Muscular dystrophy candidate gene FRG1 is critical for muscle development


  • Meredith L. Hanel,

    1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, Illinois
    Search for more papers by this author
    • Drs. Hanel and Wuebbles contributed equally to this work.

  • Ryan D. Wuebbles,

    1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, Illinois
    Search for more papers by this author
    • Drs. Hanel and Wuebbles contributed equally to this work.

  • Peter L. Jones

    Corresponding author
    1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, Illinois
    • Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, IL 61801
    Search for more papers by this author


The leading candidate gene responsible for facioscapulohumeral muscular dystrophy (FSHD) is FRG1 (FSHD region gene 1). However, the correlation of altered FRG1 expression levels with disease pathology has remained controversial and the precise function of FRG1 is unknown. Here, we carried out a detailed analysis of the normal expression patterns and effects of FRG1 misexpression during vertebrate embryonic development using Xenopus laevis. We show that frg1 is expressed in and essential for the development of the tadpole musculature. FRG1 morpholino injection disrupted myotome organization and led to inhibited myotome growth, while elevated FRG1 led to abnormal epaxial and hypaxial muscle formation. Thus, maintenance of normal FRG1 levels is critical for proper muscle development, supportive of FSHD disease models whereby misregulation of FRG1 plays a causal role underlying the pathology exhibited in FSHD patients. Developmental Dynamics 238:1502–1512, 2009. © 2008 Wiley-Liss, Inc.


Facioscapulohumeral muscular dystrophy (FSHD) is an autosomal dominant myopathy characterized by progressive atrophy of the facial, shoulder, and upper arm muscles. The FSHD genetic lesion is a contraction of the chromosome 4q35 tandem array of 3.3 kb D4Z4 repeats, below a threshold of 11 copies (Lunt et al.,1995). The leading candidate disease gene for FSHD, based on analysis of gene expression and proximity to the D4Z4 repeats, is FRG1, located 125 kb centromeric to the contracted D4Z4 (van Deutekom et al.,1996). Gross 25- to 40-fold overexpression of Frg1 specifically in the skeletal muscle produces a dystrophic muscle phenotype in transgenic mice (Gabellini et al.,2006). However, measurements of FRG1 mRNA levels from FSHD patient muscle have varied between 25-fold increased, unchanged, and 5-fold decreased compared with controls (van Deutekom et al.,1996; Gabellini et al.,2002; Jiang et al.,2003; Winokur et al.,2003; Osborne et al.,2007). Complicating the issue, FRG1 levels in FSHD patients are likely altered in numerous tissues throughout the body in addition to skeletal muscle, yet the pathology is predominantly in skeletal muscles and the underlying vasculature. Thus, correlations of FRG1 mRNA levels with FSHD pathology remain inconclusive and controversial.

Although the human 4q35 FRG1 is highly conserved in vertebrates and invertebrates (97% AA identity with mouse, 81% AA identity with Xenopus laevis, and 46% AA identity with Caenorhabditis elegans; Grewal et al.,1998), FRG1's precise biological function is still not known making it difficult to envision how changes in its expression would ultimately lead specifically to FSHD pathology. Due to FSHD, studies on FRG1 have mainly focused on its role in muscle. However, FRG1 expression has been detected in all human tissues tested, including human embryonic brain and muscle as well as placenta (van Deutekom et al.,1996), potentially indicating a role in early development. Overexpression of FRG1 suggests a function in RNA biogenesis (van Koningsbruggen et al.,2004,2007) and indeed the FSHD-model transgenic mouse found mis-splicing of muscle specific transcripts (Gabellini et al.,2006). Still, how FRG1 expression levels might affect RNA biogenesis is not known.

Here, we describe the spatiotemporal expression pattern of frg1 during early X. laevis development and characterize the developmental effects on the musculature of decreasing or increasing FRG1 levels by either translation-blocking morpholino or mRNA injections, respectively. Supporting a role for FRG1 in the muscular aspect of FSHD, we demonstrate that elevated levels of frg1 during development disrupt muscle organization, hypaxial muscle cell migration, and skeletal muscle morphology. This work supports the FRG1 overexpression disease model for FSHD and provides new insights into the mechanisms of FSHD disease pathology.


Spatiotemporal Expression of frg1 During Development Is Not Tissue Specific

The FRG1 protein, including all putative domains, is highly conserved evolutionarily among metazoans yet lacks redundant paralogues, suggesting a critical conserved function (Supp. Fig. S1, which is available online). The spatiotemporal expression of FRG1 during vertebrate development was examined by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR), whole-mount in situ hybridization, and immunostaining throughout X. laevis embryogenesis. Temporal analysis of the frg1 mRNA levels by qRT-PCR (Fig. 1A) showed peak transcript levels before the midblastula transition (MBT) that is attributed to maternal stores of frg1. However, from MBT on through stage 41, the frg1 transcript levels decreased progressively, suggesting an early developmental requirement for FRG1.

