transgenic expression of inclusion body myopathy associated mutant

Transgenic expression of inclusion bodymyopathy associated mutant p97/VCPcauses weakness and ubiquitinatedprotein inclusions in mice

Conrad C. Weihl1,2,*, Sara E. Miller1, Phyllis I. Hanson2 and Alan Pestronk1

1Department of Neurology and 2Department of Cell Biology and Physiology, Washington University School

of Medicine, 660 S. Euclid Avenue, Saint Louis, MO 63110, USA

Received December 18, 2006; Revised and Accepted February 19, 2007

Mutations in p97/VCP cause the autosomal-dominant, inherited syndrome inclusion body myopathy (IBM)associated with Paget’s disease of the bone and frontotemporal dementia (IBMPFD) (Watts, G.D., Wymer,J., Kovach, M.J., Mehta, S.G., Mumm, S., Darvish, D., Pestronk, A., Whyte, M.P. and Kimonis, V.E. (2004)Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is causedby mutant valosin-containing protein. p97/VCP is a multi-functional protein with a role in the ubiquitin-proteasome system (UPS) (Wang, Q., Song, C. and Li, C.C. (2004) Molecular perspectives on p97-VCP: pro-gress in understanding its structure and diverse biological functions. To understand how mutations inthis protein lead to a myopathy, we generated several lines of transgenic mice expressing p97/VCP-WT(TgVCP-WT) or the most common IBMPFD mutant, p97/VCP R155H (TgVCP-RH), under a muscle-specific pro-moter. TgVCP-RH animals, but not controls, became progressively weaker in a dose-dependent manner start-ing at 6 months of age. Abnormal muscle pathology, which included coarse internal architecture, vacuolationand disorganized membrane morphology with reduced caveolin-3 expression at the sarcolemma developedcoincident with the onset of weakness. These changes were not associated with alterations in sarcolemmalintegrity as measured by muscle fiber uptake of Evan’s blue dye. Even before animals displayed measurableweakness, there was an increase in ubiquitin-containing protein inclusions and high-molecular-weightubiquitinated proteins, markers of UPS dysfunction. We suggest that this early and persistent increase inubiquitinated proteins induced by IBMPFD mutations in p97/VCP may ultimately lead to animal weaknessand the observed muscle pathology. TgVCP-RH animals will be a valuable tool for understanding the patho-genesis of IBM and the role of the UPS in skeletal muscle.

INTRODUCTION

Inclusion body myopathies (IBMs) are a group of acquired andhereditary disabling muscle disorders with myopathic features,‘rimmed vacuoles’ and accumulations of aggregated proteins(1). IBM associated with Paget’s disease of the bone(PDB) and frontotemporal dementia (FTD) (IBMPFD) isan autosomal-dominantly, inherited multisystem disorderbecause of mutations in the protein p97/VCP (2). The pene-trance of each syndromic feature is variable with the mostcommon initial presentation being weakness in 84% of patients,typically with onset in the fourth decade (2,3). PDB and FTD

occur less frequently with a penetrance of 51 and 31% andmean onset of 42 and 55 years of age, respectively (2,3).Muscle pathology in IBMPFD patients has been described asa myopathy with rimmed vacuoles and no inflammation (4,5).Other features include p97/VCP containing ubiquitinatedprotein inclusions in both myonuclei and the sarcoplasm (4).Consistent with these findings, the central nervous system(CNS) pathology in IBMPFD includes intranuclear and cyto-plasmic ubiquitinated inclusions in affected neurons (6,7).

The pathogenic mechanisms leading to muscle weakness inIBM, and IBMPFD in particular, are unclear, but may relate

# The Author 2007. Published by Oxford University Press. All rights reserved.For Permissions, please email: [email protected]

*To whom correspondence should be addressed. Tel: þ1 3143626981; Fax: þ1 3143623752; Email: [email protected]

Human Molecular Genetics, 2007, Vol. 16, No. 8 919–928doi:10.1093/hmg/ddm037Advance Access published on February 28, 2007

to the accumulation and aggregation of undegraded proteins.In the case of IBM, protein aggregates have been shown tocontain b-amyloid, hyperphosphorylated tau and ubiquitin,proteins known to accumulate in other aggregate diseases(1). One hypothesis is the pathogenic pathway in IBM thatleads to muscle damage relates to a dysfunction in ubiquitin-proteasome system (UPS)-mediated protein degradationresulting in the accumulation of aggregated proteins (8,9).p97/VCP belongs to the AAA (ATPases associated with avariety of cellular activities) family of proteins. The cellularfunction of p97/VCP is complex because it is involved indiverse cellular processes including protein degradation, orga-nelle biogenesis and cell-cycle regulation (10). One cellularfunction of p97/VCP is as regulator of the UPS. p97/VCPassociates with several pathway components, including polyu-biquitinated proteins, 26S proteasome subunits and a varietyof E3 ubiquitin ligases (11–13). Depletion of p97/VCP fromcell extracts leads to an accumulation of cytosolic polyubiqui-tinated undegraded proteins (11). In addition to the degra-dation of cytosolic proteins, p97/VCP is involved in thedegradation of at least a subset of improperly folded proteinsfrom the membrane and lumen of the endoplasmic reticulum(ER), a process known as endoplasmic-reticulum-associateddegradation (ERAD) (14). The expression of dominant nega-tive p97/VCP in cultured cells impairs the degradation ofERAD substrates and leads to extensive ER-derived vacuo-lation (15). IBMPFD mutant p97/VCP expression leads toan increase in ubiquitinated protein aggregates in culturedC2C12 myoblasts (16). In addition, degradation of an ERADsubstrate is slowed in these cells (16). In this study, we over-express the most common IBMPFD mutation in p97/VCP(R155H) in mouse muscle. These animals recapitulate someof the key phenotypic and skeletal muscle features ofIBMPFD and open the way to a more detailed study of theunderlying molecular and pathogenic mechanisms in IBM.

