Myotonic dystrophy 1 (DM1) is a multi-systemic autosomal dominant genetic disease leading to a variety of neuromuscular disorders such as myotonia, muscle weakness/wasting, and cardiac conduction defects and cataracts as associated symptoms (Harper,2001). The genetic basis for DM1 is expansion of an unstable CTG triplet repeat sequence within the 3′ untranslated region of the myotonic dystrophy protein kinase gene (DMPK; Brook et al.,1992; Fu et al.,1992; Mahadevan et al.,1992). DMPK is a serine/threonine protein kinase containing coiled-coil, C-terminal membrane association, and autoregulatory domains. DMPK-1 is the primary translation product of full-length DMPK mRNA and a lower molecular mass isoform, DMPK-2, is derived from DMPK-1 by a proteolytic processing event (Bush et al.,2000).
While the physiologic function of DMPK is unknown, multiple lines of evidence have revealed decreased DMPK expression in DM1. DMPK RNA transcripts containing CUG expansions accumulate as nuclear foci (Taneja et al.,1995), reducing the expression of DMPK in affected skeletal muscle. DMPK RNA transcripts with expanded CUG repeats (1) sequester mRNA transcripts containing CAG repeats and essential RNA binding proteins such as CUG-BP1 and the human muscleblind protein homolog MBNL and (2) mediate the alternative splicing of numerous genes (Timchenko et al.,2001; Fardaei et al.,2002; Sasagawa et al.,2003). However, RNA processing defects alone cannot explain the pathophysiology of DM1, because animal models based on altered expression of CUG-BP1, MBNL, or expanded triplet repeats do not recapitulate all symptoms of the disease (Kanadia et al.,2003; Timchenko et al.,2004; Ho et al.,2005; Orengo et al.,2008; Yadava et al.,2008).
Until the function of DMPK is understood, a role for reduced DMPK levels in DM1 pathophysiology must be considered. A reduction in DMPK mRNA abundance has been observed in skeletal muscle of DM1 patients (Carango et al.,1993; Krahe et al.,1995; Wang et al.,1995). RNA studies have demonstrated that, in heterozygous DM1 individuals with one normal and one expanded allele, both DMPK alleles were transcribed into pre-mRNA. However, when mature poly(A) mRNA was examined, DMPK mRNA abundance was only 10–20% of normal (Morrone et al.,1997).
The most severe form of myotonic dystrophy, congenital DM1, is associated with numerous developmental defects including delayed muscle development. The muscle fibers of congenital DM1 patients are unfused or incompletely fused throughout development (Furling et al.,2003). Cultured myoblasts from congenital DM1 fetuses have a reduced capacity to fuse and differentiate and produce less than 50% of the normal levels of DMPK protein (Furling et al.,2001,2003), suggesting DMPK might have a specific role in embryonic myocyte development. The comparison of DM1 and DM2 (an expansion of repeats in the zinc finger protein-9 gene; Liquori et al.,2001) further suggests a key role for DMPK activity. DM1 and DM2 are caused by expanded nucleotide repeats in two different genes (Cho and Tapscott,2007). Both diseases exhibit RNA processing defects that account for a majority of DM symptoms (Mankodi et al.,2001; Fardaei et al.,2002; Orengo et al.,2008; Yadava et al.,2008). However, only DM1 manifests as a congenital disease (Day et al.,2003), suggesting a specific role for DMPK in embryonic development.
Because of the progressive muscle weakness in DM1, previous investigations of DMPK expression have focused on mature myocytes. DMPK expression has been reported to be maximal in skeletal and heart tissue, localizing to the neuromuscular junction and intercalated disc, respectively (Maeda et al.,1995), but is also found in other tissues including the ocular lens and brain. DMPK mRNA expression has been observed by in situ hybridization in tissues in which DMPK protein expression was not detected by previously available antibodies (Sarkar et al.,2004). We have recently produced a highly specific and sensitive monoclonal antibody against the coiled-coil region of DMPK. The precise specificity of this monoclonal antibody has been carefully documented (Helmke et al.,2006). Briefly, surface plasmon resonance, immunoprecipitation, MALDI-TOF mass spectrometry, Western blotting, and immunolabeling confirmed high specificity and affinity. We have thus demonstrated that the monoclonal anti-DMPK antibody specifically and reproducibly binds only DMPK with high affinity.
Genetic analysis of DMPK in mice has failed to reveal the function of the kinase in DM1 pathology. DMPK knockout mice develop a mild progressive myopathy sharing minor pathological similarities with myotonic dystrophy (Reddy et al.,1996,2002; Mounsey et al.,2000; Wansink and Wieringa,2003); however, this phenotype does not recapitulate the pathology of DM1. Transgenic mice expressing DMPK from the M creatine kinase promoter and enhancer have a hypertrophic cardiac myopathy and minor skeletal muscle defects similar to those observed in DMPK knockout mice (Jansen et al.,1996). Therefore, knockout and transgenic mice exhibit similar skeletal muscular phenotypes, suggesting DMPK is necessary for normal muscle function and is also sufficient to alter myocyte physiology.