Figure 1.

Expression of FRG1 transcript and protein during development. A: Analysis of Xenopus laevis frg1 transcript levels across developmental stages using quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR). B–G: In situ hybridization using the Xenopus laevis frg1 transcript at stage (St.) 26 (B), St. 32 (C), and St. 38 (E,F). Sense frg1 controls are shown for St. 32 (D) and St. 38 (G). Arrowheads point to migrating hypaxial muscle. H–J: FRG1 immunostaining on St. 18 (H), St. 24 (I), and St. 38 (J) embryos showing FRG1 antibody (upper) and secondary antibody alone control (lower). K,L: Stage 30 FRG1 immunostained embryos were sectioned sagitally (K) or transversely (L). Arrows point to aligned nuclei in the somite stained with FRG1 antibody. np, neural plate; ps, pronephric sinus; nt, neural tube; nc, notochord; is, intersomitic region.

The spatiotemporal expression pattern of frg1 was similarly examined during development using whole-mount in situ hybridization (Fig. 1B–G). Throughout early development (stages 18–38), frg1 expression appeared generally low and ubiquitous. Still, certain tissues and certain stages displayed distinct patterns of frg1. At stage 26, frg1 was found diffusely within the head, eyes, branchial arches, somites, and notochord (Fig. 1B). By tail bud stages, stronger frg1 staining was found in the telencephalon region of the head, eyes, and branchial arches with weaker staining in the somites. At stage 38, expression in the posterior cardinal vein was increased along with faint staining in the migrating hypaxial mesoderm, somites, and intersomitic regions (Fig. 1E,F). The staining patterns were specific to the frg1 transcript as the staining was clearly absent from the sense controls (Fig. 1C compared to sense control Fig. 1D). Overall, these expression patterns indicate that frg1 is ubiquitous early in development and likely involved in the development of multiple tissues including the musculature.

To confirm the expression of FRG1 in the somites suggested by the in situ hybridization analysis, the spatiotemporal expression patterns of FRG1 protein were further examined by immunohistochemistry using an antibody specific for X. laevis FRG1 (Fig. 1H–L, Supp. Fig. S2). Whole-mount immunostaining of stage 18 embryos revealed that, while apparently ubiquitous in tissue distribution by in situ localization, FRG1 protein is elevated specifically in the developing somite region, neural plate, and developing head (Fig. 1H). By stage 24, FRG1 protein is seen within the somites, head and eye regions, although the epithelial staining masks some of the details (Fig. 1I). This epithelial staining likely indicates that FRG1 protein is expressed in the epithelium as the standard absorption of the antisera against embryos to remove background epithelial staining resulted in considerable loss of specific FRG1 signal throughout the embryo. Stage 38 showed staining similar to stage 24, albeit with elevated staining developing within the somites (Fig. 1J).

To determine details of FRG1 localization obscured by epithelial staining, FRG1 immunostained embryos (St. 30) were embedded in paraffin and sectioned either sagitally or transversally. Sagittal sections revealed FRG1 was clearly represented in or around the aligned nuclei of the myotome (Fig. 1K, Supp. Fig. S2A and B). Transverse sectioning through the somites displayed intense FRG1 immunostaining in the tissues surrounding the developing neural tube, the notochord and also within the myotome (Fig. 1L, Supp. Fig. S2D and E). Of interest, although ubiquitously expressed early in development, by stage 30, FRG1 protein became somewhat tissue restricted, present at elevated levels in neural and muscular tissues and devoid in some areas of the embryo (Supp. Fig. S2A and E).

FRG1 Morphants Exhibit Abnormal Development

To determine the requirements for FRG1 during differentiation and tissue development, FRG1 levels were decreased using antisense translation-blocking morpholinos and development was monitored. For each set of injections, one of two independent FRG1 morpholinos (FMO) complementary to the X. laevis frg1 mRNA (Fig. 2F) or a random nonsense control morpholino (CMO) of the same length was injected asymmetrically into two-cell stage blastomeres, thus providing an injected and control uninjected side for each embryo, and development was allowed to proceed. Only embryos appearing normal through gastrulation were selected for further development and inclusion. CMO injections throughout this study were performed at 40 ng, while FMO injections were either 20 ng or 40 ng as indicated. By tail bud stages, CMO-injected embryos had developed normally (Fig. 2A); however, FMO1- or FMO2-injected embryos were smaller and had begun to show an axis curved toward the FMO-injected side with severity increasing with dosage (Fig. 2B–E). Consistently, FRG1 morphants exhibited abnormal embryonic development manifested as decreased body size, indicating that FRG1 is required for normal body growth (Fig. 2E).

Figure 2.

