Dystrophin and Dysferlin Double Mutant Mice: A Novel Model For Rhabdomyosarcoma

Abstract

While researchers are yet to establish a link a between muscular dystrophy (MD) and sarcomas in human patients, literature suggests that MD genes dystrophin and dysferlin act as tumor suppressor genes in mouse models of MD. For instance, dystrophin deficient mdx and dysferlin deficient A/J mice, models of human Duchenne Muscular Dystrophy and Limb Girdle Muscular Dystrophy type 2B, respectively, develop mixed sarcomas with variable penetrance and latency. To further establish the correlation between MD and sarcoma development, and to test whether a combined deletion of dystrophin and dysferlin exacerbates MD and augments the incidence of sarcomas, we generated dystrophin and dysferlin double mutant mice (STOCK-Dysf^prmd Dmd^mdx-5Cv). Not surprisingly, the double mutant mice develop severe MD symptoms and moreover develop rhabdomyosarcoma at an average age of 12 months, with an incidence of > 90%. Histological and immunohistochemical analyses, using a panel of antibodies against skeletal muscle cell proteins, electron microscopy, cytogenetics, and molecular analysis reveal that the double mutant mice develop rhabdomyosarcoma. The present finding bolsters the correlation between MD and sarcomas, and provides a model not only to examine the cellular origins but also to identify mechanisms and signal transduction pathways triggering development of RMS.

INTRODUCTION

Rhabdomyosarcoma (RMS) is a malignant mesenchymal neoplasm that typically involves soft tissues and shows evidence of skeletal muscle differentiation[1]. RMS is morphologically categorized into three distinct subtypes: alveolar, embryonal, and pleomorphic [2, 3]. The alveolar subtype occurs more often in adolescents and young adults and is associated with specific chromosomal translocations [t(2:13) and t(1:13)], involving PAX3-FOXO1 and/or PAX7-FOXO1 fusions [4]. The embryonal subtype is the most common soft tissue tumor in pediatric populations, typically arises from undifferentiated mesoderm and occurs most commonly in axial and visceral locations, particularly in the head, neck and genitourinary tract[3, 5–7]. The pleomorphic subtype occurs almost exclusively in adults with rare pediatric exceptions and is thought to arise from myotome derived skeletal muscle or satellite cells[2, 8].

The precise etiology of RMS is unclear, however X-ray exposure in the first trimester of pregnancy significantly increases the risk for RMS [9], and mutations in the TP53 tumor suppressor gene [7, 10], RAS oncogene [11] and cyclin dependent kinase inhibitor genes [12] are detected in childhood RMS. A more recent study found that mice with muscular dystrophy (MD) harbor genomic instabilities that could predispose them to malignant muscle-derived tumors [13].

MD genes including dystrophin, dysferlin, and calpain 3 act as tumor suppressor genes. Although a direct epidemiologic correlation between MD and RMS is not evident in human patients, mouse models of MD, including dystrophin deficient (Dmd^mdx), dysferlin deficient SJL/J and A/J, and calpain 3 null (Capn3^−/−) mice develop sarcomas albeit at different incidence rates [7, 13, 14]. Not surprisingly, double mutant mice are more vulnerable to sarcomas. Single mutant Dmd^mdx, Capn3^−/−, and Dysf^−/− show incidence rates of 39%, 23% and 5%; and double mutant Dmd^mdx Dys^−/− and Dmd^mdx Capn3^−/− show incidence rates of 47% and 44%, respectively. Dystrophic muscle has a genetic signature similar to that of sarcomas leading one to anticipate that combined lack of sarcolemmal proteins accelerates tumor formation [13].

