes (Beauchamp et al., 2000; Day et al., 2007; Zammit et al., 2006).

The view that satellite cells are the sole source of myogenic progenitors in adult muscle has been challenged in recent years by the identification of cells with a myogenic potential that are residents of the bone marrow or associated with the vasculature (Asakura et al., 2002; Dellavalle et al., 2007; Ferrari et al., 1998; Gussoni et al., 1999; LaBarge and Blau, 2002; Peault et al., 2007; Sampaolesi et al., 2006). Intramuscular cell populations that express stem cell-associated antigens such as CD34 and Stem Cell Antigen-1 (Sca1), also demonstrated to posses myogenic potential (Asakura et al., 2002; Deasy et al., 2007; Gussoni et al., 1999; Torrente et al., 2001). The significance of atypical myogenic progenitors contributed by the bone marrow and the circulation, has remained controversial and not uniformly demonstrated across different experimental models (Dreyfus et al., 2004; LaBarge and Blau, 2002; Sherwood et al., 2004). Nevertheless, the possibility that microvasculature can produce myogenic cells has gathered continuous momentum, starting with the report that the developing dorsal aorta contains multipotential cells, named mesoangioblasts, that can give rise to a range of mesodermal cell types, including myogenic cells (De Angelis et al., 1999; Minasi et al., 2002; Peault et al., 2007). During postnatal development similar cells with myogenic potential were identified within the microvascular niche, although their numbers drastically reduced with the maturation of the organism (Cusella De Angelis et al., 2003). In addition, cells displaying characteristics of skeletal muscle myoblasts were identified in a range of fetal organs (Gerhart et al., 2001) and in the thymus (Grounds et al., 1992). Vasculature is the common component of all these tissues, and it is thus compelling to propose that the atypical myogenic progenitors could either originate from the circulation or be actual residents of the vasculature wall.

The vasculature wall is comprised of a single layer of endothelium surrounded by contractile cells. The latter have been typically subdivided into smooth muscle cells and pericytes (also known as Mural cells), based on the gradual transition from the larger vessels to the microvasculature (Armulik et al., 2005; Diaz-Flores et al., 1991). We previously demonstrated that vascular-derived smooth muscle cell lines spontaneously initiated MyoD expression and produced progeny that fused into myotubes (Graves and Yablonka-Reuveni, 2000). However, because these cells had been passaged for a long time in culture, it was unclear if our findings were applicable to primary smooth muscle cells. In the present study we aimed to determine the type of microvasculature-associated cells that indeed may contribute to atypical myogenesis, and the involvement of myogenic transcription factors in the process.

Skeletal muscle is especially rich in microvasculature, with an extensive capillary network surrounding individual myofibers (Day et al., 2007). However, isolation of vascular-associated cells from muscle requires a variety of enrichment protocols and the purified cells may still contain subpopulations of residual satellite cells that express unanticipated antigens. For example, the expression of Sca1 has served as a means to enrich for murine hematopoietic stem cells, but additional studies demonstrated that Sca1 is also expressed by non-hematopoietic progenitors and endothelial cells (Day et al., 2007; Holmes and Stanford, 2007; van de Rijn et al., 1989). Moreover, rare Sca1-expressing cells, identified in association with myofibers, may in fact represent a subpopulation of activated satellite cells (Mitchell et al., 2005). This is supported by the observation that progeny of bona fide satellite cells do express Sca1 (Mitchell et al., 2005; current study). Thus, Sca1+ cells isolated from skeletal muscle and able to contribute to muscle regeneration (Asakura et al., 2002; Gussoni et al., 1999; Torrente et al., 2001) did not necessarily represent non-satellite progenitors. Earlier studies also identified other endothelial cell markers that were expressed by bona fide satellite cells (De Angelis et al., 1999). CD34 is another example of a common stem cell antigen expressed by satellite cells (Beauchamp et al., 2000), which was originally used to identify other cell types in the muscle. In the same vein, the so-called myoendothelial cells expressing both myogenic markers and the endothelial marker CD31, detected in human fetal muscle tissue (Cerletti et al., 2006), may represent a transitory compartment of typical myoblasts.

To bypass possible contribution of bona fide satellite cells to vascular cell preparations, the present study focused on cells isolated from murine retina – an organ richly supplied with capillary network and anatomically separated from skeletal muscle. Preparations of enriched vascular cells from the retina never displayed any characteristics of conventional myogenic progenitors when cultured alone or with host primary myogenic cultures. However, when cultured onto primary myogenic cultures following the initial development of myotubes, retina-derived cells spontaneously fused with host myotubes and some of the contributed donor nuclei expressed skeletal myogenic genes. These co-culture studies provide novel insights into the process of myogenic reprogramming and raise the possibility that in addition to satellite cell contribution during muscle repair, myofiber nuclei may be maintained by a non-canonical pathway that involves direct fusion of non-myogenic cells, residents of the muscle vasculature.

Materials and Methods


Animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee at the University of Washington. Chickens (White Leghorn, 8- to 9-day old) were kindly provided by Drs. O. Bermingham-McDonogh and E. Rubel (University of Washington). Mice were from colonies maintained at the University of Washington and were housed in micro-isolator cages in a pathogen-free facility under 12-hour light/dark cycle and were fed ad libidum Lab Diet 5053 (Purina Mills). Unless otherwise mentioned, mice (males and females) used to prepare primary myogenic cultures were 1.5–4 month-old and those used to prepare retinal cells were 1–2 month-old. Wild type mice of C57BL/6 background were used routinely except when the use of BALB/c mice was specifically indicated. Additionally, the following genetically manipulated mouse strains were used:

  • eGFP transgenic, C57BL/6 (C57BL/6-Tg(ACTB-EGFP)1Osb/J; Jackson Laboratory (Okabe et al., 1997); enhanced green fluorescence protein is expressed under the control of the chicken β-actin promoter and CMV enhancer in all tissues with the exception of erythrocytes and hair.
  • MLC3F-nLacZ transgenic, C57BL/6 (Beauchamp et al., 2000; Kelly et al., 1995); developed by Drs. R. Kelly and M. Buckingham (Pasteur Institute). In these mice regulatory elements of muscle specific myosin light chain 3F (MLC3F) drive LacZ expression in myofiber nuclei. Breeding pairs were kindly provided by Dr. M. Goodell (Baylor College of Medicine).
  • Myf5nLacZ/+, (Beauchamp et al., 2000; Tajbakhsh et al., 1996); developed by Drs. S. Tajbakhsh and M. Buckingham (Pasteur Institute). In these mice LacZ was knocked into the Myf5 gene and its expression reports Myf5 promoter activity in nuclei of satellite cells. Breeding pairs on enriched BALB/c background were kindly provided by Dr. M. Rudnicki (Ottawa Health Research Institute). Mice used in the present study were either on BALB/c background or crossed with C57BL/6 mice without any apparent effect on the results.
  • XLacZ4, C57BL/6 (Armulik et al., 2005; Klinghoffer et al., 2001; Tidhar et al., 2001) were used to trace vascular smooth muscle/pericytes. In this promoter trap mouse it is unknown what elements drive LacZ expression. Mice were developed in the laboratory of Dr. M. Shani (Volcani Center); breeding pairs were kindly provided by Dr. P. Soriano (Fred Hutchinson Cancer Research Center).
  • Sca1-GFP, transgenic, C57BL/10 x CBA (Ma et al., 2002; Mitchell et al., 2005) have been used in our studies to trace endothelial cells (Day et al., 2007). Mice were developed by Dr. E. Dzierzak (Erasmus University); breeding pairs were kindly provided by Dr. G. Pavlath (Emory University).
  • MyoD-GFP transgenic, homozygous, enriched FVB; GFP expression is driven by the same MyoD regulatory elements described previously for the generation of MyoD-LacZ transgenic mice in Dr. D. Goldhamer’s laboratory (Chen et al., 2001). To maintain sufficient GFP signal for detection in live cells in culture, only homozygous mice were used in the present study.
  • MyoD−/−, enriched BALB/c (Rudnicki et al., 1992; Yablonka-Reuveni et al., 1999); breeding pairs were kindly provided Dr. M. Rudnicki.

Cell culture media, plates and coating matrix

DMEM consisted of Dulbecco’s modified Eagle’s medium (high glucose, with L-glutamine, 110 mg/l sodium pyruvate, and pyridoxine hydrochloride) supplemented with antibiotics (50 U/ml penicillin and 50 mg/ml streptomycin); all from GIBCO-Invitrogen. MEM consisted of Minimal essential medium (GIBCO-Invitrogen) supplemented with 1% nonessential amino acids (Cellgro by Mediatech) and antibiotics as in DMEM. Medium supporting endothelial growth (Su et al., 2003) consisted of DMEM containing 20% fetal bovine serum (FBS, Sigma), 20mM HEPES, 1% nonessential amino acids, heparin (55 U/mL, Sigma), endothelial growth supplement (100 μg/mL, Sigma-Aldrich). Medium for mouse myogenic cultures consisted of DMEM containing 20% FBS, 10% horse serum (HS, HyClone) and 1% chicken embryo extract (Shefer et al., 2006). Medium for chicken myogenic cultures consisted of MEM containing 10% horse serum and 5% chicken embryo extract (Halevy et al., 2004). All types of cultures (apart from those for fluorescence in situ hybridization, detailed below) were maintained in standard 24-well Falcon plates (BD Biosciences). Culture wells were coated with Matrigel (growth factor reduced, BD Biosciences, diluted 1:10 in DMEM) as previously described (Shefer et al., 2006).

