RESOURCE ARTICLE

Phospholamban overexpression in mice causes a centronuclearmyopathy-like phenotypeVal A. Fajardo1,*, Eric Bombardier1,*, Elliott McMillan1, Khanh Tran1, Brennan J. Wadsworth1, Daniel Gamu1,Andrew Hopf1, Chris Vigna1, Ian C. Smith1, Catherine Bellissimo1, Robin N. Michel2, Mark A. Tarnopolsky3,4,5,Joe Quadrilatero1 and A. Russell Tupling1,‡

ABSTRACTCentronuclear myopathy (CNM) is a congenital myopathy thatis histopathologically characterized by centrally located nuclei,central aggregation of oxidative activity, and type I fiberpredominance and hypotrophy. Here, we obtained commerciallyavailablemice overexpressing phospholamban (PlnOE), awell-knowninhibitor of sarco(endo)plasmic reticulum Ca2+-ATPases (SERCAs),in their slow-twitch type I skeletal muscle fibers to determine theeffects on SERCA function. As expected with a 6- to 7-foldoverexpression of phospholamban, SERCA dysfunction wasevident in PlnOE muscles, with marked reductions in rates of Ca2+

uptake, maximal ATPase activity and the apparent affinity of SERCAfor Ca2+. However, our most significant discovery was that thesoleus and gluteus minimus muscles from the PlnOE mice displayedovert signs of myopathy: they histopathologically resembledhuman CNM, with centrally located nuclei, central aggregation ofoxidative activity, type I fiber predominance and hypotrophy,progressive fibrosis and muscle weakness. This phenotype isassociated with significant upregulation of muscle sarcolipin anddynamin 2, increasedCa2+-activated proteolysis, oxidative stress andprotein nitrosylation. Moreover, in our assessment of muscle biopsiesfrom three human CNM patients, we found a significant 53%reduction in SERCA activity and increases in both total andmonomeric PLN content compared with five healthy subjects,thereby justifying future studies with more CNM patients.Altogether, our results suggest that the commercially availablePlnOE mouse phenotypically resembles human CNM and could beused as a model to test potential mechanisms and therapeuticstrategies. To date, there is no cure for CNM and our results suggestthat targeting SERCA function, which has already been shown to bean effective therapeutic target for murine muscular dystrophy andhuman cardiomyopathy, might represent a novel therapeutic strategyto combat CNM.

KEY WORDS: SERCA, Dynamin 2, Skeletal muscle,Calcium regulation, Congenital myopathy

INTRODUCTIONThe sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) pumpscatalyze the active transport of Ca2+ into the sarcoplasmic reticulum(SR) and play a crucial role in muscle relaxation and themaintenance of resting intracellular Ca2+ [Ca2+]i, which rangesfrom 30 to 100 nM in skeletal muscle (Tupling, 2009; Schiaffinoand Reggiani, 2011). Phospholamban (PLN) is a small, 52-residueSR protein capable of physically interacting with and inhibitingSERCA activity by reducing its apparent affinity for Ca2+ (Moritaet al., 2008). Mouse (Song et al., 2004) and rabbit (Pattison et al.,2008) models overexpressing PLN in slow-twitch (type I) musclefibers (PlnOE) have been generated by attaching a Pln transgene tothe β-myosin heavy chain (β-MHC) promoter, which preferentiallydirects high levels of expression of type I skeletal-muscle-specifictransgenes (Rindt et al., 1993, 1995; Knotts et al., 1994). In thePlnOE rabbit, obvious signs of muscular dystrophy, includingcentral nucleation, severe muscle wasting, fibrosis, fatty infiltrationand muscle weakness, were observed in soleus muscles, which arerich in type I fibers (Pattison et al., 2008). In contrast, PlnOE miceexhibited soleus muscle atrophy, but no other signs of myopathywere visualized under the light microscope (Song et al., 2004).

We purchased these commercially available PlnOE mice and theirwild-type (WT) littermates (000067-MU, FVB/N background,Mutant Mouse Regional Resource Centre, Columbia, MO) to assessSERCA function because the effect of PLN overexpression onSERCA activity and Ca2+ transport in skeletal muscle remainsuncharacterized. Unlike the study by Song et al. (2004), wediscovered that the postural soleus and gluteus minimus musclesfrom PlnOE mice exhibited overt signs of myopathy. Collectively,our data suggest that the PlnOE mouse is a novel myopathy modelresembling human centronuclear myopathy (CNM), with additionaldystrophic features and potential core-like lesions.

RESULTSSERCA function is impaired in PlnOE musclePlnOE mice were generated in order to examine the functionalimportance of PLN in skeletal muscle; however, in the onlystudy published using these mice (Song et al., 2004), nomeasurements of SERCA function were reported because oftechnical limitations. In order to better define the functional roleof PLN in skeletal muscle we assessed SERCA function in thesemice, focusing our analyses on the postural muscles – soleus andgluteus minimus – since these muscles normally containapproximately 55-65% and 27% type I fibers, respectively.However, these type I fiber levels increase in response tooverexpression of PLN (discussed below). Western blotting ofwhole homogenates demonstrated clear overexpression of PLN inthese muscles because only 2.5 μg of total protein was loaded forPlnOE mice compared with the 25 μg required forWTmice (Fig. 1AReceived 26 March 2015; Accepted 21 May 2015

1Department of Kinesiology, University of Waterloo, Waterloo, Ontario N2L 3G1,Canada. 2Department of Exercise Science, Concordia University, Montreal,Quebec H4B 1R6, Canada. 3Departement of Kinesiology, McMaster University,Hamilton, Ontario L8N 3Z5, Canada. 4Department of Pediatrics, McMasterUniversity, Hamilton, Ontario L8N 3Z5, Canada. 5Department of Medicine,McMaster University, Hamilton, Ontario L8N 3Z5, Canada.*These authors contributed equally to this work

‡Author for correspondence (rtupling@uwaterloo.ca)

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

999

© 2015. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

and supplementary material Fig. S3A). Specifically, in the PlnOE

soleus relative to WT (n=7-9 per genotype), monomeric andpentameric PLN levels, normalized to actin, were 6.3-fold(P=0.00001) and 80-fold (P=0.0007) higher, respectively. In thePlnOE gluteus minimus muscles relative to WT (n=5-6 pergenotype), monomeric and pentameric PLN levels were 6.7-fold(P=0.0001) and 23-fold (P=0.0001) higher, respectively. Inaddition to the analysis in homogenates, we also performedsingle-fiber western blot analyses confirming overexpression ofPLN specific to type I fibers in this model (supplementary materialFig. S1). In contrast, PLN was barely detectable in some WT type Iand type IIA fibers (supplementary material Fig. S1). Thiscorresponds well with the fact that PLN expression in rodentskeletal muscle is far lower than that in human skeletal muscle,where PLN can be found in all type I and type IIA fibers in thevastus lateralis (Fajardo et al., 2013).In agreement with the known effects of PLN overexpression

on SERCA function in HEK 293 cells and transgenic hearts

(MacLennan and Kranias, 2003; Bhupathy et al., 2007), weobserved a reduction in SERCA’s apparent affinity for Ca2+, asindicated by a higher KCa value and a corresponding rightward shiftin the activity–pCa curve in soleus homogenates prepared fromPlnOE mice compared with WT (Table 1 and Fig. 1B). Furthermore,maximal rates of SERCA activity (−27%) and rates of Ca2+ uptake(−74%) were reduced in the PlnOE soleus compared with WT(Fig. 1B and C and Table 1). The SERCA dysfunction in PlnOE

soleus cannot be explained by altered SERCA expression sincewestern blot analysis showed that SERCA1a content is similar inWT and PlnOE muscles (Fig. 1D), whereas a significant 3-foldincrease in SERCA2a expression was observed in the soleus ofPlnOE mice (Fig. 1E). Sarcolipin (SLN), a small SR protein that isstructurally and functionally homologous with PLN (Asahi et al.,2002, 2003), was also found to be upregulated 9-fold in the soleusof PlnOE mice compared with WT (Fig. 1F). As SLN can inhibitSERCA pumps alone or in combination with PLN (Asahi et al.,2003), these results suggest that SLN may also contribute to theimpaired SERCA function we observed in PlnOE muscles. Indeed,the forced overexpression of SLN acutely in rat soleus reduces SRCa2+ transport (−31%) and causes severe contractile dysfunction(Tupling et al., 2002).