RESULTS

Generation of mice overexpressing p97/VCP-WTand p97/VCP-R155H in skeletal muscle

We modeled the muscle phenotype of IBMPFD in mice byexpressing the human IBMPFD R155H mutant transgene inthe presence of normal endogenous murine p97/VCP underthe regulation of a muscle creatine kinase (MCK) promoter.The R155H mutation is the most common IBMPFD mutationand this residue is mutated in .80% of families with IBMPFD(2). As a control, wild-type human p97/VCP was over-expressed, again along with normal endogenous murine p97/VCP. The MCK promoter allows for strong expression in skel-etal muscle with minimal expression in other tissues (17). Thispromoter construct has been used previously to drive skeletalmuscle-specific gene expression in other models of myopathy(18). Human and murine p97/VCP differ by only one aminoacid at residue 684 and are known to co-assemble in vitro(15). In order to avoid complicating interpretation ofchanges associated with the R155H mutation, we elected tonot place an epitope tag on the transgene.

Nine independent IBMPFD mutant p97/VCP-R155Hexpressing lines (TgVCP-RH) and three p97/VCP-WT

expressing lines (TgVCP-WT) were generated. Initial charac-terization included confirmation of transgene expression insubsequent offspring and routine skeletal muscle pathologyfrom four independent TgVCP-RH and two independentTgVCP-WT lines. All four TgVCP-RH lines had evidenceof disordered skeletal muscle internal architecture after .9months of age. After this initial characterization, lines RH9,RH12, WT1 and WT2 were bred further (to at least the F3generation) with C57/B6 mice and subjected to more compre-hensive analysis. All data shown are from these animals.Unless stated in the text, histologic and immunohistologicimages shown are representative of both TgVCP-WT lines(WT1 and WT2) or both TgVCP-RH lines (RH9 and RH12).For ease of readership, these animals will be referred toTgVCP-WT or TgVCP-RH, and non-transgenic littermateswill be referred to as control animals.

Murine and human p97/VCP differ by only one amino acid,migrate at the same electrophoretic mobility on sodiumdodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), and are both recognized equivalently by the mono-clonal p97/VCP antibodies used in this study. Western blots,using an anti-p97/VCP antibody, of skeletal muscle homogen-ates from transgenic animals and non-transgenic control litter-mates demonstrate an increase in p97/VCP protein in alltransgenic lines (Fig. 1A).

In order to further confirm transgene expression and assurethe correct cellular localization of our transgenic protein, weimmunostained control, TgVCP-WT and TgVCP-RH skeletalmuscle from adult animals with an antibody specific for p97/VCP. Previous reports demonstrate that p97/VCP in humanskeletal muscle is concentrated in endothelial vasculatureand is present at lower levels diffusely throughout the sarco-plasm (2,4,19). To clearly define the skeletal muscle localiz-ation of endogenous and transgene expressed human p97/VCP, we co-labeled muscle sections with anti-p97/VCP anti-body and an antibody to collagen IV which is localized tothe muscle basal lamina and endothelial vessels. Similar tohuman studies, immunostaining showed that endogenousmurine p97/VCP is located predominantly in endomysial con-nective tissue and medium-sized vessels (Fig. 1B). BothTgVCP-WT and TgVCP-RH muscle showed an increase in theamount of sarcoplasmic and subsarcolemmal p97/VCP with nochange in endomysial immunoreactivity (Fig. 1C and D).

TgVCP-RH mice are weak and have disease-associatedpathology

After breeding to the F3 generation with C57/B6 mice, a 1:1Mendelian inheritance ratio was observed for all strains[17:18 (positive:negative) and 15:16 (þ :2) for WT1 andWT2 strains, respectively, and 15:15 (þ :2) and 26:26(þ :2) for strains RH9 and RH12, respectively, suggestingthat transgene expression does not affect the pattern of inheri-tance. Control, TgVCP-WT and TgVCP-RH animals weresimilar in size and weight and had no differences in mortalityup to 15 months of age. To see if these animals had evidenceof muscle weakness, we measured mouse forelimb gripstrength using a well-characterized system (20). Grip strengthtesting consists of five separate measurements using a trapezebar attached to a force transducer that records peak generated

920 Human Molecular Genetics, 2007, Vol. 16, No. 8

force (Stoelting, WoodDale, IL). Mice instinctively grab thebar with their forepaws and continue to hold when beingpulled backwards by the tail, releasing only when unable tomaintain grip. The resulting measurement is recorded andthe three highest measurements are averaged to give thestrength score. Normal animals can generate �400–500 g offorce (20). Strength testing was performed on TgVCP-WT,TgVCP-RH and non-transgenic littermate control animals.TgVCP-RH animals were weak compared with control andTgVCP-WT lines. Weakness progressively worsened from 6to 10 months with 10-month-old animals generating one-halfthe force of age-matched controls (Fig. 1E). Interestingly,TgVCP-RH strain TgVCP-RH12 which expresses higherlevels of mutant protein (see Fig. 1A) becomes weakerearlier than TgVCP-RH9 which expresses less mutantprotein. These data suggest that there may be a dose-relatedeffect of mutant transgene expression. There was no evidenceof weakness at 3 months of age in any of the animal lines.

