ARTHRITIS & RHEUMATISMVol. 60, No. 11, November 2009, pp 3280–3289DOI 10.1002/art.24907© 2009, American College of Rheumatology

Impaired Myofibrillar Function in the Soleus Muscle ofMice With Collagen-Induced Arthritis

Takashi Yamada,1 Nicolas Place,2 Natalia Kosterina,3 Therese Ostberg,4 Shi-Jin Zhang,1

Cecilia Grundtman,5 Helena Erlandsson-Harris,4 Ingrid E. Lundberg,4 Birgitta Glenmark,6

Joseph D. Bruton,1 and Hakan Westerblad1

Objective. Progressive muscle weakness is a com-mon feature in patients with rheumatoid arthritis (RA).However, little is known about whether the intrinsiccontractile properties of muscle fibers are affected inRA. This study was undertaken to investigate musclecontractility and the myoplasmic free Ca2� concentra-tion ([Ca2�]i) in the soleus, a major postural muscle, inmice with collagen-induced arthritis (CIA).

Methods. Muscle contractility and [Ca2�]i wereassessed in whole muscle and intact single-fiber prepa-rations, respectively. The underlying mechanisms ofcontractile dysfunction were assessed by investigatingredox modifications using Western blotting and anti-bodies against nitric oxide synthase (NOS), superoxidedismutase (SOD), 3-nitrotyrosine (3-NT), carbonyl,malondialdehyde (MDA), and S-nitrosocysteine (SNO-Cys).

Results. The tetanic force per cross-sectional areawas markedly decreased in the soleus muscle of mice

with CIA, and the change was not due to a decrease inthe amplitude of [Ca2�]i transients. The reduction inforce production was accompanied by slowing of thetwitch contraction and relaxation and a decrease in themaximum shortening velocity. Immunoblot analysesshowed a marked increase in neuronal NOS expressionbut not in inducible or endothelial NOS expression,which, together with the observed decrease in SOD2expression, favors peroxynitrite formation. Thesechanges were accompanied by increased 3-NT, carbonyl,and MDA adducts content in myofibrillar proteins fromthe muscles of mice with CIA. Moreover, there was asignificant increase in SNO-Cys content in myosinheavy-chain and troponin I myofibrillar proteins fromthe soleus muscle of mice with CIA.

Conclusion. These findings show impaired con-tractile function in the soleus muscle of mice with CIAand suggest that this abnormality is due toperoxynitrite-induced modifications in myofibrillar pro-teins.

Patients with rheumatoid arthritis (RA) fre-quently have impaired muscle function, which limitsdaily activities and decreases the quality of life (1).Despite the fact that the histopathologic characteristicsof the muscle are often normal in RA patients, with noevidence of infiltrating inflammatory cells (2), a 25–50%reduction in muscle strength has been reported in up totwo-thirds of patients (3). Decreased muscle strength isgenerally associated with a loss of muscle mass (2).However, little is known about whether the intrinsiccontractile properties of the muscle fibers are affected inRA. In one study of the relationship between muscleweakness and muscle wasting in patients with RA, it wasfound that the muscle weakness was more marked thanthe muscle atrophy (4), suggesting that a loss of intrinsiccontractile performance has occurred.

Supported by grants from the Swedish Research Council, theKarolinska Institutet, the Stockholm County Council (ALF project),and the European Union Sixth Framework Programme (projectAutoCure grant LSH-018611).

1Takashi Yamada, PhD, Shi-Jin Zhang, MD, PhD, Joseph D.Bruton, PhD, Håkan Westerblad, MD, PhD: Karolinska Institutet,Stockholm, Sweden; 2Nicolas Place, PhD: Karolinska Institutet, Stock-holm, Sweden, and University of Geneva, Geneva, Switzerland; 3Na-talia Kosterina, MSc: Royal Institute of Technology, Stockholm,Sweden; 4Therese Ostberg, PhD, Helena Erlandsson-Harris, PhD,Ingrid E. Lundberg, MD, PhD: Karolinska University Hospital, Solna,Sweden, and Karolinska Institutet, Stockholm, Sweden; 5CeciliaGrundtman, PhD: Karolinska University Hospital, Solna, Sweden,Karolinska Institutet, Stockholm, Sweden, and Innsbruck MedicalUniversity, Innsbruck, Austria; 6Birgitta Glenmark, MD, PhD: Sod-ersjukhuset Hospital, and Karolinska Institutet, Stockholm, Sweden.