Soleus muscle atrophy and Ca2+-activated proteolysis inPlnOE miceSimilar to previously described results (Song et al., 2004), we alsoobserved atrophy of soleus muscles in PlnOE mice (Fig. 2Aand supplementary material Table S1). Interestingly, whenbreeding male and female PlnOE mice, soleus muscles weighing≤1 mg were observed, presumably from homozygous PlnOE

mice (+/+), because these muscles were no longer seen afterbreeding male PlnOE mice with female WT mice (Fig. 2A). Owingto this limitation of the tissue, we restricted all of our analysesin the present study to heterozygous PlnOE mice (+/−). In thestudy published by Song et al. (2004), a 35-40% reduction insoleus:body-weight ratio was observed in PlnOE mice at10-12 weeks of age. In our hands, assessment of soleus:body-weight ratio in PlnOE mice at 1 month, 4-6 months and 10-12 months of age indicates a 21%, 43% and 41% reduction,respectively, when compared with WT mice (supplementarymaterial Table S1). When we include age as a factor, absolutesoleus weight and the soleus:body-weight ratios were significantlydifferent at 4-6 months and 10-12 months, but not at 1 month. Toelucidate the mechanisms behind the diminished soleus musclemass, Song and colleagues reported an approximate 2-fold increasein calpain expression (Song et al., 2004). Calpains are a class ofproteolytic enzymes that are activated by high levels of [Ca2+]i(Murphy et al., 2006), and our results show that calpain activity was1.5-fold higher in soleus from PlnOE mice compared with WT(Fig. 2B), which probably results from reduced SERCA functionand greater [Ca2+]i in PlnOE mice. We also observed a significant1.5- to 2-fold increase in other proteolytic pathways, such ascaspase-3, cathepsin-B/L activity and protein ubiquitylation(Fig. 2C-E), which suggests that the soleus muscle atrophy inPlnOE mice is due to a general overall enhancement of proteolyticactivity in those muscles.

Central nuclei, progressive fibrosis and oxidative stress inPlnOE miceAs mentioned earlier, previous analyses of PlnOE mice revealed nosigns of myopathy under the light microscope (Song et al., 2004).Consequently, we were surprised that our histological analyses of

RESOURCE IMPACT

BackgroundPhospholamban (PLN) is awell-known inhibitor of the sarco(endo)plasmicreticulum Ca2+-ATPase (SERCA) pumps in muscle that maintain lowlevels of cytosolic Ca2+ and that play a crucial role in musclecontraction. The importance of PLN regulation of SERCA function incardiac muscle health and disease is well established but whetherPLN plays a similar role in skeletal muscle disease remains unknown.Centronuclearmyopathy (CNM) isacongenitalmyopathy characterizedbymuscle weakness, centrally located nuclei, type I fiber predominanceand central aggregation of oxidative activity in the skeletal muscles. Todate, there is no curative treatment for CNM and the exact mechanismleading to these skeletal muscle defects remains unknown. Thus, animalmodels that accurately recapitulate these histological abnormalities arerequired.

ResultsIn this study, the authors characterize a mouse model (PlnOE) in whichPLN is overexpressed in type I skeletal muscle fibers. They show thatSERCA function is greatly impaired in this model and that the soleus andgluteus minimus muscles, which normally contain many type I fibers,from PlnOE mice exhibit phenotypes that resemble human CNMincluding centrally located nuclei, type I fiber predominance, centralaggregation of oxidative activity and weakness. These muscles alsopresent with progressive atrophy, fibrosis and potential core formations.Finally, the authors report that SERCA function was on average 53%lower in muscle biopsies from three patients with CNM compared withbiopsies from five healthy individuals, whereas PLN expression seemedto be elevated in biopsies from patients with CNM compared with healthycontrols.

Implications and future directionsThese results identify a novel CNMmousemodel that can be used for theinvestigation of the mechanisms underlying CNM and for thedevelopment of therapeutic strategies. The findings also suggest thatstudies to assess the role of PLN and SERCA dysfunction in human andanimal CNM and myopathy in general might be worthwhile. To date,25-30% of CNM cases remain genetically unresolved but severalmutations in the human gene encoding PLN are known to lead toSERCA inhibition and cardiac complications. Given that many clinicalcases of CNM present with cardiomyopathy, future studies shoulddetermine whether skeletal muscle defects are also present in patientsharboring PLN mutations. Finally, these findings add to the notion thataberrant Ca2+ regulation is central to many cardiac and skeletal musclediseases and that targeting SERCA function might represent a viabletherapeutic strategy for these diseases.

1000

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

the soleus muscles from PlnOE mice showed obvious signs ofpathology that were repeatedly observed (Fig. 3). In fact, themyopathy was at least partially consistent with the dystrophic-likephenotype observed in PlnOE rabbits (Pattison et al., 2008),including greater central nucleation of fibers, fiber splitting,atrophy and progressive fibrosis, which became evident at4-6 months of age (Fig. 3A-D). Corresponding well with thedystrophic phenotype, we also observed greater whole muscleoxidative stress, as indicated by DCF (2′,7′-dichlorofluorescein)fluorescence and protein nitrosylation – effects that may be explainedby Ca2+ dysregulation (Altamirano et al., 2012). Furthermore, wefound elevated plasmaCK in 4- to 6-month-oldPlnOEmice (Fig. 3E-G); however, the relatively small increase in plasma CK suggests thatthe dystrophic-like phenotype in PlnOE mice is a direct result ofimpaired SERCA activity and elevated [Ca2+]i, and is less likely tobe caused by sarcolemmal damage per se. A similar finding has beenreported in mice overexpressing transient receptor potentialcanonical 3, which led to greater Ca2+ influx, muscular dystrophyand only slightly elevated plasma CK as a result of a generally intactsarcolemmal membrane (Millay et al., 2009).

CNM in the PlnOE mouseAlthough central nucleation might be an active marker in muscleregeneration and a common feature in muscular dystrophy (ChargeandRudnicki, 2004;Dubowitz et al., 2013; Folker andBaylies, 2013),we observed only a small proportion of fibers that were positive forembryonicMHC (supplementarymaterial Fig. S2), which is normallyexpressed transiently during muscle regeneration (Sartore et al., 1982;DiMario et al., 1991; Grady et al., 1997). These results, combinedwith the small elevation in plasma CK, suggest that very littledegeneration–regeneration cycling is occurring in PlnOE muscle.Therefore, the PlnOE mouse does not closely resemble dystrophicmyopathy; however, central nucleation in the face of low regeneration,

progressive fibrosis, and moderate elevations in CK levels is alsofound in CNM (Jungbluth et al., 2008). CNM is a congenitalmyopathy that is diagnosed by the existence of three histopathologicalfeatures found in muscle biopsies, which include: (1) centralnucleation, (2) central aggregation of oxidative activity, and (3)predominance of type I fibers and hypotrophy (Jungbluth et al., 2008;Romero, 2010). Both central aggregation of oxidative activity andtype I fiber hypotrophy were evident in soleus sections as early as1 month of age in PlnOE mice (Fig. 4A and B and Table 2). Type Ifiber predominance was established at 4-6 months of age, whereas at1 month, fibers were still transitioning towards the type I fiberdistribution as the PlnOE soleus had lower type IIA and greater typeI/IIA content compared with WT (Fig. 4A and Table 2). The type Ifiber predominance found from 4-6 months is in agreement withprevious findings (Song et al., 2004), and the increase in the slow-twitch SERCA2a isoform in PlnOE soleus (Fig. 1E) is also consistentwith type I fiber predominance. Taken together, these results indicatethat the PlnOE mouse model is a phenocopy of the histologicalfeatures found in human CNM. Importantly, the soleus muscle wasnot the only muscle affected by PLN overexpression because the

Fig. 1. SERCA function in soleusmuscles inPlnOE

mice at 4-6 months of age. (A) Western blotting forPLN in WT and PlnOE mice from soleus musclehomogenates. ForWTmice, 25 μg of total protein wasloaded, whereas only 2.5 μg was required for PlnOE

mice to detect PLN protein. (B) Ca2+-ATPase activity–pCa curves in WT (n=5) and PlnOE mice (n=6) in thepresence of the Ca2+ ionophore. (C) Ca2+ uptakeassessed in soleus muscles from WT (n=4) andPlnOE mice (n=5). (D-F) Western blotting forSERCA1a (D), SERCA2a (E) and SLN (F) in soleusmuscle from WT and PlnOE mice (n=6 per genotype).Actin was used as a loading control and all values areexpressed relative to WT. *P≤0.05 versus WT. Allvalues are presented as means±s.e. PLN (p), PLNpentamer; PLN (m), PLN monomer.

Table 1. SERCA activity in mouse soleus muscles from WT and PlnOE

mice at 4-6 months of age

Vmax KCa ΔKCa

WT 248.4±11.5 6.63±0.05 –

PlnOE 181.6±23.8* 6.47±0.04* 0.16

Values are means±s.e. Homogenates were isolated fromWT (n=6) and PlnOE

(n=5) mouse soleusmuscles and were analyzed for Ca2+-ATPase activity overCa2+ concentrations ranging from pCa 7 to pCa 4.5 to obtain KCa. KCa is theCa2+ concentration required to attain the half-maximal Ca2+-ATPase activityrate and is expressed in pCa units. *P≤0.05, compared with WT usingStudent’s t-test.

