Available online at www.sciencedirect.com

www.elsevier.com/locate/nmd

Neuromuscular Disorders 21 (2011) 132–141

Blockade of TNF in vivo using cV1q antibody reducescontractile dysfunction of skeletal muscle in response to

eccentric exercise in dystrophic mdx and normal mice

A.T. Piers a,b, T. Lavin a, H.G. Radley-Crabb b, A.J. Bakker a, M.D. Grounds b,G.J. Pinniger a,⇑

a School of Biomedical, Biomolecular and Chemical Sciences, The University of Western Australia, Crawley, WA 6009, Australiab School of Anatomy and Human Biology, The University of Western Australia, Crawley, WA 6009, Australia

Received 14 May 2010; received in revised form 22 September 2010; accepted 23 September 2010

Abstract

This study evaluated the contribution of the pro-inflammatory cytokine, tumour necrosis factor (TNF) to the severity of exercise-induced muscle damage and subsequent myofibre necrosis in mdx mice. Adult mdx and non-dystrophic C57 mice were treated withthe mouse-specific TNF antibody cV1q before undergoing a damaging eccentric contraction protocol performed in vivo on a custom builtmouse dynamometer. Muscle damage was quantified by (i) contractile dysfunction (initial torque deficit) immediately after the protocol,(ii) subsequent myofibre necrosis 48 h later. Blockade of TNF using cV1q significantly reduced contractile dysfunction in mdx and C57mice compared with mice injected with the negative control antibody (cVaM) and un-treated mice. Furthermore, cV1q treatment signif-icantly reduced myofibre necrosis in mdx mice. This in vivo evidence that cV1q reduces the TNF-mediated adverse response to exercise-induced muscle damage supports the use of targeted anti-TNF treatments to reduce the severity of the functional deficit and dystropa-thology in DMD.� 2010 Elsevier B.V. All rights reserved.

Keywords: Inflammation; Muscle damage; Isokinetic dynamometer; Duchenne muscular dystrophy; mdx mouse

1. Introduction

Dystrophin is part of the dystrophin–glycoprotein com-plex that links intra-cellular F-actin to the extra-cellularmatrix [1] and provides mechanical protection for the sar-colemma during muscular contraction [2,3]. In patientswith Duchenne muscular dystrophy (DMD) and in themdx mouse and golden retriever (GRMD) dog models ofDMD, the absence of dystrophin disrupts normal forcetransmission and reduces the stability of the sarcolemma[4]. As a result, dystrophic skeletal muscles are moresusceptible to exercise-induced muscle damage (EIMD),

0960-8966/$ - see front matter � 2010 Elsevier B.V. All rights reserved.

doi:10.1016/j.nmd.2010.09.013

⇑ Corresponding author. Tel.: +61 8 6488 3380; fax: +61 8 6488 1025.E-mail address: gavin.pinniger@uwa.edu.au (G.J. Pinniger).

particularly during lengthening (eccentric) muscle actions[5,6]. The high susceptibility to muscle damage andrepeated cycles of myofibre necrosis, especially throughoutthe growth phase [7], ultimately results in the failure ofmyofibres to regenerate and the replacement of myofibreswith fat or fibrous connective tissue. It is likely that an ele-vated inflammatory response contributes to the progressiveloss of skeletal muscle mass and function seen in the dys-trophic condition [8]. While the pro-inflammatory cytokinetumour necrosis factor (TNF) has been associated withchronic muscle wasting conditions such as cachexia [9]and sarcopenia [10], it has also been implicated in the acuteinflammatory response that exacerbates muscle damageand contractile dysfunction in mdx mice [11–14].

EIMD is characterized by an initial decrease in the forceproducing capacity of the muscle (contractile dysfunction)

A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141 133

immediately after the exercise and a subsequent secondarydecline in force producing capacity that develops over thefollowing days. Although the initial force deficit has beenattributed to the impairment of excitation–contractioncoupling [15] and/or disruption of the contractile filaments[16], this muscle damage also coincides with an inflamma-tory response including the rapid release of TNF fromthe damaged muscle and resident mast cells [17,18]. TNFattracts neutrophils and further inflammatory cells (thatalso produce TNF) to the injured site thereby increasingthe inflammatory cascade. When myofibre injury occurs,inflammation and associated cytokines are essential forcoordinating the removal of damaged tissue and for forma-tion of new skeletal muscle [19]. However, in situ experi-ments examining EIMD in mice have shown that theaccumulation of neutrophils and macrophages after injurycontributes to further muscle damage [20,21] and, alongwith ex vivo studies on cultured myotubes [22], it has beensuggested that this effect may be mediated by increasedproduction of reactive oxygen species (ROS). In the mdxmouse, an excessive state of inflammation can exacerbatemyofibre necrosis [19,23]. In normal (non-dystrophic) mus-cle, experimentally elevated TNF (in vivo for 7 days) resultsin the accumulation of inflammatory cells without necrosis[24]: however, elevated TNF (ex vivo) has rapid adverseeffects on contractile function of normal myofibres withearly loss of force due (at least in part) to associated oxida-tive stress [25]. Therefore, the immediate acute increases inTNF may contribute to the initial muscle weakness follow-ing EIMD, whereas a prolonged inflammatory responsemay lead to the myofibre necrosis and progressive musclewasting that is characteristic of the dystrophic pathology.