A human DMPK family member, LATS1, regulates mitosis and cytokinesis in mammalian cells (Tao et al.,1999; Yang et al.,2004; Bothos et al.,2005); however, LATS1 knockout mice are viable and display mild phenotypes (Yang et al.,2004). The regulation of mitosis and cytokinesis by DMPK family members is clearly a complex process and involves several potentially compensatory pathways. The fact that DMPK knockout mice are viable does not preclude a role for DMPK in development. Furthermore, previous studies in LATS1 suggest DMPK may also have roles that are undetected in standard genetic knockout strategies and can only be discovered through in vitro analysis. We hypothesized that DMPK has a key role in the regulation of proliferation and differentiation in developing myocytes. Our studies have revealed specific expression of DMPK in postmitotic mammalian and avian embryonic myocytes. We have also observed a migration of DMPK during cultured cardiac myocyte maturation. DMPK overexpression is sufficient to disrupt sarcomeric structure and induce apoptosis in cardiac myocytes. Furthermore, we have discovered DMPK is necessary for the differentiation of cultured skeletal myoblasts and is sufficient to induce apoptosis.
Embryonic DMPK Expression
DMPK is expressed in developing mouse myocytes and cultured myogenic cell lines.
Due to the lack of a specific antibody against DMPK, expression studies have previously focused on only mRNA expression patterns. Furthermore, the symptoms of DM1 have directed a majority of expression analysis in mature adult tissues (Sarkar et al.,2004). However, the symptoms of congenital DM1 suggest a role for DMPK in embryonic development. We, therefore, analyzed DMPK expression during murine embryogenesis by immunohistochemistry. DMPK is expressed in E10.5, E12.5, and E14.5 somites (Fig. 1A,B and data not shown). Skeletal muscle fibers distinctly express DMPK at E14.5 and E16.5 (Fig. 1D,E). DMPK expression is also observed in cardiac myocytes at E10.5, E12.5, E14.5, E16.5, and E18.5 (Fig. 1C and data not shown). At E16.5, DMPK expression is also seen in the interdigital region of the limb, which undergoes apoptosis during embryogensis. The expression pattern of DMPK not only parallels the organ systems affected in DM1, but also suggests a role for DMPK in development. Negative control immunolabeling without primary antibody had no detectable staining (data not shown).
Due to the distinct expression of DMPK during myogenesis, we were curious whether DMPK was also expressed in immortalized myogenic cell culture lines. Expression of both full-length DMPK (DMPK-1) and cleaved DMPK (DMPK-2) was observed in all myogenic cell lines analyzed, whether grown under proliferating or differentiating conditions (H9C2, C2C12, BC3H1, Fig. 1G).
A DMPK homolog is expressed during chick embryogenesis in myocytes and postmitotic cells.
To confirm the DMPK expression pattern observed in another vertebrate model, we used chicken embryos. Furthermore, immunostaining in the chick embryo allow specific staining with less background than resulted when using a mouse antibody on mouse sections. It was first necessary to confirm DMPK expression in the chick. Western blotting with our monoclonal DMPK antibody confirmed the presence of a DMPK-like protein in the chick (Fig. 2A). The chick homolog migrates at the same molecular weight as mammalian DMPK (70 kDa). Two protein bands were observed, confirming processing of chick DMPK into two isoforms in the chick: DMPK-1 and DMPK-2. Western blotting confirmed the expression of chick DMPK throughout embryonic development in Hamburger- Hamilton stage (St.) 12, 18, 29, and 36 embryos (Fig. 2A).
To determine the pattern of DMPK expression during chick development, we stained sectioned embryos with the monoclonal DMPK antibody (Fig. 2B–G). Overall, the expression of chick DMPK mirrored DMPK expression in the murine embryo. However, specific DMPK staining possible in the chick embryo revealed distinct patterns of DMPK expression in the vertebrate embryo. DMPK is expressed in the St. 17 and St. 22 embryonic notochord, aorta, neural tube, floorplate, and dermamyotome (Fig. 2B–D). These tissues represent the first postmitotic cells in the chick embryo (Kahane and Kalcheim,1998). DMPK expression was also detected in the mouse notochord, neural tube, and ocular lens (data not shown). This distinct expression pattern suggests DMPK has a role in the transition of proliferating cells into postmitotic, differentiating cells.