Given that TgVCP-RH animals were weak between 6 and10 months of age, we elected to examine the histopathologicchanges in their skeletal muscle. We performed routine histo-chemical analysis on mouse skeletal muscle sections frombiceps, triceps, quadriceps/hamstrings, soleus/gastrocneumiusand diaphragm. In addition to myopathic features whichinclude abnormal variation in muscle fiber size and anincrease in endomysial connective tissue, the most distinctivefeature of IBM muscle pathology is rimmed vacuoles seen byhematoxylin and eosin (H&E) or modified Gomori trichrome(mGT) stains. Although our TgVCP-RH animal skeletalmuscle sections stained with H&E did not show classicrimmed vacuoles, they did show prominent small linear baso-philic ‘rimmed cracks.’ in the sarcoplasm and subsarcolemmalregions at both 6 and 10 months of age (Fig. 2B and C) thatwere not present in control or TgVCP-WT animals (Fig. 2A).Staining of similar muscle sections with mGT demonstrateddisordered internal architecture with blue staining inclusionsand linear red rimmed cracks between myofibrils inTgVCP-RH animals (Fig. 2E). These changes were notpresent in age-matched TgVCP-WT (Fig. 2D) or controlanimals. TgVCP-RH animals also demonstrated abnormalvariation in muscle fiber size and an increase in endomysialconnective tissue that was more severe in the diaphragmand most prominent after 12 months of age (compareFig. 2F with G). Sarcoplasmic and occasionally myonuclearcongo red positive staining was seen at .15 months of agein only one of the characterized TgVCP-RH lines (RH12)and was again most prominent in the diaphragm (Fig. 2H).No congo red staining was seen in TgVCP-WT or controlanimals at similar time points.

Skeletal muscle tissue from patients with IBM and IBMPFDhas prominent ubiquitin containing inclusions of both sarco-plasmic and myonuclear (1,4). We looked for this in ourTgVCP-RH animals using the FK2 antibody that recognizesubiquitinated proteins. FK2 immunofluorescence revealed pro-minent ubiquitinated protein inclusions that were sarcoplas-mic, subsarcolemmal and intramyonuclear in skeletal muscletissue from 6-month-old TgVCP-RH animals (Fig. 2J). Age-matched control and TgVCP-WT animals demonstrated FK2immunoreactivity that was less intense around myonuclei(Fig. 2I).

Figure 1. Generation of TgVCP-WT and TgVCP-RH transgenic mouselines. (A) Total quadriceps lysates from 3-month-old control (Non-trg),TgVCP-WT (wtTrg1 and wtTrg2) and TgVCP-RH (mutTrg9, mutTrg6and mutTrg12) animals were immunoblotted for p97/VCP. Controllanes represent the amount of endogenous p97/VCP. All transgeniclines had an increase in p97/VCP protein levels. Notably wtTrg2 andmutTrg9 transgenic lines have comparable levels of protein and wereused in subsequent experiments. Actin loading control for each sampleis shown below. (B–D) Dual immunofluorescence of 3-month-old quad-riceps muscle from (B) control, (C) TgVCP-WT and (D) TgVCP-RHanimals using anti-p97/VCP (red) and anti-collagen IV (green) anti-bodies. Endogenous murine p97/VCP is predominantly localized to endo-mysial connective tissue with a low level in the sarcoplasm. BothTgVCP-WT and TgVCP-RH animals have an increase in sarcoplasmicp97/VCP immunofluorescence that does not occur in the endomysium.White bar is 25 mm. (E) Quantitative grip strength testing in grams, ofcontrol, TgVCP-WT (WT1, WT2) and TgVCP-RH (RH9, RH12) at 3,6 and 10 months of age. TgVCP-RH12 animals have statistically signifi-cant weakness beginning at 6 months that is worsened by 10 months ofage and TgVCP-RH9 animals develop statistically significant weaknessat 10 months of age. At least five animals from each transgenic linefor each time point was measured. �P-value of ,0.05; ��P-value of,0.005.

Human Molecular Genetics, 2007, Vol. 16, No. 8 921

IBMPFD and IBM patient skeletal muscle has p97/VCPcontaining inclusions (2,4,19). We looked for this in our trans-genic animals using two different antibodies to p97/VCP. Nosimilar change was seen in any of the animal lines.

Longitudinal toludine blue semi-thin plastic-embedded sec-tions of 6-month-old TgVCP-RH skeletal muscle showedintermyofibril vacuoles consistent with sarcotubular dilationthat was not present in age-matched control or TgVCP-WTanimals (compare Fig. 3A with 4A). Higher resolution analy-sis by transmission EM highlighted disruption betweenmyofibrils in TgVCP-RH animal skeletal muscle thatwas not evident in TgVCP-WT animals (compare Fig. 3Bwith C). Other features seen in TgVCP-RH muscle includedsarcotubular dilatation that was in continuity with myeloidbodies and vacuoles containing proteinaceous debris(Fig. 3D and E). In addition, there were prominent autophago-somes in TgVCP-RH muscle (Fig. 3D).

TgVCP-RH mice have disorganized sarcolemmal structure

In TgVCP-RH animals, 6 months of age or older sarcolemmalstructure was markedly abnormal with prominent invagina-tions and protrusions that were not apparent in control orTgVCP-WT animals. This pathology was seen best onlongitudinal toludine blue plastic-embedded sections(compare Fig. 4A with B). These changes prompted us toexplore the localization of proteins normally found in thesarcolemma. Consistent with IBMPFD patient skeletalmuscle, there was normal sarcolemmal localization of dystro-phin, a2-laminin and a-sarcoglycan (5) in TgVCP-WT andTgVCP-RH animals (Fig. 4C–H). However, abnormal sarco-lemmal morphology is still apparent in TgVCP-RH animals’skeletal muscle sections (Fig. 4 D, F and H). In contrast,immunostaining for caveolin-3 was reduced on the sarco-lemma and increased in the sarcoplasm of TgVCP-RH butnot control or TgVCP-WT animals (compare Fig. 5A withB). Immunoblots showed that total amounts of caveolin-3protein in muscle were unchanged in TgVCP-RH animalscompared to control and TgVCP-WT animals (Fig. 5C). Todetermine whether the changes in sarcolemmal structurewere associated with changes in sarcolemmal membraneintegrity, we injected 9-month-old control, TgVCP-WT andTgVCP-RH animals with Evan’s blue dye (EBD). EBDassociates with albumin and is taken up into muscle cellswhen the sarcolemma is damaged (21). There was littleEBD dye present in myofibers from control, TgVCP-WTand TgVCP-RH quadriceps muscle (Fig. 5D–F). The increasein EBD around muscle fibers in the TgVCP-RH tissues isconsistent with the increased connective tissue seen inTgVCP-RH animals (22).