Here, we generated mice doubly deficient in dystrophin and dysferlin on a mixed B6 and A/J strain background. The F2 mice develop a severe MD phenotype and display a high incidence (>90%) of RMS starting at ~8 months of age mainly involving the front and rear limbs. Histologic, immunohistochemical, ultrastructural, cytogenetic and molecular analyses reveal that the double mutant mice have RMS. This model has clinical significance as the double mutant mice on B6 and A/J background largely develop RMS with high penetrance and short latency, and further the chromosomal translocations found in the RMS cells can be valuable in elucidating the mechanisms and/or identifying genes in human RMS.

MATERIALS AND METHODS

Animals

C57BL/6J, A/J, and B6Ros.Cg-Dmd^mdx-5Cv/J mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were bred and maintained under SPF conditions at The Jackson Laboratory. Food (LabDiet® 5K52/5K67, Richmond, Indiana; 6% fat, autoclaved) and acidified water were provided ad libitum. The A/J strain lacks dysferlin due to a spontaneous retrotransposon insertion within intron 4 of the dysferlin gene resulting in a null mutation (Dysf^prmd) at the dysferlin locus that is fixed in this strain[15]. The Duchenne muscular dystrophy (Dmd) locus is on the X Chromosome while the dysferlin gene maps to Chromosome 6. We initiated the cross by mating female B6Ros.Cg-Dmd^mdx-5Cv/J to male A/J mice. The F1 offspring were intercrossed and progeny that genotyped as double heterozygous were used to generate a stock of mice that were null at both alleles. To identify mutant dysferlin and dystrophin genes, PCR genotyping was performed as described previously[15, 16]. These mice were designated STOCK-Dysf^prmd Dmd^mdx-5Cv. Mice were monitored for development of MD and tumors. Tumor bearing mice were euthanized and subjected to standard necropsy procedures. The Animal Care and Use Committee at The Jackson Laboratory approved all the experimental procedures performed in this study.

Cell culture reagents and antibodies

PBS, DMEM, penicillin-streptomycin and HyClone fetal bovine serum (FBS) were obtained from Thermo Scientific (Waltham, MA). TrypLE™ Select and DMSO were obtained from Invitrogen (Carlsbad, CA), and Sigma-Aldrich (St. Louis, MO), respectively. Antibodies to desmin (ab6322) and dysferlin (ab75571) were purchased from Abcam (Cambridge, MA). Ki67-SP6 was obtained from Lab Vision (Fremont, CA). Antibodies to dystrophin (NB120-3149) and smooth muscle actin (SMA) were purchased from Novus Biologicals (Littleton, CO). Antibody to mouse Myog (556358) was purchased from BD Pharmingen (San Diego, CA). Vectastain Elite ABC kit (Rabbit IgG), Vectastain M.O.M. peroxidase kit and ImmPACT DAB peroxidase substrate were purchased from Vector Laboratories (Burlingame, CA). Trp53 mouse monoclonal antibody (2524) and anti-mouse secondary (7076) were purchased from Invitrogen (Carlsbad, CA).

Isolation and culture of RMS cells

Tumors excised from STOCK-Dysf^prmd Dmd^mdx-5Cv mice were finely minced using scalpel blades in PBS containing penicillin-streptomycin solution. The tissues were incubated in TrypLE™ Select for 15 minutes at room temperature (RT). Cells were then resuspended in DMEM containing 10% FBS and penicillin-streptomycin (culture medium), transferred to a 15 ml tube and centrifuged for 5 minutes at 1200 rpm. Cell supernatant was discarded, the pellet was resuspended in 1ml culture medium and cells were counted using a Beckman coulter counter (Miami, FL). Cells seeded at a density of 5 × 10⁶ in T75 tissue culture flasks were grown in a humidified chamber at 37°C. Aliquots of the primary tumors were cryopreserved in 10% DMSO and 90% FBS.

Histology

Preparation of specimens for histology and immunohistochemistry was performed as previously described [17].