Address correspondence and reprint requests to TakashiYamada, PhD, Department of Physiology and Pharmacology, Karo-linska Institutet, S-171 77 Stockholm, Sweden. E-mail: Takashi.Yamada@ki.se.

Submitted for publication March 27, 2009; accepted in revisedform July 17, 2009.

3280

Significant skeletal muscle weakness also appearsto develop in a variety of inflammatory conditions,including chronic heart failure (5–7), sepsis (8), andchronic infection (9). Although the basis of muscledysfunction in inflammatory disease is multifactorial,there is good evidence to indicate a major role ofintrinsic impairment of the contractile apparatus (10). Insome studies, impaired contractility has been associatedwith an altered intracellular Ca2� transient (11,12).However, many other investigators have reported de-pressed contractility, even in skinned muscle fiber prep-arations (6,13), suggesting that a significant componentof contractile dysfunction arises as a result of abnormal-ities in the contractile apparatus. Results of severalprevious studies have suggested that the alterations inmyofibrillar functions in inflammatory disease may bedirectly related to redox modifications in the key myo-fibrillar proteins myosin heavy-chain (MyHC) (5,7,14),actin (5), and tropomyosin (5).

Recent evidence indicates that most of the cyto-toxicity attributed to excessive nitric oxide (NO) pro-duction is due to the activities of the peroxynitriteanion, which is a product of the diffusion-controlledreaction of NO with the superoxide radical (15,16).Several potentially harmful modifications are producedby peroxynitrite-derived radicals, including nitration ofaccessible tyrosine residues to form 3-nitrotyrosine (3-NT), S-nitrosylation of cysteine residues to formS-nitrosocysteine (SNO-Cys), formation of protein car-bonyls, and peroxidation of unsaturated fatty acid–containing phospholipids (15). Of these modifications,the formation of 3-NT is a major peroxynitrite-mediatedprotein modification that is different from any modifi-cation mediated by reactive oxygen species (15). In-creased levels of 3-NT are found in the joints of RApatients (17) and in an animal model of RA (18).Intriguingly, accumulation of 3-NT has been associatedwith reduced production of skeletal muscle force inpathologic conditions (8). Furthermore, experimentsanalyzing skinned muscle fibers have shown that exoge-nous peroxynitrite has a strong inhibitory effect onmaximal Ca2�-activated force (19). Thus, modificationsin myofibrillar proteins caused by peroxynitrite-derivedradicals could be a mechanism underlying the contractiledysfunction associated with RA.

One widely used animal model of RA is collagen-induced arthritis (CIA), which displays many of thepathologic characteristics of human RA, including sim-ilar patterns of synovitis, pannus formation, erosion ofcartilage and bone, fibrosis, and loss of joint function(20). The susceptibility to both human RA and murine

CIA is associated with genes encoding class II majorhistocompatibility complex (MHC) molecules (21). Inboth RA and CIA, there is an increase in NO levelsgenerated by NO synthase (NOS) (16) and activationof multiple inflammatory signaling cascades, with theensuing release of bioactive circulating molecules, suchas cytokines (e.g., tumor necrosis factor � [TNF�] andinterleukin-1� [IL-1�]) (22). Moreover, treatment withanti-TNF� antibodies has been shown to reduce theseverity of disease in both RA and CIA (23). Interest-ingly, application of TNF� results in impaired myo-fibrillar function and decreased force production inassociation with increased intracellular oxidant activ-ity (24).