1001

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

postural gluteus minimus muscle, which normally comprises roughly27% type I fibers in WT mice, also displayed signs of impairedSERCA function, CNM and endomysial fibrosis (supplementarymaterial Fig. S3, Tables S2,S3). Despite a similar level of monomericPLN overexpression, the relative impairment in SERCA function inthe gluteus minimus muscles compared with the PlnOE soleus waslower, with only a 20% reduction in the rates of Ca2+ uptake(supplementary material Fig. S3C). This might be explained by thefact that the gluteusminimusmuscles have a greater number of type IIfibers that do not overexpress PLN, which could mask the inhibitionof SERCA pumps occurring in the type I fibers. In any event, boththe PlnOE soleus and gluteus minimus muscles resembled thehistopathological features associated with CNM.To date, CNMs are known to be genetically heterogeneous

(Dowling et al., 2014; Jungbluth and Gautel, 2014) and have beenattributed to X-linked recessive mutations in the MTM1 geneencoding myotubularin (Laporte et al., 1996), autosomal-dominantmutations in the DNM2 gene encoding dynamin-2 (Bitoun et al.,2005) and the BIN1 gene encoding amphiphysin-2 (Bohm et al.,2014), and autosomal-recessive mutations in BIN1 (Nicotet al., 2007), RYR1 encoding the SR ryanodine receptor (Jungbluthet al., 2007; Wilmshurst et al., 2010) and TTN encoding titin(Ceyhan-Birsoy et al., 2013).Many different structural abnormalitieshave been identified that aid in distinguishing between the varioushuman CNMs (Romero, 2010; Jungbluth and Gautel, 2014). Forexample, the autosomal-dominant DNM2-related CNMs presentwith radiating SR strands and marked contrast of the diameters

between type I and type II fibers (Romero, 2010). In addition to thesefeatures, increases in connective tissue and the presence of coreformations can also be observed inDNM2-related CNMs (Jungbluthand Gautel, 2014). Sincewe observed radiating SR strands (Fig. 4C),type II fiber hypertrophy (Table 2 and supplementary materialTable S3), progressive fibrosis (Fig. 3C and D) and potential core-like formations (Fig. 5) in thePlnOEmice, our results suggest that theCNM phenotype found in these mice resembles autosomal-dominant DNM2-related CNM. However, because increases inendomysial connective tissue and the presence of core formationsmay also be observed in RYR1-related and TTN-related CNM(Jungbluth et al., 2007; Jungbluth and Gautel, 2014), the PlnOE

mouse also shares a resemblance to these forms of CNM.We use theterm ‘potential core-like lesions’ because future studies with electronmicroscopy are required to verify the presence and characterize thestructure of cores in the PlnOE mouse.

Muscle function in the PlnOE mouseGeneral muscle weakness is common in cases of CNM and ouranalyses of contractile function in isolated intact soleus muscle, aswell as whole-body endurance during treadmill exercise,demonstrate that muscle function is impaired in PlnOE mice(supplementary material Fig. S4A-E). Specifically, we found thatPlnOE mice generate lower normalized force at both submaximaland maximal stimulation frequencies (30-100 Hz) (supplementarymaterial Fig. S4D) and reached exhaustion much sooner than WTmice during a treadmill exercise test (supplementary material

Fig. 2. Soleus muscles of PlnOE mice are atrophied at 4-6 months of age. (A) Representative images of soleus muscles extracted fromWT and PlnOE mice.+/−, heterozygous; +/+, homozygous. (B) Calpain activity in soleus homogenates from WT (n=4) and PlnOE mice (n=5). (C) Caspase-3 activity in soleusmuscle homogenates from WT (n=4) and PlnOE mice (n=7). (D) Cathepsin-B/L activity in soleus homogenates from WT and PlnOE mice (WT, n=4; PlnOE, n=6).Calpain activity, caspase-3 activity, and cathepsin-L activity are in arbitrary units normalized to mg protein and are presented relative to WT. (E) Total proteinextracts from soleus muscles (15 μg) were immunoblotted with anti-ubiquitin (Ub) antibody. Ponceau stain was used as a loading control. The sum of opticaldensities from detectable ubiquitylated proteins (p-Ub) as well as the optical density of monomeric ubiquitin (m-Ub, 10 kDa) was measured and comparedbetween genotypes (n=4 per group). Values were normalized to the sum of optical densities of bands visualized through Ponceau stain. *P≤0.05 versus WT.All values are presented as means±s.e.

1002

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

Fig. 3. See next page for legend.

1003

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

Fig. S4E). In agreement with previous findings (Song et al., 2004),we found that normalized twitch force generation of soleus musclewas not different between PlnOE and WT mice, whereas the twitch

kinetics were significantly slower in PlnOE compared with WT(supplementary material Fig. S4A-C).

PLN and SERCA dysfunction in human CNMThus far, it is evident that PLN overexpression in mouse skeletalmuscle leads to CNM; however, whether PLN plays a primarypathogenic role in CNM remains unknown. In support of thisnotion, PLN protein was found to be upregulated in the tibialisanterior muscles from mice lacking microRNA 133a-1 and 133a-2,where CNM was observed (Liu et al., 2011). In addition, weobtained muscle biopsies from three patients with CNM and whencompared with biopsies from five healthy controls, we found thatthe maximal rate of SERCA activity was significantly reducedby 53% (supplementary material Fig. S5A) and that monomericPLN (P=0.12) and total PLN (P=0.08) content were greater(supplementary material Fig. S5B). Therefore, we provide the firstindication of a potential role for SERCA dysfunction and elevatedPLN in human CNM; however, caution should be taken wheninterpreting these results. Recently, we established through single-fiber western blotting and immunohistochemical analyses that PLNexpression follows SERCA2a and MHC I expression, whereas SLN

Fig. 4. Centronuclear myopathy in the soleus muscles from PlnOE mice. (A) Representative images of soleus muscles showing type I fiber predominanceand hypotrophy. MHC immunofluorescence stained sections of the soleus muscles at 1 month, 4-6 months and 10-12 months of age. Cross sections werestained with antibodies against MHC to identify type I (blue), type IIA (green), type IIB (red) and type IIX (unstained) fibers. (B) Representative images of soleusmuscles showing central accumulation of oxidative activity. Succinate dehydrogenase (SDH)-stained sections display central aggregation of oxidative activityin the PlnOE mice at 1 month, 4-6 months and 10-12 months of age. (C) NADH-TR-stained cross sections demonstrating radiating SR strands in soleus musclesfrom PlnOE mice (arrows). Asterisks represent the same fiber across cryosections. Scale bars: 50 μm.

Fig. 3. Dystrophic features in soleus muscles from PlnOE mice. (A) H&E-stained sections of the soleus muscles from WT and PlnOE mice at 1 month,4-6 months and 10-12 months of age. (B) Percentage of fibers containingcentral nuclei in the soleus at 1 month, 4-6 months and 10-12 months of age(n=3 per genotype at each age with 300-600 fibers counted per mouse).(C) Van-Geison-stained sections of the soleus muscles from WT and PlnOE

mice at 1 month, 4-6 months and 10-12 months of age. (D) Quantification offibrotic area in the soleus at 1 month, 4-6 months and 10-12 months of age(n=3-4 per group at each age). ImageJ software was used to quantify fibroticarea. (E) Reactive oxygen species generation in whole soleus determinedusing DCF assay at 4-6 months of age (WT, n=4; PlnOE, n=5). (F) Total proteinextracts from soleus muscles (15 μg) were immunoblotted with anti-nitrotyrosine antibody. The sumof optical densities fromdetectable nitrosylatedproteins (50-100 kDa) was measured and compared between genotypes at4-6 months of age (n=6 per group). Values were normalized to the sum ofoptical densities of bands visualized through Ponceau stain. (G) Plasma CKlevels in WT (n=5) and PlnOE mice (n=9) at 4-6 months of age. Scale bars:50 μm (A,C). *P≤0.05 versus WT using two-way ANOVA and Tukey’s post hocanalysis for % central nuclei and % fibrosis; Student’s t-test for DCF, proteinnitrosylation and plasma CK. All values are presented as means±s.e.

1004

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

follows SERCA1a and MHC II expression in healthy human vastuslateralis (Fajardo et al., 2013). Since we observed a trend for higherMHC I (supplementary material Fig. S5C, P=0.10) and SERCA2a(supplementary material Fig. S5D, P=0.10) expression, andsignificant reductions in both SERCA1a (supplementary materialFig. S5E) and SLN (supplementary material Fig. S5F) in CNMmuscles compared with healthy controls, it is possible that thesedifferences, including upregulated PLN, are due mostly to theexpected type I fiber predominance in CNM patients. Furthermore,in addition to the low sample size used in the present study, theCNM cases obtained were of notable age variability and weregenetically unrelated, with one patient harboring a DNM2 mutation(R369W) and the other two patients remaining geneticallyunresolved. Importantly, despite these limitations, we were stillable to show a 53% reduction in SERCA activity and the expressionpatterns of the SERCA isoforms, SLN and PLN are consistent withthe expected type I fiber predominance often found in CNM.