Despite some equivocal reports [26], numerous studiesindicate that TNF levels are elevated in mdx mice andhuman DMD patients [27–30]. Elevated local TNF, whencombined with the increased susceptibility to EIMD, con-tributes to myofibre necrosis and, ultimately, the fibrosisinherent in DMD, GRMD and mdx muscle [23]. In vivoexperiments on mdx mice have demonstrated that blockadeof TNF [11,12,14], or depletion of host neutrophils [13],reduces the severity of the dystrophic pathology. The twodrugs used to block TNF activity, Remicade� (antibodyto human TNF, also known as Infliximab) and Enbrel�

(soluble receptor to TNF, also known as Etanercept), arewidely used clinically to treat inflammatory disorders suchas arthritis and Crohn’s disease [31]. To increase the effi-cacy of TNF blockade in mice and to avoid potential prob-lems of immune response to the human constant domainsequences of infliximab, a rat monoclonal antibody specificfor mouse TNF [32] was modified to generate a chimericalmonoclonal antibody with rat variable and mouse IgG2a,kappa constant domains (cV1q; [11,33]).

This study used a custom built mouse dynamometer [34]to quantify precisely the initial damage of dystrophic mus-cle and the impact of cV1q mediated TNF blockade on theinitial contractile dysfunction and subsequent myofibrenecrosis in response to a bout of EIMD performed in situ.

2. Materials and methods

2.1. Animals

All experiments were performed on 10–13 week oldadult male dystrophic C57BL/10ScSnmdx/mdx and controlC57Bl/10ScSn mice, hereafter referred to as mdx andC57, respectively. Adult mdx mice were used in these exper-iments since the level of myofibre necrosis and regenerationhas stabilised to a relatively low level at this age [35,36].The mice, obtained from a specific pathogen free colonyat the Animal Resource Centre, (Murdoch, Western Aus-tralia), were housed in cages, supplied with food and waterwithout restriction, and maintained in a 12 h light/dark air-conditioned (20–25 �C) environment. All animal proce-dures were approved by the Animal Ethics and Experimen-tation Committee of the University of Western Australia inaccordance with the guidelines of the National Health andMedical Research Council of Australia.

2.2. Experimental outline

Experiments were performed using a custom builtmouse dynamometer [34] to induce controlled and consis-tent eccentric EIMD in vivo and to assess the contractileparameters of muscles from mdx and C57 mice. Initial con-tractile dysfunction was quantified by comparing the iso-metric torque production before and after the EIMDprotocol (initial damage), while histological techniqueswere used to quantify subsequent (secondary) myofibrenecrosis. The effect of cV1q mediated blockade of TNFon initial and secondary muscle damage was comparedwith un-treated and sham injected (cVaM) control mice.

2.3. Experimental procedures

Mice were anaesthetised by inhalation of a gaseous mix-ture of isoflurane (isoflurane, 0.4 L/min N2O, 0.4 L/minO2) which was maintained for the duration of the experi-ment (approximately 1 h) via a flow-through facemask overthe mouse’s head. Anaesthetised mice were connected tothe dynamometer for the measurement of contractile func-tion and to induce EIMD. The dynamometer protocol per-formed in this experiment is described in detail by Hameret al. [37] and explained briefly below.

2.4. Dynamometer protocol for eccentric exercise-inducedmuscle damage

Before performing the EIMD protocol, optimal stimula-tion parameters were determined based on the torque–voltand torque–frequency relationships for isometric contrac-tions (200 ms train) recorded at a neutral ankle angle (i.e.the foot positioned perpendicular to the tibia). The optimalstimulation voltage from all experiments ranged from 1 to10 V and the optimal stimulation frequency ranged from125 to 300 Hz. There were no significant differences in these

134 A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141

parameters between mouse strain or treatment group, nordid the EIMD protocol significantly affect these parameters.These stimulation parameters were then used when record-ing the torque–angle relationship and when performing theEIMD protocol consisting of 20 eccentric contractions.After a 10 min recovery period following EIMD, the initialprocedures of torque–volt, torque–frequency and torque–angle were repeated for assessment of initial contractile dys-function. Specific details of these procedures are as follows.

2.5. Torque–angle relationship

A torque–angle relationship for the anterior crural mus-cles was determined from isometric contractions elicited at5� increments between 15� dorsiflexion and 55� plantarflexion. The foot was passively rotated to 15� dorsiflexionand a 200 ms train of pulses delivered to the common pero-neal nerve to elicit an isometric contraction. Immediatelyafter stimulation had ceased the ankle was rotated 5�towards plantar flexion and rested at this angle for 30 sprior to recording of isometric torque at the new jointangle. The procedure was repeated at 5� increments up to55� plantar flexion, with the isometric torque beingrecorded at each position.

Each isometric contraction was analysed for the meanpeak torque during the final 80 ms of activation. The peaktorque was plotted against joint angle and a Gaussian fitwas applied using Curve Expert v1.36. The Gaussian modelfitted was of the form:

y ¼ a�ðx�bÞ2

2c2

The coefficients of this fit are: a = the amplitude of theGaussian fitted curve, equivalent to the peak joint torque;b = the angle at which peak torque was generated and cis the width at half the peak.

2.6. Exercise-induced muscle damage

After the torque–angle relationship was established, anexercise protocol of 20 eccentric contractions was per-

Fig. 1. Mouse dynamometer apparatus with the ankle of the right hind limbprotocol. The knee was clamped and the foot attached to a foot plate with thmuscles were stimulated to contract via hook electrodes connected to the commhead of the fibular. During the EIMD protocol the anterior crural muscles wrotated at 1000 �s�1 to 55� PF during the contraction.

formed between ankle angles of 15� and 55� plantar flexionwith a 30 s rest interval between each trial. The anteriorcrural muscles were stimulated initially at an ankle angleof 15� plantar flexion to produce an isometric contraction(Fig. 1A). After 100 ms of stimulation the ankle wasrapidly rotated to 55� plantar flexion (Fig. 1B) at anangular velocity of 1000 �s�1, thus producing an eccentriccontraction in the anterior crural muscles. Immediatelyfollowing the eccentric contraction, the position of the footwas passively returned to the starting position at 10 �s�1

under servomotor control. After a 10 min rest period, thetorque–volt, torque–frequency and torque–angle proce-dures were repeated to determine the decrease in the peakjoint torque resulting from the EIMD protocol (Fig. 2).At completion of the final procedure the incision wassutured and the mouse was allowed to recover. All micerecovered quickly from anaesthesia, were fully mobilewithin minutes of regaining consciousness and displayedno signs of abnormal limb movement.