RA patients show significantly greater posturalsway (25) and slower walking speeds (26). These prob-lems are likely to be related to dysfunction of thepostural muscles, such as the soleus muscle, which isimportant in stabilization of the ankle joint at rest andduring walking. In this study, we used mice with CIA asa model for acute RA and studied the contractilefunction of the soleus muscle. Our results showed im-paired contractile function in the soleus muscle of micewith CIA, and the findings suggest that this abnormalitycould be attributed to contractile protein modificationsthat were induced by peroxynitrite-derived radicals.

MATERIALS AND METHODS

Induction and evaluation of CIA. Female DBA/1 mice,weighing 18–22 grams, were supplied by Taconic (LilleSkensved, Denmark). Mice were housed 5 per cage, with freeaccess to food and water in a 12-hour light/dark cycle.

Type II collagen (CII) was obtained from bovine nasalcartilage and mixed in a collagen emulsion (27). A preparationof 100 �g of CII and 300 �g of Mycobacterium tuberculosisin 0.1 ml of emulsion was injected subcutaneously into thebase of the tail of the DBA/1 mice. On day 28, the micereceived a booster injection of CII in Freund’s incompleteadjuvant (100 �g subcutaneously). Control mice were injectedwith 0.1 ml of saline on both occasions.

Mice were observed daily for the development oferythema and swelling of the metatarsophalangeal and anklejoints. Individual paws were scored for inflammation on a scaleof 0–3, as follows: 0 � normal, 1 � 1 joint affected, 2 � 2 jointsaffected, and 3 � whole paw affected. Mice were killed byrapid neck disarticulation when the inflammation score for asingle hind limb had reached the maximum (score of 3), whichoccurred a mean � SEM 16 � 1 days (n � 8) after the secondinjection. This resembles an acute phase of severe RA inhumans. The soleus muscle from the limbs of saline-treatedcontrol mice and from the limbs with an inflammation score of3 in mice with CIA were excised; in some experiments, we alsoexcised fast-twitch extensor digitorum longus (EDL) muscles.

PEROXYNITRITE AND MUSCLE WEAKNESS IN MURINE ARTHRITIS 3281

All experimental procedures were approved by the StockholmNorth Ethics Committee.

Measurements of force and shortening velocity. Intactsoleus muscles were mounted between a force transducer(Dual-Mode Muscle Lever System; Aurora Scientific, Toronto,Ontario, Canada) and an adjustable holder, and superfusedwith Tyrode solution supplemented with 0.1% fetal calf serum,bubbled with 5% CO2/95% O2, and kept at 30°C. Stainless-steel hooks were tied to the tendons as close as possible to theends of the muscle fibers. Muscles were stimulated withsupramaximal current pulses of 0.5 msec. Muscles were mon-itored during contraction, with the use of a microscope with agraticule fitted to a 12.5�-magnification lens. No stretching orslippage of the tendons in the soleus muscles of control mice ormice with CIA was observed during contraction. The musclelength was adjusted to the length (L0) that yielded maximumtetanic force.

The time to peak force and half-relaxation time of thetwitch were measured. The force–frequency relationship wasdetermined by evoking tetani at different frequencies (10–120 Hz, 600 msec long) at 1-minute intervals. In some exper-iments, the force–frequency relationship was also determinedin the EDL muscle.

In the soleus muscle, the shortening velocity at zeroload (V0) was measured with slack tests (28). Rapid releases ofat least 5 different amplitudes were applied during tetani of1-minute intervals. The release amplitudes were plottedagainst the time needed to take up the slack, and a straight linewas fitted to the data points. The V0 was then obtained bydividing the slope of this fitted line by L0.

Muscle weight was measured after these experiments.Force was normalized to the cross-sectional area, calculated asthe muscle weight divided by L0 and the density of the muscle(1,056 kg/m3) (14).

Measurement of myoplasmic free Ca2� concentration([Ca2�]i). Intact single fibers were dissected from the soleusmuscle, mounted at optimum length in a stimulation chamber,and microinjected with the Ca2� indicator indo-1 (MolecularProbes, Eugene, OR). Force–[Ca2�]i curves were generated byplotting the mean force of whole soleus muscle against themean [Ca2�]i at different frequencies (29,30).