Isolation and culturing microvascular cells from mouse retina

Retinal microvasculature preparations were obtained from enucleated mouse eyes following previously described procedures (Gitlin and D’Amore, 1983; Su et al., 2003) with some modifications. Eyes were placed in DMEM, fixed from the side of the optic nerve using a fine-tipped forceps and the cornea was bisected with micro scissors and removed. The lens was then removed and the retina was carefully pulled out with the forceps, separated from the ciliary body, cleaned from the pieces of pigmented epithelium and rinsed extensively with DMEM to minimize any possible debris. Control studies (detailed in Results) demonstrated that the preparation was free of any residual extraocular muscle (EOM) fragments. Retinas from 2–4 mice were pooled together, cut into small pieces and subjected to enzymatic digestion for 2 hrs at 37°C in a solution containing collagenase-dispase (1 mg/ml, Roche) and DNase type I (0.5 mg/ml, Sigma) to release single cells. Enzymes were dissolved in DMEM containing 2%FBS; at this level FBS did not inhibit enzyme activity, but increased viability of cells during dissociation. The resulting suspension was gently triturated with a 21g needle to disengage cells from the digested tissue. The suspension was filtered through a 40 μm nylon mesh (Falcon) to remove undigested material. FBS was added to the suspension to the final concentration of 10%. The filtered material was subjected to low speed centrifugation (800–1000g for 10 min). The final pellet was washed once with DMEM/10%FBS and re-suspended in the medium for cell growth. Retinal cell preparations were eventually dispensed into 24-well plates; each well received cell equivalence of 2 retinas in 0.5 ml medium. In experiments where retinal cells were seeded into primary myogenic cultures, the final cell pellet was suspended in the complete myogenic medium.

The aforementioned retinal digestion procedure minimized survival of neuronal cells. Only about 1% of the total cells recovered by the isolation procedure survived and adhered within the initial days in culture. The majority of these cells appeared to be vascular cells as revealed by immunostaining (more details in Results). The remaining 99% of the cells in the final digest, which were of neuronal origin, were dead or damaged. They did not adhere to the culture plate and were typically removed two days after culturing, upon the first medium change.

In some studies, retinal cells from GFP-expressing mice were sorted by FACS into GFP+ and GFP populations. Digested retinal cells were re-suspended in DMEM/2%FBS and subjected to FACS using an Influx Cell Sorter (Cytopeia Incorporated) with UV (351–364nm) and 488 nm argon lasers as previously described (Day et al., 2007). Gating of GFP positive events was set to at least 10 times the fluorescence intensity of negative events. Sorted cells were collected directly into rich growth medium for primary myogenic cultures, harvested by low speed centrifugation and further analyzed in cell culture. The percentage of GFP+ sorting events depended on the mouse model and was below 1% of total events (specific details are provided in Results). The percentage of GFP sorting events was ~30% and the remaining ~70% events accounted for debris that appeared to be contributed by dead retinal neurons. Also, within the ~30% GFP population, the majority of cells did not adhere to the culture plate and died during the initial two days in culture, appearing to represent additional damaged neuronal cells.

Mouse and chicken myogenic cultures

Unless otherwise noted, mouse primary myogenic cultures were prepared from the pooled hindlimb muscles tibialis anterior, extensor digitorum longus, and gastrocnemius by digestion with 0.1% Pronase (Calbiochem) as previously described (Shefer et al., 2006). Cells were cultured into 24-well plates coated with Matrigel at a density of 2×104 per well using our standard growth medium for mouse myogenic cultures detailed above. Single myofibers were isolated from hindlimb EDL muscles of adult mice and cultured individually as previously described (Shefer et al., 2006). When primary myogenic cultures were prepared from extraocular muscles, muscles were harvested from enucleated eyes and cultures were prepared as described for limb muscles. Chicken primary myogenic cultures were prepared from breast muscle according to our previous publication (Halevy et al., 2004) and cultured into Matrigel-coated wells, using standard growth medium for chicken myogenic cultures.

Isolation and culture of bone marrow-derived and blood borne cells

Bone marrow was isolated from tibia bones as described previously (Goodell et al., 1996) and single cell suspensions were obtained by trituration with a 22g needle. Blood was drawn from the retro-orbital plexus with a heparinized capillary and depleted from erythrocyte via hypotonic shock with distilled water (Terstappen et al., 1989). Cells were washed by low-speed centrifugation and used for further analysis in co-cultures with primary myoblasts.

Immunofluorescent staining

Cultures were fixed with paraformaldehyde, permeabilized with Triton-X100 (permeabilization was omitted when labeling for Sca1), blocked with normal goat serum, immunolabeled, and counterstained with DAPI as previously described (Shefer et al., 2006). Cultures were analyzed by single or double immunofluorescence using mouse, rat and rabbit primary antibodies followed by species-specific Alexa Fluor 488 and 568 species-specific antibodies (Molecular Probes, 1:1000).

Primary antibodies were against the following antigens:

  • Pax7 (mouse IgG1, ascites fluid, Developmental Studies Hybridoma Bank [DSHB], 1:2000 dilution)
  • MyoD (mouse IgG1, clone 5.8A, BD Biosciences, 1:800)
  • myogenin (mouse IgG1, clone F5D, hybridoma supernatant, DSHB, 1:2)
  • sarcomeric myosin heavy chain (MF20, mouse hybridoma supernatant, DHSB, 1:2)
  • beta-galactosidase (β-gal) (JIE7, mouse hybridoma supernatant, DSHB, 1:16)
  • GFP (rabbit, Novus Biologicals, 1:10,000)
  • Pax6 (mouse IgG1, DHSB, 1:2; provided by T. Reh, University of Washington).
  • Sca1 (rat IgG, Pharmingen, 1:100; provided by J. Chamberlain, University of Washington)
  • CD31 (rat IgG, Pharmingen, 1:100)
  • Tie2 (mouse IgG1a, Pharmingen, 1:50)

. Biotinylated isolectin-B4 (Vector Laboratories) was also used for endothelial cell identification as previously detailed (Day et al., 2007).

Histochemical detection of β-galactosidase activity

β-gal activity was detected by X-Gal staining (Beauchamp et al., 2000) omitting detergents from the staining solution. Cultures from LacZ-expressing mice were fixed with 2% paraformaldehyde for 10 min, rinsed with PBS, incubated in X-Gal solution for 2–16 hours at 37°C and then rinsed in PBS. When cultures were also analyzed for GFP expression, GFP detection was enhanced with anti-GFP antibody since X-Gal staining often resulted in reduced GFP fluorescence intensity.

Y-chromosome in situ detection

Y-chromosome FISH was performed as previously described (Gussoni et al., 1999; Muskiewicz et al., 2005) to detect male retinal donor nuclei when co-cultured with primary myogenic cells of female origin. Co-cultures were generated in 8-well chamber glass slides Labtech II (Nalgene Nunc) precoated with diluted Matrigel as described above. Positive and negative control myogenic cultures from males and females, respectively, were grown and processed in parallel with the co-cultures. Cultures were fixed in Histochoice tissue fixative (Amresco) for 30 minutes, rinsed in PBS, dehydrated through graded alcohols (50%, 70%, 90%, 100%, 5 minutes each) and stored at −20°C. Y-chromosome in situ hybridization was detected with sheep anti-digoxigenin-rhodamine (Roche Applied Science; diluted 1:100) and slides were mounted with Vectashield containing DAPI.

Microscope and imaging system

Y-chromosome FISH was analyzed with a Nikon Eclipse E1000 equipped with a Hamamatsu Orca ER digital camera and images were acquired using Openlab software (Improvision). In studies where β-gal+ cells were identified by X-Gal reaction, images were acquired with Nikon Coolpix 4500 camera. All other observations were made with an inverted fluorescent microscope (Nikon Eclipse, TE2000-S). Images were acquired with a Qimaging Retiga 1300i Fast 1394 monochrome CCD camera. The CCD camera drive and color acquisition were controlled by MetaView Imaging System (Universal Imaging Corporation). Composites of digitized images were created using Adobe Photoshop software.