Our studies revealed a distinct pattern of DMPK expression in many postmitotic cells. In addition to early floor plate expression, DMPK is expressed in St. 22 cardiac myocytes (Fig. 2E), skeletal muscle (Fig. 2F), and St. 26 ocular lens (Fig. 2G). The expression of DMPK in the developing lens demonstrated expression restricted to postmitotic, differentiated lens cells. The well-characterized transition region in the developing lens serves as a distinct boundary of DMPK expression (Friedman et al.,1989). Parallel immunostaining without primary antibody revealed no positive staining, confirming the specificity of our staining pattern (data not shown). Similar expression patterns of DMPK in murine and avian embryos suggest a highly conserved role for DMPK in development. Conserved expression patterns between the mouse and chicken embryo further confirms the distinct expression patterns observed. Moreover, the specific expression of DMPK in many postmitotic cell populations suggests a novel role for DMPK during cell differentiation.
In summary, DMPK is specifically expressed in cardiac and skeletal myocytes during murine and chick embryogenesis. Analysis of the avian dermamyotome revealed restricted DMPK expression in postmitotic myocytes. Specific postmitotic DMPK expression is also observed in the neural tube and lens, suggesting a more global role for DMPK in cellular differentiation. Because of the specific myocyte defects observed in adult and congenital DM1, we further explored the function of DMPK in cardiac and skeletal myocytes.
DMPK Localization and Function in Cardiac Myocytes
Changes in subcellular localization of DMPK in mammalian cardiac myocytes.
Our immunohistochemical analysis revealed expression of DMPK in developing mouse cardiac myocytes. To elucidate the developmental role of DMPK, we were interested in determining kinase localization in cardiac myocytes during development. Immunofluorescent imaging of cultured neonatal rat cardiac myocytes revealed distinct localization of DMPK at the nuclear envelope (Fig. 3A). Previously, perinuclear localization of myc-tagged human DMPK was also reported after the adenoviral infection of neonatal rat cardiac myocytes (Kaliman et al.,2004). However, this subcellular localization pattern did not persist into adult cardiac myocytes. In cultured adult mouse cardiac myocytes, DMPK localizes to the cellular membrane (Fig. 3C). Treatment of neonatal cardiac myocytes with leukemia inhibitory factor (LIF) promotes characteristics of differentiation in neonatal rat cardiac myocytes, including myocyte elongation and mature sarcomere assembly (Nicol et al.,2001). When neonatal cardiac myocytes are treated with 1,000 U/ml LIF for 24 hr, DMPK begins to migrate away from the nucleus toward the cellular membrane (Fig. 3B). DMPK migration was observed in 91.2% ± 3.3% of neonatal rat cardiac myocytes after LIF addition (n = 3 experiment replications), but was not observed in any untreated cells analyzed. Together, these data suggest a distinct shift in DMPK subcellular localization during murine myogenesis.
Changes in subcellular localization of DMPK in differentiating chick cardiac myocytes.
Our analysis of cultured mammalian cardiac myocytes suggested a migration in DMPK localization during myocyte development. To confirm this shift, we tracked DMPK localization in a cardiac culture system supporting cardiac myocyte differentiation in vitro. Chick cardiac myocytes present an ideal culturing model of myocytes that more completely differentiate in culture than mammalian myocytes. Cardiac myocytes from 6-day chick embryos were cultured for 11 days, and DMPK localization was assayed by immunofluorescent imaging (Fig. 4). In chick myocytes cultured for 2 days, DMPK was found specifically around the nuclear envelope, similar to previous observations in neonatal rat cardiac myocytes. From days 4 to 7 of maturation, DMPK migrated away from the nuclear envelope toward the cell membrane. This migration pattern mimicked what was observed in neonatal rat cardiac myocytes after LIF addition. Within 11 days of differentiation, DMPK was found at the cell membrane of mature chick cardiac myocytes. This mature localization pattern resembled the localization pattern of adult mouse cardiac myocytes. Our analysis of DMPK in chick cardiac myocytes confirmed a shift in DMPK localization during cardiac myocyte differentiation. Furthermore, these changes suggest the possibility of varied roles for DMPK during myogenesis. Different localization patterns of this kinase may represent different roles for the kinase throughout the maturation of cardiac myocytes.
Overexpression of DMPK in chick cardiac myocytes leads to disruption of sarcomeres and apoptosis.
To elucidate the role of DMPK in developing cardiac myocytes, we overexpressed tagged forms of DMPK in cultured chick cardiac myocytes (Fig. 5). We transfected chick cardiac myocytes with green fluorescent protein (GFP) -tagged DMPK lacking C-terminal membrane association and autoinhibitory domains (DMPKΔMA). DMPKΔMA corresponds to DMPK-2 and possesses increased enzymatic activity as compared to full-length DMPK (Bush et al.,2000). We examined chick cardiac myocytes cells expressing DMPKΔMA-GFP, full-length DMPK-GFP (FL-DMPK-GFP), or GFP alone by confocal microscopy. Within 24 hr of transfection, sarcomeric structure is disrupted in chick cardiac myocytes (Fig. 5A). Immunofluorescent imaging of α-actinin showed normal striated sarcomeres when GFP is transfected; however, tagged DMPKΔMA and FL-DMPK disrupted normal α-actinin staining patterns. Sarcomeric structure was lost and aggregations of α-actinin were found in DMPK overexpressing cells. Sarcomeric disruption was confirmed by phalloidin staining of F-actin. Overexpression of DMPK induces the formation of large F-actin foci, similar to the α-actinin foci seen in Figure 5 (data not shown). DMPK overexpression (DMPKΔMA-GFP and FL-DMPK-GFP) caused sarcomere disruption in 95.1% ± 3.3% of chick cardiac myocytes (n = 5); however, transfection of GFP alone only caused sarcomere disruption in 5.4% ± 3.3% of cells (n = 5). Clearly DMPK is sufficient to influence cardiac sarcomeric structure, suggesting a possible role for DMPK in sarcomere formation, stabilization, or maintenance.