Dynamin 2 levels in CNMAs a result of the dominant inheritance of the DNM2 mutation, onetheory is thatDNM2-related CNMmay be caused by a toxic gain-of-

function in dynamin 2 (Cowling et al., 2014; Dowling et al., 2014).In support of this, overexpression of WT dynamin 2 in skeletalmuscle results in features associated with CNM (Cowling et al.,2011; Liu et al., 2011). In addition, upregulation of dynamin 2 wasobserved in skeletal muscles from mtm1–/y mice, which accuratelymodels human X-linked myotubular myopathy (XLCNM), andheterozygous knockdown of dynamin 2 in the mtm1–/y mouseimproves muscle pathology and function (Cowling et al., 2014). Weanalyzed dynamin 2 protein content in PlnOE muscles and founda 5-fold (supplementary material Fig. S6A) and 3-fold(supplementary material Fig. S6B) higher level in the soleus andgluteus minimus muscles, respectively, compared with WT.Whether dynamin 2 plays a pathological role in the CNM foundin the PlnOE mouse and the exact mechanism leading to itsaugmented expression requires further investigation.

Elevated levels of dynamin 2 are thought to also play apathological role in human CNM because results from Laporte’slaboratory have shown an approximate 1.5-fold increase in dynamin2 expression in muscle lysates from three XLCNM patientscompared with two healthy controls (Liu et al., 2011; Cowlinget al., 2014). To confirm whether elevated dynamin 2 contributesto uman CNM pathology, we assessed the dynamin 2 proteincontent in the three CNM muscle biopsies obtained here; however,in contrast to what was previously found in human CNM (Liuet al., 2011; Cowling et al., 2014), we found a 35% reduction indynamin 2 compared with healthy subjects (supplementary materialFig. S6C). Notwithstanding the aforementioned sample size andother limitations with our human CNM cases, these findings maysuggest that reducing dynamin 2 as a therapeutic strategy may not beappropriate for all forms of human CNM as previously suggested(Demonbreun and McNally, 2014) and may be limited to XLCNM.

DISCUSSIONTo date, most studies concerning the physiological significance ofPLN inhibition of SERCA have focused solely on its role in cardiacmuscle health and disease (for a review, see MacLennan andKranias, 2003). To our knowledge, our current study is the first todemonstrate an important role for PLN and SERCA dysfunction inskeletal muscle health and CNM pathology. Specifically, weobtained commercially available PlnOE mice to determine PLN’seffect on SERCA function in skeletal muscle and uncovered amouse model that phenotypically resembles the histopathologicalfeatures associated with human CNM. In addition, our results withthree human CNM patients suggest that SERCA dysfunction,possibly through increased PLN expression, contributes to CNMpathology. Although it is possible that the predominance of type Ifibers in human CNM may explain the elevations in PLNexpression, it is important to consider our results from the PlnOE

mouse, which indicate that overexpression of PLN specifically intype I fibers can cause type I fiber predominance and CNM. Thus,future studies with more CNM patients that are similar in age andgenetically related are required to better investigate whether PLNplays a primary pathological role in human CNM.

We cannot fully explain why we repeatedly observed myopathyin the PlnOE mouse whereas Song and co-workers did not reportsuch a finding (Song et al., 2004). Comparisons are difficult to makebecause SERCA function was not assessed and histology resultswere not shown in their study. Strain differences seem unlikely toaccount for the discrepant results since both transgenic lines weregenerated on an FVB/N background; however, it is possible thatgenetic drift and the level of PLN overexpression may be significantfactors. Nevertheless, several results are in agreement between the

Table 2. Quantitative analysis of fiber type distribution and cross-sectional area in soleus muscles from WT and PlnOE mice

1 month 4-6 months 10-12 months

Fiber distribution (%)Type IWT 56.2±0.8 58.9±6.9 64.1±2.6PlnOE 60.7±2.5b 91.7±2.9*,a 94.4±1.8*,a

Type I/IIAWT 6.4±1.1b 0.5±0.5a 0.6±0.3a

PlnOE 13.0±1.5*,b 0.6±0.2a 0.7±0.2a

Type IIAWT 34.0±1.4 35.5±3.1 31.3±1.8PlnOE 21.9±1.7*,b 7.7±2.6*,a 4.9±1.7*,a

Fiber CSA (μm2)Type IWT 1895±158 2051±273 2584±343PlnOE 1291±93* 871±28* 1190±90*

Type IIAWT 1260±89 1915±126 2225±159a

PlnOE 1617±103b 2922±180*,a 5342±703*,a,b

Values are means±s.e. (n=3 per genotype). *Significantly different from WTwithin the same age; aSignificantly different from 1 month within the samegenotype; bSignificantly different from 4-6 months within the same genotype;P≤0.05 using two-way ANOVA including genotype and age as factors and aTukey’s post hoc test when necessary.

Fig. 5. Potential core-like lesions in PlnOE mouse at 4-6 months of age.SDH-stained sections from the soleus muscle show areas devoid of oxidativestaining representing potential core-like lesions (yellow arrows).Corresponding H&E-stained serial section shows that the lack of oxidativestaining with SDH is not due to the presence of a vacuole or an artifact in themuscle fiber.

1005

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

two studies, including type I fiber predominance, atrophy, increasedproteolytic markers and slower force kinetics in soleus muscle.It has been estimated that 25-30% of human CNM cases remain

genetically unresolved (Romero, 2010; Dowling et al., 2014) and ourresults raise the possibility that mutations in Pln leading to increasedSERCA inhibition could be involved in those cases. Although Plnmutations either resulting in elevated PLN expression (Minamisawaet al., 2003) or reduced PLN phosphorylation (Schmitt et al., 2003;Haghighi et al., 2006) have been causally linked to humancardiomyopathy, it is unknown whether these mutations can affectskeletal muscle health. For example, deletion of arginine 14 (R14del)in the coding region of human PLN results in increased SERCAinhibition through increased monomeric PLN and reduced PLNphosphorylation at Ser16 by PKA (Haghighi et al., 2006).Interestingly, one patient with an R14del mutation in PLNdeveloped late-onset mild dilated cardiomyopathy, but was initiallyevaluated in a muscle dystrophy clinic for a 25-year history of slowlyprogressive muscle weakness (DeWitt et al., 2006). Since there wereno significant abnormalities in the staining patterns of sarcolemmalproteins, muscular dystrophy was discounted; however, to ourknowledge, the existence of CNM was not tested. Recent reports ofhuman cases of CNM coexisting with cardiomyopathy (Ceyhan-Birsoy et al., 2013; Agrawal et al., 2014; Gal et al., 2015) furtherhighlight the importance of assessing the existence of CNM in skeletalmuscles from patients with cardiomyopathy arising from PLNmutations.Currently, the mechanisms leading to CNM are not completely

understood; however, several mechanisms have been suggested,including abnormalities of triad structure and function (Al-Qusairiet al., 2009; Al-Qusairi and Laporte, 2011; Bohm et al., 2014;Dowling et al., 2014; Gibbs et al., 2014). As a result of these triaddefects, aberrant Ca2+ handling and excitation–contraction (EC)coupling are also thought to be important for CNM pathogenesis.Indeed, the identification of mutations in RYR1 leading to CNM(Jungbluth et al., 2007; Wilmshurst et al., 2010) further supports thenotion that triad dysfunction, Ca2+ dysregulation and EC couplingare key pathogenic drivers of CNM (Dowling et al., 2014). Since theSERCA pumps regulate SR Ca2+ load and, thus, contractility, theSERCA pumps are also crucial for Ca2+ regulation and ECcoupling. Thus, our results with the PlnOE mice add further supportto the hypothesis that defects in triad function, Ca2+ handling andEC coupling are important for CNM pathogenesis and extend thehypothesis by adding SERCA dysfunction as a potential pathogenicmechanism. Since abnormalities in triad structure are often found inhuman and animal CNM, future studies using electron microscopyshould examine the triad structure in the PlnOE mouse to determinewhether structural abnormalities also contribute to the Ca2+

dysregulation seen in these mice. Furthermore, although areasdevoid of oxidative (SDH) staining that cannot be explained byartifacts or the presence of vacuoles may represent core formationsin the PlnOE mouse, future studies with electron microscopy willconfirm and characterize these potential core-like lesions.One other interesting question raised by our study pertains to the

role of SLN in CNM. Similar to PLN, SLN is a SERCA pumpinhibitor (Asahi et al., 2002, 2003; Gorski et al., 2013), whichsuggests that the 7- to 9-fold upregulation in SLN protein found in thePlnOE soleus and gluteus minimus muscles may contribute to theSERCA dysfunction, Ca2+ dysregulation and CNM pathology.Interestingly, this response in SLN expression is consistent withother mouse models of myopathy (Nakagawa et al., 2005; Ottenheijmet al., 2008; Al-Qusairi et al., 2009; Liu et al., 2011; Calvo et al., 2012;Schneider et al., 2013), but the role of SLN remains unknown. Our

finding that SLN content was reduced in human CNM reveals apotential species difference and suggests that this SERCA regulatormay haveminimal contribution to the SERCA dysfunction and overallCNM pathology. To test whether or not SLN actively contributes tothe CNM phenotype found in the PlnOE mice, future studies targetingthe Sln gene are required. Furthermore, a novel micropeptide,myoregulin (MLN), was found to be structurally and functionallyhomologous with both PLN and SLN (Anderson et al., 2015). Whennew antibodies targetingMLN protein expression become available, itwill be interesting to determine its response and potential involvementin murine and human CNM and myopathy in general.