Cells present in retinal vasculature preparations can fuse with myotubes

To isolate cells of the retinal vascular wall we modified a published collagenase digestion method (Gitlin and D’Amore, 1983; Su et al., 2003) and used a combination of collagenase, dispase and DNase (as described in Materials and Methods). These enzymes are used to release microvascular cells from the brain (Song and Pachter, 2003) and do not support survival of mature neuronal cells of the retina (Shen et al., 1995). As shown below, the dissociated retinas yielded cell cultures comprised predominantly of endothelial and smooth muscle cells.

These retinal vasculature preparations, referred to in the present study as “retinal preparations” or “retinal cells”, were investigated for their capacity to give rise to myogenic cells. Retinal cells were first cultured alone in media supporting endothelial or myogenic cell growth or in conditioned media from primary myogenic cultures. Cell morphology and growth characteristics varied under the different medium conditions, but none of the cells in these cultures were confirmed to be myogenic. This conclusion was based on immunostaining analyses, demonstrating the absence of cells expressing characteristic transcription factors of the myogenic lineage (i.e., Pax7, MyoD and myogenin) and the absence of mono- or multi-nucleated cells expressing sarcomeric myosin, while parallel control primary myogenic cultures contained positive cells (data now shown).

We then asked if the retinal preparations can produce myogenic cells upon their exposure to a myogenic niche. The retinal cells were cultured onto primary myogenic cultures that were isolated from hindlimb muscles of wildtype (C57BL/6) mice and grown in Matrigel-coated 24-well plates. Each co-culture was established in 3–5 parallel wells with retinal preparations equivalent to 1–2 retinas per well. To distinguish the retinal “donor” cells and the “host” myogenic cells in the co-cultures, retinal preparations were isolated from eGFP mice. Retinal cells were added to the myogenic cultures either before myotubes were formed (semi-confluent culture, day 4) or following the initial development of myotubes (confluent culture, day 7). Co-cultures were then followed for an additional 7–12 days. In both cases, the co-cultures demonstrated individual eGFP+ cells of various morphologies, some of which showed clonal growth with time. The majority of these eGFP+ cells displayed morphological characteristics of endothelial cells and formed endothelial tubes, which expressed CD31 (Figs. 1A–A″) and Tie2 (not shown), and displayed endothelium-characteristic binding of isolectin-B4 (Day et al., 2007). None of the mononuclear eGFP+ cells expressed myogenic markers such as Pax7, MyoD, myogenin or sarcomeric myosin, while the “host” myogenic cells (i.e., eGFP cells) in the same co-cultures expressed these proteins (e.g., Figs. 1B′ and C′).

Fig. 1
Retinal cells from eGFP mouse co-cultured onto a day 7 primary myogenic culture with developing myotubes. Endothelial cells represent the predominant cell type in retinal-derived cells, as revealed by CD31 immunolabeling; endothelial tube networks develop ...

Myotubes expressing eGFP were also detected in the co-cultures (Figs. 1B–C″), but such myotubes were found only when the retinal cells were added to host cultures that begun forming myotubes (i.e., when host myogenic cultures were 7-day old, but not when host cultures were 4-day old and lacked myotubes). Each co-culture displayed 1–2 eGFP+ myotubes within the initial 1–2 days of co-culture. The number of eGFP+ myotubes gradually increased during the next several days, culminating in 20–50 visible eGFP+ myotubes per well within the 12 days of co-culturing. We report only on the absolute numbers of eGFP+ myotubes per well arising from 1–2 retinas, as it has been impossible to accurately quantify the ratio of eGFP+ myotubes to total myotube number in the elaborate myotube networks. eGFP+ myotubes were indistinguishable from the rest of the myotubes present in the co-cultures with regard to morphology, expression of muscle specific markers such as sarcomeric myosin and MyoD (Figs. 1B–C″), and ability to contract spontaneously (data not shown).

Notably, the finding that the addition of retinal cells to myogenic cultures prior to formation of myotubes (i.e., when host cultures were 4 days old) did not yield eGFP+ mononucleated cells expressing myogenic traits indicated that the retinal derived cells do not enter myogenic differentiation on their own. Also, the addition of eGFP+ retinal cells when myotubes were “older” (i.e., when host myogenic cultures were 9- to 12-day old) did not yield eGFP+ myotubes, as determined by direct fluorescent analysis or with an antibody against GFP (not shown). This observation indicates that fusion of retinal cells with myotubes can take place only during a specific time window, potentially limited to the earlier phase of myotube development.

Fusing retinal cells contribute donor nuclei to host myotubes

The observed eGFP expression by host myotubes could be due to cytoplasmic contribution of retinal cells or could reflect nuclear contribution. To gain insight into this aspect, retinal cells were harvested from eGFP male mice and cultured onto a day 7 primary myogenic cultures isolated from wildtype female muscles. Fluorescence in situ hybridization (FISH) with a mouse-specific Y-chromosome probe was then performed and confirmed the presence of 1–2 male retinal nuclei in host myotubes containing a total of 10–50 myonuclei (Figs. 2A–C). Thus, this Y-chromosome analysis indicated that donor retinal cells fused with the host-derived myotubes, but did not fuse with each other to form de novo myotubes, as evident by the absence of myotubes containing only male nuclei. Notably, the original eGFP signal was drastically reduced and often fully lost following the processing of the cultures for FISH analysis. Also, a non-specific general fluorescent background was generated by the processing. Therefore, we also correlated eGFP expression in myotubes with the presence of Y-chromosome by marking areas with eGFP+ myotubes before processing.

Fig. 2
FISH Y-chromosome analysis of retinal cells from male eGFP mouse in host myotubes from wildtype female mouse following co-culturing. Three representative fields are shown (A–C). Typically, 1–2 donor retinal nuclei (pointed to with arrows) ...

We ruled out the possibility that the fusing retinal cells could have potentially derived from residual satellite cells originating from the extraocular muscles (EOM) during the isolation of the retina. First, microscopy examination did not reveal any obvious muscle fragments in isolated retinas. Second, there was no evidence for muscle-specific LacZ reporter expression in retinas isolated from Myf5nLacZ/+ or MLC3F-nLacZ mice, in contrast to extraocular muscle, which was positive for both markers (data not shown). Myf5- and MLC3F-driven LacZ expression identifies satellite cells and myofiber nuclei, respectively, as previously demonstrated in hindlimb muscles (Beauchamp et al., 2000; Zammit et al., 2006).

Fusing retinal cells from MLC3F-nLacZ transgenic mice initiate reporter expression in host myotubes

Following the demonstration that fusing retinal cells contributed nuclei to myotubes, we asked if such donor nuclei were induced to express myogenic genes by the myotube environment. Using the same co-culturing assay system described above, we demonstrated that fusing retinal cells from MLC3F-nLacZ mice indeed were able to activate LacZ expression (Fig. 3). In standard primary myogenic cultures from hindlimb or EOM muscles of MLC3F-nLacZ mice, expression of β-gal was observed first in differentiated myoblasts, starting on day 3, followed by expression in myotubes that were first seen on day 5 (Figs. 3A–D). In contrast, in the retina-hindlimb co-cultures β-gal expression was seen only in myotubes (Figs. 3E–F) and only in 5–10% of myonuclei within a myotube. This observation indicates direct fusion of donor cells with host myotubes, rather than de novo myotube formation. Furthermore, the intensity of the β-gal signal in retina-hindlimb co-cultures was far reduced compared to that seen in myotubes developed in primary myogenic cultures. An overnight β-gal reaction was required to observe positive nuclei in host myotubes compared to just a brief development that was needed for observing β-gal+ nuclei in standard primary cultures from MLC3F-nLacZ mice. As an additional control, myoblasts were isolated from hindlimb muscles of MLC3F-nLacZ mice and co-cultured at a low density (similar to the density of donor retinal cells) onto a host primary culture. As expected, MLC3F-nLacZ myoblasts fused with host wildtype myotubes to form hybrid myotubes with some β-gal+ nuclei and additionally formed de novo myotubes with all myonuclei exhibiting strong β-gal signal (Figs. 3G and H).

Fig. 3
Transgenic MLC3F promoter is activated in retinal-derived cells isolated from MLC3F-nLacZ mice upon their fusion with myotubes. Primary myogenic cultures derived from the hindlimb (A, B) or extraocular muscles (C, D) of MLC3F-nLacZ mouse exhibit MLC3F ...