2.7. Evans Blue Dye injections

Intra-peritoneal (IP) injections of a 1% sterile solution(w/v) of Evans Blued Dye (EBD) in phosphate–bufferedsaline (PBS, pH 7.5) at a volume of 1% of body weight wereadministered 24 h before the EIMD protocol. Prior toinjection, the EBD solution was sterilised by passagethrough a Millex�-GP 0.22 mm filter and stored at 4 �C.After the injection, animals were returned to their cageand allowed food and water ad libitum.

2.8. Histological assessment of myofibre necrosis and

sarcolemmal damage

Mice were sacrificed by cervical dislocation while underterminal anaesthesia at 48 h after the EIMD protocol.Body weight was recorded prior to sacrifice. The TA mus-cles of the test limb (subjected to EIMD) and the non-tested (contralateral) limb were excised, sliced transverselythrough the mid-belly of the muscle, mounted vertically in

at (A) 15� plantar flexion and (B) 55� plantar flexion during the EIMDe ankle aligned with the axis of the torque transducer. The anterior cruralon peroneal nerve via a small incision in the skin and muscle fascia near theere stimulated to contract at an initial ankle angle of 15� PF and rapidly

Fig. 2. The joint torque–angle relationships before (solid symbols) and after (open symbols) the dynamometer EIMD protocol in mdx and C57 mice. Dataare presented as the mean (±SE) normalised joint torque at ankle joint angles from �15� (dorsiflexion) to 55� (plantar flexion) for (A) untreated C57 mice,(B) C57 mice treated with cV1q, (C) untreated mdx mice, and (D) mdx mice treated with cV1q. Vertical dashed line indicates the neutral ankle angle (0�,when the foot is perpendicular to the leg).

A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141 135

tragacanth gum on cork blocks and snap frozen in isopen-tane, cooled by liquid nitrogen, and stored at �80 �C.Alternate frozen sections (8 lm) were cut for EBD andHaemotoxylin and Eosin (H&E) slides on a Leica(CM3050) cryostat at �21 �C.

The H&E stained muscle sections were observed withbright-field light microscopy, non-overlapping tiled imagesof transverse muscle sections were taken at 10� magnifica-tion to produce an overall image of the entire muscle cross-section. Images were acquired using a Leica DM RBEmicroscope, Stage Pro movement software and a Q ImagingMicropublisher 3.3 RTV digital camera. Muscle morphol-ogy was analysed from the tiled images using Image ProPlus 6.2 software. Areas of myofibre necrosis were identifiedby the presence of fragmented sarcoplasm and the infiltra-tion of inflammatory cells. Myofibre necrosis was calculatedas a percentage of the whole muscle cross-sectional area.

The unstained EBD sections were viewed by red auto-fluorescence microscopy (Fluoro filter N.2.1: Green) at10� magnification and tiled the same as for H&E. Sectionswith less than 20 EBD positive fibres (less than 1% of totalarea) were not tiled. EBD positive fibres were measuredusing Image Pro Plus 6.2 and expressed as a percentageof total muscle cross-sectional area.

2.9. Use of cV1q antibody to block TNF

This study investigated the effectiveness of blockingTNF using the mouse specific anti-TNF antibody cV1q

(Centocor USA), which is a rat/mouse chimeric, specificfor murine TNF [11]. A negative control solution of cVaMwas also administered, which is an isotope matched controlantibody for cV1q. These reagents were generously pro-vided to us by Centocor (USA). In the initial experiments(cV1q-1w) both mdx and C57 mice received two intra-per-itoneal injections of either cV1q or cVaM made up in PBSat 1 week and 1 h prior to the EIMD at a concentration of20 lg/g body weight. In additional experiments (cV1q-4 h),mdx mice received a single injection of cV1q (20 lg/g bodyweight) at 4 h prior to EIMD (dosages based on [11]).

2.10. Data analysis

All data are presented as mean ± SE, unless statedotherwise. Unpaired student t-tests were performed fordirect comparison of untreated C57 and mdx mice. Multi-ple comparisons between all groups and between tested andun-tested (contralateral) limb muscles were made using uni-variate analysis of variance (ANOVA). A least significantdifference (LSD) post-hoc t-test was used to identify differ-ences between groups. Significance was set at P 6 0.05 forall groups.

As a quantitative analysis of the benefit of cV1q treat-ment, a ‘recovery score’ for the correction of the mdxdefect was calculated as ([cV1q-untreated mdx]/[C57-untreated mdx]) � 100 where a score of 100% indicates thatthe parameter in treated mdx mice was equal to that ofcontrol C57 mice, and 0% indicates that no gain was

Fig. 3. Mean (±SE) for (A) peak isometric torque production in mdx andC57 mice before the dynamometer EIMD protocol. (B) Optimal jointangle, at which the peak torque was recorded (data presented as ankleangle in degrees plantar flexion). (C) The torque deficit (contractiledysfunction) in mdx and C57 mice after the EIMD protocol. (D) Musclenecrosis in TA muscles from the dynamometer tested and control (contra-lateral) limbs of mdx mice sampled 48 h after the EIMD protocol.*significantly different compared to mdx mice (P < 0.05); # significantlydifferent to test limb (P < 0.05).

136 A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141

obtained compared with untreated mdx mice [38]. Recov-ery scores were generated for the initial torque deficit andsubsequent myofibre necrosis for mdx mice from thecV1q-1w and cV1q-4h protocols.

3. Results

The number of animals in each group, along with meanage and body weight, are presented in Table 1. When thebody weights of mice in each treatment were grouped bymouse strain, the mean body weight for mdx mice(29.3 ± 0.53 g) was significantly higher than for C57 mice(24.9 ± 0.38 g, P < 0.05). Due to the significant differencein body weights between mdx and C57 mice, peak isometrictorque was normalised to body weight and presented inunits Nmm/kg.