FRG1 morpholino injections affect development. A–C: Stage 28 embryos injected with 40 ng of CMO (A), 40 ng of FMO1 (B), and 40 ng of FMO2 (C). D: Increased magnification of embryo injected on the right side in B with fluorescein-labeled morpholino side. White arrow points to embryo is curved to the right in B and D. E: Comparison of sizes of representative embryos injected with 40 ng of CMO, 40 ng of FMO1, and 20 ng of FMO1. F: Sequence of the 5′ untranslated region (UTR) of the Xenopus laevis frg1 mRNA with the morpholino sequences for FMO1 and FMO2 underlined and the translational start codon (atg) indicated by *.

FMO Injection Leads to Smaller Myotome and Altered Epaxial Muscle Morphology

Curved axis phenotypes are often associated with altered somite morphogenesis. To analyze the somites and particularly the myotome, stage 34–36 embryos injected with FRG1 morpholino were immunostained for differentiated muscle (12/101 mAb). Both nonoverlapping FRG1 morpholinos (Fig. 2F) produced a reduction in the dorsal–ventral expansion of the myotome on the morpholino-injected side in the majority of embryos showing 87% for 40 ng of FMO1, 89% for 20 ng of FMO1, and 43% for 40 ng of FMO2 compared to 10% for 40 ng of CMO (Figs. 3A, 4E,O). In stage 36 tadpoles, the 12/101 staining also stains the hypaxial muscle, which has differentiated from myoblasts that migrate from the ventral–lateral lip of the myotome. Not surprisingly, all stage 36 tadpoles exhibiting myotomes with a decreased width also showed a lack of differentiated hypaxial muscle (Fig. 3F).

Figure 3.

FRG1 depletion leads to decreased size of myotome and decreased muscle differentiation. A–C: The 12/101 staining of stage 34 embryos injected with 20 ng of FMO1 (A), 40 ng of FMO2 (B), and 40 ng of CMO (C). D,E: Close-view of myotome in A and B. F: Dorsal view of stage 36 embryo injected with 40 ng of FMO1, showing absence of hypaxial muscle on the injected side. G–I: Transverse (G,H) and coronal (I) sections of embryo injected with 40 ng of FMO1. Diagram above G shows approximate locations of sections. Arrows in G and H point to pronephric tubules and duct, respectively. J,K: Dorsal view of stage 18 embryos injected with 40 ng of FMO1 showing severe and mild phenotypes, respectively. Arrows mark the posterior end of the 12/101 staining. L: Dorsal view of stage 18 embryo injected with 40 ng of CMO. M,N: Transverse sections of embryos J and K, respectively. O: Percentage of morpholino-injected embryos displaying decreased width of myotome. P: Diagram of anatomy present in section M, pointing out the location of dermomyotome (dm) and orientation of cells in the somites. i, injected side.

Figure 4.

Decreased expression of myotome markers, decreased segmentation, and loss of hypaxial muscle in FRG1-depleted embryos. A,B: Stage 18 embryos injected (i, injected side) with either 40 ng of FMO1 or 40 ng of CMO respectively, stained for pax3 (dm dermomytome). C: Transverse section of a pax3-stained embryo injected with 40 ng of FMO1 showing less organized pax3 expression on the injected side (i) in the dermomyotome (arrowheads). D,E:myoD staining of stage 18 embryos injected with 20 ng of FMO1 and 40 ng of CMO, respectively. F: Graph showing percent stage 20 embryos with reduced pax3 in the dermomyotome and diffuse myoD staining. G,H:pax3 staining of injected (i) and uninjected sides of stage 34 embryos injected with 20 ng of FMO1, with close-up view of hypaxial muscle region. J,K:myoD staining of injected and uninjected sides of stage 34 embryos injected with 20 ng of FMO1. I,L:pax3 and myoD staining, respectively, of embryos injected with 40 ng of CMO. M,N: Vimentin antibody staining of injected (20 ng of FMO1) and uninjected side embryo (pm, pronephric mesenchyme). O: Percentage of embryos with defects in segmentation, hypaxial muscle, and mesenchyme with total numbers of embryos analyzed above each bar. P: Embryo injected with 40 ng of FMO1 and probed for myf5 shows no change in myf5 expression levels, despite a clear reduction in myotome width (double head arrow).

To further investigate the lack of proper growth and the gross morphological differences in the myotome due to FRG1 knockdown, histological sections were analyzed. In anterior sections, the myotome of the uninjected side is a slightly concave, continuous sheet of cells of relatively uniform thickness tapering at the dorsal tip (Fig. 3G, left) compared with the FMO1-injected side that displays a convex pattern, is thicker overall, particularly at both the dorsal and ventral regions, and contains less densely packed myotubes. There appears to be a lack of dorsal–ventral extension and the myotome is positioned more medially than the uninjected myotome. In the more posterior regions the shape of the myotome is more normal and the intensity gradient of 12/101 staining from the lateral side of the myotome is preserved, but the size difference between the uninjected and FMO1 side of the myotome is clearly evident (Fig. 3H). Of interest, the FMO1 morphants also showed an effect beyond the myotome, although within another mesoderm-derived structure, of disruption of pronephrous development. Both the pronephric tubules (Fig. 3G) and pronephric duct (Fig. 3H) have formed normally on the uninjected but appear small and underdeveloped in the morphant side.