We further analyzed the fusion of retinal cells from double transgenic eGFP/MLC3F-nLacZ mice with host myotubes (Fig. 4). The use of the double transgenic mouse permitted tracking live cultures for the appearance of eGFP+ myotubes before fixing the cultures forβ-gal detection. In agreement with studies described in Fig. 1, eGFP+ myotubes first appeared on days 1–2 of co-culture, and their numbers increased thereafter. Co-cultures fixed on days 3–12 demonstrated β-gal+ nuclei only in myotubes, while eGFP+ mononucleated cells were always negative for β-gal. The eGFP signal was reduced after β-gal staining, especially in β-gal+ myotubes. Therefore, to enhance the sensitivity of eGFP detection, β-gal stained cultures were immunolabeled with an anti-GFP antibody. This immunolabeling step showed that not all eGFP+ myotubes contained β-gal+ nuclei (Fig. 4; depicts three representative fields in a day 5 co-culture). The number of myotubes containing β-gal+ nuclei and the intensity of β-gal staining gradually increased at later time points in the co-cultures. For example, in a typical study the number of myotubes containing β-gal+ nuclei was 14, 37, and 64 (per well) on co-culture days 5, 9, and 12, respectively. While the number of β-gal+ nuclei increased in more advanced co-cultures, the total number of eGFP+ myotubes (detected by immunostaining) was always much higher than the number of myotubes with β-gal+ nuclei. Since β-gal is being produced in the cytoplasmic compartment of the myotube and subsequently may enter one or more myonuclei (possibly of both donor and host origin), there may be a need for a critical level of LacZ to be produced before it can be detected based on enzymatic reaction. It is also possible that more than one cell type from the retinal preparations can fuse with host myotubes (resulting in all cases in eGFP+ myotubes), but only a specific cell type can undergo myogenic reprogramming. Indeed, the latter possibility has been confirmed by us as described below based on retinal preparations from Sca1-GFP mice.

Fig. 4
Three representative fields in a co-culture or retina cells from double-transgenic eGFP/MLC3F-nLacZ mice and wildtype primary myogenic cultures. Panels depict co-localization of fusion (GFP+ myotubes) and myogenic reprogramming (β-gal+ nuclei) ...

Retinal endothelial cells fuse with myotubes but do not undergo myogenic reprogramming

To specifically investigate endothelial contribution to fusion and reprogramming we analyzed retinal preparations from Sca1-GFP transgenic mice. Sca1 is a recognized marker of endothelial cells (van de Rijn et al., 1989). We previously showed strong GFP expression driven by the Sca1 promoter in the vasculature of skeletal muscle from Sca1-GFP transgenic mice (Day et al., 2007). Here we demonstrate that Sca1-GFP transgene is strongly expressed in the vasculature of the retina (Fig. 5A). Retinal cells from Sca1-GFP mice contributed GFP expression to host myotubes when co-cultured with primary myogenic cells (Figs. 5B–D). The first GFP+ myotubes were observed on day 1 of co-culture, increasing up to 10–20 myotubes per well by day 3. Single GFP+ cells were also observed in these co-cultures (Figs. 5B′) and their endothelial cell identity was confirmed based on immunostaining with an antibody against CD31 (Figs. 5C–C′″). Additionally, these retinal-derived Sca1-GFP+ cells reacted with anti-Tie2 antibody and bound isolectin-B4, recognized endothelial cell markers (data not shown). Immunostaining with anti-Sca1 antibody confirmed co-localization of the Sca1-GFP transgene with endogenous Sca1 expression in these GFP+ cells (Fig. 5D). This immunostaining analysis also showed that many of the host-derived myoblasts in the co-culture also expressed the Sca1 antigen, but the myotubes were negative for Sca1 (Fig. 5D). After 3–5 days of co-culture, GFP was not detectable any more in myotubes, indicating that Sca1-driven GFP production could not be supported by the transcriptional milieu of the myotubes.

Fig. 5
GFP+ cells in retinal preparation from Sca1-GFP mice can fuse with host myotubes but do not undergo myogenic reprogramming. (A) Live retina displays Sca-1-GFP transgene expression in the vascular bed, including larger vessels and capillaries. (B–D) ...

We further assessed participation of the retinal Sca1+ cell population in nuclear reprogramming by establishing co-cultures of retinal cells from double transgenic mice Sca1-GFP/MLC3F-LacZ and primary myogenic cultures. Myotubes that were initially positive for GFP no longer expressed GFP by day 5 when β-gal+ nuclei appeared in the co-cultures, and it was therefore impossible to determine if GFP and β-gal signals co-localized within myotubes. To circumvent this difficulty, retinal cells from double transgenic Sca1-GFP/MLC3F-nLacZ mouse were first sorted by FACS into GFP+ and GFP populations (1.5×103 and 1.9×106 events, respectively). Each population was then separately co-cultured with primary myogenic cells and analyzed for contribution of lacZ signal in myotubes. Unexpectedly, β-gal+ myonuclei were detected only in the co-cultures of Sca1-GFP retinal cells but not in the Sca1-GFP+ population (Fig. 5E and F), although GFP positive myotubes were observed in the latter co-culture. These results indicated that the Sca1-GFP+ (endothelial) cells from the retinal vascular bed could participate in fusion with myotubes, but the population of cells capable of myogenic reprogramming following fusion was not contained within the Sca1-GFP+ cell fraction. Thus, the capacity to undergo myogenic reprogramming was restricted to a second retinal-derived cell type that was fusing with host myotubes. The study described next aimed to determine if the second cell type that was able to fuse with myotubes was contained within the smooth muscle component of the vasculature.

Retinal-derived smooth muscle/pericytes contribute nuclei to host myotubes

Contractile cells engulfing the endothelium have been subdivided into smooth muscle and pericytes. The retinal vascular network consists mainly of microvasculature and the associated contractile cells have been considered to be pericytes (Armulik et al., 2005; Klinghoffer et al., 2001). To analyze possible presence of pericytes in our retinal preparations and their possible fusion with host myotubes, we utilized the XLacZ4 mouse in which the retinal smooth muscle cells and pericytes were shown to express β-gal in their nuclei (Armulik et al., 2005; Klinghoffer et al., 2001). Donor retinal cells from XLacZ4 mice were co-cultured with primary myogenic cultures, fixed at different time points and reacted with X-Gal (Fig. 6). Cultures displayed β-gal+ mononucleated cells, appearing alone or in groups, in each observed arbitrary field (Fig. 6A and B). Myotubes containing one, and in rare cases two, β-gal+ nuclei were also identified in the co-cultures (Fig. 6C–E). On average, 15–20 myotubes with β-gal+ nuclei were found in each well on the initial two days of co-culture. β-gal+ nuclei inside myotubes comprised 5–10% of all β-gal+ nuclei during the initial two days. The average number of myotubes containing β-gal+ nuclei declined to less than 5 per well by day 3. We suggest that this decline in XLacZ4 transgene expression in myotubes is likely due to the termination of transgene expression in the myotube environment.

Fig. 6
Retina-derived XlacZ4 expression in host myotubes. Pericytes/smooth muscle cells from the retina of Xlacz4 mouse contribute mononucleated cells (single cells and cell clusters) (A and B) and also fuse with myotubes (C–E) upon co-culturing of retinal ...

We also analyzed co-cultures of retinal cells derived from XLacZ4/Sca1-GFP double transgenic mice. This analysis demonstrated that expression of GFP and β-gal was in separate cells (Fig. 6F). In view of the aforementioned observation that Sca1-GFP+ cells fused with myotubes, but did not contribute to myogenic reprogramming (based on MLC3F-nLacZ expression), and the finding that retinal cells derived from XLacZ4 mice fused with myotubes and did not express Sca1-driven GFP, we propose that the pericytes are in fact the cells that contribute to reprogramming upon fusion with myotubes.

Circulating cells do not contribute to the pool of retinal cells that can fuse with host myotubes

We evaluated the possibility that blood borne cells may fuse with host myotubes. This analysis was important in view of the published reports implicated circulating and bone marrow derived cells in fusion with myotubes (detailed in Introduction) and the fact that the presence of residual blood cells in retinal preparations could not be avoided. Nucleated blood cells obtained from the retro-orbital plexus of double transgenic eGFP/MLC3F-nLacz or Sca1-GFP mice were co-cultured with host primary myoblasts in the same manner as the retinal preparations, applying 105 donor cells per well. Neither GFP+ myotubes nor β-gal+ nuclei were detected. Furthermore, co-cultures of donor-derived whole bone marrow from double transgenic eGFP/MLC3F-nLacZ or Sca1-GFP mice (106 cells/well) and host myogenic cultures developed very rare eGFP+ myotubes (1–2 versus 30–50 in the retina co-cultures), but β-gal+ nuclei were not detected (data not shown).

Fusing retinal cells initiate MyoD, but not Myf5, gene expression

In the studies discussed above, myogenic reprogramming of fusing retinal cells was determined by initiation of MLC3F-nLacZ expression, an efficient transgenic marker of myogenic differentiation. During myogenesis of satellite cells, MyoD and Myf5 are expressed before differentiation and upregulation of MLC3F-nLacZ transgene (Zammit et al., 2006). Moreover, MyoD and Myf5 are involved in the induction of the myogenic lineage during embryogenesis (Ludolph and Konieczny, 1995). Hence, we were interested to determine if reprogramming of fusing retinal cells involved expression of the latter two myogenic regulatory factors.