In summary, the commercially available PlnOE mousehistopathologically resembles human CNM. To date, there is nocure for CNM and many unresolved questions remain, including themechanisms leading to skeletal muscle defects in patients.Mechanistically, our results from the PlnOE mouse are consistentwith the hypothesis that triad dysfunction, aberrant Ca2+ handlingand EC coupling are important for CNM pathogenesis; but whetherit is higher cytosolic Ca2+ and/or altered Ca2+ dynamics that lead toCNM histopathology, and how the muscle translates thisdysregulated Ca2+ into the CNM phenotype remains unknownand future studies using the PlnOE mouse will prove valuable.Furthermore, treatment strategies aimed at improving SERCAfunction have already shown promise in murine muscular dystrophy(Goonasekera et al., 2011; Gehrig et al., 2012; Mazala et al., 2015)and human cardiomyopathy (Horowitz et al., 2011; Jessup et al.,2011; Zsebo et al., 2014). Thus, future studies that improve SERCAfunction through various modes, including alterations in PLNcontent and/or inhibitory function, in the PlnOE mouse and otheranimal models of CNM, could lead to the development of viabletherapeutic strategies for CNM patients.

MATERIALS AND METHODSPlnOE micePlnOE mice were resuscitated from cryopreserved embryos by the mmRRC(000067-MU, FVB/N background) to generate a breeding colony with WTFVB/N mice in our facility. The Pln transgene was attached to the β-MHCpromoter so that these mice overexpress PLN in their slow-twitch type Iskeletal muscle fibers. Animals were housed in an environmentallycontrolled room with a standard 12 h:12 h light:dark cycle and allowedaccess to food and water ad libitum. Experiments were performed onlittermate heterozygous PlnOE and WT males that were between the ages of1 and 12 months. Reintroduction of homozygous PlnOE mice was avoidedby breeding heterozygous male PlnOE mice with female WT mice. Analysisof 287 newborn mice showed that WT and heterozygous PlnOE mice wereborn at the expected mendelian frequency with a ratio of 137:150. Allanimal procedures were reviewed and approved by the Animal CareCommittee of the University of Waterloo and are consistent with theguidelines established by the Canadian Council on Animal Care.

Human subjectsFive untrained university students were used as healthy controls. Allsubjects were fully informed of all experimental procedures and allassociated risks before written consent was obtained. Written approval forthe research was granted by the Human Research Ethics Committee at theUniversity of Waterloo. Three patients with CNM from McMasterUniversity Medical Centre were used in this study. One female (55 yearsold) patient had CNM due to a missense mutation in dynamin 2 (DNM2,1105C→T, R369W) whereas the other two male CNM patients (17 and44 years old) remain genetically unresolved. Of the two unresolved cases,one underwent 163-gene next-generation sequencing, which includedMTM1, DNM2, RYR1, CACNA1S and TTN. The other unresolved patientwas screened forMTM1 and DNM2; however, he did not return to completea screen for additional genes. Written approval for the use of material from

1006

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

previously consented and diagnosed patients was granted by the Chair of theHuman Research Ethics Committee at McMaster University. All muscletissue samples (∼100 mg) were obtained from the vastus lateralis using theneedle biopsy technique under suction (Tarnopolsky et al., 2011). Thesesamples were homogenized and then frozen at−80°C prior to being used forwestern blotting and SERCA activity assays.

SERCA activity and Ca2+ uptakeCa2+-dependent SERCA activity was assessed in homogenates preparedfrom mouse (WT and PlnOE) soleus and gluteus minimus muscles andhuman vastus lateralis muscles over Ca2+ concentrations ranging from pCa7.5 to 4.5 in the presence of the Ca2+ ionophore A23187 (Sigma C7522)using a spectrophotometric plate reader assay that has been describedpreviously (Duhamel et al., 2007). SERCA activity–pCa curves weregenerated with GraphPad Prism™ by non-linear regression curve fittingusing an equation for a general cooperative model for substrate activation.Ca2+ uptake was measured in homogenates in the presence of theprecipitating anion, oxalate, using the fluorescent dye Indo-1 and aspectrofluorometer equipped with dual-emission monochromators (Tuplingand Green, 2002). Rates of Ca2+ uptake assessed at a free cytosolic Ca2+

concentration of 1500 nM are reported.

AntibodiesPrimary antibodies against SERCA2a (2A7-A1), PLN (2D12), dynamin 2(PA5-19800) and RyR (MA3-925) were obtained from Pierce Antibodies.The primary antibody for SERCA1a (A52) was a kind gift from Dr DavidMacLennan (University of Toronto) (Zubrzycka-Gaarn et al., 1984). Theprimary antibody directed against SLN was generated by LampireBiological Laboratories (Fajardo et al., 2013). Anti-ubiquitin (P4D1) andanti-nitrotyrosine (189542) antibodies were obtained from Cell SignalingTechnology and Cayman Chemicals, respectively. The primary antibodyagainst α-actin (A4700) was obtained from Sigma-Aldrich. The primaryantibodies against MHCI (BA-F8), MHCIIa (SC-71), MHCIIb (BF-F3),embryonic MHC (BF-F6) (Schiaffino et al., 1989; Lucas et al., 2000) anddystrophin (3B7) (Nguyen thi et al., 1990) were obtained fromDevelopmental Studies Hybridoma Bank. Secondary antibodies forwestern blotting, goat anti-mouse IgG (peroxidase conjugated) and goatanti-rabbit IgG (peroxidase conjugated) were obtained from Santa CruzBiotechnology. Secondary antibodies for immunofluorescence staining,Alexa Fluor 350 anti-mouse IgG2b, Alexa Fluor 488 anti-mouse IgG1 andAlexa Fluor 555 anti-mouse IgM, were obtained from Molecular Probes.

Western blot analysisSingle-fiber western blots were performed as previously described (Fajardoet al., 2013) to determine fiber-type specificity of PLN overexpression.Briefly, single fibers extracted from soleus muscles of WT and PlnOE micewere placed into 1× solubilizing buffer. Solubilized proteins were thenseparated using Tricine-based SDS-PAGE (6-13% layered gel). Separatedproteins were then transferred onto 0.2 μm polyvinylidene difluoride(PVDF) membranes. Membranes were cut at the 75 kDa band (Western CPrecision Plus™, Bio-Rad, CA, USA) and were immunoprobed with PLN(<75 kDa strip) and withMHCI (>75 kDa strip). Following this, membraneswere immunoprobed with horseradish-peroxidase-conjugated secondaryantibodies and antigen–antibody complexes were detected by LuminataForte™ (Millipore, MA, USA) for PLN, and ECL Western Blot Substrate(BioVision, MA, USA) for MHCI. After detection of MHCI, the membranewas stripped and re-probed with MHCIIa and antigen–antibody complexeswere detected using ECLWestern Blot Substrate.

Similarly, western blot analysis was performed to determine expressionlevels of SLN, PLN, SERCA isofroms, dynamin 2, and RyR aswell as levels ofprotein ubiquitylation (Ub) and nitrosylation (Ny) in mouse (WT and PlnOE)and human CNM muscles after homogenates were placed into 1× solubilizingbuffer. Solubilized proteins from tissue homogenates, were separated usingTricine based SDS-PAGE (13% total acrylamide for PLN and SLN) or astandard SDS-PAGE (7.5% total acrylamide for SERCA isoforms anddynamin 2 and a 7.5-15% layered gel for Ub and Ny). Separated proteins werethen transferred onto 0.2 μmPVDFmembranes (PLN, SERCAs,Ub, Ny, RyR)

or nitrocellulose membranes (SLN). Membranes were then immunoprobedwith their corresponding primary antibodies. Following this, membranes wereimmunoprobed with horseradish-peroxidase-conjugated secondary antibodiesand signals were detected by SuperSignal West Femto™ substrate (Pierce,Thermo Fisher Scientific Inc.) for SLN; Luminata Forte™ for SERCA2a,PLN, Ub, Ny and RyR; and ECL Western Blot Substrate for SERCA1a anddynamin 2. Quantification of optical densities was performed using GeneTools(Syngene, MD, USA) and values were normalized to total protein or α-actin.