Ubiquitinated protein accumulation precedes muscleweakness and IBM pathology in TgVCP-RH animals

At time points earlier than 6 months of age, there was noevidence of animal weakness or prominent histopathologicchanges in skeletal muscle from TgVCP-RH, TgVCP-WTor control animals. In order to identify a marker of diseasethat might precede these symptomatic changes, we elected

Figure 2. TgVCP-RH mice have IBM-associated disease pathology.Unless stated in the text, there was no qualitative difference betweenTgVCP-WT (WT1 or WT2) strains or TgVCP-RH (RH9 or RH12) strains,and a representative micrographs of skeletal muscle from TgVCP-WT orTgVCP-RH animals is shown. Diaphragmatic muscle from (A, D, F)TgVCP-WT or (B, C, E, G, H) TgVCP-RH (B, I, J) at 6 months, (C–E)10 months or (F and G) 12 months of age. Frozen muscle sections arestained (A–C) with hematoxlin and eosin and (D and E) modified gomori-trichrome while (F and G) are semi-thin plastic-embedded sections stainedwith toludine blue. White arrows demonstrate basophilic (B and C) or red(E) rimmed cracks. Black arrows (G) highlight variability in muscle fibersize characteristic of a myopathy. (H) Congo red staining visualized byTexas red filters demonstrates positive staining of fibers both in the sarcoplasmand myonuclei of 15-month-old TgVCP-RH12 animal skeletal muscle.Panels I and J are frozen tissue sections of quadriceps skeletal muscle from6-month-old TgVCP-WT (I) and TgVCP-RH animals immunostained forubiquitinated proteins with FK2 antibody. Closed white arrow highlightslarge sarcoplasmic and open white arrow highlights myonuclear congo red(H) or ubiquitinated protein (J) inclusions. Black line at bottom of eachfigure is 25 mm.


to look at FK2 immunoreactivity in 3 months oldTgVCP-RH animals. There was an increase in FK2 immu-noreactivity in these animals. The increase in FK2 immunor-eactivity became more pronounced as the animals aged(Fig. 6C). The FK2 immunoreactivity appeared to be moreprominent in myonuclei in 3-month-old animals and wasboth sarcoplasmic and myonuclear in older animals. Interest-ingly, there was an age-dependent increase in ubiquitinatedproteins in both control and TgVCP-WT animals (Fig. 6Aand B). The sarcolemmal FK2 signal seen in skeletalmuscle sections is not consistently changed across all linesand ages, suggesting that the increase in FK2 immunostain-ing is myonuclear and sarcoplasmic. Western blots ofskeletal muscle lysates using the FK2 antibody showed aparallel increase in high-molecular-weight ubiquitinated

proteins that was evident at time points as early as 30 daysand persisted at 10 months of age (Fig. 6D and E forhistogram of the densitometry three independent results).

DISCUSSION

We generated a transgenic animal model using skeletalmuscle-specific expression of IBMPFD mutant p97/VCP.These animals develop progressive weakness, vacuolation,sarcolemmal pathology and congo red positive and ubiquiti-nated inclusions. Importantly, an abnormal level of ubiquiti-nated proteins in affected mouse skeletal muscle isdetectable at time points as early as 1 month of age prior toanimal weakness and other pathology.

Notably, transgenic expression of p97/VCP-R155H failed togenerate classic rimmed vacuoles but instead had evidence ofrimmed cracks. We elected to look at skeletal muscle pathol-ogy coincident with and before the onset of animal weakness.This allowed us to correlate changes in muscle pathology withdisease progression in our animals. An additional reason isthat aged inbred mouse strains have skeletal muscle pathology

Figure 4. TgVCP-RH mice have abnormal sarcolemmal membranes. Semi-thin plastic embedded sections of 9-month-old (A) TgVCP-WT or (B)TgVCP-RH diaphragm tissue demonstrating distorted sarcolemma and darkstaining linear inclusions. Dystrophy-related immunostaining of quadricepsfrom 6-month-old (C, E and G) TgVCP-WT and (D, F and H) TgVCP-RHanimals. (C–D) dystrophin, (E and F) a2-laminin, (G and H) a-sarcoglycan.Line at bottom is 25 mm.

Figure 3. Ultrastructure of TgVCP-RH animal skeletal muscle. (A) Toludineblue-stained semi-thin plastic sections of 6-month-old diaphragm tissue fromTgVCP-RH animals. Arrows denote sarcotubular swellings between myofi-brils. (B and C) Electron micrograph of 9-month-old TgVCP-WT (B) orTgVCP-RH (C–E) diaphragm muscle. Note autophagosomes in continuitywith sarcotubular vacuoles (white arrow) containing membranous and protein-aceous debris. White bars in (B–E) are 500 nm.


such as tubular aggregates and mitochondrial changes thatmay confound the pathologic examination of older ages(23). It may be that rimmed vacuole formation occurs attime points later than analyzed in our study and may explainthe variability and occasional rarity of rimmed vacuoles seenin IBMPFD patient biopsies (4). Moreover, it is intriguingthat animal weakness can occur with ubiquitinated proteinaccumulation in the absence of rimmed vacuoles, suggestingthat UPS dysfunction is the principal cause of muscle weak-ness in these animals.