3.1. Muscle contractile properties of untreated mdx and C57

mice

3.1.1. Peak torque production and optimal joint angle

Peak isometric torque normalised to body weight(Fig. 3A) was significantly lower in mdx mice comparedto C57 mice (P = 0.029). The muscles of mdx mice were�25% weaker than the non-dystrophic C57 mice, whichis consistent with the significantly lower specific force inTA recorded by Dellorusso et al. [6], from in situ experi-ments in mdx and C57 mice. Furthermore, the optimalangle at which the peak torque was recorded was signifi-cantly greater for mdx mice compared to C57 mice(Fig. 3B). This indicates a shift in maximal force produc-tion to longer muscle lengths which may reflect an increasein series compliance within the muscle [39] due to theunderlying dystropathology in mdx mice.

3.1.2. Susceptibility to EIMD

The effects of the EIMD protocol on torque–angle rela-tionships for C57 and mdx mice are presented in Fig. 2Aand C, respectively. Initial contractile dysfunction wasquantified physiologically as the percentage decrease inpeak isometric torque after the EIMD protocol (Fig. 3C).Both mdx and C57 mice experienced a marked decreasein mean peak joint torque after the 20 eccentric contrac-tions. However, the deficit in mean peak joint torqueresulting from the EIMD protocol was twofold greater in

Table 1Age and body weight of mdx and C57 mouse groups. Data are presentedas mean ± SE

n Age (weeks) Body weight (g)

mdx untreated 7 12.0 ± 0.3 30.2 ± 1.1mdx cVaM 4 12.2 ± 0.2 29.0 ± 0.7mdx cV1q (1 week) 7 11.7 ± 0.7 29.2 ± 2.1mdx cV1q (4 h) 5 11.3 ± 0.8 29.2 ± 0.9C57 untreated 4 11.0 ± 0.6 25.4 ± 0.9C57 cVaM 4 11.7 ± 0.7 24.9 ± 0.5C57 cV1q (1 week) 5 10.2 ± 0.2 24.4 ± 0.6

mdx compared to C57 mice (P = 0.02). The significantlygreater torque deficit in mdx compared to C57 mice con-firms that dystrophic muscles are more susceptible toEIMD resulting from controlled eccentric contractions,as performed here in vivo with the mouse dynamometer.This is consistent with previous EIMD studies in mdx miceusing isolated EDL muscles [40] and also for in vivo dyna-mometer studies in GRMD dogs [5].

3.1.3. Histological assessment of myofibre necrosis and

sarcolemmal damage

Myofibre necrosis was not evident in any C57 muscles,in accordance with previous observations in our laboratoryfollowing voluntary wheel running or horizontal treadmillrunning (Radley-Crabb, unpublished data). Therefore, toobtain a reliable measure of the extent of myofibre necrosisresulting from EIMD in mdx mice, comparisons were madebetween the TA muscles of the test limb and the un-testedlimb (contralateral control). The mean area of necrosisincreased about fourfold in tested mdx muscle comparedto un-tested (contralateral) muscle (P = 0.002) (Fig. 3D).Evidence of necrosis in the un-tested control leg gives an

A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141 137

indication of the severity of the background dystropathol-ogy in both legs prior to the EIMD protocol, although it isrecognised that there can be biological variation evenbetween both legs of an individual mdx mouse [35].

As for the analysis of myofibre necrosis, there was noevidence of EBD positive fibres in any of the tested musclefrom C57 mice. For the TA muscles of mdx mice, however,EBD positive fibres accounted for 5.6 ± 1.3% of total mus-cle cross-sectional area (Fig. 8A).These results indicate thatthe muscles of the mdx mice are weaker (lower peak tor-que) and more susceptible to EIMD than non-dystrophicC57 mice. The absence of myofibre necrosis and EBD pos-itively stained fibres in any muscles from the C57 micereflects the capacity for normal (non-dystrophic) mice totolerate EIMD.

3.2. The effect of cV1q treatment on mdx and C57 musclesafter EIMD

3.2.1. The effect of cV1q on peak torque and optimal joint

angle

Peak isometric torque prior to EIMD normalised tobody weight for all mdx and C57 groups is presented inFig. 4A. There was a significant main effect of mouse strainon peak torque (F1,28 = 10.6, P < 0.001) (mdx versus C57)and post-hoc analysis revealed that the mdx mice were sig-nificantly weaker than the non-dystrophic C57 mice(P < 0.01). There was no significant main effect of treat-ment (F2,28 = 8.6, P = 0.08) on the peak torque. The jointangle at which the peak torque was recorded (optimalangle) for all mdx and C57 groups is presented inFig. 4B. The optimal angle for untreated mdx and C57mice that was reported in Fig. 3B is also included herefor comparison with the treatment groups. There was nosignificant main effect of treatment (F2,28 = 1.09,P = 0.35) on the optimal joint angle.

Fig. 5. Mean (±SE) peak torque deficit (contractile dysfunction) intreated and untreated mdx and C57 mice after the dynamometer EIMDprotocol. # significantly different from untreated; *significantly differentfrom corresponding mdx group; P < 0.05.

3.2.2. Effect of cV1q treatment on susceptibility to EIMD

The torque–angle relationships recorded before andafter the EIMD protocol in cV1q-treated C57 and mdxmice are presented in Fig. 2B and D, respectively. The ini-tial contractile dysfunction as determined by torque deficitafter EIMD is presented in Fig. 5 for all mdx and C57

Fig. 4. Mean (±SE) data for treated and untreated mdx and C57 mice recordebody weight. (B) Optimal joint angle. *significantly different from correspond

groups. There was a significant main effect of mouse strain(F1,28 = 43.3, P < 0.001) (mdx versus C57) on the torquedeficit after EIMD. Post-hoc analysis revealed that theC57 mice experienced significantly less torque deficit thanthe mdx mice (P < 0.001). Furthermore, there was a signif-icant main effect of treatment (F2,28 = 8.6, P = 0.002) withcV1q-1w significantly reducing the torque deficit in bothmdx and C57 mice (P < 0.001). The acute cV1q-4h treat-ment, examined in mdx mice, also significantly reducedthe torque deficit following EIMD (P < 0.05).