To investigate possible MyoD expression, we asked if donor retinal cells from a MyoD-GFP transgenic mouse could activate transgene expression upon their fusion with host myotubes. In this mouse, MyoD promoter/enhancer elements drive cytoplasmic GFP expression in myofibers. We first established that the reporter was expressed in myogenic cultures from MyoD-GFP mice. In both single myofiber cultures and primary myogenic cultures, all myotubes and some of the mononucleated cells expressed MyoD-GFP (Figs. 7A–B′; depicted images are of live cultures). In addition, when cells were cultured at clonal density, only myogenic clones expressed MyoD-GFP while transgene expression was absent in non-myogenic clones (not shown). These observations indicated that MyoD-GFP expression was specific to the myogenic lineage, permitting us to utilize this model to investigate myogenic reprogramming of retinal cells.

Fig. 7
MyoD-GFP transgene expression in primary myogenic cultures and in retina-hindlimb co-cultures. MyoD-GFP is expressed by some mononucleated cells and by all myotubes in a single myofiber culture (A and A′) and in a primary myogenic culture from ...

Co-cultures of retinal cells from MyoD-GFP mice onto primary myogenic cultures displayed about 20–50 GFP+ myotubes per well by day 12 of co-culture (Figs. 7C–D′). MyoD-GFP signal in host myotubes was much weaker than that seen in myotubes formed in primary myogenic cultures isolated from MyoD-GFP mice. While some of the MyoD-GFP+ myotubes in the retina-hindlimb co-cultures could be observed directly based on green fluorescence, the majority of the MyoD-GFP+ myotubes in the co-cultures were detected only after immunostaining with anti-GFP antibody. The low level cytoplasmic GFP expression following fusion of donor retinal cells from MyoD-GFP mice prevented additional studies on possible co-expression of MyoD-GFP and MLC3F-nLacZ (or XLacZ4) in fusing retinal cells from double transgenic mice. To ensure sufficient donor-based MyoD-GFP signal in host myotubes, we had to utilize only homozygous MyoD-GFP mice as the retinal source.

Notably, we detected rare weakly GFP+ mononucleated cells during enzymatic digestion of retinas of MyoD-GFP mice and during the initial co-culture days. These GFPlow cells were negative for MyoD or Pax7, and positive for the neuronal marker Pax6 (Hitchcock et al., 1996; Klassen et al., 2004) based on immunostaining. It is possible that some of the regulatory elements of the MyoD construct are responsive to the transcriptional milieu in the rare sub-population of cells that display such apparently “ectopic” MyoD-GFP expression. To exclude participation of these MyoD-GFPlow cells in formation of GFP+ myotubes, freshly isolated retinal cell preparations from MyoD-GFP mice were sorted into GFP+ (0.009% of the total sorted events) and GFP cells and co-cultured with host myogenic cultures. Only the GFP cell population contributed to the development of MyoD-GFP+ myotubes (data not shown).

The detection of MyoD-GFP reporter expression by host myotubes following co-culturing with retinal cells suggested that donor nuclei can initiate MyoD promoter function. We further demonstrated that fusing retinal cells from mouse could contribute endogenous mouse MyoD expression in host myotubes when using chicken muscle for the primary myogenic cultures and employing mouse-specific antibody against MyoD (Fig. 8). The appearance of eGFP+ myotubes, starting on day 1 of co-culture indicated that retinal cells of mouse origin fused with the host chicken myotubes. The mouse monoclonal antibody against rodent MyoD labeled several adjacent myonuclei only within eGFP+ myotubes, but did not recognize the host endogenous chicken MyoD, (Figs. 8C–C′″). In contrast, the rabbit anti-chicken MyoD labeled strongly nearly all nuclei in chicken myogenic cultures (Figs. 8B and B′). While the first eGFP+ myotubes were observed after one day of co-culture and their number increased in subsequent days, nuclei expressing mouse MyoD were first seen by day 6 of co-culture and were only in some of the eGFP+ myotubes. We reproducibly detected 3–4 eGFP+ myotubes per well with several adjacent nuclei positive for mouse MyoD on days 6, 7 and 9 of co-culturing at the time that the total number of eGFP+ host chicken myotubes was about 50 per well. The finding that only some of the eGFP+ myotubes displayed nuclear expression of mouse MyoD is compatible with the observed reduced number of host mouse myotubes containing MLC3F-nLacZ nuclei compared to total eGFP+ myotubes in the co-cultures (Fig. 4).

Fig. 8
Fusion with myotubes and myogenic reprogramming of donor retinal cells from eGFP mouse upon their co-culturing with host chicken primary myogenic cultures. Rabbit antibody against chicken MyoD detects positive nuclei of myoblasts and myotubes in primary ...

Notably, we typically detected a higher number of MyoD+ nuclei of mouse origin in host chicken myotubes, or MLC3F-driven LacZ+ myonuclei in host mouse myotubes (discussed above), compared to the 1–2 fusing cells identified with the Y-chromosome probe. This apparent discrepancy in fact indicates that the myogenic-specific nuclear markers MyoD and MLC3F-driven LacZ also enter nuclei adjacent to the original donor-derived nucleus. Furthermore, since MyoD and MLC3F-driven LacZ are made in the cytoplasmic compartment of the myotube and subsequently may enter one or more myonuclei (of both donor and host origin), there might be a need for a critical level of protein to be produced before it could be detected in host myotubes. Our observed translocation of donor contributed nuclear markers MyoD and nLacZ to host-derived myonuclei is in accordance with previous observations made with nLacZ-expressing donor myoblasts, where a higher number of LacZ-expressing nuclei was detected compared to the true number of donor cells fused with host myofibers (Yang et al., 1997). Such translocation of a nuclear protein likely requires strong and continuous expression of the donor contributed gene product. Differently, the aforementioned nLacZ expression contributed by fusing pericytes was eventually shut off in host myotubes and translocation was not observed.

We also asked if retinal cells from Myf5nLacZ/+ mice could contribute β-gal+ myofiber nuclei when co-cultured with mouse primary myogenic cultures. For live distinction of myotubes with fused retinal cells, we used retinas from Myf5nLacZ/+ mice that were crossed with eGFP mice. We did not detect any β-gal+ nuclei within myofibers in the co-cultures. The results were consistent when detection was performed either by immunostaining with an antibody against β-gal or by X-Gal reaction. In contrast, myofiber nuclei of primary myogenic cultures from Myf5nLacZ/+ mice displayed β-gal label, although the signal was weaker than in myoblasts and not uniformly maintained in all myotubes (data not shown; summarized schematically in Zammit et al., 2006). Myonuclei in some of the myotubes present in culture day 8 were β-gal+ but such positive myonuclei were no longer visible on culture day 11. We suggest that Myf5-driven nLacZ expression is initiated at the myoblast level and that the host myotube environment may not be able to promote initiation of Myf5 promoter activity by the fusing retinal cells.

Primary myogenic cultures from mice lacking MyoD support fusion of donor retinal cells, but do not permit their full myogenic reprogramming

To gain further insight into the role of the host myotube transcriptional milieu in supporting myogenic reprogramming, we investigated if myotubes in primary myogenic culture from MyoD−/− mice could support fusion and activation of myogenic promoters in fusing retinal cells. We previously demonstrated that MyoD−/− myoblasts did not fail to form myotubes but displayed a longer proliferative period (Yablonka-Reuveni et al., 1999). Co-cultures of donor retinal cells from eGFP mice and MyoD−/− host myogenic cultures demonstrated expression of GFP in myotubes but no MyoD+ nuclei based on immunofluorescent analysis. Also, even though MyoD+ nuclei were not detected, donor retinal cells from MyoD-GFP mice contributed GFP expression to host myotubes from MyoD−/− mice (data not shown). Co-cultures of retinal cells from the double transgenic eGFP/MLC3F-nLacZ mice and MyoD−/− host myogenic (Fig. 9) or eGFP/Myf5nLacZ/+ mice and MyoD−/− host myogenic cultures (not shown) also demonstrated eGFP+ myotubes, but no MLC3F-driven or Myf5-driven β-gal protein expression or MyoD protein in myofiber nuclei or cytoplasm. Control donor myoblasts from eGFP mice cultured onto host MyoD−/− primary myogenic cultures started to express MyoD before fusion and maintained its expression after fusion. In all studies described here, wildtype primary myogenic cultures from strain matched (BALB/c) mice or from MyoD+/+ siblings (generated by crossing MyoD+/− mice) were used as positive controls for MyoD expression and for supporting MLC3F-nLacZ expression following fusion of retinal-derived cells with host myotubes. Collectively, these studies suggest that the transcriptional milieu in MyoD−/− myotubes is sufficient to activate MyoD reporter expression as seen by MyoD-GFP but insufficient for activating endogenous MyoD expression or MLC3F promoter function.

Fig. 9
Myogenic cultures from MyoD−/− mice support fusion of donor retinal cells with host myotubes but not myogenic reprogramming. Co-culturing of retina-derived cells from double transgenic eGFP/MLC3F-nLacZ mice with myogenic cultures from ...