Immunohistochemistry

Deparaffinized and hydrated slides were subjected to antigen unmasking using 10mM sodium citrate buffer (pH 6.0). Briefly, slides were heated in a microwave to boiling temperature in sodium citrate buffer for 20 minutes and then cooled at RT for 30 minutes before blocking the endogenous peroxidase activity with 3% H₂O₂ for 15 minutes at RT. Slides were washed with TBST (2.42g Trizma base and 8g sodium chloride to 1L dH20; 0.1% Tween-20; pH 7.6) and blocked with 10% goat serum at RT for 1h. A Vector M.O.M immunodetection kit was used to stain for Myog (1:50), desmin (1:100), and dysferlin. The Vectastain Elite ABC kit was used for Ki67, SMA and dystrophin staining. Tissue sections were incubated with primary antibodies either overnight at 4°C (myogenin, dysferlin and dystrophin) or for 1h at 37°C (Ki67 and SMA), followed by incubation with biotinylated secondary antibodies. Slides were washed twice in TBST and incubated with ABC reagent, followed by another wash with TBST for 10 minutes, and application of DAB peroxidase substrate and hemotoxylin counterstain.

Electron Microscopy

RMS tumors were retrieved from paraffin sections [8] and fixed overnight in 2% glutaraldehyde and the procedure described previously [17] was followed.

Spectral Karyotyping (SKY)

Cryopreserved RMS cells were thawed and grown in 75cm² culture flasks in a humidified chamber at 37°C for 48h. Metaphase spreads were prepared from cultures incubated with fresh medium containing 50 µg/mL of colchicine for 45 minutes at 37°C. SKY was performed as previously described [18] prior to counterstaining with DAPI.

RNA isolation and qPCR

Total RNA from C57BL/6J gastrocnemius muscle and from STOCK-Dysf^prmd Dmd^mdx-5Cv RMS tumors was isolated according to the manufacturer’s instructions using the Qiagen RNeasy mini kit (Valencia, CA). The Agilent 2100 Bioanalyzer was used to determine the quality and concentration of total RNA. The MessageSensor™ RT Kit (Austin, TX) enabled cDNA synthesis. The following pre-made primers and probes for 18S rRNA (Mm03928990-g1), Igf2 (Mm00439564_m1), Vcan (Mm01283063_m1), Pparg (Mm01184322_m1), Myo1e (Mm01173838_m1), Cdkn2a (Mm00494449_m1), Met (Mm01156972_m1), Nf1 (Mm00812424_m1), Pax3 (Mm00435493_m1), Pax7 (Mm03053796_s1), Foxo1 (Mm00490671_m1), Myod1 (Mm00440387_m1), Myog (Mm00446195_g1), Ptch1 (Mm00436026_m1) and Rb1 (Mm00485586_m1) were purchased from Applied Biosystems (Carlsbad, CA). The Comparative Cycle Threshold (ddCt) method was used to calculate relative changes in gene expression. The PCR cycle involved an initial 2 minutes hold at 50°C, and then a 10 minutes hold at 95°C, followed by 40 cycles of denaturation at 95°C for 15sec and primer annealing for 1min at 60°C.

Data analysis

Kaplan-Meier analysis and Student’s t-test were performed using GraphPad Prism v4 (GraphPad Software, San Diego, CA, USA).

RESULTS

Muscular dystrophy and RMS in dysferlin and dystrophin double mutant mice

A loss of function mutation in either the dysferlin or the dystrophin gene results in MD[15, 19]. Dystrophin and dysferlin deficient (STOCK-Dysf^prmd Dmd^mdx-5Cv) F2 mice developed pronounced MD symptoms as expected from their genetic constitution. In congruity with a recent study in which dystrophin and dysferlin deficient double mutant mice developed various skeletal muscle derived tumors[13], we observed that lack of dysferlin increases the incidence of spontaneous RMS development in adult (~1 year old) dystrophin-deficient STOCK-Dysf^prmd Dmd^mdx-5Cv F2 mice. The tumors predominantly involved the upper rear and front limbs (one tumor per mouse) and occurred at a median age of 53 weeks (Fig 1). The incidence of RMS in double mutant mice was greater than 90% (18 of 19 F2 mice). One female mouse without a visible lump was found dead at 404 days of age.