3.2.3. Assessment of cV1q treatment on necrosis and

sarcolemmal damage after EIMD

Necrosis was quantitated on H&E stained sections(Fig. 6): since necrosis was not observed in any C57 mice,only data from mdx mice are presented (Fig. 7). The % areaof necrosis in the TA muscle of tested limbs was comparedto the non-tested contralateral limb. Significant maineffects of leg (F1,28 = 19.0, P < 0.001) (whether dynamome-ter tested or contralateral un-tested limb) and treatment(F2,28 = 4.0, P = 0.029) were observed on muscle necrosisin mdx mice. Post-hoc analysis revealed that for all treat-ment groups the level of necrosis was significantly greaterin the TA muscle of the tested limb compared to the un-tested limb (P < 0.05). Furthermore, exposure to cV1qfor 1 week (cV1q-1w) significantly reduced the level ofmyofibre necrosis in the TA muscles of both the tested limb(Fig. 6) and the un-tested limb, compared to untreated mdx

d prior to the EIMD protocol for (A) peak isometric torque normalised toing mdx group; P < 0.05.

Fig. 6. Typical examples of myofibre necrosis in H&E stained transverse sections of TA muscles subjected to the dynamometer EIMD protocol. (A)Untreated mdx muscle showing myofibre necrosis, fragmented sarcomeres and inflammatory cell accumulation. Similar necrosis is observed in (B) cVaM(control IgG) treated mdx muscle. (C) mdx muscle treated with cV1q (anti-TNF IgG) for 1 week prior to EIMD protocol showing areas of centrallynucleated myofibres but far less myofibre necrosis. Scale bars represent 0.1 mm.

Fig. 7. Mean (±SE) area of myofibre necrosis in dynamometer tested andun-tested (control) TA muscles of mdx mice sampled 48 h after thedynamometer EIMD protocol. *significantly different from correspondingtest leg; # significantly different from untreated group; P < 0.05.

Fig. 8. TA cross sections showing positive staining for Evans Blue Dye (EBD)mice treated with cV1q for 1 week, and (C) mdx mice treated with cVaM for 1 wstained fibres in TA muscles of mdx mice sampled 48 h after the dynamometerin C57 mice.

138 A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141

mice (P < 0.05). However, the level of necrosis in the acutecV1q-4h treated mdx mice was not significantly differentfrom the un-treated mdx mice (in either tested or un-testedlimbs). These results show that 1 week exposure to cV1q inmdx mice significantly reduced necrosis in both the testedand un-tested (control) limbs compared to untreated mdxmice. The effects of cV1q-1w and cVaM treatments on sar-colemmal damage were evaluated by the analysis of EBDpositively stained myofibres (Fig. 8). Neither cV1q-1wnor cVaM treatments had any significant effect on the num-ber of EBD positive fibres in mdx mice (F2,15 = 0.52,P = 0.60). EBD staining was not performed in cV1q-4htreated mice.

in myofibres 48 h after the EIMD protocol for (A) untreated mdx, (B) mdxeek. Scale bars represent 500 lm. (D) Mean (±SE) area of EBD positively

EIMD protocol. Note: There were no (<1%) EBD positive fibres observed

A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141 139

Calculation of the ‘recovery score’ shows a correction bycV1q-1w treatment of the mdx values (control untreatedvalue as 0%) towards those for normal C57 mice (consid-ered as 100%) of 78% for torque deficit and 56% for myo-fibre necrosis. The recovery score for the torque deficit inthis study is similar to the 81% recovery reported by Gillis[38] for isolated EDL muscles of transgenic mdx mice over-expressing a truncated form of utrophin. The recoveryscore for mdx mice exposed to a single cV1q-4h injectionwas 56% for torque deficit but only 5% for necrosis.

4. Discussion

The contractile dysfunction of dystrophic mdx muscles(compared with normal controls) measured by the dyna-mometer in this study is similar to that for the GRMDdog model of DMD [5] and for transgenic mdx mice thatover express insulin-like growth factor-1 [34]. The mainfindings of this study support our hypothesis that blockadeof TNF (using cV1q antibody) would significantly reducethe extent of contractile dysfunction and (in some cases)subsequent myofibre necrosis in mdx mice following a boutof damaging eccentric exercise. These data further supportthe use of cV1q administration to protect dystrophic mus-cles from initial and longer-term damage.

Exposure to cV1q for 1 week before EIMD significantlyreduced the initial contractile dysfunction (recoveryscore = 78%) and the subsequent myofibre necrosis (recov-ery score = 56%) in response to EIMD. The beneficialeffects of cV1q in reducing myofibre necrosis has also beendemonstrated in mdx mice exposed to voluntary wheel run-ning [11] with similar effects observed for other anti-TNFtreatments (e.g. Remicade and Enbrel [12–14]). Interest-ingly, TNF blockade was also beneficial and reduced theinitial contractile dysfunction in C57 mice, although to alesser extent than for mdx mice. This observation suggeststhat, in addition to its contribution to myofibre necrosis,TNF also plays a crucial role in mediating the initial func-tional weakness following EIMD in both normal and dys-trophic muscles: this accords with ex vivo studies on non-dystrophic muscles [25].