It is unclear how the accumulation of ubiquitinated proteinsleads to skeletal muscle dysfunction. It may be that unde-graded proteins are intrinsically toxic to muscle fibers, andas ubiquitinated proteins accumulate over time, disorderedinternal architecture, vacuolation, sarcolemmal changes andweakness ensue. It has been suggested that perturbations inthe UPS may be involved in the pathogenesis of IBM (8).This speculation was based on immunohistochemical analysisof acquired and hereditary forms of IBM in which ubiquiti-nated protein inclusions are a prominent feature (24). Theidentification of mutations in p97/VCP, a protein essentialfor UPS function in IBMPFD patients, further supportsthe concept that aberrant protein degradation throughthe UPS is pathogenic in skeletal muscle and leads to IBM.The CNS changes in IBMPFD also suggest involvement

of the UPS. The dementia in IBMPFD is accompanied byubiquitin-positive cytoplasmic and nuclear inclusions similarto those in the human IBMPFD myopathies and ourTgVCP-RH animal skeletal muscle (4,6,7; and unpublishedobservations).

Our animal model demonstrated intermyofiber vacuolation.Ultrastructural analysis suggests that vacuolation may be sar-cotubular in origin. Dominant negative p97/VCP expression incultured cells causes extensive ER-derived vacuolation (15),however the role of p97/VCP in the extensively organizedskeletal muscle sarcotubular system is unknown. IBMPFDmutants in cell culture cause dilation of the ER network aswell (16). These changes may relate to impairment in p97/VCP’s role in ERAD (14). It is interesting that the loss ofTrim32, an E3 ligase important in the ubiquitination of skel-etal muscle-specific substrates, causes sarcotubular swellingin patients with LGMD2H (25). In addition, loss of sil1, anER resident co-chaperone essential to ERAD, leads to ubiqui-tinated protein accumulation in cells and a rimmed vacuolemyopathy (26).

We and others have described sarcoplasmic accumulationsof p97/VCP in IBMPFD patient skeletal muscle (2,4,19). Inaddition, our previous study found that transient overexpres-sion in cell culture of IBMPFD mutant p97/VCP containinga C-terminal GFP tag led to the presence of cytoplasmic,

Figure 5. TgVCP-RH mice have a reduction of caveolin-3 at the sarcolemma with normal sarcolemmal integrity. Caveolin-3 immunostaining of quadricepsmuscle from 6 -month-old (A) TgVCP-WT and (B) TgVCP-RH quadriceps muscle. Note the decrease in caveolin-3 immunoreactivity at the sarcolemmaand the accumulation of punctuate sarcoplasmic accumulations. (C) Immunoblot for caveolin-3 from quadriceps lysates from 6-month-old control,TgVCP-WT (WT1 and WT2), TgVCP-RH (RH9 and RH12). (D–F) Evan’s blue dye (EBD) uptake in quadriceps muscle from 9-month-old (D) control, (E)TgVCP-WT and (F) TgVCP-RH animals 16 h post-intraperitoneal injection. Note there is minimal dye uptake in all myofibers (arrow denotes one single positivemyofiber in TgVCP-RH animal). The increase in EBD fluorescence surrounding myofibers is secondary to the increase in connective tissue in TgVCP-RH animalmuscle. Line at bottom is 25 mm.


Figure 6. TgVCP-RH mice have an increase in ubiquitinated proteins. Immunostaining for ubiquitinated proteins using the FK2 antibody of quadricepsmuscle from 3-, 6- or 9-month-old (A) control, (B) TgVCP-WT, and (C) TgVCP-RH animals. Note the increase in ubiquitinated protein inclusions thatare sarcoplasmic and nuclear. Inset box highlights an FK2 lined rimmed vacuole. Line at the bottom is 25 mm. (D) Immunoblot of quadriceps musclelysates from control, TgVCP-WT (WT1) and TgVCP-RH (RH9 and RH12) with the FK2 antibody. Note the increase in high-molecular-weight ubiquitinatedproteins (HMW-Ub) in TgVCP-RH muscle lysates. In order to visualize a quantitative difference in the HMW-Ub smear, a lighter exposure of the 10-monthtime point is shown as well. Myosin loading control is below blot. (E) Densitometric quantitation of immunoblots developed with ECL substrate fromthree independent experiments is shown. Error bars represent standard deviation and � denotes P-value , 0.01 when compared to control animals of thesame age.


endoplasmic-reticulum-associated p97/VCP inclusions (16).This is in contrast to our current data in which we see fewp97/VCP inclusions in skeletal muscle from our transgenicmouse lines. It may be that transient overexpression ofGFP-tagged IBMPFD mutants in cell culture exaggerates apropensity for these mutants to aggregate that is not detectedas readily in our transgenic animals.

p97/VCP has a clear role in the UPS and associates withseveral pathway components including polyubiquitinated pro-teins. Currently, it is unclear whether IBMPFD mutations inp97/VCP disrupt its normal function in this pathway andresult in a loss of function or perhaps mutations confer atoxic gain of an unrelated function leading to weakness.However, current evidence suggests that these mutationsmay confer a gain of function. In particular, previous studiesdemonstrate that IBMPFD mutant p97/VCP has normalATPase activity and normal binding to three essentialco-factors, Ufd1, Npl4 and ataxin-3 (4,16). Our data demon-strate worsening skeletal muscle weakness with higherexpression of IBMPFD mutant p97/VCP, suggesting thatIBMPFD mutant p97/VCP may behave in a gain of functionmanner. Moreover, the highest expressing transgenic line,TgVCP-RH12, was the only line that demonstrated congophi-lic amyloid inclusions (although at 15 months of age).

The prominent sarcolemmal membrane abnormalities in ourmouse model have not been previously described in aggregatemyopathies. Interestingly, these changes happen in theabsence of fiber necrosis and are not associated with an altera-tion in membrane permeability as assessed by EBD uptake.This is consistent with skeletal muscle biopsies fromIBMPFD patients which do not show prominent muscle fiberdegeneration and have modestly elevated serum creatinekinase (2,4,5). Whether the sarcolemmal changes are associ-ated with alterations in membrane polarization and calciumrelease remains to be determined, but has been seen in othermouse models of IBM (27).