Published studies suggest that progenitors other than satellite cells might be able to contribute to adult myogenesis (Peault et al., 2007). However, there is a lack of clear consensus about the origin of these cells, and the extent of their capacity to contribute to myogenesis in a cell autonomous manner. Different from satellite cells that perform as myogenic progenitors both in vivo and in culture, atypical myogenic cell sources may contribute to myogenesis exclusively after recruitment by myotubes, as was shown for bone marrow progenitors (Camargo et al., 2003; Lee et al., 2005) and adult mesenchymal stem cells (Schulze et al., 2005). However, other studies claimed the ability of non-satellite cells to assume myogenic phenotype prior to fusion with the existing myofibers and in some cases even to occupy the satellite cell niche (Breton et al., 1995; Dellavalle et al., 2007); (Dreyfus et al., 2004; LaBarge and Blau, 2002; Sampaolesi et al., 2003). There is also inconsistency between different reports on the potential of non-myogenic cells to upregulate myogenic genes in host myofibers (Ferrari et al., 1998; Gussoni et al., 1999; Lapidos et al., 2004; Wernig et al., 2005). These differences between studies concerning myogenic outcome might be dependent on the specific type of donor cells being analyzed, strain background, and myogenic reprogramming markers being used. Moreover, atypical myogenic cell sources might function only during certain types of extensive muscular activities and depletion in bona fide satellite cells (Deasy et al., 2007; Palermo et al., 2005), which further complicates their identification in a reproducible manner.

Here, we investigated the potential of vascular preparations from juvenile and adult mouse retina to undergo myogenesis. The choice of retina, an organ richly supplied with capillary network that is also anatomically separated from skeletal muscles, ensured that the vasculature preparations were devoid of satellite cells. Our studies did not detect spontaneous skeletal myogenesis in cells comprising the retinal preparation. However, we discovered that certain cells in the retinal vascular preparations could fuse with developing myotubes. The fusing retinal cells contributed their nuclei to the host myotubes, and in some instances these nuclei underwent myogenic reprogramming, as determined by the expression of the muscle-specific MLC3F-nLacZ transgene. This process of myogenic reprogramming involved the expression of MyoD by the donor retinal cells as demonstrated by the utilization of MyoD-GFP transgenic mice and by detection of mouse MyoD in chicken myotubes upon co-culturing of mouse retinal cells with chicken myogenic cultures. However, Myf5 expression, as judged by Myf5 promoter activity using retinal donor cells from Myf5nLacZ/+ mice was not detected in the donor nuclei. Likewise, we previously reported that the spontaneous transition of smooth muscle cell lines into skeletal myogenesis followed a program where MyoD, but not Myf5, expression was initiated (Graves and Yablonka-Reuveni, 2000). Using host primary cultures from MyoD null mice, we also demonstrated that full reprogramming of the fusing retinal nuclei did not occur in the absence of MyoD expression in host myotubes. While donor retinal cells fused with MyoD null myotubes and further initiated MyoD-GFP transgene expression, neither endogenous MyoD expression nor activation of the MLC3F-nLacZ transgene by the fused retinal nuclei was supported. Thus, we identified the muscle-specific transcription factor MyoD as a key regulator of the process of myogenic conversion of fusing non-myogenic cells.

Our findings that MyoD but not Myf5(nLacZ) gene activity can be detected upon myogenic reprogramming provide new insight regarding distinctions between the function of Myf5 and MyoD. These two myogenic regulatory factors are involved in establishing the myogenic lineage during early development. Genetic ablation of each one of them does not impair muscle development, suggesting overlapping roles for Myf5 and MyoD during muscle development (Rudnicki et al., 1993). Nevertheless, based on our data, it seems that MyoD, but not Myf5, gene activity participates in the myogenic reprogramming process. This distinction raises an interesting frame for investigating the role of myogenic reprogramming in vivo, as it is possible that such a process would not take place in MyoD null mice. Furthermore, the availability of Myf5 null mice that survive to adulthood (Ustanina et al., 2007), offers a means to investigate the capacity of myotubes lacking Myf5 to support myogenic reprogramming. Similarly, establishing whether non-myogenic donor cells from MyoD or Myf5 null mice would undergo myogenic reprogramming might shed further light on the role of these myogenic transcription factors in adult life. The detection of MyoD-expressing cells in non-muscle tissues (Gerhart et al., 2001; Grounds et al., 1992; Walker et al., 2001) further raises the possibility that MyoD might play a yet undefined role in certain non-myogenic cell types. The relationship between the latter MyoD-expressing cells and perictyes/smooth muscle cells warrants further investigation.

We found more than one population within retinal preparations capable of fusing with myotubes. Sca1-GFP+ population, which represented retinal endothelial cells fused with myotubes in retina-limb co-cultures but did not activate myogenic genes (as determined by MLC3F-nLacZ transgene expression) after fusion. The Sca1-GFP mouse was initially introduced in connection with transgene expression in adult mouse hematopoietic stem cells (Ma et al., 2002). However, our analysis of blood preparations showed no fusion with host myotubes and the minimal fusion observed with bone marrow cells was never followed by myogenic reprogramming. Thus, we were able to rule out the possible contribution of blood borne and endothelial cells to nuclear reprogramming upon fusion with myotubes. Our findings that a) cells capable of myogenic conversion are contributed by the Sca1-GFP population, and b) pericytes whose nuclei found in myotubes, are Sca1-GFP, suggest that the cell type undergoing myogenic conversion is contained within the perictye population. However, available mouse models did not allow us to analyze more directly the expression of myogenic markers by fusing pericytes. The MLC3F-nLacZ mouse, used throughout this study to demonstrate myogenic reprogramming, could not be used in conjunction with XLacZ4 mouse due to common reporter. Furthermore, MyoD-GFP mouse could not be used in combination with XlacZ4 due to very weak GFP expression in heterozygous setting and differences in expression timing of the reporters in host myotubes. Nevertheless, the mesenchymal origin and plasticity of pericytes (i.e., able to give rise to bone and fat cells) (Farrington-Rock et al., 2004), provide strong support for their potential reprogramming capacity.

Early heterokaryon studies on the capacity of non-myogenic cells to undergo myogenic reprogramming involved forced fusion between non-myogenic and myogenic cells (Pavlath and Blau, 1986). In the present study we demonstrated spontaneous fusion between non-myogenic cells and host myotubes and subsequent myogenic reprogramming. While fusion of non-myogenic cells with myotubes might be a stochastic event that can involve a number of different cell types, myogenic reprogramming might take place only upon fusion of specific cell types. Recognizing the difference between fusion per se and fusion followed by nuclear reprogramming may assist in clarifying existing contradictions about the types of non-satellite cells that can or cannot contribute to myogenesis.

We acknowledge that studies by others reported on the capacity of both pericytes and endothelium-related cells to assume a myogenic fate without fusion with host myotubes (Dellavalle et al., 2007; Peault et al., 2007). However, within the experimental models used in the present study, we did not observe cell autonomous myogenic activity of retina-derived cells of juvenile and mature mice. It is indeed possible that amplifying vascular cells in culture by passaging can induce transition into the skeletal myogenic phenotype, as shown by us previously for established smooth muscle lines derived from rat vasculature (Graves and Yablonka-Reuveni, 2000). Likewise, pericytes derived from human skeletal muscle underwent skeletal myogenesis following in vitro expansion (Dellavalle et al., 2007). Interestingly, in the latter study skeletal myogenesis was observed upon fusion of the cultured cells into multinucleated syncytia, whereas prior to fusion the mononuclear cells did not express typical markers of myoblasts (Dellavalle et al., 2007). However, in the present study we avoided long term cell passaging to minimize phenotypic changes resulting from cell culturing. It is also possible that vasculature cells isolated from skeletal muscle might be already triggered in vivo to enter myogenesis due to the cues from the muscle environment.

Satellite cells were shown to participate in repair of damaged myofibers (Collins et al., 2005) and their absence in postnatal life seems to impair postnatal muscle development (Kuang et al., 2006). However, it remains unclear if cells of non-satellite cell origin can participate in myofiber maintenance during daily routine activities. We determined that fusion and myogenic reprogramming ability of the retina-derived cells and the capacity of the host myotubes to support this processes was present in juvenile, young adult and mature mice (mice up to 1 year of age were analyzed). Shedding light on alternative cellular origins for replenishing myofiber nuclei may be beneficial for enhancing muscle performance during conditions of muscle inactivity due to injury and disease, which result in muscle atrophy. Additionally, a capacity to recruit myofiber nuclei from alternative cells may facilitate reduction of age-linked muscle atrophy, which seems to correlate with a declining performance of satellite cells with age (Conboy et al., 2005; Shefer et al., 2006). We conclude with the proposal that fusion of non-myogenic cells with myofibers, whether or not followed by myogenic reprogramming, may reflect a biologically significant process involved in replenishing myofiber nuclei during skeletal muscle maintenance throughout the lifespan.