Figure 1.

Kaplan-Meier analysis showing tumor-free survival among 18 mice that developed RMS. We followed 19 STOCK-Dysf^prmd Dmd^mdx-5Cv mice from weaning onwards to tumor development. Gross lesions (lumps) were first observed starting at 32 weeks. Eighteen mice developed RMS, whereas one female mouse without visible tumor was found dead. There was no gender bias in RMS formation.

Histologically, MD was characterized by myofiber degeneration, myofiber necrosis, inflammation, myofiber atrophy, and myofiber regeneration with centrally placed nuclei (Fig 2). The tumors often appeared adjacent to foci of regeneration within dystrophic muscle and featured infiltrative arrays, fascicles and sheets of pleomorphic neoplastic cells that varied from round to polygonal to fusiform. Elongated rhabdomyoblasts with eosinophilic cytoplasm and a row of central nuclei (tadpole or strap cells) and multinucleated giant cells were frequently noted. Areas of tumor necrosis and inflammation were also observed. In some sections, the tumors displayed certain features of the spindle cell variant of embryonal RMS such as densely packed bundles of elongated fusiform cells with cigar shaped nuclei, prominent nucleoli and eosinophilic cytoplasm. In other areas, the tumors featured hallmarks common to pleomorphic RMS such as haphazardly oriented and loosely arranged pleomorphic, round or often large and bizarre cells with more eosinophilic cytoplasm (Fig 2). The tumors lacked conspicuous characteristics of alveolar RMS, such as ill-defined aggregates of poorly differentiated round cells with the formation of alveolar spaces separated by fibrovascular septa. To further characterize the RMS in double mutant mice, immunohistochemistry and cytogenetic analyses were performed.

Figure 2.

Muscular dystrophy and RMS in STOCK-Dysf^prmd Dmd^mdx-5Cv mice (a 47-week-old male (a, d and i), a 48-week-old female (b, c and e) and a 36-week-old male (f, g and h). Dystrophic muscle displays marked variation in myofiber size, areas of myofiber degeneration and necrosis (arrow) (a), inflammatory cell infiltrates (thick arrows), atrophic myofibers (arrow head), internalized nucleus (asterisk) and regenerating fibers (b), foci of fibrosis (star) and dystrophic mineralization (M) (c). The dystrophic skeletal muscle (DM) is infiltrated by RMS (RMS) (d), which consists of aggregates of round, fusiform and polygonal cells with occasional multinucleated giant cells (slender arrows) (e) and tadpole or strap cells - elongated rhabdomyoblasts with eosinophilic cytoplasm and a row of central nuclei (S) (f). Tumor necrosis (N) is evident in some fields (g). The RMS features densely packed bundles of fusiform cells with cigar shaped nuclei (asterisk) in some sections (h) and loosely arranged pleomorphic to round cells with eosinophilic cytoplasm and occasional bizarre nuclei (arrowheads) in other areas (i). H&E stain; original magnification, ×50 (d); ×400 (a, b, c, e, f, g, h and i). Scale bars represent 50µm.

Positivity for markers of skeletal muscle differentiation

Myog and Myod1 are markers of skeletal muscle differentiation [20, 21] that are expressed during the early embryonic stage and play major roles in skeletal muscle development [22]. They are commonly used as markers for diagnosis of RMS. On immunohistochemical analysis, RMS cells from double mutant mice exhibited positive staining for the skeletal muscle markers Myog, and staining for the proliferation-related Ki-67 antigen (Fig 3A). Tumors were negative for smooth muscle actin (SMA) and indeed for sarcolemmal expression of dysferlin, but positive for the skeletal muscle marker desmin and paired-box protein 7 (Pax7) (Fig 3B).