Greater detail of the overall myotome segmentation was obtained through coronal histological sections of embryos used for immunohistochemistry. As shown in Figure 3I, it is difficult to discern individual myotomal segments on the FMO1-injected side and the numbers of myotubes is drastically less. The intersomitic boundaries are not very well defined because myotubes are not strictly aligned at their center as they are on the uninjected side. This is interesting because expression of FRG1 in the myotome is localized to the centrally aligned nuclei (Fig. 1K). Still, all myotubes are basically oriented along the anterior–posterior axis demonstrating that the somites have in fact rotated.

To assess earlier stages of myogenesis, the status of the differentiated muscle was assayed in stage 18–22 embryos. Overall, 12/101 staining was less intense on the FMO-injected side of the embryos and consistently extended less posterior compared with the controls (Figs. 3J, 4L). In the more severe cases, the 12/101 staining indicated a lack of somite segmentation on the injected side (Fig. 3J). In less severe cases, 12/101 staining on the injected side occurred as individual somites in the anterior with more posterior somite boundaries becoming blurred (Fig. 3K). Because muscle differentiation occurs in an anterior to posterior gradient, this suggests that loss of FRG1 results in a delay in muscle differentiation.

Further analysis of the neurulation stage myotome revealed that, similar with later stage embryos (Fig. 3A–F), the size of the myotome differed between the two sides with the medial–lateral width of the myotome being narrow on the FMO1–injected side. In addition, the shape of the somites on the uninjected side showed individual rounded somites, whereas individual somites on the injected side have a straight lateral edge. To determine the underlying structural differences within the myotome, histological transverse sections of 12/101 immunostained embryos were analyzed, revealing that the differences between the injected and uninjected sides are likely related to defects in both myotome and neural plate morphogenesis (Fig. 3M,N). In early neurogenesis, the myotome is initially ventral to the neural plate. As the presomitic mesoderm and lateral deep neural plate cells elongate, the neural folds are pushed up and later fuse, and the myotome becomes positioned lateral to the neural tube. In a section of the more severely affected embryo (Fig. 3M), the uninjected side shows the myotome is lateral to the neural fold that has folded up to the midline. The injected side resembles an earlier developmental stage with the more flattened neural plate and the ventral myotome position. In addition, the myotubes on the uninjected side are cut transverse, but the myotubes on the injected side are cut longitudinally (Fig. 3M,P), consistent with younger, unsegmented somites (Hamilton,1969). In the less severely affected embryo (Fig. 3N), FMO1 injection led to decreased levels of 12/101 staining compared with the uninjected side and the myoceolic space is still visible on the injected side. Overall FMO1 injection produces a delay in development of the somites and neural tube. However, at later stages, the neural tube appears to have formed normally as evidenced by normal immunostaining of FMO1-injected embryos at stage 34 for the neural marker NCAM (Supp. Fig. S3). The persistence of improper myotome segmentation later on in development suggests a somitogenesis defect rather than simply a developmental delay.

FMO Injection Affects pax3 and myoD Expression Domains

To determine whether the alterations in myotome development caused by FMO injection were due to alterations in expression of critical muscle transcription factors, we performed in situ hybridizations on late neurula (St. 18–22) and tadpoles (St. 34–36) with pax3 and myoD probes. In the somites pax3 is a marker of the proliferating cells of the outer dermomyotome, a laterally positioned epithelial cell layer that contributes to the growing myotome (Fig. 3P). At late neurula stage, by whole-mount examination, the pax3 staining appeared decreased on the injected side of the embryos in the dermomyotome but not in the prospective neural crest (Fig. 4A; 20 ng of FMO1, 81%, n = 21). Consistent with the more lateral location of the edge of the neural plate shown on the developmentally delayed injected side in sections (Fig. 3M), the line of pax3 staining was displaced laterally from the midline on the injected side. Thus, the apparent decreased pax3 staining may be the result of a delay in development of the dermomyotomal layer. In support of this, transverse sections of FMO1-injected embryos shows expression of pax3 on the injected side in the dermomyotome region, but they are dispersed and not organized into a thin line of cells as on the uninjected side (Fig. 4C, arrowheads).