The structural changes induced by DMPK overexpression in chick cardiac myocytes are accompanied by an induction of apoptotic cell death after 48 hr. Terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling (TUNEL) staining revealed apoptosis in a majority of cell transfected with DMPKΔMA-GFP (Fig. 5B). Transfection with DMPKΔMA-GFP caused TUNEL staining in 91.9% ± 4.8% chick cardiac myocytes (n = 3). GFP expression, however, did not induce TUNEL staining in any transfected cells analyzed (n = 3). This data demonstrates DMPK is sufficient to disrupt cardiac sarcomeres and can induce apoptotic cell death. Strong interdigital expression of DMPK in the mouse limb (Fig. 1F) suggests a developmental role of DMPK in apoptosis induction. These cell culture studies confirm a pro-apoptotic role for DMPK. We propose the tight regulation of DMPK protein levels and subcellular localization may control varied roles for DMPK in vertebrate development: myogenesis, sarcomere regulation, cardiac maturation, and pro-apoptosis.
In summary, DMPK localization dramatically shifts during the differentiation of both mammalian and chick cardiac myocytes. Furthermore, DMPK is sufficient to disrupt cardiac sarcomere structure and cause cell death in chick cardiac myocytes. We were next interested in determining the role of DMPK in skeletal myoblast differentiation.
DMPK Localization and Function in Skeletal Myoblasts
Overexpression of DMPK in chick skeletal myoblasts causes cell rounding and apoptosis.
Immunohistochemical staining showed DMPK expression in mouse and chick skeletal muscle fibers. To manipulate DMPK levels in fusing myoblasts, we cultured primary skeletal myoblasts from 13-day chick embryos. Transfection of skeletal myoblasts with DMPK revealed the kinase is sufficient to cause cell rounding and apoptosis (Fig. 6A). Transfection of chick myoblasts with GFP did not affect cell morphology; however, a 24-hr transfection with DMPKΔMA-GFP or FL-DMPK-GFP induced cell rounding, indicative of apoptosis. Circular morphology was only observed in 9.7% ± 5.0% of chick skeletal myoblasts after GFP transfection (n = 3); however, transfection with either DMPK construct induced circular morphology in 79.4% ± 2.4% of cells (n = 3). Positive TUNEL staining confirmed the induction of apoptosis after DMPKΔMA-GFP overexpression (Fig. 6B). TUNEL-positive nuclei were observed in 4.7 %± 2.6% of GFP-transfected cells (n = 3) and in 97.2% ± 2.8% of DMPKΔMA-GFP-transfected cells (n = 3). In summary, overexpression of DMPK is sufficient to alter cell morphology and induce programmed cell death in primary skeletal myoblasts.
Endogenous DMPK localization in immortalized mouse myoblasts.
Our studies of chick skeletal myoblasts suggested DMPK has a key role in myoblast differentiation. We were interested in further characterizing this role in a more accessible culture model. We, therefore, examined DMPK in C2C12 immortalized mouse myoblasts, C2C12 cells. During C2C12 cell division, we observed a change in subcellular DMPK localization (Fig. 7). In interphase cells, DMPK was found throughout the cytoplasm, with aggregation around the nuclear envelope. At the onset of mitosis, DMPK is found throughout the cell, with highest concentration at the cell membrane. This membranous localization of DMPK persists throughout mitosis (metaphase, anaphase). Once cell division commences, DMPK is localized to both the cell membrane and the reforming nuclear envelope. The dramatic shift in DMPK localization from the nuclear envelope to the cell membrane resembles those changes observed during cardiac myocyte differentiation. This suggests DMPK migration is associated with both myocyte differentiation and cell division.
DMPK overexpression induces cell rounding and apoptosis in C2C12 myoblasts.