Figure 3.

Immunohistochemistry of RMS. A. a, d, g, and j depict normal skeletal muscle from a 22-week-old female C57BL/6J control mouse. b, e, h and k illustrate skeletal muscle from a 13-week-old male A/J control mouse. c, f, i, and l depict RMS from a 47-week-old, male STOCK-Dysf^prmd Dmd^mdx-5Cv mouse. Neoplastic cells exhibit strong Ki-67 staining (c), and modest cytoplasmic (nonspecific/artifactual) and nuclear Myog staining (f). There is no sarcolemmal dysferlin staining in skeletal muscle from the dysferlin-deficient A/J mouse (h) or in neoplastic cells and adjacent skeletal muscle (i). Arrows indicate skeletal muscle and arrowhead indicates sarcolemmal dysferlin expression. B. Skeletal muscle (arrow) and neoplastic cells are strongly positive for desmin (j), negative for SMA (k), and modestly positive for Pax7 (l). Immunohistochemistry: Anti Ki-67, a–c; Anti Myog, d–f; Anti-Dysferlin, g–i; Anti-Desmin, j; Anti-SMA, k; Anti-Pax7, l. Original magnification, ×400.

Ultrastructural features of rhabdomyosarcoma

On ultrastructural evaluation we found evidence of cytoplasmic filament aggregates, parallel arrays of filaments with sarcomere-like organization, and nodes of electron dense material suggestive of Z-band formation (Fig 4). The sarcomere-like arrays were short and haphazardly oriented and were suggestive of rudimentary or abortive sarcomeric differentiation. These features are reiterative of those seen in rhabdomyosarcomas, especially of the pleomorphic and embryonal subtypes.

Figure 4.

Electron microscopy of RMS; (a) depicts cytoplasmic aggregates of thick and thin filaments (f); (b) depicts short and haphazardly oriented rudimentary sarcomere-like structures (s) and nodes of electron dense material indicative of Z-band formation (z). Transmission electron micrographs; original magnification, ×8000 (a); ×25000 (b). Scale bars represent 2 µm (a) and 50nm (b).

Absence of characteristic chromosomal translocations

In addition to immunohistochemistry, cytogenetic analysis is used to distinguish different RMS variants. RMS cells typically exhibit variable copy numbers of chromosomes in different cells [4, 23]. Of the different human RMS subtypes, only alveolar RMS is associated with specific chromosomal translocations, t(2:13) and t(1:13), resulting in PAX3-FOXO1 and PAX7-FOXO1 fusions, respectively [4]. In contrast, embryonal and pleomorphic RMS have no consistent chromosomal translocations and show rather complex karyotypes [24, 25] including extra copies of Chromosomes 2, 8 and 13[26]. Many cases of embryonal RMS are also associated with loss of heterozygosity involving multiple loci at the 11p15.5 region [27] containing a number of imprinted genes, including insulin-like growth factor-2 (IGF2), cyclin-dependent kinase inhibitor 1C (CDKN1C) and H19.

When we applied spectral karyotyping to metaphase spreads of normal and RMS cells to determine genomic instability, there were no chromosomal translocations similar to those that are characteristic of alveolar RMS. Moreover there was no evidence of Pax-Foxo1 fusion gene products. The data here was gathered from examination of spectral karyotype (SKY) images from 51 metaphase spreads isolated from three different RMS tumors. Chromosomes 2 and 14 showed greater than 90% of translocations [t(2:11), t(2:14), t(2:16), t(14:2), t(14:7) and t(14:11)] and Chromosomes 7 and 9 were associated with greater than 50% of translocations [t(7:14), t(7:16), t(9:6) and t(9:11)] in two of three tumors analyzed. Moreover, somatic Chromosomes 2, 6, 8, 10, 13 and 14, and the X Chromosome had on average more than 2 copies, with averages ranging from 2.5 copies per chromosome to 4.125 copies (trisomy, tetrasomy and pentasomy) (Fig 5). We also observed predominant monosomies involving Chromosomes 3, 9, 10, 11, 13, 16, 17, 18 and 19. Our SKY karyotyping suggests that these RMS cells carry multiple variable chromosomal translocations and table 1 summarizes karyotype analyses from three different STOCK-Dysf^prmd Dmd^mdx-5Cv mice.