Preparation of muscle samples. Muscle samples werefractionated into myofibrillar and cytosolic fractions (31).Samples were homogenized in pyrophosphate buffer with amotor-driven glass homogenizer and then centrifuged at 1,000gfor 10 minutes. The supernatant was removed (designated thecytosolic fraction). The pellet was then washed 4 times with10-volume low-salt buffer followed by 1 wash with low-saltbuffer containing 0.02% (volume/volume) Triton X-100 and 1wash with 0.02% (weight/volume) sodium deoxycholate. Thepellet was then washed twice more in low-salt buffer andsubsequently suspended in pyrophosphate buffer (designatedthe myofibrillar fraction). Protein concentrations of musclefractions were measured using the Bradford technique (32).

Determination of MyHC and actin content in myo-fibrillar proteins. To separate the myofibrillar proteins, so-dium dodecyl sulfate–polyacrylamide gel electrophoresis(SDS-PAGE) was performed using a 4–12% Bis-Tris gel(Invitrogen, Carlsbad, CA). Aliquots of myofibril extractscontaining 5 �g of protein were subjected to electrophoresisand stained with SimplyBlue Safestain (Invitrogen). Images of

the gels were acquired using the GelDoc imaging system(Bio-Rad, Philadelphia, PA). The relative content of MyHC oractin in total myofibrillar proteins was evaluated densitometri-cally using ImageJ (National Institutes of Health, Bethesda,MD; http://www. rsb.info.nih.gov/).

Separation of MyHC isoforms. Using a 6% polyacryl-amide slab gel, electrophoresis was performed (14). Aliquotsof the extracts containing 0.5 �g of myofibrillar protein wereapplied to the gel. Electrophoresis was run at 4°C for 48 hoursat 160V. Gels were silver-stained, and images of the gels wereacquired using the GelDoc imaging system. The percentdistribution of various isoforms was densitometrically evalu-ated with ImageJ.

Immunoblotting. Immunoblotting was performed us-ing anti–3-NT (1:1,000; Cayman Chemical, Ann Arbor, MI),anti–2,4-dinitrophenylhydrazone (1:150, Oxyblot kit; Chemi-con International, Temecula, CA), antimalondialdehyde (anti-MDA) (1:500; Academy Bio-Medical, Houston, TX), anti–SNO-Cys (1:1,000; Sigma, St. Louis, MO), anti–neuronal NOS(anti-nNOS), anti–endothelial NOS (anti-eNOS) and anti–inducible NOS (anti-iNOS) (1:1,000; BD Biosciences, Lexing-ton, KY), anti–copper/zinc superoxide dismutase (anti-SOD1)(1:1,000; Abcam, Cambridge, UK), anti–manganese SOD (anti-SOD2) (1:1,000; Upstate Biotechnology, Lake Placid, NY),anti–troponin I (1:1,000; Millipore, Billerica, MA), anti–dihydropyridine receptor (1:500; Abcam), and anti-actin(1:1,000; Sigma).

The immunoblots for redox modification were per-formed on unstimulated muscle samples, to avoid the compli-cation of changes that could result from contractile activity.Aliquots of the fractionated proteins (10 �g of myofibrillarprotein or 20 �g of cytosolic protein) were subjected toSDS-PAGE (4–12% or 12% Bis-Tris gels; Invitrogen). Pro-teins were transferred onto polyvinylidene difluoride mem-branes. Membranes were blocked in 5% (w/v) nonfat milk withTris buffered saline containing 0.05% (v/v) Tween 20, followedby incubation overnight at 4°C with a primary antibody madeup in 5% (w/v) nonfat milk. Membranes were then washed andincubated for 1 hour at 22°C with a secondary antibody(donkey anti-rabbit or donkey anti-mouse, 1:5,000; Bio-Rad).Immunoreactive bands were visualized using enhanced chemi-luminescence (SuperSignal; Pierce, Rockford, IL). The redoxmodification was assessed with densitometry, involving mea-surement of the total density in a box covering the width andlength of each lane, with results analyzed using ImageJ. Inaddition, for SNO-Cys, the density of MyHC and troponin Iwas measured.