The mislocalization of caveolin-3 in TgVCP-RH animals isassociated with the sarcolemmal changes and temporally cor-relate with animal weakness. Why these changes occur inassociation with the accumulation of ubiquitinated proteinsis unclear? Mutations in caveolin-3 cause a variety of skeletalmuscle disorders including LGMD1C, a distal myopathy andrippling muscle disease (28). The mislocalization ofcaveolin-3 and structural changes in sarcolemmal membranescould play a role in the weakness in our TgVCP-RH mice.Patients with IBMPFD have a clinical picture with both prox-imal and distal muscle weakness and overlap the phenotypicspectrum of caveolinopathies (4,5). In addition, a similar mis-localization of caveolin-3 is seen in IBMPFD patient musclebiopsies (unpublished observations). Dysregulation of theUPS has been implicated in caveolin-3-associated muscledisease. In particular, caveolin-3 has been identified as anERAD substrate and disease-causing mutations in caveolin-3cause the protein to aggregate, sequester normal caveolin-3and mislocalize to the Golgi (29). These changes are correctedupon proteasome inhibition (29). Although a direct interactionbetween caveolin-3 and p97/VCP has not been demonstrated,it is conceivable that p97/VCP regulation of ERAD substrates,such as caveolin-3, causes the sarcolemmal membranechanges in our transgenic animals.

In summary, we find that transgenic expression of IBMPFDmutant p97/VCP results in an increase in ubiquitinated pro-teins in skeletal muscle. The accumulation of these ubiquiti-nated proteins precedes other skeletal muscle pathology suchas vacuolation and sarcolemmal membrane changes, whichinstead appear to coincide with the onset of measurableanimal weakness. We suggest that the accumulation of ubiqui-tinated proteins over time results in skeletal muscle weaknessand myopathic pathology. The study of IBMPFD muscledisease using our transgenic mouse model will lend insightinto this unique disorder and should also shed light on otherdiseases with UPS dysfunction and prominent ubiquitinpathology.

MATERIALS AND METHODS

Transgenic animal construction

Human p97/VCP cDNA was obtained from Drs VirginiaKimonis and Giles Watts. The R155H mutation was generatedusing QuickChange site-directed mutagenesis (Stratagene) andconfirmed by direct DNA sequencing. Both p97/VCP-WT andp97/VCP-R155H cDNAs were subcloned directly into a1256MCK-CAT transgenic targeting vector (obtained fromDr Stephen Hauschka, University of Washington) using aSeamless Cloning kit (Stratagene). The promoter and codingsequence were confirmed by DNA sequence analysis. Alinear fragment containing the MCK-p97/VCP sequence wasisolated by digesting the targeting vector with HindIII andKpnI and subsequent gel purification. This fragment wassent to the mouse genetics core facility at Washington Univer-sity for transgenic animal production. Animals were screenedfor transgene insertion using PCR amplification of tail DNA.Animals were bred to C57/B6 (Jackson Laboratories) to atleast the F3 generation for phenotypic analysis. Controlanimals were non-transgenic littermates.

Skeletal muscle immunoblot

Freshly isolated skeletal muscle tissue was flash-frozen inliquid nitrogen and stored at 2808C. Frozen tissue wasadded to 1� skeletal muscle sample buffer [4% SDS,125 mM Tris pH 8.8, 40% glycerol, 0.5 mM phenylmethyl-sulfonyl fluoride (PMSF), 100 mM dithiothreitol (DTT)] andhomogenized on ice with a mortar and pestle and thenbriefly sonicated with a Branson Sonifier with 50% cyclewith an output of six on ice for 5 min. The lysate was thenspun at 16 000g for 10 min and supernatant was collected.Equivalent amounts of protein lysates were added to eachlane after the addition of 2� Laemmli buffer and boiling for5 min. Immunoblots were performed as previously described(16). The following antibodies were used: mouse anti-p97(BD Biosciences), rabbit anti-caveolin-3 (ABcam), FK2(biomol), mouse anti-actin (Sigma).

Skeletal muscle histochemistry and immunohistochemistry

Freshly isolated skeletal muscle was mounted using tragacanthgum and quick frozen in liquid-nitrogen-cooled isopentane.Samples were stored at 2808C until sectioning. Frozen


biopsy samples were sectioned into 7 mm thick sections. Skel-etal muscle was taken from biceps, triceps, quadriceps/ham-strings, soleus/gastrocneumius, diaphragm and axial muscle.Histochemistry and immunohistochemistry were performedas previously described (2,5). Congo red staining was visual-ized using a Texas red filter set as previously described forskeletal muscle tissue (30). In addition to the antibodiesused above for immunoblots, rabbit polyclonal-anti-p97/VCP(obtained from Virginia Kimonis) rabbit anti-collagen IV(ABcam), mouse monoclonal anti-dystrophin (Dys-1)(ABcam), mouse monoclonal anti-a2-laminin (ABcam), andmouse monoclonal anti-a-sarcoglycan (ABcam) were used.In the case of FK2 immunohistochemistry, fixed muscle sec-tions were incubated in 1 M KCl, 30 mM HEPES, 65 mM

PIPES, 10 mM EDTA, 2 mM MgCl2, pH 6.9 for 1 h at roomtemperature to remove soluble proteins prior to immunolabel-ing as has been described previously for immunohistochemicaldetection of aggregate prone proteins (31).

For thin-section electron microscopy, freshly isolated skel-etal muscle was placed in Karnovski’s fixative overnight.Fixed tissue was embedded, sectioned and stained accordingto standard procedures.