Figure 5.

Spectral karyotype analysis of RMS tumors in STOCK-Dysf^prmd Dmd^mdx-5Cv mice (32-week-old female (B) and 59-week-old female (C)). Representative metaphase spreads from C57BL/6J (A) and STOCK-Dysf^prmd Dmd^mdx-5Cv (B and C), which are representative of two independent tumors, reveal that RMS tumors harbor complex chromosomal abnormalities: monosomy of Chromosomes 3, 4, 16,19 and X; trisomy of Chromosomes 1, 5, 6, 8, 13 and 18; tetrasomy of Chromosomes 3, 4, 6, 10, 11, 12, 14, 15, 19 and X; pentasomy of Chromosomes 2 and 17, and translocations t(2;4), t(2;16), t(9;15), t(14;15), t(14;11), t(14;18), 2t(16;4) and t(18;16). Slender arrows indicate aneuploidy and stout arrows indicate chromosomal translocations.

Table 1.

Chromosomal translocations in RMS tumors from three different STOCK-Dysf^prmd Dmd^mdx-5Cv mice.

Translocations involving chromosome(s)	RMS#1 (%) n=30	RMS#2 (%) n=13	RMS#3 (%) n=8
Chr 14	100	0	100
Chr 2	93.75	5.3	100
Chr 7	62.5	0	75
Chr 9	56.25	0	87.5
Chr 18	50	0	37.5
Chr 4	50	0	12.5
Chr 16	43.75	0	100
Chr 11	37.5	5.3	75
Chr 3	37.5	5.3	12.5
Chr 15	18.75	0	37.5
Chr 1	18.75	0	0
Chr 6	0	0	25
Chr 13	0	0	12.5

n indicates the number of metaphase spreads analyzed per tumor. The numbers in the table represent the percentage of metaphase spreads in individual tumor cells showing translocations associated with specific chromosomes.

Interestingly, a recent study suggested that patterns of genomic instability common to soft tissue sarcomas are present in dystrophic skeletal muscle from mice without overt tumors. But some of these muscles purportedly contained foci of incipient tumors, thus implying an association of these genomic changes with a pre-neoplastic process. Most notable among these genetic changes were increases in copy numbers of chromosomes 8 and 15[13]. These changes along with the absence of chromosomal translocations are noticeably different from the cytogenetic profiles of tumors in our study.

Expression of markers of embryonal RMS

In addition to immunohistochemistry, electron microscopy and cytogenetics, we performed molecular analysis of RMS cells for further characterization of the tumors. We examined total RNA extracted from RMS tumors (n=6) and from C57BL/6J normal gastrocnemius muscle (n=3) for the expression of paired box transcription factors Pax3 and Pax7, forkhead box O1 (Foxo1), Myog and Myod1. Expressed in skeletal muscle satellite cells, both Pax3 and Pax7 are essential for myogenic development [28, 29]. Alveolar RMS variants typically show significant increases in the mRNA levels of Pax3, Pax7 and fusion products of Pax3/Pax7-Foxo1, whereas only Pax7 message is increased in embryonal RMS cases [30]. Recently, Goldstein et al. reported overexpression of both PAX3 and PAX7 in primary human embryonal RMS [31]. Consistently, we found that both Pax7 and Pax3 were overexpressed by more than 10-fold when compared to C57BL/6J controls and normalized to 18S rRNA, suggesting that both variants show considerable upregulation of satellite cell markers. Expectedly, we detected a dramatic increase in the transcript levels of the skeletal muscle markers Myog and Myod1 (Fig 6).