Ultimately, expression of MYOD is required for myogenic differentiation. At stage 20, myoD expression was decreased on the FMO1-injected side consistent with a decrease in the amount of differentiating cells in the myotome. Transcripts from these cells remain predominantly near the nuclei in the center of individual somites, as seen on the uninjected side (Fig. 4D), myoD was still expressed in the FMO1-injected side, despite being unable to discern individual somites, indicating that either the cells within the somites do not have their nuclei aligned or that intersomitic boundaries have not been properly established. Decreased and diffuse myoD staining was found in 100% (n = 12) and 57% (n = 14) of embryos injected with 40 ng and 20 ng of FMO1, respectively (Fig. 4F). At stage 34, FMO injection does not affect levels of myoD expression within the epaxial myotome; however, because myoD staining is highest near the nuclei of the myotubes, it is clear on the injected side that the epaxial muscle nuclei are not aligned properly (Fig. 4J,K). Staining for myoD revealed segmentation defects in 53% (n = 19) and 100% (n = 21) embryos injected with 20 ng and 40 ng, respectively, of FMO1 (Fig. 4O).

During the tadpole stages analyzed, the expanding epaxial and hypaxial myotomes are marked by expression of muscle determination factor myf5 at the dorsomedial and ventrolateral tips (Hopwood,1991). Embryos injected with either 40 ng or 20 ng of FMO1 showed normal expression levels of myf5 localized properly at the dorsal and ventral tips of the myotome (72%, n = 25 and 100%, n = 19, respectively), despite that many embryos clearly showed a decreased myotome width (Fig. 4P). Therefore, the decrease in myotome growth in the FMO1-injected embryos is not due to a block in the ability of pax3 progenitors to differentiate into the myogenic lineage.

FMO Injection Affects Hypaxial Muscle Formation and Decreases Mesenchymal Tissue

Hypaxial muscle forms from a de-epithelialization at the ventral–lateral lip of the myotome leading to the ventral migration of muscle cell precursors. The absence of myoD in the hypaxial region on the injected side (Fig. 4J–L) is consistent with the lack of hypaxial muscle development as visualized with the 12/101 mAb (Fig. 3F). At stage 34, pax3 staining is also decreased within the migrating hypaxial muscle precursors on the FMO1-injected side (Fig. 4G–I) and can be seen at the ventral edge of the myotome, but proper delamination of these hypaxial muscle precursors has not occurred. Overall, 20 ng (67%, n = 27) and 40 ng (100%, n = 27) of FMO1 morphant embryos exhibited partial to complete absence of hypaxial muscle precursors as visualized with pax3 and myoD staining (Fig. 4O).

Dermomyotome de-epithelialization events also form the mesenchymal dermatome lateral to the myotome (Scaal and Christ,2004). Using vimentin as a mesenchymal cell marker (Dent et al.,1989), FMO1-injected embryos were analyzed for the distribution of this intermediate filament protein by immunostaining. Vimentin-expressing mesenchymal cells normally found on the lateral surface of the somites are missing from the FMO1-injected tadpoles (Fig. 4M,N). Interestingly the vimentin-expressing pronephric mesenchyme cells are also absent in the FMO1-injected tadpoles. The pronephric mesenchyme is derived from the lateral mesoderm and is the precursor to the pronephrous. A lack of pronephric tubules and an abnormal duct were seen in sections of stage 26 FMO1-injected embryos; however, the head mesenchyme appeared normal. Therefore, FMO1 appears to specifically inhibit the formation of mesenchyme cells from lateral mesoderm and somites.

Elevated Developmental FRG1 Levels Lead to Altered Epaxial Muscle Morphology

FRG1 is critical to the proper growth of the myotome; thus, the effects of overexpression of FRG1 on epaxial muscle development were analyzed. Analysis of differentiated muscle by 12/101 immunostaining for embryos overexpressing FRG1 showed altered epaxial morphology. At 500 pg (50%, n = 34) and 1 ng (67%, n = 73), injected frg1 mRNA the epaxial muscle on the injected side was not uniformly stained due to gaps between fibers, nonparallel misaligned muscle fibers, and variation in fiber size and shape (Fig. 5A,B,E,F compared to 5C,D,G,H). Background levels in mRNA tracer-injected embryos of this phenotype were negligible (3%, n = 34). The anterior somite region of the injected side was observed to be slightly more narrow then the uninjected side, however, much less severe then in FMO injections. To understand how the cells within the epaxial muscle were organized, coronal and transverse histological sections from 12/101 stained embryos were analyzed. In coronal sections, the FRG1 overexpressing side displayed darker staining, potentially representing more differentiated cells, with a large variation of fiber sizes and shapes leading to the appearance of disrupted somite organization (Fig. 5J). Additionally, rounded intensely stained cells were seen laterally separated from the myotome. Similar separation of intensely stained cells was observed in transverse sections (Fig. 5K–M). Anterior somite transverse sections also displayed decreased dorsal–ventral myotome length and increased medial–lateral width on the injected side (Fig. 5K,L). Anterior somites displayed an expanded region of lightly stained cells. In addition, the space between the myotome and the surface ectoderm into which detached 12/101 positive cells appear to have migrated was larger on the frg1-injected side and contained an increased number of unstained, possibly mesenchymal cells (Fig. 5K). As the sections move more posterior, the injected side no longer has these aberrantly expanded regions (Fig. 5L,M). The ventral region of the somite is bifurcated growing both medially and laterally (Fig. 5M) rather then progressing ventrally. The loose, detached organization of myofibers from the myotome potentially indicates a loss of normal cell–cell contacts. Thus, both knock down and overexpression of FRG1 resulted in defective growth and morphogenesis of the myotome indicating that precise levels of FRG1 must be maintained for normal muscle development.