It is worth noting that the exposure to cV1q for 1 weekreduces not only necrosis after EIMD (in the test leg) butalso the background pathology in the mdx mice as indi-cated by a significant (albeit small) reduction in necrosisin the un-tested limb. To distinguish between possiblelonger-term beneficial effects of cV1q on the muscle envi-ronment compared with more short-term effects of block-ing TNF, we conducted experiments with a single cV1qinjection administered 4 h before EIMD. We chose a 4 htime-point to ensure that the cV1q had sufficient time totake effect following the IP injection. This short-term expo-sure to cV1q in mdx mice also reduced the contractile dys-function following EIMD (recovery score = 56%).However, this protective effect on force production didnot extend to myofibre necrosis, which was still elevatedin the TA of the tested mdx limb (recovery score = 5%).

Furthermore, the level of necrosis in the un-tested limbwas not significantly different from that of untreated mdxmice which would indicate that the background level ofpathology was unaffected. Therefore, at least the protectiveeffect of cV1q on contractile dysfunction must be mediatedby the short-term blockade of TNF rather than some majoradaptation of the mdx environment. This is consistent withthe beneficial effects of cV1q in normal C57 mice in reduc-ing the contractile dysfunction after EIMD. Therefore, wepropose that elevated levels of TNF in response to EIMDhave direct effects on contractile function.

TNF has been widely implicated as a possible mediatorof the muscle weakness observed in conditions of chronicand acute inflammation (e.g. muscle damage/trauma,chronic obstructive pulmonary disorder, AIDS, congestiveheart failure etc.). Elevated systemic TNF in mice results inloss of body weight (by 3 weeks) and decreased muscle fibrecross-sectional area (by 5 weeks) [9] and, in tissue culturedmyotubes, chronic TNF exposure (up to 3 days) results inmuscle protein catabolism [41]. Acute exposure of isolatedmuscle preparations to TNF (for 4 h), however, causescontractile dysfunction and weakness, without evidenceof catabolism [25]. Our dynamometer results are in accordwith the experiments of Reid et al. [25] who demonstratedthat TNF administration to isolated mouse muscles in vitro

depressed tetanic force without altering the intra-cellularCa2+ concentration ([Ca2+]i) in diaphragm and limb mus-cles of non-dystrophic mice. Other in vivo experiments byHardin et al. [42] show that IP administration of TNF actsquickly (within 1 h) via the TNF-1 receptor to increaseROS production and decrease specific force in non-dystro-phic mouse diaphragm. These effects were abolished bytreatment with the antioxidant Trolox leading to the sug-gestion that muscle-derived oxidants are essential post-receptor mediators of TNF-induced force loss [42].

Altered cytosolic Ca2+ homeostasis has also been impli-cated as a putative pathway in development of the dystro-pathology [43]. Elevated [Ca2+]i may arise from transienttears in the fragile sarcolemma of dystrophic myofibres,and/or through mechanosensitive (stretch-activated) cal-cium channels. The uptake of EBD following the in vivo

eccentric exercise protocol was limited to �5% of myofibresin mdx mice, which is slightly less than that reported byWhitehead et al. (8.6%) from in vitro experiments onEDL muscles from mdx mice [44]. Interestingly, the levelof EBD staining was unaffected by cV1q treatment suggest-ing that the beneficial effects of cV1q reported in this studyare not mediated by reducing the sarcolemmal tearing indystrophic myofibres. This does not preclude the possibilitythat the contractile dysfunction is mediated by altered cal-cium homeostasis as numerous recent studies have impli-cated the contribution of transient receptor potentialchannels (TRPC) to the dystropathology [45,46] and ele-vated expression of TRPC1 has been reported in mdx mus-cle [47,48]. Regardless of the source of entry, the influx ofCa2+ into cells has been linked to ROS production andactivation of the nuclear factor-kappa B (NF-jB) pathway

140 A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141

[43,49] that regulates pro-inflammatory cytokine expres-sion, in particular TNF [43]. Thus, a putative mechanisminvolved in the exercise-induced contractile dysfunction inmdx mice may involve a TNF mediated positive feedbackrelationship with NF-jB [30,50] and ROS [51]. The protec-tive effect of TNF blockade using cV1q might be mediated(at least in part) by altering this positive feedback cycle,although further quantitative studies are required to clarifythis hypothesis.

5. Summary

We have shown that cV1q administration (1 week beforeEIMD) protects dystrophic skeletal myofibres against ini-tial contractile dysfunction and subsequent myofibre necro-sis in adult mdx mice subjected to a damaging eccentriccontraction protocol in vivo using a custom built mousedynamometer. This technique enables the precise in situ

quantitation of the initial physiological damage of musclesand evaluation of the impact of cV1q mediated TNF block-ade on both contractile dysfunction and myofibre necrosis.The results from this study indicate that TNF acts quickly(within 30 min) to cause muscle weakness. Short term (4 h)administration of cV1q protects against TNF-mediatedcontractile dysfunction but not necrosis of mdx muscles,indicating that different mechanisms contribute to theseevents. The controlled EIMD produced by the dynamom-eter, with precise timing of damage combined with the abil-ity to repeatedly measure functional parameters in vivo

(including force production, muscle fatigue and adaptation[34]) makes this a powerful tool to test the efficacy of var-ious therapeutic interventions and to more precisely definethe molecular mechanisms responsible for initial andlonger-term damage of dystrophic muscles.

Acknowledgements

We gratefully acknowledge Centocor (USA) for provid-ing the cV1q anti-mouse TNF antibody and the negativecontrol antibody (cVaM) and Dr. David Shealy for his sup-port and helpful comments regarding this work. We alsothank Professor Peter Hamer (The University of NotreDame, Australia) for his contribution to the dynamometerwork. This research was supported by the Raine MedicalResearch Foundation, UWA, the Medical Health ResearchInfrastructure Fund of WA and the National Health andMedical Research Council of Australia.

References

[1] Ervasti JM. A role for the dystrophin–glycoprotein complex as atransmembrane linker between laminin and actin. J Cell Biol1993;122:809–23.