Figure 6.

Quantitative RT-PCR analysis of B6 normal muscle tissue (n=3) and STOCK-Dysf^prmd Dmd^mdx-5Cv RMS tumors (n=6, including four female and two male mice ranging from 33 to 64 weeks of age). Assessed mRNA fold change of RMS tumors was normalized to the corresponding C57BL/6J control. 18S rRNA served as an internal control. Data represent mean ± sem. *p<0.05; **p<0.01; ***p<0.001.

To test if the tumors expressed markers for liposarcoma and fibrosarcoma we measured transcript levels of peroxisome proliferator activated receptor gamma (Pparg), myosin IE (Myo1e), and versican (Vcan). There was a subtle but significant decrease in the levels of Vcan and Pparg transcripts in comparison to wild type controls. This observation concurred with the absence of fibrosarcoma or liposarcoma on morphologic evaluation. Furthermore, there was no evidence of amplification of oncogene Met or loss of cyclin-dependent kinase inhibitor 2A (Cdkn2a) or neurofibromatosis 1 (Nf1). Notably, there was a significant increase in the message for insulin-like growth factor-2 (Igf2). There was no change in the transcript levels for Foxo1, tumor suppressor gene patched homolog 1 (Ptch1) and retinoblastoma (Rb1) (Fig 6).

DISCUSSION

Mutations in dystrophin and dysferlin genes cause Duchenne (DMD) and Limb Girdle type 2B (LGMD2B) muscular dystrophy, respectively. While dystrophin, a sarcolemmal protein, with other components of the dystrophin-glycoprotein complex (DGC), connects the intracellular actin cytoskeleton with the extracellular matrix [32] and contributes to membrane integrity, dysferlin regulates membrane repair at the site of injury [33]. The dysferlin gene, a member of the ferlin family, encodes a type II transmembrane protein that has a C-terminal domain anchored to the plasma membrane and six conserved C2 domains in the cytoplasm. Interaction of C2 domains with phospholipids regulates vesicle trafficking and membrane fusion of vesicles at the plasma membrane disruption site [34], which is impaired in dysferlin null mice. Damage to the plasma membrane in dysferlin competent mice causes an increase in dysferlin expression that helps seal the membrane at the site of injury [35]. In addition, dysferlin regulates calcium entry at the disruption site [32]. Literature suggests that mouse models of MD develop RMS. Dystrophin-deficient mice develop RMS, although at an older age (>16 months) and with a low incidence (<10%) [36]. Female dysferlin-deficient A/J mice show a moderate (34%) to a very low (12.5%) incidence of RMS [37, 38]. Recently, Sher et al. found that by 20 months of age A/J mice develop a higher frequency of pleomorphic RMS (79% of females and 70% of males) [14]. Our data demonstrates that dystrophin and dysferlin double mutant mice develop RMS with high penetrance and short latency. Taken together, these results suggest that disruption of DGC or impairment of calcium homeostasis results in RMS tumorigenesis. Further, a combined lack of dystrophin and dysferlin accelerates MD symptoms and augments tumor formation.

One possible explanation for high penetrance and short latency of RMS in double mutant mice could be that the calcium imbalances might regulate proliferation of mature myoblasts [39, 40]. A significant mediator of pathogenesis in dystrophin-deficient mdx mice is stretch-induced damage of dystrophic muscles [41]. The direct consequences of this damage include activation of stretch-gated calcium channels, and consequently a prolonged increase in intracellular calcium and production of reactive oxygen species. Given that repairing of a disrupted skeletal muscle membrane is a calcium-dependent process [42], loss of dysferlin in dystrophin deficient mdx mice could possibly exacerbate muscle degeneration from dysregulated calcium homeostasis and membrane fusion. Notably, the association between RMS and hypercalcemia is not rare [43]. Another protein involved in membrane repair and cytoarchitecture along with dysferlin is calpain 3. Mutations in calpain 3, a calcium-dependent muscle specific cysteine protease, cause Limb Girdle Muscular Dystrophy type 2A (LGMD2A) [33]. Similar to dysferlin deficiency in dystrophin deficient mdx mice, lack of calpain 3 in mdx mice increases the incidence of sarcomas [13]. Further investigation is needed to establish whether calcium-dependent mechanisms underlie abnormal proliferation of myoblasts and transformation to RMS.