Statistical analysis. Data are presented as the mean �SEM. Student’s unpaired t-tests were used to establish signif-icant differences between control and CIA muscles. Two-wayrepeated-measures analysis of variance (ANOVA) was usedwhen comparing repeated measurements in 2 groups. If theANOVA showed a significant difference between groups, aBonferroni post hoc test was performed. P values less than 0.05were considered significant.

RESULTS

Body and muscle weights. After the mice werekilled, the body and muscle weights of the mice were

3282 YAMADA ET AL

determined. The body weights of the mice with CIAwere significantly lower than those of the control group(mean � SEM 17.1 � 0.5 gm versus 20.0 � 0.4 gm [n �6–8 per group]; P � 0.001). In addition, the soleusmuscle weights were also lower in the mice with CIAcompared with controls (3.1 � 0.2 mg versus 4.3 � 0.4 mg[n � 6–8 per group]; P � 0.01). However, there was nosignificant difference in the normalized soleus muscleweight (determined as the ratio of muscle weight to bodyweight) between mice with CIA and control mice(0.18 � 0.01 mg/gm versus 0.21 � 0.02 mg/gm; P � 0.05).Similar results were obtained for the EDL muscles, inwhich the EDL muscle weights were significantly lowerin mice with CIA than in control mice (5.3 � 0.3 mgversus 6.5 � 0.3 mg [n � 6 per group]; P � 0.05),whereas there was no difference in the normalized EDLmuscle weight between the 2 groups (0.31 � 0.02 mg/gmversus 0.32 � 0.01 mg/gm; P � 0.05).

Contractile speed of CIA soleus muscles. Therewas no significant difference in the specific twitch forceof the soleus muscle between mice with CIA and con-trol mice (mean � SEM 117 � 12 kN/m2 versus 143 �17 kN/m2 [n � 6–7 per group]; P � 0.05). In con-trast, time to peak tension (44 � 2 msec versus 35 � 3msec [n � 6–7 per group]; P � 0.05) and half-relaxationtime of the twitch contraction (58 � 4 msec versus 34 �2 msec [n � 6–7 per group]; P � 0.001) were signifi-cantly increased in CIA soleus muscles compared withcontrol soleus muscles.

V0 was assessed by slack tests (Figures 1A–C).Consistent with the slowed twitch characteristics, V0

showed a significant decrease in CIA muscles comparedwith control muscles (4.5 � 0.2 L0/second versus 5.6 �0.4 L0/second [n � 6–7 per group]; P � 0.05).

The intercept on the y-axis in Figure 1C reflectsthe series elasticity of the muscles (33). The elasticity ofthe soleus muscle in mice with CIA (0.57 � 0.06 mm)was not different from that in control mice (0.75 � 0.05mm). This, together with the fact that visual inspectionrevealed no stretching or slippage of the tendons,showed that the changes in contractile function betweenCIA and control muscles observed under our experi-mental conditions were not related to altered tendonproperties.

Intrinsic contractile properties of CIA soleusmuscles. There was a marked reduction in the specifictetanic force in CIA soleus muscles at stimulation fre-quencies from 30 Hz to 120 Hz (at 30 Hz, P � 0.01versus controls; at 50–120 Hz, P � 0.001 versus controls)(Figure 2A). We observed a similar picture in the EDLmuscles from mice with CIA, with specific forces being

significantly lower in the CIA muscles than in the controlmuscles at 20–150-Hz stimulation (P � 0.05) and being�35% lower than in control muscles at 100 Hz (237 �20 kN/m2 versus 364 � 30 kN/m2 [n � 5–7 per group]).Thus, the changes in muscle weight and force productionobserved in mice with CIA were similar in the soleus andEDL muscles, despite major differences in the compo-sition of the fiber type (i.e., mixed fiber types [asdescribed below] versus fast-twitch fibers) and in theactivation pattern (mostly tonic versus phasic activation)(34). No further experiments were performed on theEDL muscle.