[2] Petrof BJ, Shrager JB, Stedman HH, Kelly AM, Sweeney HL.Dystrophin protects the sarcolemma from stresses developed duringmuscle contraction. Proc Natl Acad Sci USA 1993;90:3710–4.

[3] Grounds MD, Sorokin L, White J. Strength at the extracellularmatrix–muscle interface. Scand J Med Sci Sports 2005;15:381–91.

[4] Matsumura K, Campbell KP. Dystrophin–glycoprotein complex: itsrole in the molecular pathogenesis of muscular dystrophies. MuscleNerve 1994;17:2–15.

[5] Childers MK, Okamura CS, Bogan DJ, Bogan JR, Petroski GF,McDonald K, et al.. Eccentric contraction injury in dystrophic caninemuscle. Arch Phys Med Rehabil 2002;83:1572–8.

[6] Dellorusso C, Crawford RW, Chamberlain JS, Brooks SV. Tibialisanterior muscles in mdx mice are highly susceptible to contraction-induced injury. J Muscle Res Cell Motil 2001;22:467–75.

[7] Grounds M. Two-tiered hypotheses for Duchenne muscular dystro-phy. Cell Mol Life Sci 2008;65:1621–5.

[8] Porter JD, Khanna S, Kaminski HJ, Rao JS, Merriam AP,Richmonds CR, et al.. A chronic inflammatory response dominatesthe skeletal muscle molecular signature in dystrophin-deficient mdxmice. Hum Mol Genet 2002;11:263–72.

[9] Coletti D, Moresi V, Adamo S, Molinaro M, Sassoon D. Tumornecrosis factor-alpha gene transfer induces cachexia and inhibitsmuscle regeneration. Genesis 2005;43:120–8.

[10] Phillips T, Leeuwenburgh C. Muscle fiber specific apoptosis andTNF-a signaling in sarcopenia are attenuated by life-long calorierestriction. FASEB J 2005;19:668–70.

[11] Radley HG, Davies MJ, Grounds MD. Reduced muscle necrosis andlong-term benefits in dystrophic mdx mice after cV1q (blockade ofTNF) treatment. Neuromuscul Disord 2008;18:227–38.

[12] Grounds MD, Torrisi J. Anti-TNFalpha (Remicade) therapy protectsdystrophic skeletal muscle from necrosis. FASEB J 2004;18:676–82.

[13] Hodgetts S, Radley H, Davies M, Grounds MD. Reduced necrosis ofdystrophic muscle by depletion of host neutrophils, or blockingTNFalpha function with Etanercept in mdx mice. NeuromusculDisord 2006;16:591–602.

[14] Pierno S, Nico B, Burdi R, Liantonio A, Didonna MP, Cippone V,et al.. Role of tumour necrosis factor [alpha], but not of cyclo-oxygenase-2-derived eicosanoids, on functional and morphologicalindices of dystrophic progression in mdx mice. a pharmacologicalapproach. Neuropathol Appl Neurobiol 2007;33:344–59.

[15] Warren GL, Ingalls CP, Lowe DA, Armstrong RB. Excitation–contraction uncoupling: major role in contraction-induced muscleinjury. Exerc Sport Sci Rev 2001;29:82–7.

[16] Lieber RL, Thornell LE, Friden J. Muscle cytoskeletal disruptionoccurs within the first 15 min of cyclic eccentric contraction. J ApplPhysiol 1996;80:278–84.

[17] Gordon JR, Galli SJ. Mast cells as a source of both preformed andimmunologically inducible TNF-alpha/cachectin. Nature1990;346:274–6.

[18] Collins RA, Grounds MD. The role of tumor necrosis factor-alpha(TNF-a) in skeletal muscle regeneration: studies in TNF-a(�/�) andTNF-a(�/�)/LT-a(�/�) mice. J Histochem Cytochem 2001;49:989–1002.

[19] Tidball JG. Inflammatory processes in muscle injury and repair. Am JPhysiol Regul Integr Comp Physiol 2005;288:345–53.

[20] Pizza FX, Peterson JM, Baas JH, Koh TJ. Neutrophils contribute tomuscle injury and impair its resolution after lengthening contractionsin mice. J Physiol 2005;562:899–913.

[21] Lapointe BM, Frenette J, Cote CH. Lengthening contraction-inducedinflammation is linked to secondary damage but devoid of neutrophilinvasion. J Appl Physiol 2002;92:1995–2004.

[22] Pizza FX, McLoughlin TJ, McGregor SJ, Calomeni EP, GunningWT. Neutrophils injure cultured skeletal myotubes. Am J Physiol CellPhysiol 2001;281:C335–41.

[23] Grounds MD, Radley HG, Gebski BG, Bogoyevitch MA, Shav-lakadze T. Implications of cross-talk between TNF and IGF-1signalling in skeletal muscle. P Aus Physl 2008;39:7–13.

[24] Peterson JM, Feeback KD, Baas JH, Pizza FX. Tumor necrosisfactor-a promotes the accumulation of neutrophils and macrophagesin skeletal muscle. J Appl Physiol 2006;101:1394–9.

[25] Reid MB, Lannergren J, Westerblad H. Respiratory and limb muscleweakness induced by tumor necrosis factor-a: involvement of musclemyofilaments. Am J Respir Crit Care Med 2002;166:479–84.

A.T. Piers et al. / Neuromuscular Disorders 21 (2011) 132–141 141

[26] Lundberg I, Brengman JM, Engel AG. Analysis of cytokineexpression in muscle in inflammatory myopathies, Duchenne dystro-phy, and non-weak controls. J Neuroimmunol 1995;63:9–16.

[27] Porreca E, Guglielmi MD, Uncini A, Di Gregorio P, Angelini A, DiFebbo C, et al.. Haemostatic abnormalities, cardiac involvement andserum tumor necrosis factor levels in X-linked dystrophic patients. JThromb Haemost 1999;81:543–6.