Conversely, lack of dysferlin might negatively impact apoptotic signaling in RMS cells. Apoptosis is compromised in a tumor microenvironment, allowing proliferating cells to evade programmed cell death. Although apoptosis contributes to myofiber degeneration in dystrophin-deficient and sarcoglycan-deficient mice [44, 45], it is not clear whether apoptosis mediates MD in dysferlin-deficient mice. De Luna et al. examined the functional role of dysferlin in normal and pathological human satellite cells and reported that dysferlin is upregulated in activated satellite cells, and further, using amiloride, an inhibitor of myoblast fusion, the authors suggested that expression of dysferlin is increased in differentiated multinucleated fibers or myotubes compared to undifferentiated myoblasts [46]. Therefore it is possible that lack of dysferlin in different cell types including activated satellite cells might disrupt the apoptotic machinery, and contribute to aberrant proliferation of satellite cells, whose primary function is to replace the damaged muscle fibers. A zinc finger transcription factor Fos regulates apoptotic signaling. Trp53, Fos double knockout mice develop RMS with 90% incidence, however ectopic expression of Fos not only increases apoptosis but also decreases expression of muscle markers such as Myod1, Myog and Pax7 [47]. Future studies need to determine whether dysferlin plays a similar role to influence apoptosis.

RMS can be a consequence of aberrant proliferation of a subpopulation of skeletal muscle progenitor cells, or it can emanate from mesenchymal progenitor cells [48], given that myogenic cells arise from embryonal mesoderm. To identify the cellular origins of RMS, Rubin et al. investigated the effect of conditional deletion of Trp53 and/or retinoblastoma 1 (Rb1) with or without patched homolog 1 (Ptch1) haploinsufficiency in satellite cells as well as in proliferating and differentiating myoblasts [49]. They found that loss of Trp53 in differentiating myoblasts rather than satellite cells gives rise to a higher proportion of embryonal RMS [49]. Based on the evidence presented by Rubin et al., it is likely that lack of dystrophin and dysferlin results in aberrant proliferation of maturing myoblasts in the degenerating muscle, and RMS emanates from this subset of cells. In support of this hypothesis, the more severe dystrophy in the double mutant mice and subsequent increased rate of myoblast proliferation correlates with the earlier age of onset and increased incidence of RMS compared to that in mice singly mutant for dystrophin or dysferlin. We also noticed significant increases in transcript levels of Pax3 and Pax7, implicating that in addition to differentiating myoblasts a subpopulation of activated satellite cells/proliferating myoblasts might also contribute to RMS [49]. It should be noted that Pax3 and Pax7 are markers of satellite cells and are downregulated in differentiating myoblasts.

In summary, this study shows that dysferlin and dystrophin double mutant mice on a segregating B6 and A/J background presented RMS that shares immunohistochemical and genetic features with the human embryonal and pleomorphic variants. There is no indication of alveolar RMS, as confirmed by morphologic, cytogenetic and molecular analyses. While we notice RMS, Schmidt et al. and Rubin et al. reported mixed sarcomas in double mutant MD mice [13, 49]. The differences in the tumor type and incidence observed in our study can be attributed to differences in strain specific backgrounds. The double mutant mice described herein provide a mouse model for high incidence of RMS with relatively short latency, and further studies in these mice can prove useful to identify cellular origins and novel treatment strategies such as gene therapy for RMS.