The maximal force of the soleus muscle wasassessed by producing 100 Hz tetani in the presence ofcaffeine (100 �M), which is a widely used stimulator ofthe Ca2� release channels of sarcoplasmic reticulum(SR) (35). Caffeine was applied for 3 minutes beforeproduction of the tetanus. The caffeine-stimulated teta-nus was markedly lower in CIA soleus muscles than incontrol soleus muscles (228 � 21 kN/m2 versus 357 � 23kN/m2 [n � 7–8 per group]; P � 0.01), which impliesthat impaired myofibrillar function, reduced Ca2� re-lease, and/or reduced total SR Ca2� content couldaccount for the defects in specific force production inthe CIA soleus muscle.

To determine the site of excitation–contractioncoupling failure in CIA soleus muscle, we measured the[Ca2�]i using the Ca2� indicator indo-1. The resultsshowed a significantly higher tetanic [Ca2�]i in CIAsoleus fibers at high stimulation frequencies (at 100 Hz,P � 0.01 versus controls; at 120 Hz, P � 0.001 versus

Figure 1. Shortening velocity at zero load (V0) in mouse soleusmuscle, as measured by the slack test. A and B, Changes in musclelength (top) and muscle force (bottom) during application of 5consecutive tetani with rapid releases, in soleus muscles from control(Cont) mice (A) and mice with collagen-induced arthritis (CIA) (B).The initial 40 msec of force redevelopment was fitted by a singleexponential function. The time value at zero force in each fitted curvewas used in measurements of the time to take up the slack. C,Amplitude of the shortening step plotted against the time to take upthe slack. Data points were fitted by a straight line, and the slope of thisline was divided by the optimal length to obtain the V0.

PEROXYNITRITE AND MUSCLE WEAKNESS IN MURINE ARTHRITIS 3283

controls) (Figure 2B). The resting [Ca2�]i was similarbetween control muscle fibers and CIA muscle fibers(101 � 10 nM versus 102 � 18 nM). The time taken forthe tetanic [Ca2�]i to decay by 90% from its final plateauvalue was 234 � 37 msec in control muscle fibers and

198 � 28 msec in CIA muscle fibers (P � 0.05). Thus,the force depression and slowed twitch kinetics in CIAsoleus muscle cannot be ascribed to abnormal SR Ca2�

release or reuptake or to decreased Ca2� content in theSR.

Force–[Ca2�]i curves were constructed from themean data (Figure 2C). Analyses of these curves showedthat the calculated maximum Ca2�-activated force(Pmax) was reduced by �40% in CIA soleus musclesrelative to control soleus muscles (mean 234 kN/m2

versus 374 kN/m2). In contrast, there was no differencein [Ca2�]i at 50% of Pmax (Ca50) between CIA andcontrol muscles (0.26 �M in both groups) or in thesteepness of the relationship (1.85 versus 2.29). Thus,the marked force depression in the CIA soleus musclewas due mainly to the impaired ability of the cross-bridge to generate force.

MyHC and actin content. The decreased forceproduction in the CIA soleus muscle could result from adecreased amount of contractile proteins. We thereforeassessed the expression levels of MyHC and actin.

Figure 2. Decreased muscle force production due to impaired cross-bridge function in CIA soleus muscles. A, Force–frequency relation-ship in control and CIA soleus muscles. B, Myoplasmic free Ca2�

concentration ([Ca2�]i)–frequency relationship in control and CIAsoleus muscle fibers. C, Force–[Ca2�]i curves, generated by plottingthe mean muscle force against the mean [Ca2�]i at different frequen-cies. Bars show the mean � SEM results from 6–8 muscles per group.�� � P � 0.01; ��� � P � 0.001 versus controls. See Figure 1 for otherdefinitions.