[28] Saito K, Kobayashi D, Komatsu M, Yajima T, Yagihashi A, IshikawaY, et al.. A sensitive assay of tumor necrosis factor a in sera fromDuchenne muscular dystrophy patients. Clin Chem 2000;46:1703–4.

[29] Huang P, Zhao XS, Fields M, Ransohoff RM, Zhou L. Imatinibattenuates skeletal muscle dystrophy in mdx mice. FASEB J2009;23:2539–48.

[30] Messina S, Bitto A, Aguennouz Mh, Minutoli L, Monici MC,Altavilla D, et al.. Nuclear factor kappa-B blockade reduces skeletalmuscle degeneration and enhances muscle function in Mdx mice. ExpNeurol 2006;198:234–41.

[31] Kageyama Y, Takahashi M, Torikai E, Suzuki M, Ichikawa T,Nagafusa T, et al.. Treatment with anti-TNF-a antibody infliximabreduces serum IL-15 levels in patients with rheumatoid arthritis. ClinRheumatol 2007;26:505–9.

[32] Echtenacher B, Falk W, Mannel DN, Krammer PH. Requirement ofendogenous tumor necrosis factor/cachectin for recovery fromexperimental peritonitis. J Immunol 1990;145:3762–6.

[33] Scallon B, Cai A, Radewonuk J, Naso M. Addition of an extraimmunoglobulin domain to two anti-rodent TNF monoclonal anti-bodies substantially increased their potency. Mol Immunol2004;41:73–80.

[34] Ridgley JA, Pinniger GJ, Hamer PW, Grounds MD. The physiolog-ical effects of IGF-1 (class 1:Ea transgene) over-expression onexercise-induced damage and adaptation in dystrophic muscles ofmdx mice. Pflugers Arch Eur J Physiol 2009;457:1121–32.

[35] Grounds MD, Radley HG, Lynch GS, Nagaraju K, De Luca A.Towards developing standard operating procedures for pre-clinicaltesting in the mdx mouse model of Duchenne muscular dystrophy.Neurobiol Dis 2008;31:1–19.

[36] Lynch GS. Role of contraction-induced injury in the mechanisms ofmuscle damage in muscular dystrophy. Clin Exp Pharmacol Physiol2004;31:557–61.

[37] Hamer PW, McGeachie JM, Davies MJ, Grounds MD. Evans BlueDye as an in vivo marker of myofibre damage: optimising parametersfor detecting initial myofibre membrane permeability. J Anat2002;200:69–79.

[38] Gillis J-M. Multivariate evaluation of the functional recoveryobtained by the overexpression of utrophin in skeletal muscles ofthe mdx mouse. Neuromuscul Disord 2002;12:S90–4.

[39] Gregory JE, Morgan DL, Allen TJ, Proske U. The shift in muscle’slength–tension relation after exercise attributed to increased seriescompliance. Eur J Appl Physiol 2007;99:431–41.

[40] Consolino CM, Brooks SV. Susceptibility to sarcomere injuryinduced by single stretches of maximally activated muscles of mdxmice. J Appl Physiol 2004;96:633–8.

[41] Li Y-p, Schwartz RJ, Waddell ID, Holloway BR, Reid MB. Skeletalmuscle myocytes undergo protein loss and reactive oxygen-mediatedNF-jB activation in response to tumor necrosis factor a. FASEB J1998;12:871–80.

[42] Hardin BJ, Campbell KS, Smith JD, Arbogast S, Smith J, Moylan JS,et al.. TNF-a acts via TNFR1 and muscle-derived oxidants to depressmyofibrillar force in murine skeletal muscle. J Appl Physiol2008;104:694–9.

[43] Whitehead NP, Yeung EW, Allen DG. Muscle damage in mdx(dystrophic) mice. Role of calcium and reactive oxygen species. ClinExp Pharmacol Physiol 2006;33:657–62.

[44] Whitehead NP, Pham C, Gervasio OL, Allen DG. N-acetylcysteineameliorates skeletal muscle pathophysiology in mdx mice. J Physiol2008;586:2003–14.

[45] Millaya DP, Goonasekeraa SA, Sargenta MA, Mailleta M, AronowaBJ, Molkentin JD. Calcium influx is sufficient to induce musculardystrophy through a TRPC-dependent mechanism. PNAS2009;106:19023–8.

[46] Sabourin J, Lamiche C, Vandebrouck A, Magaud C, Rivet J,Cognard C, et al.. Regulation of TRPC1 and TRPC4 cation channelsrequires an a1-syntrophin-dependent complex in skeletal mousemyotubes. J Biol Chem 2009;284:36248–61.

[47] Rolland J-F, De Luca A, Burdi R, Andreetta F, Confalonieri P,Conte Camerino D. Overactivity of exercise-sensitive cationchannels and their impaired modulation by IGF-1 in mdx nativemuscle fibers: beneficial effect of pentoxifylline. Neurobiol Dis2006;24:466–74.

[48] Zanou N, Iwata Y, Schakman O, Lebacq J, Wakabayashi S, Gailly P.Essential role of TRPV2 ion channel in the sensitivity of dystrophicmuscle to eccentric contractions. FEBS Lett 2009;583:3600–4.

[49] Arthur PG, Grounds MD, Shavlakadze T. Oxidative stress as atherapeutic target during muscle wasting: considering thecomplex interactions. Curr Opin Clin Nutr Metab Care 2008;11:408–16.

[50] Li YP, Reid MB. NF-kB mediates the protein loss induced by TNF-ain differentiated skeletal muscle myotubes. Am J Physiol Regul IntegrComp Physiol 2000;279:1165–70.

[51] Rando TA, Disatnik M-H, Yu Y, Franco A. Muscle cells from mdxmice have an increased susceptibility to oxidative stress. NeuromusculDisord 1998;8:14–21.