Our immunohistochemical staining detected DMPK expression in developing embryonic muscle cells, suggesting that DMPK may have a role in regulation of myocyte differentiation. To investigate the role of DMPK in myocyte development, we manipulated DMPK levels in C2C12 mouse myoblasts. We examined C2C12 cells expressing DMPKΔMA-GFP, FL-DMPK-GFP, or GFP alone by confocal microscopy. We observed induction of apoptosis in a majority of transfected cells. Within 24 hr of gene transfection, C2C12 cells rounded and lifted from the cell culture plate. A small number were observed before detachment (Fig. 8A). Circular morphology was observed in 83.4% ± 2.3% of DMPK-transfected cells (n = 5), whereas circular morphology was only observed in 2.5% ± 2.5% of GFP-transfected cells (n = 5). TUNEL-positive staining confirmed DNA fragmentation indicative of apoptosis (Fig. 8B). TUNEL-positive nuclei were observed in 3.5% ± 1.7% of GFP-transfected C2C12 cells (n = 3) and in 91.7% ± 3.0% of DMPKΔMA-GFP-transfected cells (n = 3). Similar to what was seen in primary chick myoblasts, overexpression of DMPK is sufficient to alter cell morphology and induce programmed cell death in C2C12 myoblasts.
Depletion of DMPK in C2C12 cells prevents myotube differentiation.
Our immunohistochemical staining detected DMPK expression in developing embryonic muscle cells. To observe whether DMPK was necessary for myoblast differentiation, we conducted RNAi-mediated knockdown experiments in C2C12 cells. We obtained an shRNA construct capable of depleting DMPK transcripts (shDMPK, SuperArray). This knockdown plasmid contains GFP, allowing identification of transfected cells. We confirmed a 36.7 ± 5.5% reduction (n = 4) of DMPK by Western blotting of fluorescence-activated cell-sorted GFP-positive cells (Fig. 9B). This reduction is comparable to the level of DMPK protein reduction observed in congenital DM1 myoblasts (Furling et al.,2001,2003).
We used confocal microscopy to examine the effects of knocking down DMPK in myoblasts. C2C12 cells depleted of DMPK did not undergo normal differentiation into myotubes. As can be seen in Figure 9A, myoblasts transfected with GFP or a random shRNA construct (shControl) fused into myotubes after incubation in differentiation media. However, a majority of cells transfected with shDMPK did not fuse into myotubes, but instead exhibited a disorganized cytoskeleton and abnormal cellular processes. Uncharacteristic cellular processes were observed in 82.1% + 3.8% C2C12 myoblasts after DMPK depletion (n = 6). This suggests that DMPK is necessary for myoblast differentiation and cytoskeletal organization.
Molecular analysis of C2C12 cells confirmed the necessity of DMPK for myocyte differentiation. Reduction in DMPK leads to a significant decrease in myogenin after 2 days of differentiation (Fig. 9C). Densitometric quantification confirmed a 23.4 ± 5.3% reduction (n = 5) in myogenin after DMPK depletion. However, expression of the early differentiation markers MyoD, p21, and Rb remained unchanged. Thus, the morphological changes observed after shDMPK transfection correlate with reduced expression of myogenic genes in C2C12 cells. These data support the hypothesis that DMPK is necessary for the coordinated differentiation of C2C12 myoblasts into myotubes.
In summary, DMPK is necessary for skeletal myoblast differentiation. The depletion of DMPK in mouse myoblasts prevents myotube formation and specifically reduces the expression of myogenin. Furthermore, the overexpression of DMPK causes cell rounding and apoptosis in both chick and mouse myoblasts.
Our studies have identified a novel role for DMPK in myogenesis. The adult symptoms of DM1 have focused much of the previous research to define DMPK function on mature striated muscle. Because of this focus, the possible roles of DMPK in development have been overlooked. For the first time, we have reported DMPK protein expression during skeletal and cardiac myogenesis. The lack of a suitable antibody against DMPK has prevented the previous identification of this specific pattern of DMPK protein expression. Previous research has revealed DMPK mRNA is expressed in several tissue types including skeletal muscle, brain, and lens and was reported to be most abundant in cardiac muscle (Maeda et al.,1995; Whiting et al.,1995; Jansen et al.,1996). Our specific antibody challenges the previously reported tissue distribution and suggests that DMPK protein expression is widespread in developing mammalian muscle cells.
During embryogenesis, DMPK is specifically expressed in postmitotic cell populations, including muscle cells, differentiated lens cells, the notochord, and neural floor plate. This specific expression pattern suggests DMPK has a role in the transition to or stabilization of the postmitotic state during cell differentiation. However, DMPK is expressed in proliferating cell culture lines (Fig. 1). Therefore, DMPK expression alone is not sufficient to stop cell proliferation.
We observed a shift in the subcellular localization of DMPK during cardiac myocyte differentiation. A long-standing question in cardiac research has focused on the distinct changes in cell shape and structure during differentiation. It is unknown what initiates the morphological changes to achieve the rod shape of mature cardiac myocytes. During cardiac myocyte maturation, the localization of DMPK dramatically shifts from the nucleus to the cell membrane. The addition of LIF to neonatal rat cardiac myocytes induces changes in both cell morphology and DMPK localization, as DMPK localization changes with the maturation of cardiac myocytes.