Histological, histochemical and immunofluorescence stainingSoleus and gluteus minimusmuscles fromWTand PlnOEmicewere removedand embedded in OCT compound (Tissue-Tek), frozen in liquid nitrogen-cooled isopentane, stored at −80°C, and cut into 10-μm-thick cryosectionswith a cryostat (Thermo Electronic) maintained at −20°C. Histologicalstaining included H&E and Van Gieson and histochemical staining includedsuccinate dehydrogenase (SDH) and NADH-TR activity. Images wereacquired with a brightfield Nikon microscope linked to a PixeLink digitalcamera and quantified with Image-Pro PLUS analysis and ImageJ software.Immunofluorescence analysis of MHC expression was previously described(McMillan and Quadrilatero, 2011; Bloemberg and Quadrilatero, 2012;Fajardo et al., 2013) and performed with primary antibodies against MHCI,MHCIIa and MHCIIb. Additional immunofluorescence analysis ofembryonic MHC expression was conducted with primary antibodiesagainst embryonic MHC along with primary antibodies against dystrophin.Slides were visualized with an Axio Observer Z1 fluorescent microscopeequipped with standard red, green, blue filters, an AxioCam HRm camera,and AxioVision software (Carl Zeiss). Quantification of fibers and cross-sectional area (CSA) was performed using ImageJ software.

Proteolytic activity and DCF assaysCalpain, caspase-3 and cathepsin-B/L activity were determined in soleusmuscle homogenates using the substrates, Suc-LLVY-AMC (Enzo-LifeSciences), Ac-DEVD-AMC (Alexis Biochemicals) and z-FR-AFC (EnzoLife Sciences), respectively (McMillan and Quadrilatero, 2011; Bloemberget al., 2014). These fluorogenic substrates are weakly fluorescent but yieldhighly fluorescent products following proteolytic cleavage by theirrespective proteases. Fluorescence was measured using a SPECTRAmaxGemini XS microplate spectrofluorometer (Molecular Devices, Sunnyvale,CA) with excitation and emission wavelengths: 360 nm and 440 nm,respectively for caspase-3; 380 nm and 460 nm, respectively for calpain;400 nm and 505 nm, respectively for cathepsin. Calpain activity was takenas the difference in fluorescence from homogenate incubated with andwithout calpain inhibitor. Calpain, caspase-3 and cathepsin activities werenormalized to total protein content and expressed as fluorescence intensityin arbitrary units per mg of protein.

Generation of reactive oxygen species in soleus muscle homogenates fromWT and PlnOE mice was determined as previously described (McMillan andQuadrilatero, 2011) using 2′,7′-dichlorodihydrofluorescein-diacetate (DCFH-DA, Invitrogen, Carlsbad, CA) at 37°C. Cellular esterases hydrolyze DCFH-DA to the non-fluorescent DCFH, which can then be oxidized by a variety ofROS to form highly fluorescent DCF. Fluorescence was measured using aSPECTRAmaxGemini XSmicroplate spectrofluorometer with excitation andemission wavelengths of 490 nm and 525 nm, respectively. Fluorescenceintensity was normalized to total protein content and expressed as arbitraryunits per mg protein.

Plasma CK analysisWT and PlnOE mice were anesthetized using somnotol (0.65 mg/kg bodyweight) and blood from the left ventricle was drawn into a heparinizedsyringe. Blood was centrifuged at 5000 g for 8 min and the plasma wasdecanted and stored at −80°C until analysis. CK activity was measuredusing a kinetic fluorometric assay as previously described (Szasz et al.,1976).

Electrical stimulation and muscle contractility measurementsExperiments were performed on adult (4-6 month) WT and PlnOE mice.Mice were killed by cervical dislocation, and the intact soleus muscles were

1007

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

removed and placed into a bath with oxygenated Tyrode solution (95% O2,5%CO2) containing 121 mMNaCl2, 5 mMKCl, 24 mMNaHCO3, 1.8 mMCaCl2, 0.4 mM NaH2PO4, 5.5 mM glucose, 0.1 mM EDTA and 0.5 mMMgCl2, pH 7.3 (Lännergren et al., 2000), and was maintained at 25°C.Muscles were situated between flanking platinum electrodes driven by abiphasic stimulator (Model 710B, Aurora Scientific) and electrically evokedmuscle force was assessed across a range of stimulation frequencies from 1to 100 Hz. Data were analyzed using Dynamic Muscle Control DataAcquisition software (Aurora Scientific). Specifically, peak isometric forceamplitude (mN) and the maximal rates of force development (+dF/dt) andrelaxation (−dF/dt) were determined during a twitch and across the range ofstimulation frequencies. Peak isometric forcewas then normalized to muscleweight (mN/g).

Treadmill exerciseWTand PlnOE mice were assessed for muscle performance by running themto exhaustion on an enclosed motorized treadmill. All animals exercised at arunning speed of 8 m/min for 10 min followed by 10 min at 16 m/min andthen allowed to reach exhaustion at 24 m/min at a 5° incline.

StatisticsAll values are presented as means±s.e. Statistical significance was set toP≤0.05. Comparisons between WT and PlnOE mice were performed usingStudents t-test; however, a two-way repeated-measures ANOVA was usedfor force–frequency analysis, and when age was included as a factor. Posthoc testing was done using Tukey’s HSD. Comparisons between healthycontrols and human CNM patients were performed using Student’s t-test.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsA.R.T., V.A.F. and E.B. conceived the study idea. V.A.F. coordinated western blotexperiments and with K.T. performed the western blot analyses. E.B. and C.B.performed the histological and histochemical analysis. E.B., C.B. and E.M.performed the immunofluorescence staining experiments. V.A.F., D.G. and A.H.performed SERCA ATPase activity. V.A.F. and C.V. performed Ca2+-uptakemeasures. E.M. performed proteolytic activity and oxidative stress assays. V.A.F.and E.B. performed plasma creatine kinase analysis. B.J.W. and I.C.S. conductedsoleus contractile analysis. V.A.F. conducted treadmill running. J.Q. and A.R.T.contributed reagents, materials and analysis tools. M.A.T. provided the musclebiopsies from CNM patients. A.R.T. and V.A.F. wrote the manuscript that wasreviewed by all authors.

FundingThis work was supported by research grants from the Canadian Institutes of HealthResearch (CIHR) [grant numbers MOP 86618 andMOP 47296 to A.R.T.]. I.C.S. andC.V. were supported by postgraduate scholarship doctoral awards from the NaturalSciences and Engineering Research Council of Canada. V.A.F. was supported by adoctoral award from CIHR.

Supplementary materialSupplementary material available online athttp://dmm.biologists.org/lookup/suppl/doi:10.1242/dmm.020859/-/DC1

ReferencesAgrawal, P. B., Pierson, C. R., Joshi, M., Liu, X., Ravenscroft, G.,Moghadaszadeh, B., Talabere, T., Viola, M., Swanson, L. C., Haliloglu, G.et al. (2014). SPEG interacts with myotubularin, and its deficiency causescentronuclear myopathy with dilated cardiomyopathy. Am. J. Hum. Genet. 95,218-226.

Al-Qusairi, L. and Laporte, J. (2011). T-tubule biogenesis and triad formation inskeletal muscle and implication in human diseases. Skelet. Muscle 1, 26.

Al-Qusairi, L., Weiss, N., Toussaint, A., Berbey, C., Messaddeq, N., Kretz, C.,Sanoudou, D., Beggs, A. H., Allard, B., Mandel, J.-L. et al. (2009). T-tubuledisorganization and defective excitation-contraction coupling in muscle fiberslacking myotubularin lipid phosphatase. Proc. Natl. Acad. Sci. USA 106,18763-18768.

Altamirano, F., Lopez, J. R., Henriquez, C., Molinski, T., Allen, P. D. andJaimovich, E. (2012). Increased resting intracellular calcium modulates NF-kappaB-dependent inducible nitric-oxide synthase gene expression in dystrophicmdx skeletal myotubes. J. Biol. Chem. 287, 20876-20887.

Anderson, D. M., Anderson, K. M., Chang, C.-L., Makarewich, C. A., Nelson,B. R., McAnally, J. R., Kasaragod, P., Shelton, J. M., Liou, J., Bassel-Duby, R.et al. (2015). A micropeptide encoded by a putative long noncoding RNAregulates muscle performance. Cell 160, 595-606.

Asahi, M., Kurzydlowski, K., Tada, M. and MacLennan, D. H. (2002). Sarcolipininhibits polymerization of phospholamban to induce superinhibition of sarco(endo)plasmic reticulum Ca2+-ATPases (SERCAs). J. Biol. Chem. 277,26725-26728.

Asahi, M., Sugita, Y., Kurzydlowski, K., De Leon, S., Tada, M., Toyoshima, C.and MacLennan, D. H. (2003). Sarcolipin regulates sarco(endo)plasmicreticulum Ca2+-ATPase (SERCA) by binding to transmembrane helices aloneor in association with phospholamban. Proc. Natl. Acad. Sci. USA 100,5040-5045.

Bhupathy, P., Babu, G. J. and Periasamy, M. (2007). Sarcolipin andphospholamban as regulators of cardiac sarcoplasmic reticulum Ca2+ ATPase.J. Mol. Cell. Cardiol. 42, 903-911.