Figure 5.

Elevated FRG1 causes epaxial muscle abnormalities. Stage 36 tadpoles injected with 1 ng of frg1 were immunostained with12/101 antibody. A–H: Injected sides displayed uneven staining (A,B,E,F) compared to the uninjected sides (C,D,G,H). I–M: These embryos were then sectioned either sagittally (J) or transversally (K,L,M), with relative positions depicted in I. The injected sides are denoted by “i”. N,O: Stage 20 embryos were immunostained with 12/101 (N) and sectioned transversally (O). P: A graphic summery of epaxial muscle abnormalities for 500 pg of frg1, 1 ng of frg1, and tracer is shown.

Elevated FRG1 Leads to Expansion of Myotome and Neural Plate

Overexpression of FRG1 was analyzed in late neurulation when muscle differentiation has not progressed in the more posterior somites. At this stage, an increased width of 12/101 staining was readily apparent on the frg1-injected side (Fig. 5N). However, unlike stage 36, neurulation stage embryos displayed decreased intensity of 12/101 staining with a less defined, diffuse appearance of the somites (Fig. 5N). Transverse sections clearly showed a large expansion of lightly to nonstained cells in the lateral portion of the myotome (Fig. 5O). Additionally, the expansion did not seem to be confined to the myotome as the neural plate on the injected side was also much larger then on the uninjected side.

Elevated FRG1 Leads to Elevated Neural pax3 Levels and Thickened Dermomyotome

To determine whether the morphology of the elevated FRG1 myotome was a result of altered levels of muscle transcription factors, pax3 and myoD were analyzed by in situ hybridization. No alterations in pax3 staining were observed in the dermomyotome of stage 18–22 embryos; however, an expansion of staining was seen in the neural plate (Fig. 6A) in 90% of embryos (n = 32). Similar to 12/101 staining, 100% (n = 15) of myoD stained neurulation stage embryos displayed a lack of proper somite formation and a dispersed staining pattern (Fig. 6B). Of interest, transverse sections of embryos analyzed for pax3 and myoD displayed a thickened region of cells of he dermomyotome (Fig. 6C,D) and both displayed the same increased myotome size at this early stage as was seen when analyzed by 12/101 immunostaining. Thus, at neurulation stages, FRG1 levels positively correlate with myotome size and any alteration of FRG1 expression level adversely affects early somite organization.

Figure 6.

Elevated FRG1 leads to somite and hypaxial muscle disruptions. A,B: Stage 20 embryos injected with 1 ng of frg1 were analyzed by in situ hybridization with pax3 (A) or myoD (B). C,D: Transverse sections of pax3 and myoD stained embryos are shown in C and D, respectively. The injected sides are denoted by “i”. E–J: Somite disruptions and hypaxial muscle were analyzed in stage 36 embryos injected with 500 pg of frg1 (E,I) and 1 ng of frg1 (G) compared to the uninjected sides F, J, and H, respectively, by myoD (E–H) or pax3 (I,J) in situ hybridization. A graphic summary of somite and hypaxial abnormalities for 500 pg of frg1, 1 ng of frg1, and tracer is shown in K.

Elevated FRG1 Causes Abnormal Hypaxial Muscle Formation

Both myoD and pax3 levels appear normal in stage 36 FRG1 overexpressing embryos; however, increasing the levels of frg1 led to somite segmentation errors. Analysis of myoD expression pattern showed dose-dependent segmentation defects in 64% (500 pg, n = 31) and 89% (1 ng, n = 27; Fig. 6E,G, compare to uninjected sides 6F,H) of embryos and altered hypaxial muscle precursor cell migration (Fig. 6E,G,I compare to 6F,H,J). Control injections (Fig. 6K) using the tracer RNA alone represent background levels of hypaxial fusion (15%) and somite segmentation defects (16%). While knock down of FRG1 resulted in a uniformly deformed myotome, elevated FRG1 displayed discrete patches of disrupted organization, usually found in the more anterior somites. Unlike FMO, frg1 mRNA injections typically had hypaxial muscle precursor separation but the individual segments were fused for both 500-pg and 1-ng injections, 67% (n = 49) and 78% (n = 55), respectively (Fig. 6K). As the hypaxial muscle typically delaminates from the ventral–lateral lip of somites, the fusion may be caused by the abnormal somite segmentation.