We have also observed that overexpression of DMPK initiates immediate changes in cardiac and skeletal myocytes. Hence, DMPK is sufficient to change myocyte morphology. When DMPK is overexpressed in chick cardiac myocytes, sarcomere organization is lost. DMPK overexpression induces cell rounding and apoptosis in chick skeletal myoblasts and C2C12 cells. Our data thus support the supposition that DMPK has a role in the regulation of cell morphology during myogenesis.
Changes in DMPK subcellular localization not only occur during cardiac maturation, but also during skeletal myoblast division and migration. We observed a shift in DMPK from the nucleus to the cell membrane during mitosis of C2C12 myoblasts. Furthermore, DMPK strongly localized to the leading edge of migrating chick myoblasts and C2C12 cells (data not shown). This reveals that DMPK migration is a part of normal cellular function, and not only cell shape changes observed during cardiac differentiation. The aggregation of DMPK at the outer edges of migrating, fusing, and dividing myocytes suggests DMPK could have an undetected role in cytoskeletal rearrangement. It is worth noting that homozygous loss of the DMPK family member Warts in cells of the adult fly leads to cellular hypertrophy and hyperplasia, suggesting that DMPK could also regulate cellular morphogenesis and proliferation (Justice et al.,1995). Studies of the Hippo/Warts signal transduction cascade have further implicated the DMPK family member Warts in the pathway linking cytoskeletal dynamics to cell proliferation (Morisaki et al.,2002; Iida et al.,2004; Edgar,2006). In addition, a human DMPK family member, LATS1, also promotes cytokinesis (Bothos et al.,2005). The study of these DMPK family members further suggests DMPK may function in the regulation of cell morphology and proliferation.
The induction of apoptosis in myoblasts is consistent with the premature expression of myogenic genes in immature cells. Myocyte differentiation is characterized by ordered developmental stages that are replicated in vitro by cultured myoblasts (Walsh and Perlman,1997; Kitzmann and Fernandez,2001; Shiokawa et al.,2002). Myogenesis initiates when proliferating myoblasts first exit the cell cycle and concomitantly express the myogenic factors MyoD and myogenin (Joulia et al.,2003; Shen et al.,2003). Expression of the cdk inhibitor p21 indicates an irreversible stabilization of the postmitotic state (Andres and Walsh,1996; Puri et al.,1997). A large fraction of cells are eliminated by apoptosis during this step in myogenesis. Apoptosis occurs in a subset of unstable cells expressing myogenin but not p21 (Wang and Walsh,1996; Kamradt et al.,2002). The phosphorylation of Rb by p21 coincides with terminal differentiation in myocytes (Porrello et al.,2000).
We hypothesized that DMPK induced apoptosis by inducing the premature expression of myogenic genes in C2C12 cells. However, we did not observe significant changes in myogenic protein levels (myogenin, p21, Rb, MHC, MyoD) after DMPK overexpression (data not shown). Previous studies in our laboratory had demonstrated that DMPK could drive myogenin expression in BC3H1 cells (Bush et al.,1996). Unlike C2C12 cells, BC3H1 cells do not express MyoD, and thus, myogenin expression is sufficient to induce differentiation after DMPK overexpression. However, DMPK overexpression may induce apoptosis by a mechanism unrelated to myogenic differentiation. Expression of DMPK in the interdigital region of the mouse limb suggests a pro-apoptotic role of DMPK in normal embryonic development. Our cell culture analyses have confirmed this pro-apoptotic role for DMPK.
DMPK is clearly necessary for myogenin expression during C2C12 differentiation into myotubes. When DMPK is depleted in C2C12 cells, myogenin expression is absent after 2 days of differentiation. Morphologically, these depleted cells do not fuse to form myotubes, but instead form abnormal cellular processes. These results suggest that the reduction of DMPK observed in congenital DM1 may account for the observed lack of muscle development. Specifically, the muscle fibers of congenital DM1 patients are unfused or incompletely fused (Furling et al.,2003). Our experimentation in C2C12 cells reveals a novel molecular mechanism that may explain the lack of myotube fusion seen in these patients.
The dramatic results we observed in C2C12 cells are somewhat surprising considering the relatively mild muscle phenotypes observed in DMPK knockout mice (Reddy et al.,1996,2002; Wansink and Wieringa,2003). However, genetic redundancy among DMPK family members could explain the development of mild myopathy in homozygous null DMPK mice. However, cell culture analysis has revealed the importance of DMPK for muscle function. A similar inconsistency was previously observed with another DMPK family member, LATS1. While LATS1 knockout mice are viable and have mild phenotypic abnormalities (Yang et al.,2004), cell culture has uncovered important roles for LATS1 in regulation of the cell cycle (Tao et al.,1999; Yang et al.,2004; Bothos et al.,2005).