Quantitative grip strength measurement

Strength testing consisted of five separate measurements usinga trapeze bar attached to a force transducer that recorded peakgenerated force (Stoelting, WoodDale, IL). Mice instinctivelygrab the bar with their forepaws and continue to hold whilebeing pulled backwards by the tail, releasing only whenunable to maintain grip. The resulting measurement wasrecorded, and the three highest measurements were averagedto give the strength score. For each time point and strain, atleast five animals were used. P-values were determined bypaired Student’s t-test.

Evan’s blue dye injection

To evaluate muscle cell membrane integrity, EBD (10 mg/mlin phosphate-buffered saline) was injected intraperitoneallyinto mice (0.1 ml/10 g body weight) as described previously(21). The mice were killed 12–16 h after injection, and theirmuscles were sectioned and examined under a fluorescencemicroscope for uptake of dye.

Densitometric analysis

Autoradiographs of immunoblots from three independent exper-iments were scanned using an Epson 636 Expression Scanner.Densitometry of high-molecular-weight ubiquitinated proteins(HMW-Ub) was analyzed using Scion Image analysis software(Scion Corporation, Frederick, ML, USA) and normalized tothe myosin loading standard, so as not to bias samples withmore or less protein. For each experimental condition, the rela-tive change in HMW-Ub from TgVCP-WT or TgVCP-RHmuscle lysates was determined by comparing it with theamount of HMW-Ub in control muscle lysates from the sameexperiment. The average amount of HMW-Ub present incontrol muscle lysates was arbitrarily set to 1. The plotted histo-gram represents the average change in HMW-Ub from three

independent experiments. Error bars represent standard devi-ation and P-values were determined by paired Student’s t-test.

ACKNOWLEDGEMENTS

We thank Drs Virginia Kimonis and Giles Watts for thought-ful discussion regarding the construction of transgenicanimals. Plasmid 1256MCK-CAT transgenic targeting vectorwas a kind gift from Dr Stephen Hauschka, University ofWashington. We thanks Dr David Harris for help withanimal husbandry. This research was supported through NIHR01NS050717 (P.I.H.) and NIH K08 AG026271 (C.C.W.).

Conflict of Interest statement. None declared.

REFERENCES

1. Askanas, V. and Engel, W.K. (2005) Molecular pathology andpathogenesis of inclusion-body myositis. Microsc. Res. Tech., 67,114–120.

2. Watts, G.D., Wymer, J., Kovach, M.J., Mehta, S.G., Mumm, S., Darvish,D., Pestronk, A., Whyte, M.P. and Kimonis, V.E. (2004) Inclusion bodymyopathy associated with Paget disease of bone and frontotemporaldementia is caused by mutant valosin-containing protein. Nat. Genet., 36,377–381.

3. Kimonis, V.E. and Watts, G.D. (2005) Autosomal dominant inclusionbody myopathy, Paget disease of bone, and frontotemporal dementia.Alzheimer Dis. Assoc. Disord., 19 (Suppl. 1), S44–S47.

4. Hubbers, C.U., Clemen, C.S., Kesper, K., Boddrich, A., Hofmann, A.,Kamarainen, O., Tolksdorf, K., Stumpf, M., Reichelt, J., Roth, U. et al.(2006) Pathological consequences of VCP mutations on human striatedmuscle. Brain, 130, 381–393.

5. Kovach, M.J., Waggoner, B., Leal, S.M., Gelber, D., Khardori, R.,Levenstien, M.A., Shanks, C.A., Gregg, G., Al-Lozi, M.T., Miller,T. et al. (2001) Clinical delineation and localization to chromosome9p13.3-p12 of a unique dominant disorder in four families: hereditaryinclusion body myopathy, Paget disease of bone, and frontotemporaldementia. Mol. Genet. Metab., 74, 458–475.

6. Forman, M.S., Mackenzie, I.R., Cairns, N.J., Swanson, E., Boyer, P.J.,Drachman, D.A., Jhaveri, B.S., Karlawish, J.H., Pestronk, A., Smith, T.W.et al. (2006) Novel ubiquitin neuropathology in frontotemporal dementiawith valosin-containing protein gene mutations. J. Neuropathol. Exp.Neurol., 65, 571–581.

7. Schroder, R., Watts, G.D., Mehta, S.G., Evert, B.O., Broich, P.,Fliessbach, K., Pauls, K., Hans, V.H., Kimonis, V. and Thal, D.R. (2005)Mutant valosin-containing protein causes a novel type of frontotemporaldementia. Ann. Neurol., 57, 457–461.

8. Fratta, P., Engel, W.K., McFerrin, J., Davies, K.J., Lin, S.W. and Askanas,V. (2005) Proteasome inhibition and aggresome formation in sporadicinclusion-body myositis and in amyloid-beta precursorprotein-overexpressing cultured human muscle fibers. Am. J. Pathol.,167, 517–526.

9. Vattemi, G., Engel, W.K., McFerrin, J. and Askanas, V. (2004)Endoplasmic reticulum stress and unfolded protein response in inclusionbody myositis muscle. Am. J. Pathol., 164, 1–7.

10. Wang, Q., Song, C. and Li, C.C. (2004) Molecular perspectives onp97-VCP: progress in understanding its structure and diverse biologicalfunctions. J. Struct. Biol., 146, 44–57.

11. Dai, R.M. and Li, C.C. (2001) Valosin-containing protein is amulti-ubiquitin chain-targeting factor required in ubiquitin-proteasomedegradation. Nat. Cell Biol., 3, 740–744.

12. Hatakeyama, S., Matsumoto, M., Yada, M. and Nakayama, K.I. (2004)Interaction of U-box-type ubiquitin-protein ligases (E3s) with molecularchaperones. Genes Cells, 9, 533–548.

13. Zhong, X., Shen, Y., Ballar, P., Apostolou, A., Agami, R. and Fang,S. (2004) AAA ATPase p97/valosin-containing protein interacts withgp78, a ubiquitin ligase for endoplasmic reticulum-associateddegradation. J. Biol. Chem., 279, 45676–45684.