Although FRG1 overexpression has been a leading candidate for the mechanism mediating FSHD pathology, very few studies on FRG1 have been performed. Here, we have uncovered a role for FRG1 in the development of muscular structures. We found that elevated levels of FRG1 disrupted skeletal muscle, consistent with previous work in an adult mouse model for FSHD (Gabellini et al.,2006) and clinical findings of muscle from FSHD patients consisting with variable fiber size (Fitzsimons et al.,1987; Padberg et al.,1995a,b; Osborne et al.,2007). FRG1 depletion resulted in overall reduced size of the Xenopus embryo. At neurulation stages, elevated and depleted FRG1 had opposing effects on myotome size. A conserved growth function for FRG1 is supported by RNAi knockdown of C. elegans FRG1 (ZK1010.3) resulting in slow growth (Kamath and Ahringer,2003). Together these results strongly suggest an evolutionarily conserved function for FRG1 in growth.

Beyond cellular growth, many of the defects observed in this study involve cellular transitions between epithelial and mesenchymal states. The Xenopus frg1 overexpression phenotype displayed improper lateral detachment of muscle cells from the myotome, suggesting that increased FRG1 is promoting delamination and migration, even in tissues where this process should not be taking place. Similarly, whereas hypaxial muscle precursors normally delaminate from the ventral–lateral edge of the somite in distinct blocks, in the FRG1 overexpression model, the entire lateral edge of the myotome appears to be undergoing delamination and the hypaxial myoblast migration is disorganized. Conversely, FMO-injected embryos show an inability to delaminate hypaxial muscle. Cell migration is dependent on the permissiveness of the environment, which involves extracellular matrix components such as hyaluronan. In Xenopus, the morpholino knock down of XHas2, a hyaluronan biosynthesis enzyme, produced a similar phenotype to the FRG1 knock down, displaying disrupted segmentation, less muscle differentiation, and inhibition of hypaxial muscle formation (Ori et al.,2006). Further study will reveal FRG1's role in the cells' ability to delaminate and migrate during development.

Both overexpression and depletion of FRG1 resulted in abnormal somite segmentation suggesting FRG1 may function in the mesenchymal to epithelial transition (MET) that occurs during early somite formation and/or in the maintenance of the epithelial characteristics of the dermomytome portion of the somite (Hay,2005). Addition of differentiated muscle to the epaxial myotome occurs by de-epithelialization from the dorsomedial myotome lip (Hay,2005). If depletion of FRG1 reduces the ability of cells to de-epithelialize, the addition of muscle cells to the myotome would be inhibited. Indeed, we observed a reduction in the amount of muscle cells in the myotome on the FRG1 morphant embryos. A related process, where an epithelial to mesenchymal transition (EMT) from the dermomyotome contributes mesenchyme to the formation of connective tissue, was also affected by FMO injection as seen by the overall reduction of vimentin, a marker of mesenchymal cells. Thus, proper MET and EMT both have a requirement for FRG1.


A role for FRG1 overexpression in FSHD is still controversial due in part to inconsistencies between gene expression studies on FSHD muscle and cell lines, the lack of information on the precise function of FRG1, and recent studies supporting potential roles for alternative candidate genes. Our study strongly supports FRG1 misexpression as being capable of producing the musculature pathology seen in FSHD and provides evidence for the critical role of FRG1 in vertebrate muscular development.


Frog Husbandry

Adult X. laevis were purchased from Xenopus Express. All procedures were carried out in accordance with established UIUC IACUC approved protocols for animal welfare.

Probe Constructs

PCRs for cloning used Triplemaster polymerase enzyme mixture (Eppendorf); RT-PCRs used SuperScript III HiFi one-step RT-PCR kit (Invitrogen); restriction enzymes were purchased from New England Biolabs. All oligonucleotide primers are listed in Supplementary Table S1. X. laevis total RNA was extracted from ovary with Trizol (Invitrogen) as per manufacturer's instructions and used for RT-PCR (primers #1/#2 for frg1, #3/#4 for myoD, #5/#6 for pax3, and #7/#8 for myf5) to produce cDNA. All products were cloned into pGEM T-Easy (Promega) and sequenced. Plasmid pcDNA tdTomato was made by digestion of ptdTomato-N1 (GIFT) and pcDNA3.1 (Invitrogen) with NheI/EcoRI and ligation. Plasmid pEGFP-N1 (Clonetech) was similarly cloned into MS2 vector. EGFP was amplified by PCR (#9/#10) from pEGFP-N1 (Clonetech) and cloned into pGEM. EGFP was then digested out with NheI/SacI and cloned into MS2 digested with XbaI/SacI.