Bitoun, M., Maugenre, S., Jeannet, P.-Y., Lacene, E., Ferrer, X., Laforêt, P.,Martin, J.-J., Laporte, J., Lochmuller, H., Beggs, A. H. et al. (2005). Mutationsin dynamin 2 cause dominant centronuclear myopathy. Nat. Genet. 37,1207-1209.

Bloemberg, D. and Quadrilatero, J. (2012). Rapid determination of myosin heavychain expression in rat, mouse, and human skeletal muscle using multicolorimmunofluorescence analysis. PLoS ONE 7, e35273.

Bloemberg, D., McDonald, E., Dulay, D. and Quadrilatero, J. (2014). Autophagyis altered in skeletal and cardiac muscle of spontaneously hypertensive rats. ActaPhysiol 210, 381-391.

Bohm, J., Biancalana, V., Malfatti, E., Dondaine, N., Koch, C., Vasli, N., Kress,W.,Strittmatter, M., Taratuto, A. L., Gonorazky, H. et al. (2014). Adult-onsetautosomal dominant centronuclear myopathy due to BIN1 mutations. Brain 137,3160-3170.

Calvo, A. C., Manzano, R., Atencia-Cibreiro, G., Olivan, S., Munoz, M. J.,Zaragoza, P., Cordero-Vazquez, P., Esteban-Perez, J., Garcıa-Redondo, A.and Osta, R. (2012). Genetic biomarkers for ALS disease in transgenic SOD1(G93A) mice. PLoS ONE 7, e32632.

Ceyhan-Birsoy, O., Agrawal, P. B., Hidalgo, C., Schmitz-Abe, K., DeChene,E. T., Swanson, L. C., Soemedi, R., Vasli, N., Iannaccone, S. T., Shieh, P. B.et al. (2013). Recessive truncating titin gene, TTN, mutations presenting ascentronuclear myopathy. Neurology 81, 1205-1214.

Charge, S. B. P. and Rudnicki, M. A. (2004). Cellular and molecular regulation ofmuscle regeneration. Physiol. Rev. 84, 209-238.

Cowling, B. S., Toussaint, A., Amoasii, L., Koebel, P., Ferry, A., Davignon, L.,Nishino, I., Mandel, J.-L. and Laporte, J. (2011). Increased expression of wild-type or a centronuclear myopathy mutant of dynamin 2 in skeletal muscle of adultmice leads to structural defects and muscle weakness. Am. J. Pathol. 178,2224-2235.

Cowling, B. S., Chevremont, T., Prokic, I., Kretz, C., Ferry, A., Coirault, C.,Koutsopoulos, O., Laugel, V., Romero, N. B. and Laporte, J. (2014). Reducingdynamin 2 expression rescues X-linked centronuclear myopathy. J. Clin. Invest.124, 1350-1363.

Demonbreun, A. R. and McNally, E. M. (2014). Dynamin 2 the rescue forcentronuclear myopathy. J. Clin. Invest. 124, 976-978.

DeWitt, M. M., MacLeod, H. M., Soliven, B. and McNally, E. M. (2006).Phospholamban R14 deletion results in late-onset, mild, hereditary dilatedcardiomyopathy. J. Am. Coll. Cardiol. 48, 1396-1398.

DiMario, J. X., Uzman, A. and Strohman, R. C. (1991). Fiber regeneration is notpersistent in dystrophic (MDX) mouse skeletal muscle. Dev. Biol. 148, 314-321.

Dowling, J. J., Lawlor, M. W. and Dirksen, R. T. (2014). Triadopathies: anemerging class of skeletal muscle diseases. Neurotherapeutics 11, 773-785.

Dubowitz, V., Sewry, C. and Oldfors, A. (2013). Muscle Biopsy: A PracticalApproach. Oxford: Saunders.

Duhamel, T. A.,Green,H. J., Stewart, R.D., Foley, K. P., Smith, I. C. andOuyang, J.(2007). Muscle metabolic, SR Ca(2+) -cycling responses to prolonged cycling, withand without glucose supplementation. J. Appl. Physiol. 103, 1986-1998.

Fajardo, V. A., Bombardier, E., Vigna, C., Devji, T., Bloemberg, D., Gamu, D.,Gramolini, A. O., Quadrilatero, J. and Tupling, A. R. (2013). Co-expression ofSERCA isoforms, phospholamban and sarcolipin in human skeletal muscle fibers.PLoS ONE 8, e84304.

Folker, E. S. and Baylies, M. K. (2013). Nuclear positioning in muscle developmentand disease. Front. Physiol. 4, 363.

Gal, A., Inczedy-Farkas, G., Pal, E., Remenyi, V., Bereznai, B., Geller, L., Szelid,Z., Merkely, B. and Molnar, M. J. (2015). The coexistence of dynamin 2 mutationand multiple mitochondrial DNA (mtDNA) deletions in the background of severecardiomyopathy and centronuclear myopathy. Clin. Neuropathol 34, 89-95.

Gehrig, S. M., van der Poel, C., Sayer, T. A., Schertzer, J. D., Henstridge, D. C.,Church, J. E., Lamon, S., Russell, A. P., Davies, K. E., Febbraio, M. A. et al.(2012). Hsp72 preserves muscle function and slows progression of severemuscular dystrophy. Nature 484, 394-398.

Gibbs, E. M., Davidson, A. E., Telfer, W. R., Feldman, E. L. and Dowling, J. J.(2014). The myopathy-causing mutation DNM2-S619L leads to defectivetubulation in vitro and in developing zebrafish. Dis. Model. Mech. 7, 157-161.

1008

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms

Goonasekera, S. A., Lam, C. K., Millay, D. P., Sargent, M. A., Hajjar, R. J.,Kranias, E. G. and Molkentin, J. D. (2011). Mitigation of muscular dystrophy inmice by SERCA overexpression in skeletal muscle. J. Clin. Invest. 121,1044-1052.

Gorski, P. A., Glaves, J. P., Vangheluwe, P. and Young, H. S. (2013). Sarco(endo)plasmic reticulum calcium ATPase (SERCA) inhibition by sarcolipin is encoded inits luminal tail. J. Biol. Chem. 288, 8456-8467.

Grady, R. M., Teng, H., Nichol, M. C., Cunningham, J. C., Wilkinson, R. S. andSanes, J. R. (1997). Skeletal and cardiacmyopathies in mice lacking utrophin anddystrophin: a model for Duchenne muscular dystrophy. Cell 90, 729-738.

Haghighi, K., Kolokathis, F., Gramolini, A. O., Waggoner, J. R., Pater, L., Lynch,R. A., Fan, G.-C., Tsiapras, D., Parekh, R. R., Dorn, G. W., II et al. (2006). Amutation in the human phospholamban gene, deleting arginine 14, results inlethal, hereditary cardiomyopathy. Proc. Natl. Acad. Sci. USA 103, 1388-1393.

Horowitz, J. D., Rosenson, R. S., McMurray, J. J. V., Marx, N. and Remme, W. J.(2011). Clinical Trials Update AHA Congress 2010. Cardiovasc. Drugs Ther. 25,69-76.

Jessup, M., Greenberg, B., Mancini, D., Cappola, T., Pauly, D. F., Jaski, B.,Yaroshinsky, A., Zsebo, K. M., Dittrich, H., Hajjar, R. J. et al. (2011). Calciumupregulation by percutaneous administration of gene therapy in cardiac disease(CUPID): a phase 2 trial of intracoronary gene therapy of sarcoplasmic reticulumCa2+-ATPase in patients with advanced heart failure. Circulation 124, 304-313.

Jungbluth, H. and Gautel, M. (2014). Pathogenic mechanisms in centronuclearmyopathies. Front. Aging Neurosci. 6, 339.

Jungbluth, H., Zhou, H., Sewry, C. A., Robb, S., Treves, S., Bitoun, M.,Guicheney, P., Buj-Bello, A., Bonnemann, C. and Muntoni, F. (2007).Centronuclear myopathy due to a de novo dominant mutation in the skeletalmuscle ryanodine receptor (RYR1) gene. Neuromuscul. Disord. 17, 338-345.

Jungbluth, H., Wallgren-Pettersson, C. and Laporte, J. (2008). Centronuclear(myotubular) myopathy. Orphanet. J. Rare Dis. 3, 26.

Knotts, S., Rindt, H., Neumann, J. andRobbins, J. (1994). In vivo regulation of themouse beta myosin heavy chain gene. J. Biol. Chem. 269, 31275-31282.

Lannergren, J., Bruton, J. D. and Westerblad, H. (2000). Vacuole formation infatigued skeletal muscle fibres from frog and mouse: effects of extracellularlactate. J. Physiol. 526, 597-611.

Laporte, J., Hu, L. J., Kretz, C., Mandel, J.-L., Kioschis, P., Coy, J. F., Klauck,S. M., Poustka, A. and Dahl, N. (1996). A gene mutated in X-linked myotubularmyopathy defines a new putative tyrosine phosphatase family conserved in yeast.Nat. Genet. 13, 175-182.