A plausible molecular mechanism for DM1 is a general defect of RNA metabolism. This would explain the pleiotropic effects and dominant inheritance pattern of the disease. Alternately, a loss of DMPK expression might be responsible for some DM1 symptoms, with haploinsufficiency of DMPK producing dominant inheritance. We propose these models are not mutually exclusive. It is possible that most DM1 symptoms result from dominant-negative effects upon mRNA processing and others from loss of DMPK activity. Current DM1 research has focused on aberrant mRNA processing in DM1 muscle, yet animal models of DM1 based on defective splicing through MBNL or CUG-BP1 cannot fully explain the pathophysiology of DM1. This suggests a decrease in the abundance of DMPK mRNA in the skeletal muscle of DM1 patients may account for some DM1 symptoms. Our data have revealed expression of DMPK during myogenesis. In addition, we demonstrated that DMPK is required for myocyte differentiation. The reduction of DMPK in DM1 would, therefore, be expected to contribute to DM1 pathology.
Tissue Preparation and Staining
Paraffin sections of fixed mouse embryos were obtained from Zyagen (San Diego, CA). Chick embryos were fixed in 4% paraformaldehyde (PFA) for 2 hr before dehydrating in graded ethanol solutions. Dehydrated embryos were incubated in liquid paraffin overnight before embedding. Embryos were sectioned at 7 μm. Sectioned mouse tissue was probed for DMPK using the Mouse-On-Mouse kit (Vector, Burlingame, CA). Chick embryo sections were stained using the Vector ABC Elite kit. For all immunohistochemistry, monoclonal DMPK was diluted at 1:100. Immunohistochemical staining was visualized by using the Vector 3,3′-diaminobenzidine (DAB) kit (Vector). All sections were counterstained with Hematoxylin-2 (Richard-Allan, Kalamazoo, MI) and mounted in Permount (Fisher, Fair Lawn, NJ). Stained sections were photographed on an Olympus IX71 microscope (Center Valley, PA) with an Olympus DP70 camera. Chick embryo lysates were staged (Hamburger and Hamilton,1951), homogenized in RIPA buffer and sonicated before Western blotting (see below). RIPA buffer contained 50 mM Tris (pH 7.5), 150mM NaCl, 1% Triton-X, 0.5% deoxycholate, 0.1% sodium dodecyl sulfate and Complete protease inhibitor cocktail (Roche, Indianapolis, IN).
Primary Cell Culturing
Chick cardiac myocytes were cultured from 6-day embryos as described (Gregorio and Fowler,1995). Hearts were dissected and shaken in 0.045% collagenase type II (Worthington, Freehold, NJ) in Hanks' balanced salt solution (HBSS) for 5 min at 37°C, followed by a 10 second low-speed vortexing. After settling, the supernatant containing cells was saved and new collagense was added to the heart tissue. These steps were repeated 8 times. The first two supernatant fractions were discarded, and fractions 3–8 were centrifuged and resuspended in DMEM with 10% fetal bovine serum (FBS; Sigma, St. Louis, MO) and 100 U/ml penicillin/streptomysin (Hyclone, Logan, UT). Cells were preplated at 37°C for 1 hr to remove fibroblasts. Cardiac myocytes were subsequently centrifuged and resuspended in DMEM with 5% FBS and 1% penicillin/streptomycin. Cells were plated on coverslips coated with 5 μg/ml fibronectin (Fisher). Cells were transfected with Fugene-HD (Roche) at a 7:2 ratio (Fugene:DNA) after 5 days of culturing. We observed an approximate transfection efficiency of 40%. Medium was refreshed every 2 days of culturing.
Chick skeletal myoblasts were prepared from 13-day embryos as described (Machida et al.,2003). Thigh muscle was dissected and digested with dispase (2,000 U/ml; Sigma) for 15 min at 37°C. The cell suspension was centrifuged and resuspended in M-199 medium (Sigma) containing 15% calf serum. The suspended cells were preplated 3 times in uncoated culture dishes for 40 min each to allow fibroblast attachment and plated on gelatin-coated coverslips. M-199 medium was replaced every 2 days. Cells were transfected after 5 days of culturing with Fugene-HD at a 6:2 ratio. We observed an approximate transfection efficiency of 20%.
Neonatal rat cardiac myocytes were isolated as previously detailed (Simpson,1985). Rat cardiac myocytes were plated on gelatin-coated coverslips for 24 hr before manipulation. LIF (Millipore, Billerica, MA) was added to culture medium at a concentration of 1,000 U/ml for 24 hr. Adult mouse cardiac myocytes were isolated as described (O'Connell et al.,2003; Huang et al.,2007). Briefly, hearts were cannulated and perfused with collagenase type II (Worthington). Myocytes were plated on laminin-coated coverslips, and were cultured in MEM with HBSS, 1 mg/ml bovine serum albumin, and 10 mM 2,3-butanedione monoxime. All primary cultures (chick and mammalian) were grown at 37°C in a 5% CO2 tissue culture incubator. All transfections, immunostaining, and imaging were done as described below.