We are grateful to Thong Pham and Dr. Peter Rabinovitch for their valuable assistance with cell sorting (performed at the core facility of the University of Washington Nathan Shock Center of Excellence) and to our team member, Dr. Kenneth Day, for his important input during discussions of the project. We additionally thank Dr. Tom Reh and his team members (University of Washington) for valuable input on the retina studies. This study could not be possible without the mouse models and antibodies kindly provided to us by many colleagues, who are listed by their respective contribution in Materials and Methods. This work was supported by a grant to Z.Y.R. from the National Institute on Aging (AG021566). Z.Y.R. acknowledges additional support during the course of this study from the National Institute on Aging (AG013798) and the USDA Cooperative State Research, Education and Extension Service (NRI, 2003-35206-12843).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Armulik A, Abramsson A, Betsholtz C. Endothelial/pericyte interactions. Circ Res. 2005;97:512–23. [PubMed]
  • Asakura A, Seale P, Girgis-Gabardo A, Rudnicki MA. Myogenic specification of side population cells in skeletal muscle. J Cell Biol. 2002;159:123–34. [PMC free article] [PubMed]
  • Beauchamp JR, Heslop L, Yu DS, Tajbakhsh S, Kelly RG, Wernig A, Buckingham ME, Partridge TA, Zammit PS. Expression of CD34 and Myf5 defines the majority of quiescent adult skeletal muscle satellite cells. J Cell Biol. 2000;151:1221–34. [PMC free article] [PubMed]
  • Breton M, Li ZL, Paulin D, Harris JA, Rieger F, Pincon-Raymond M, Garcia L. Myotube driven myogenic recruitment of cells during in vitro myogenesis. Dev Dyn. 1995;202:126–36. [PubMed]
  • Camargo FD, Green R, Capetanaki Y, Jackson KA, Goodell MA. Single hematopoietic stem cells generate skeletal muscle through myeloid intermediates. Nat Med. 2003;9:1520–7. [PubMed]
  • Cerletti M, Molloy MJ, Tomczak KK, Yoon S, Ramoni MF, Kho AT, Beggs AH, Gussoni E. Melanoma cell adhesion molecule is a novel marker for human fetal myogenic cells and affects myoblast fusion. J Cell Sci. 2006;119:3117–27. [PMC free article] [PubMed]
  • Chen JC, Love CM, Goldhamer DJ. Two upstream enhancers collaborate to regulate the spatial patterning and timing of MyoD transcription during mouse development. Dev Dyn. 2001;221:274–88. [PubMed]
  • Collins CA, Olsen I, Zammit PS, Heslop L, Petrie A, Partridge TA, Morgan JE. Stem cell function, self-renewal, and behavioral heterogeneity of cells from the adult muscle satellite cell niche. Cell. 2005;122:289–301. [PubMed]
  • Conboy IM, Conboy MJ, Wagers AJ, Girma ER, Weissman IL, Rando TA. Rejuvenation of aged progenitor cells by exposure to a young systemic environment. Nature. 2005;433:760–4. [PubMed]
  • Cusella De Angelis MG, Balconi G, Bernasconi S, Zanetta L, Boratto R, Galli D, Dejana E, Cossu G. Skeletal myogenic progenitors in the endothelium of lung and yolk sac. Exp Cell Res. 2003;290:207–16. [PubMed]
  • Day K, Shefer G, Richardson JB, Enikolopov G, Yablonka-Reuveni Z. Nestin-GFP reporter expression defines the quiescent state of skeletal muscle satellite cells. Dev Biol. 2007;304:246–59. [PMC free article] [PubMed]
  • De Angelis L, Berghella L, Coletta M, Lattanzi L, Zanchi M, Cusella-De Angelis MG, Ponzetto C, Cossu G. Skeletal myogenic progenitors originating from embryonic dorsal aorta coexpress endothelial and myogenic markers and contribute to postnatal muscle growth and regeneration. J Cell Biol. 1999;147:869–78. [PMC free article] [PubMed]
  • Deasy BM, Lu A, Tebbets JC, Feduska JM, Schugar RC, Pollett JB, Sun B, Urish KL, Gharaibeh BM, Cao B, Rubin RT, Huard J. A role for cell sex in stem cell-mediated skeletal muscle regeneration: female cells have higher muscle regeneration efficiency. J Cell Biol. 2007;177:73–86. [PMC free article] [PubMed]
  • Dellavalle A, Sampaolesi M, Tonlorenzi R, Tagliafico E, Sacchetti B, Perani L, Innocenzi A, Galvez BG, Messina G, Morosetti R, Li S, Belicchi M, Peretti G, Chamberlain JS, Wright WE, Torrente Y, Ferrari S, Bianco P, Cossu G. Pericytes of human skeletal muscle are myogenic precursors distinct from satellite cells. Nat Cell Biol. 2007;9:255–67. [PubMed]
  • Diaz-Flores L, Gutierrez R, Varela H, Rancel N, Valladares F. Microvascular pericytes: a review of their morphological and functional characteristics. Histol Histopathol. 1991;6:269–86. [PubMed]
  • Dreyfus PA, Chretien F, Chazaud B, Kirova Y, Caramelle P, Garcia L, Butler-Browne G, Gherardi RK. Adult bone marrow-derived stem cells in muscle connective tissue and satellite cell niches. Am J Pathol. 2004;164:773–9. [PMC free article] [PubMed]
  • Farrington-Rock C, Crofts NJ, Doherty MJ, Ashton BA, Griffin-Jones C, Canfield AE. Chondrogenic and adipogenic potential of microvascular pericytes. Circulation. 2004;110:2226–32. [PubMed]
  • Ferrari G, Cusella-De Angelis G, Coletta M, Paolucci E, Stornaiuolo A, Cossu G, Mavilio F. Muscle regeneration by bone marrow-derived myogenic progenitors. Science. 1998;279:1528–30. [PubMed]
  • Gerhart J, Bast B, Neely C, Iem S, Amegbe P, Niewenhuis R, Miklasz S, Cheng PF, George-Weinstein M. MyoD-positive myoblasts are present in mature fetal organs lacking skeletal muscle. J Cell Biol. 2001;155:381–92. [PMC free article] [PubMed]
  • Gitlin JD, D’Amore PA. Culture of retinal capillary cells using selective growth media. Microvasc Res. 1983;26:74–80. [PubMed]
  • Goodell MA, Brose K, Paradis G, Conner AS, Mulligan RC. Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo. J Exp Med. 1996;183:1797–806. [PMC free article] [PubMed]
  • Graves DC, Yablonka-Reuveni Z. Vascular smooth muscle cells spontaneously adopt a skeletal muscle phenotype: a unique Myf5(−)/MyoD(+) myogenic program. J Histochem Cytochem. 2000;48:1173–93. [PubMed]
  • Grounds MD, Garrett KL, Beilharz MW. The transcription of MyoD1 and myogenin genes in thymic cells in vivo. Exp Cell Res. 1992;198:357–61. [PubMed]
  • Gussoni E, Soneoka Y, Strickland CD, Buzney EA, Khan MK, Flint AF, Kunkel LM, Mulligan RC. Dystrophin expression in the mdx mouse restored by stem cell transplantation. Nature. 1999;401:390–4. [PubMed]
  • Halevy O, Piestun Y, Allouh MZ, Rosser BW, Rinkevich Y, Reshef R, Rozenboim I, Wleklinski-Lee M, Yablonka-Reuveni Z. Pattern of Pax7 expression during myogenesis in the posthatch chicken establishes a model for satellite cell differentiation and renewal. Dev Dyn. 2004;231:489–502. [PubMed]
  • Hitchcock PF, Macdonald RE, VanDeRyt JT, Wilson SW. Antibodies against Pax6 immunostain amacrine and ganglion cells and neuronal progenitors, but not rod precursors, in the normal and regenerating retina of the goldfish. J Neurobiol. 1996;29:399–413. [PubMed]
  • Holmes C, Stanford WL. Stem Cell Antigen-1: Expression, Function, and Enigma. Stem Cells 2007 [PubMed]
  • Kelly R, Alonso S, Tajbakhsh S, Cossu G, Buckingham M. Myosin light chain 3F regulatory sequences confer regionalized cardiac and skeletal muscle expression in transgenic mice. J Cell Biol. 1995;129:383–96. [PMC free article] [PubMed]
  • Klassen HJ, Ng TF, Kurimoto Y, Kirov I, Shatos M, Coffey P, Young MJ. Multipotent retinal progenitors express developmental markers, differentiate into retinal neurons, and preserve light-mediated behavior. Invest Ophthalmol Vis Sci. 2004;45:4167–73. [PubMed]
  • Klinghoffer RA, Mueting-Nelsen PF, Faerman A, Shani M, Soriano P. The two PDGF receptors maintain conserved signaling in vivo despite divergent embryological functions. Mol Cell. 2001;7:343–54. [PubMed]