Figure 3. Decrease in myosin heavy-chain (MyHC) content, but nochanges in MyHC isoform expression in CIA soleus muscles. A, Ex-pression levels of the myofibrillar proteins MyHC and actin wereassessed in control and CIA soleus muscles by SimplyBlue Safestain.B, The percentage distribution of MyHC and actin content in totalmyofibrillar proteins was compared between control and CIA soleusmuscles. C, Fast-MyHC isoforms IIa, IId/x, and IIb and slow-MyHCisoform I were electrophoretically separated, and their expressionlevels were assessed in control and CIA soleus muscles. D, Thepercentage distribution of the 4 MyHC isoforms was comparedbetween control and CIA soleus muscles. Bars in B and D show themean and SEM results from 3–4 muscles per group. � � P � 0.05versus controls. See Figure 1 for other definitions.

3284 YAMADA ET AL

Figure 3A shows a typical expression pattern of thesemyofibrillar proteins in control and CIA muscles. Therewas a small, but statistically significant, reduction (�7%)in the MyHC content in CIA muscles as compared withcontrol muscles (Figure 3B). In contrast, there was nodifference in actin content between the 2 groups. Thus,a minor part of the force deficit might be attributed to areduction in contractile protein levels.

Expression of MyHC isoforms. The soleus mus-cle in DBA mice is composed of both slow- and fast-twitch fibers. The MyHC-I, MyHC-IIa, and MyHC-IId/xisoforms, but not the MyHC-IIb isoform, were observedin both control mice and mice with CIA (Figure 3C).There were no significant differences in the distributionof the MyHC isoforms between the 2 groups (Figure3D). Thus, the altered contractile properties, includingthe decreased V0, in CIA soleus muscle, were not due tochanges in the expression of different MyHC isoforms.

Expression levels of NOS and SOD. Posttransla-tional modifications of proteins, such as oxidative andnitrosative modifications, have been implicated in theintrinsic contractile dysfunction of skeletal muscle ininflammatory diseases (10). We therefore investigatedwhether the defects in contractile properties were ac-companied by changes in the redox status in the CIAsoleus muscle. There was a significant increase in theexpression of nNOS, but not eNOS, in CIA musclescompared with control muscles (Figures 4A and B). TheiNOS isoform was not detected in either group.

The expression of SOD2 was significantly re-duced by �20% in CIA muscles compared with controlmuscles (Figures 4A and C). There was no difference inSOD1 expression between the 2 groups (Figure 4C).

Redox modifications in myofibrillar proteins.Peroxynitrite is a strong oxidant that is formed in vivo viathe diffusion-limited reaction between NO and thesuperoxide anion (15). The findings of altered NOS andSOD expression suggest that accelerated production ofperoxynitrite has occurred, since decreased expressionof SOD2 could result in an increased steady-state con-centration of superoxide, and the overproduction ofNO via nNOS further favors the formation of peroxy-nitrite (15). To assess the involvement of peroxynitrite inimpaired contractile function in the CIA soleus muscle,myofibrillar extracts were analyzed for redox modifica-tions, using specific antibodies.

Of note, 3-NT is considered to be a major endproduct of peroxynitrite-derived radical interaction withproteins. The soleus muscles from mice with CIAshowed significantly higher levels of total 3-NT contentthan did control soleus muscles (Figures 5A and D).

In addition to the modifications of tyrosine resi-dues, peroxynitrite-derived radicals can produce proteincarbonyls (36) and lipid peroxidation (15). The totalcarbonyl content was increased in CIA soleus musclescompared with control soleus muscles (Figures 5Band D).

The presence of MDA-protein adducts is anindicator of lipid peroxidation (Figure 5C). The soleusmuscles from mice with CIA showed higher levels of

Figure 4. Alterations in the metabolism of reactive nitrogen andreactive oxygen species in CIA soleus muscles. A, The levels of nitricoxide synthase (NOS) and superoxide dismutase (SOD) isoforms incontrol and CIA soleus muscles were determined by Western blotting.The inducible NOS isoform was not detected in either group. B and C,The levels of NOS isoforms (B) and SOD isoforms (C) were quanti-fied, with results normalized to the dihydropyridine receptor (DHPR)content. Bars show the mean and SEM results from 5–10 muscles pergroup. � � P � 0.05; ��� � P � 0.001 versus controls. nNOS �neuronal NOS; eNOS � endothelial NOS; SOD1 � copper/zinc SOD;SOD2 � manganese SOD (see Figure 1 for other definitions).