Immortal Cell Culture, Transfection, Staining, and Protein Analysis
All cell culture lines (C2C12, H9c2, BC3H1) were purchased from ATCC (Manassas, VA). Cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM, ATCC) supplemented with 10% FBS, penicillin/streptomysin (100 U/ml) in 5% CO2 at 37°C. H9c2 cells were differentiated into myocytes as previously described in 1.0% FBS with or without 10 nM retinoic acid (RA, Sigma; Menard et al.,1999; Pagano et al.,2004). BC3H1 cells were differentiated in DMEM supplemented with 1.0% FBS (Lathrop et al.,1985). C2C12 cells were similarly differentiated into myocytes as described with 3.0% horse serum (Invitrogen) and 0.4× Insulin-Transferrin-Selenium-G Supplement (Invitrogen; Cooper et al.,2004). C2C12 cells were grown on glass coverslips coated with 0.1% gelatin and were transfected with Fugene-HD at a 4:2 ratio. We observed an approximate transfection efficiency of 60%.
Transfection constructs DMPKΔMA-GFP and FL-DMPK-GFP were created by cloning DMPKΔMA or FL-DMPK (Bush et al.,2000) into pEGFP-C2 (BD Biosciences, San Jose, CA) at EcoRI restriction sites. pEGFP-C2 was transfected in parallel as a control. Short hairpin RNA plasmids (shControl, shDMPK) were purchased from SuperArray (Frederick, MD).
Before staining, cells were fixed in 4% PFA for 5 min and permeabilized in methanol for 5 min. Cells were blocked for 1 hr in 10% horse serum, 1% bovine serum albumin (BSA), and 0.02% sodium azide in PBS before incubation with antibodies in 3.0% BSA in PBS: rabbit anti-β-tubulin (1:100; NeoMarkers, Fremont, CA) or rabbit anti-α-actinin (Sigma; 1:100). Secondary anti-rabbit Texas Red antibodies (Invitrogen) were diluted 1:100. All antibodies were directly diluted from purchased stock reagents. F-actin was stained by a 15-min incubation with Alexa Fluor 488 or Texas Red conjugated phalloidin (Invitrogen; 1:100). Cells were incubated with 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI; 0.5 μg/ml bisBenzimide; Sigma) for 5 min before mounting in Fluoromount (Electron Microscropy Sciences, Hatfield, PA). TUNEL staining was used to detect apoptotic cells (TMR Red In Situ Cell Death Detection Kit, Roche). All fluorescent images were obtained on an Olympus FV1000 confocal microscope with an Olympus 60× PlanApo 60×/1.40 oil immersion objective. Three-dimensional images were reconstructed with Imaris 5.0 (Bitplane, St. Paul, MN). The z-axis was rotated in all images to allow visualization of cell morphology. To quantify events documented by confocal imaging, cells were counted from 8 to 10 high-power (×60) microscopic fields for each experiment. Each experiment was replicated 3 to 6 times (n), as noted. Statistics are represented as an average ± standard error of the mean (SEM).
For protein analysis, transfected cells were sorted on a Becton Dickinson FACSVantage SE flow cytometry system (Franklin Lakes, NJ). Sorted GFP-positive cells were lysed in RIPA and analyzed by Western blot using the Western Breeze kit (Invitrogen). Blots were probed with antibodies against DMPK (mouse monoclonal; 1:4,000), glyceraldehydes-3-phosphate dehydrogenase (GAPDH; mouse monoclonal, Ambion, Austin, TX; 1:8,000), LATS1 (goat polyclonal; Santa Cruz Biotechnology, Santa Cruz, CA; 1:200), myogenin (mouse monoclonal, Santa Cruz; 1:100), MyoD (rabbit polyclonal, Santa Cruz; 1:200), p21 (mouse monoclonal, Pharmingen; 1:500), Rb (mouse monoclonal, BD Biosciences; 1:200). Membranes were washed with TBS-T and incubated with horse radish peroxidase-conjugated secondary antibodies for 1 hr at room temperature: anti-mouse (Western Breeze, Invitrogen), anti-rabbit (1:1,000; Biomeda, Burlingame, CA), anti-goat (1:4,000; Zymed, San Francisco, CA). The blot was again washed with TBS-T before visualization using the Western Breeze (mouse) or ECL Plus Western Blotting Detection System (rabbit and goat; GE Healthcare, Buckinghamshire, UK). Chemilluminescent images were captured and quantified using a VersaDoc 3000 (Bio-Rad, Hercules, CA).
This work was supported by National Institutes of Health Grants HL064136 (M.B.P.) and P20-RR-017662 (A. M. Gerdes) and the South Dakota 2010 Initiative Research Center Program. The authors would like to thank Stephen Armstrong and Timothy O'Connell for helpful discussions.