  • Kuang S, Charge SB, Seale P, Huh M, Rudnicki MA. Distinct roles for Pax7 and Pax3 in adult regenerative myogenesis. J Cell Biol. 2006;172:103–13. [PMC free article] [PubMed]
  • LaBarge MA, Blau HM. Biological progression from adult bone marrow to mononucleate muscle stem cell to multinucleate muscle fiber in response to injury. Cell. 2002;111:589–601. [PubMed]
  • Lapidos KA, Chen YE, Earley JU, Heydemann A, Huber JM, Chien M, Ma A, McNally EM. Transplanted hematopoietic stem cells demonstrate impaired sarcoglycan expression after engraftment into cardiac and skeletal muscle. J Clin Invest. 2004;114:1577–85. [PMC free article] [PubMed]
  • Lee JH, Kosinski PA, Kemp DM. Contribution of human bone marrow stem cells to individual skeletal myotubes followed by myogenic gene activation. Exp Cell Res. 2005;307:174–82. [PubMed]
  • Ludolph DC, Konieczny SF. Transcription factor families: muscling in on the myogenic program. Faseb J. 1995;9:1595–604. [PubMed]
  • Ma X, Robin C, Ottersbach K, Dzierzak E. The Ly-6A (Sca-1) GFP transgene is expressed in all adult mouse hematopoietic stem cells. Stem Cells. 2002;20:514–21. [PubMed]
  • Minasi MG, Riminucci M, De Angelis L, Borello U, Berarducci B, Innocenzi A, Caprioli A, Sirabella D, Baiocchi M, De Maria R, Boratto R, Jaffredo T, Broccoli V, Bianco P, Cossu G. The meso-angioblast: a multipotent, self-renewing cell that originates from the dorsal aorta and differentiates into most mesodermal tissues. Development. 2002;129:2773–83. [PubMed]
  • Mitchell PO, Mills T, O’Connor RS, Kline ER, Graubert T, Dzierzak E, Pavlath GK. Sca-1 negatively regulates proliferation and differentiation of muscle cells. Dev Biol. 2005;283:240–52. [PubMed]
  • Muskiewicz KR, Frank NY, Flint AF, Gussoni E. Myogenic potential of muscle side and main population cells after intravenous injection into sub-lethally irradiated mdx mice. J Histochem Cytochem. 2005;53:861–73. [PubMed]
  • Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y. ‘Green mice’ as a source of ubiquitous green cells. FEBS Lett. 1997;407:313–9. [PubMed]
  • Palermo AT, Labarge MA, Doyonnas R, Pomerantz J, Blau HM. Bone marrow contribution to skeletal muscle: a physiological response to stress. Dev Biol. 2005;279:336–44. [PubMed]
  • Pavlath GK, Blau HM. Expression of muscle genes in heterokaryons depends on gene dosage. J Cell Biol. 1986;102:124–30. [PMC free article] [PubMed]
  • Peault B, Rudnicki M, Torrente Y, Cossu G, Tremblay JP, Partridge T, Gussoni E, Kunkel LM, Huard J. Stem and Progenitor Cells in Skeletal Muscle Development, Maintenance, and Therapy. Mol Ther. 2007;15:867–877. [PubMed]
  • Rudnicki MA, Braun T, Hinuma S, Jaenisch R. Inactivation of MyoD in mice leads to up-regulation of the myogenic HLH gene Myf-5 and results in apparently normal muscle development. Cell. 1992;71:383–90. [PubMed]
  • Rudnicki MA, Schnegelsberg PN, Stead RH, Braun T, Arnold HH, Jaenisch R. MyoD or Myf-5 is required for the formation of skeletal muscle. Cell. 1993;75:1351–1359. [PubMed]
  • Sampaolesi M, Blot S, D’Antona G, Granger N, Tonlorenzi R, Innocenzi A, Mognol P, Thibaud JL, Galvez BG, Barthelemy I, Perani L, Mantero S, Guttinger M, Pansarasa O, Rinaldi C, Cusella De Angelis MG, Torrente Y, Bordignon C, Bottinelli R, Cossu G. Mesoangioblast stem cells ameliorate muscle function in dystrophic dogs. Nature. 2006;444:574–9. [PubMed]
  • Sampaolesi M, Torrente Y, Innocenzi A, Tonlorenzi R, D’Antona G, Pellegrino MA, Barresi R, Bresolin N, De Angelis MG, Campbell KP, Bottinelli R, Cossu G. Cell therapy of alpha-sarcoglycan null dystrophic mice through intra-arterial delivery of mesoangioblasts. Science. 2003;301:487–92. [PubMed]
  • Schulze M, Belema-Bedada F, Technau A, Braun T. Mesenchymal stem cells are recruited to striated muscle by NFAT/IL-4-mediated cell fusion. Genes Dev. 2005;19:1787–98. [PMC free article] [PubMed]
  • Shefer G, Van de Mark DP, Richardson JB, Yablonka-Reuveni Z. Satellite-cell pool size does matter: defining the myogenic potency of aging skeletal muscle. Dev Biol. 2006;294:50–66. [PMC free article] [PubMed]
  • Shen J, Watanabe S, Kaneko A. Cell dissociation with papain reduces the density of cGMP-activated channels of the retinal rod. Jpn J Physiol. 1995;45:151–64. [PubMed]
  • Sherwood RI, Christensen JL, Conboy IM, Conboy MJ, Rando TA, Weissman IL, Wagers AJ. Isolation of adult mouse myogenic progenitors: functional heterogeneity of cells within and engrafting skeletal muscle. Cell. 2004;119:543–54. [PubMed]
  • Song L, Pachter JS. Culture of murine brain microvascular endothelial cells that maintain expression and cytoskeletal association of tight junction-associated proteins. In Vitro Cell Dev Biol Anim. 2003;39:313–20. [PubMed]
  • Su X, Sorenson CM, Sheibani N. Isolation and characterization of murine retinal endothelial cells. Mol Vis. 2003;9:171–8. [PubMed]
  • Tajbakhsh S, Bober E, Babinet C, Pournin S, Arnold H, Buckingham M. Gene targeting the myf-5 locus with nlacZ reveals expression of this myogenic factor in mature skeletal muscle fibres as well as early embryonic muscle. Dev Dyn. 1996;206:291–300. [PubMed]
  • Terstappen LW, Meiners H, Loken MR. A rapid sample preparation technique for flow cytometric analysis of immunofluorescence allowing absolute enumeration of cell subpopulations. J Immunol Methods. 1989;123:103–12. [PubMed]
  • Tidhar A, Reichenstein M, Cohen D, Faerman A, Copeland NG, Gilbert DJ, Jenkins NA, Shani M. A novel transgenic marker for migrating limb muscle precursors and for vascular smooth muscle cells. Dev Dyn. 2001;220:60–73. [PubMed]
  • Torrente Y, Tremblay JP, Pisati F, Belicchi M, Rossi B, Sironi M, Fortunato F, El Fahime M, D’Angelo MG, Caron NJ, Constantin G, Paulin D, Scarlato G, Bresolin N. Intraarterial injection of muscle-derived CD34(+)Sca-1(+) stem cells restores dystrophin in mdx mice. J Cell Biol. 2001;152:335–48. [PMC free article] [PubMed]
  • Ustanina S, Carvajal J, Rigby P, Braun T. The Myogenic Factor Myf5 Supports efficient skeletal muscle regeneration by enabling transient myoblast amplification. Stem Cells. 2007 2007 May 10; Epub ahead of print. [PubMed]
  • van de Rijn M, Heimfeld S, Spangrude GJ, Weissman IL. Mouse hematopoietic stem-cell antigen Sca-1 is a member of the Ly-6 antigen family. Proc Natl Acad Sci U S A. 1989;86:4634–8. [PMC free article] [PubMed]
  • Walker GA, Guerrero IA, Leinwand LA. Myofibroblasts: molecular crossdressers. Curr Top Dev Biol. 2001;51:91–107. [PubMed]
  • Wernig G, Janzen V, Schafer R, Zweyer M, Knauf U, Hoegemeier O, Mundegar RR, Garbe S, Stier S, Franz T, Wernig M, Wernig A. The vast majority of bone-marrow-derived cells integrated into mdx muscle fibers are silent despite long-term engraftment. Proc Natl Acad Sci U S A. 2005;102:11852–7. [PMC free article] [PubMed]
  • Yablonka-Reuveni Z, Rudnicki MA, Rivera AJ, Primig M, Anderson JE, Natanson P. The transition from proliferation to differentiation is delayed in satellite cells from mice lacking MyoD. Dev Biol. 1999;210:440–55. [PubMed]
  • Yang J, Ontell MP, Kelly R, Watkins SC, Ontell M. Limitations of nls beta-galactosidase as a marker for studying myogenic lineage or the efficacy of myoblast transfer. Anat Rec. 1997;248:40–50. [PubMed]
  • Zammit PS, Partridge TA, Yablonka-Reuveni Z. The skeletal muscle satellite cell: the stem cell that came in from the cold. J Histochem Cytochem. 2006;54:1177–91. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...