Liu, N., Bezprozvannaya, S., Shelton, J. M., Frisard, M. I., Hulver, M. W.,McMillan, R. P., Wu, Y., Voelker, K. A., Grange, R. W., Richardson, J. A. et al.(2011). Mice lacking microRNA 133a develop dynamin 2-dependentcentronuclear myopathy. J. Clin. Invest. 121, 3258-3268.

Lucas, C. A., Kang, L. H. D. and Hoh, J. F. Y. (2000). Monospecific antibodiesagainst the three mammalian fast limb myosin heavy chains. Biochem. Biophys.Res. Commun. 272, 303-308.

MacLennan, D. H. and Kranias, E. G. (2003). Phospholamban: a crucial regulatorof cardiac contractility. Nat. Rev. Mol. Cell Biol. 4, 566-577.

Mazala, D. A. G., Pratt, S. J. P., Chen, D., Molkentin, J. D., Lovering, R. M. andChin, E. R. (2015). SERCA1 overexpression minimizes skeletal muscle damagein dystrophic mouse models. Am. J. Physiol. Cell Physiol. 308, C699-C709.

McMillan, E. M. and Quadrilatero, J. (2011). Differential apoptosis-related proteinexpression, mitochondrial properties, proteolytic enzyme activity, and DNAfragmentation between skeletal muscles. Am. J. Physiol. Regul. Integr. Comp.Physiol. 300, R531-R543.

Millay, D. P., Goonasekera, S. A., Sargent, M. A., Maillet, M., Aronow, B. J. andMolkentin, J. D. (2009). Calcium influx is sufficient to induce muscular dystrophythrough a TRPC-dependent mechanism. Proc. Natl. Acad. Sci. USA 106,19023-19028.

Minamisawa, S., Sato, Y., Tatsuguchi, Y., Fujino, T., Imamura, S.-i., Uetsuka, Y.,Nakazawa, M. and Matsuoka, R. (2003). Mutation of the phospholambanpromoter associated with hypertrophic cardiomyopathy. Biochem. Biophys. Res.Commun. 304, 1-4.

Morita, T., Hussain, D., Asahi, M., Tsuda, T., Kurzydlowski, K., Toyoshima, C.and MacLennan, D. H. (2008). Interaction sites among phospholamban,sarcolipin, and the sarco(endo)plasmic reticulum Ca(2+)-ATPase. Biochem.Biophys. Res. Commun. 369, 188-194.

Murphy, R. M., Verburg, E. and Lamb, G. D. (2006). Ca2+ activation of diffusibleand bound pools of μ-calpain in rat skeletal muscle. J. Physiol. 576, 595-612.

Nakagawa, O., Arnold, M., Nakagawa, M., Hamada, H., Shelton, J. M., Kusano,H., Harris, T. M., Childs, G., Campbell, K. P., Richardson, J. A. et al. (2005).

Centronuclear myopathy in mice lacking a novel muscle-specific protein kinasetranscriptionally regulated by MEF2. Genes Dev. 19, 2066-2077.

Nguyen thi, M., Cartwright, A. J., Morris, G. E., Love, D. R., Bloomfield, J. F. andDavies, K. E. (1990). Monoclonal antibodies against defined regions of themuscular dystrophy protein, dystrophin. FEBS Lett. 262, 237-240.

Nicot, A.-S., Toussaint, A., Tosch, V., Kretz, C., Wallgren-Pettersson, C.,Iwarsson, E., Kingston, H., Garnier, J.-M., Biancalana, V., Oldfors, A. et al.(2007). Mutations in amphiphysin 2 (BIN1) disrupt interaction with dynamin 2 andcause autosomal recessive centronuclear myopathy. Nat. Genet. 39, 1134-1139.

Ottenheijm, C. A. C., Fong, C., Vangheluwe, P., Wuytack, F., Babu, G. J.,Periasamy, M., Witt, C. C., Labeit, S. and Granzier, H. (2008). Sarcoplasmicreticulum calcium uptake and speed of relaxation are depressed in nebulin-freeskeletal muscle. FASEB J. 22, 2912-2919.

Pattison, J. S., Waggoner, J. R., James, J., Martin, L., Gulick, J., Osinska, H.,Klevitsky, R., Kranias, E. G. and Robbins, J. (2008). Phospholambanoverexpression in transgenic rabbits. Transgenic Res. 17, 157-170.

Rindt, H., Gulick, J., Knotts, S., Neumann, J. and Robbins, J. (1993). In vivoanalysis of the murine beta-myosin heavy chain gene promoter. J. Biol. Chem.268, 5332-5338.

Rindt, H., Knotts, S. and Robbins, J. (1995). Segregation of cardiac and skeletalmuscle-specific regulatory elements of the beta-myosin heavy chain gene. Proc.Natl. Acad. Sci. USA 92, 1540-1544.

Romero, N. B. (2010). Centronuclear myopathies: a widening concept.Neuromuscul. Disord. 20, 223-228.

Sartore, S., Gorza, L. and Schiaffino, S. (1982). Fetal myosin heavy chains inregenerating muscle. Nature 298, 294-296.

Schiaffino, S. and Reggiani, C. (2011). Fiber types in mammalian skeletalmuscles. Physiol. Rev. 91, 1447-1531.

Schiaffino, S., Gorza, L., Sartore, S., Saggin, L., Ausoni, S., Vianello, M.,Gundersen, K. and Lømo, T. (1989). Three myosin heavy chain isoforms in type2 skeletal muscle fibres. J. Muscle Res. Cell Motil. 10, 197-205.

Schmitt, J. P., Kamisago, M., Asahi, M., Li, G. H., Ahmad, F., Mende, U., Kranias,E. G., MacLennan, D. H., Seidman, J. G. and Seidman, C. E. (2003). Dilatedcardiomyopathy and heart failure caused by a mutation in phospholamban.Science 299, 1410-1413.

Schneider, J. S., Shanmugam, M., Gonzalez, J. P., Lopez, H., Gordan, R.,Fraidenraich, D. and Babu, G. J. (2013). Increased sarcolipin expression anddecreased sarco(endo)plasmic reticulum Ca2+ uptake in skeletal muscles ofmouse models of Duchenne muscular dystrophy. J. Muscle Res. Cell Motil. 34,349-356.

Song, Q., Young, K. B., Chu, G., Gulick, J., Gerst, M., Grupp, I. L., Robbins, J.and Kranias, E. G. (2004). Overexpression of phospholamban in slow-twitchskeletal muscle is associated with depressed contractile function and muscleremodeling. FASEB J. 18, 974-976.

Szasz, G., Gruber, W. and Bernt, E. (1976). Creatine kinase in serum:1. Determination of optimum reaction conditions. Clin. Chem. 22, 650-656.

Tarnopolsky, M. A., Pearce, E., Smith, K. and Lach, B. (2011). Suction-modifiedBergstrom muscle biopsy technique: experience with 13,500 procedures.MuscleNerve 43, 716-725.

Tupling, A. R. (2009). Excitation-contraction coupling. In Encyclopedia ofNeuroscience (ed. M. Binder, N. Hirokawa and U. Windhorst), pp. 1479-1483.Berlin; Heidelberg: Springer.

Tupling, R. and Green, H. (2002). Silver ions induce Ca2+ release from the SR invitro by acting on the Ca2+ release channel and the Ca2+ pump. J. Appl. Physiol.92, 1603-1610.

Tupling, A. R., Asahi, M. andMacLennan, D. H. (2002). Sarcolipin overexpressionin rat slow twitch muscle inhibits sarcoplasmic reticulum Ca2+ uptake and impairscontractile function. J. Biol. Chem. 277, 44740-44746.

Wilmshurst, J. M., Lillis, S., Zhou, H., Pillay, K., Henderson, H., Kress, W.,Muller, C. R., Ndondo, A., Cloke, V., Cullup, T. et al. (2010). RYR1mutations area common cause of congenital myopathies with central nuclei. Ann. Neurol. 68,717-726.

Zsebo, K., Yaroshinsky, A., Rudy, J. J., Wagner, K., Greenberg, B., Jessup, M.and Hajjar, R. J. (2014). Long-term effects of AAV1/SERCA2a gene transfer inpatients with severe heart failure: analysis of recurrent cardiovascular events andmortality. Circ. Res. 114, 101-108.

Zubrzycka-Gaarn, E., MacDonald, G., Phillips, L., Jorgensen, A. O. andMacLennan, D. H. (1984). Monoclonal antibodies to the Ca2++Mg2+-dependentATPase of sarcoplasmic reticulum identify polymorphic forms of the enzyme andindicate the presence in the enzyme of a classical high-affinity Ca2+ binding site.J. Bioenerg. Biomembr. 16, 441-464.

1009

RESOURCE ARTICLE Disease Models & Mechanisms (2015) 8, 999-1009 doi:10.1242/dmm.020859

Disea

seModels&Mechan

isms