14. Rabinovich, E., Kerem, A., Frohlich, K.U., Diamant, N. and Bar-Nun,S. (2002) AAA-ATPase p97/Cdc48p, a cytosolic chaperone required forendoplasmic reticulum-associated protein degradation. Mol. Cell. Biol.,22, 626–634.

15. Dalal, S., Rosser, M.F., Cyr, D.M. and Hanson, P.I. (2004) Distinct rolesfor the AAA ATPases NSF and p97 in the secretory pathway. Mol. Biol.Cell, 15, 637–648.

16. Weihl, C.C., Dalal, S., Pestronk, A. and Hanson, P.I. (2006) Inclusionbody myopathy-associated mutations in p97/VCP impair endoplasmicreticulum-associated degradation. Hum. Mol. Genet., 15, 189–199.

17. Shield, M.A., Haugen, H.S., Clegg, C.H. and Hauschka, S.D. (1996)E-box sites and a proximal regulatory region of the muscle creatine kinasegene differentially regulate expression in diverse skeletal muscles andcardiac muscle of transgenic mice. Mol. Cell Biol., 16, 5058–5068.

18. Sugarman, M.C., Yamasaki, T.R., Oddo, S., Echegoyen, J.C., Murphy,M.P., Golde, T.E., Jannatipour, M., Leissring, M.A. and LaFerla, F.M.(2002) Inclusion body myositis-like phenotype induced by transgenicoverexpression of beta APP in skeletal muscle. Proc. Natl. Acad. Sci.USA, 99, 6334–6339.

19. Guyant-Marechal, L., Laquerriere, A., Duyckaerts, C., Dumanchin, C.,Bou, J., Dugny, F., Le Ber, I., Frebourg, T., Hannequin, D. and Campion,D. (2006) Valosin-containing protein gene mutations: clinical andneuropathologic features. Neurology, 67, 644–651.

20. Connolly, A.M., Keeling, R.M., Mehta, S., Pestronk, A. and Sanes, J.R.(2001) Three mouse models of muscular dystrophy: the natural history ofstrength and fatigue in dystrophin-, dystrophin/utrophin-, and lamininalpha2-deficient mice. Neuromuscul. Disord., 11, 703–712.

21. Matsuda, R., Nishikawa, A. and Tanaka, H. (1995) Visualization ofdystrophic muscle fibers in mdx mouse by vital staining with Evans blue:evidence of apoptosis in dystrophin-deficient muscle. J. Biochem. (Tokyo),118, 959–964.

22. Straub, V., Rafael, J.A., Chamberlain, J.S. and Campbell, K.P. (1997)Animal models for muscular dystrophy show different patterns ofsarcolemmal disruption. J. Cell Biol., 139, 375–385.

23. Chevessier, F., Marty, I., Paturneau-Jouas, M., Hantai, D. andVerdiere-Sahuque, M. (2004) Tubular aggregates are from whole

sarcoplasmic reticulum origin: alterations in calcium binding proteinexpression in mouse skeletal muscle during aging. Neuromuscul. Disord.,14, 208–216.

24. Askanas, V. and Engel, W.K. (2002) Inclusion-body myositis andmyopathies: different etiologies, possibly similar pathogenic mechanisms.Curr. Opin. Neurol., 15, 525–531.

25. Frosk, P., Weiler, T., Nylen, E., Sudha, T., Greenberg, C.R., Morgan, K.,Fujiwara, T.M. and Wrogemann, K. (2002) Limb-girdle musculardystrophy type 2H associated with mutation in TRIM32, a putativeE3-ubiquitin-ligase gene. Am. J. Hum. Genet., 70, 663–672.

26. Zhao, L., Longo-Guess, C., Harris, B.S., Lee, J.W. and Ackerman, S.L.(2005) Protein accumulation and neurodegeneration in the woozy mutantmouse is caused by disruption of SIL1, a cochaperone of BiP. Nat. Genet.,37, 974–979.

27. Moussa, C.E., Fu, Q., Kumar, P., Shtifman, A., Lopez, J.R., Allen, P.D.,LaFerla, F., Weinberg, D., Magrane, J., Aprahamian, T. et al. (2006)Transgenic expression of beta-APP in fast-twitch skeletal muscle leads tocalcium dyshomeostasis and IBM-like pathology. FASEB J., 20,2165–2167.

28. Woodman, S.E., Sotgia, F., Galbiati, F., Minetti, C. and Lisanti, M.P.(2004) Caveolinopathies: mutations in caveolin-3 cause four distinctautosomal dominant muscle diseases. Neurology, 62, 538–543.

29. Galbiati, F., Volonte, D., Minetti, C., Bregman, D.B. and Lisanti, M.P.(2000) Limb-girdle muscular dystrophy (LGMD-1C) mutants ofcaveolin-3 undergo ubiquitination and proteasomal degradation: treatmentwith proteasomal inhibitors blocks the dominant negative effect ofLGMD-1C mutanta and rescues wild-type caveolin-3. J. Biol. Chem., 275,37702–37711.

30. Askanas, V., Engel, W.K. and Alvarez, R.B. (1993) Enhanced detection ofcongo-red-positive amyloid deposits in muscle fibers of inclusion bodymyositis and brain of Alzheimer’s disease using fluorescence technique.Neurology, 43, 1265–1267.

31. Davies, J.E., Wang, L., Garcia-Oroz, L., Cook, L.J., Vacher, C.,O’Donovan, D.G. and Rubinsztein, D.C. (2005) Doxycycline attenuatesand delays toxicity of the oculopharyngeal muscular dystrophy mutationin transgenic mice. Nat. Med., 11, 672–677.


transgenic expression of inclusion body myopathy associated mutant

Documents