In Situ Hybridizations

Embryos were staged according to Nieuwkoop and Faber (1994), fixed 1–2 hr in MEMFA (0.1 M MOPS pH 7.4, 2 mM ethylenediaminetetraacetic acid [EDTA], 3.7% Formaldehyde), washed 2 × 30 min in 100% methanol and stored in 100% methanol at −20°C until use. The frg1, myoD, pax3, and myf5 cDNA pGEM clones were linearized with NcoI and NdeI and transcribed with SP6 or T7 RNA polymerase incorporating digoxigenin (DIG) -11-UTP (Roche Diagnostics) to generate antisense or sense RNA probes. In situ hybridizations were performed according to standard methods (Harland,1991) and detected with alkaline phosphatase (AP) linked anti-DIG antibody (Roche Diagnostics) and the chromogenic substrates BCIP (5-bromo-4-chloro-3-indolyl phosphate, toluidine salt) and NBT (nitro blue tetrazolium chloride; Roche Diagnostics). Embryos were fixed overnight in Bouin's fixative, followed by washing in 70% ethanol/30% phosphate buffered saline (PBS) -Tween 0.1%, and pigment was removed by treatment for 1 to 2 hr in 1% H2O2, 5% formamide, and 0.5× standard saline citrate (SSC) under bright light. Embryos were then washed in methanol 10 min and transferred to 1 mM EDTA in PBS or glycerol for analysis and photography.

X. laevis FRG1 Antibody

A rabbit polyclonal antibody for X. laevis FRG1 was generated (GenScript Corp) against a unique peptide sequence EREAKRDDDIPNED near the FRG1 C-terminus. Anti-serum was affinity purified against immobilized peptide and specificity was confirmed by Western blotting against whole embryo extract (Supp. Fig. S3). Before use in whole-mount immunostaining the antibody was preabsorbed against a broad range of embryo stages to remove background polyclonal epithelial staining.


Embryos were staged and fixed as above, rehydrated in PBS-DT (1% dimethyl sulfoxide [DMSO], 1% Tween-20) and washed for 15 min in PBS-DT. Samples were blocked in 0.1 M glycine, 2% milk, 1% bovine serum albumin, 1% Tween-20 and 1% DMSO for 4 hr at room temperature or overnight at 4°C. Primary antibodies were diluted in blocking solution as follows: Skeletal muscle marker (12/101) 1/2, Vimentin 1/20, or XL FRG1 1/200 and incubated with embryos overnight at 4°C. 12/101 was detected using a horseradish peroxidase (HRP) secondary (GE Healthcare) with a diaminobenzidine staining kit from (Roche). Vimentin antibody was detected with AP-conjugated goat anti-mouse secondary antibodies (Jackson Immuno Research) (1/5,000) and the FRG1 antibody was detected with AP-conjugated purified donkey anti-rabbit (Jackson Immuno Research) (1/10,000) and developed as above for in situ hybridizations. The 12/101 and vimentin monoclonal antibodies, developed by Jeremy P. Brockes and Michael Klymkowsky, respectively, were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences (Iowa City, IA).

RNA and MO Microinjections

pcDNA3.1-tdTomato, MS2-EGFP, and MS2-xt-frg1 were linearized with AvrII, EcoRI, and AflII respectively, and in vitro transcribed with T7 (tdTomato) or SP6 (EGFP and xt-frg1) mMessage mMachine (Ambion) respectively. Xt-frg1 mRNA was injected at 0.5 pg or 1 ng, with 500 pg of tdTomato or 1 ng of egfp mRNA coinjected. Control tdTomato or egfp mRNA injections were performed at 500 pg and 2 ng, respectively. Fluoroscein-labeled FMO1 and FMO2 were designed against the MGC84293 within the 5′untranslated region through the first nine coding nucleotides (Fig. 2F), and fluorescein-labeled standard control MO were purchased from Gene Tools, LLC. Microinjections were performed on two cell embryos in 1× MMR with 3% Ficoll and incubated at 19°C. Approximately 3–6 hr after injection, embryos were transferred to 0.1× MMR with 3% Ficoll. After 24–36 hr, embryos were either peeled and fixed for stage 18–22 embryos or cultured in 0.1× MMR until the desired stage. After neural tube closure, all injected embryos were sorted based on left, right, or bilateral fluorescence. Embryos displaying gastrulation defects or inconsistent developmental abnormalities were removed from the analysis.


All authors contributed to the conceptual design of experiments, interpretation of the data, and writing of the manuscript. M.L.H. and R.D.W. performed all experiments. We thank Jon Henry and Phil Newmark, UIUC, for technical support. The authors declare no competing financial interests. P.L.J. was funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and M.L.H. was funded by the FSH Society Landsman Charitable Trust Fellowship.