PEROXYNITRITE AND MUSCLE WEAKNESS IN MURINE ARTHRITIS 3285

total MDA-protein adducts than did control soleusmuscles (Figure 5D).

Among the protein modifications induced byredox stress, structural alterations in cysteine residuesare most strongly implicated in muscle function (37).Previous studies have demonstrated that the reaction ofperoxynitrite-derived radicals with cysteine could resultin the formation of SNO-Cys (38). In the present study,there was a significant increase in total SNO-Cys contentin CIA muscles compared with control muscles (Figures6A and B). Careful inspection of the detected bandsshowed that MyHC, the most abundant contractileprotein in the myofibrils, had higher SNO-Cys content inCIA muscles compared with control muscles (Figures6A and B). Moreover, there was a marked increase inSNO-Cys content in the regulatory contractile protein

troponin I in CIA muscles compared with control mus-cles (Figures 6C and D).

DISCUSSION

In the present study, we showed a marked de-crease in muscle force per cross-sectional area in thepostural soleus muscle of mice with CIA, which is acommonly used model for investigating factors contrib-uting to RA in humans (20–23,27). The reduction inforce production was accompanied by the slowing offorce development and relaxation and a decreased V0.Measurements of [Ca2�]i in intact single fibers indicatedthat impaired cross-bridge function, rather than alter-ations in SR Ca2� handling, is responsible for thecontractile dysfunctions in CIA soleus muscle. More-over, CIA muscles displayed higher redox modifications

Figure 5. Increased total 3-nitrotyrosine (3-NT), carbonyl, and mal-ondialdehyde (MDA)–protein adducts content in myofibrillar proteinsfrom collagen-induced arthritis (CIA) soleus muscles. The levels of3-NT (A), carbonyl (B), and MDA-protein adducts (C) were assessedin myofibrillar proteins from control (Cont) and CIA soleus muscles byWestern blotting, and results were quantified as arbitrary units nor-malized to actin content (D). Densitometry was used to measure thetotal density of each lane. Bars show the mean and SEM results from3–4 muscles per group. � � P � 0.05 versus controls.

Figure 6. Increased S-nitrosocysteine (SNO-Cys) content in myosinheavy-chain (MyHC), troponin I (TnI), and other myofibrillar proteinsfrom CIA soleus muscles. A, Representative Western blots illustratingthe levels of SNO-Cys in myofibrillar proteins from control and CIAsoleus muscles. The densities of all bands in a box covering each laneand the MyHC band were densitometrically measured. B, Quantifica-tion of the levels of SNO-Cys in total myofibrillar proteins and inMyHC in CIA and control soleus muscles, with results normalized toactin content. C, The same membrane as used in the experiment shownin A, developed for a longer time to reveal the less abundant lowmolecular weight bands reacting with anti–SNO-Cys antibody. Mem-branes were then stripped and reprobed with anti–troponin I antibody.D, Quantification of the Western blotting results shown in C, with thedensity of the SNO-Cys band at �23-kd protein normalized to thecontent of total troponin I. Bars in B and D show the mean and SEMresults from 4 muscles per group. � � P � 0.05; �� � P � 0.01 versuscontrols. See Figure 5 for other definitions.

3286 YAMADA ET AL

in myofibrillar proteins, which are most likely the con-sequence of increased peroxynitrite production. Over-all, these results suggest a deleterious effect ofperoxynitrite-derived radicals on cross-bridge functionin the soleus muscle of mice with CIA.

The marked depression in isometric forces, evenafter forces were normalized to the cross-sectional area,indicates that intrinsic contractile defects are responsiblefor CIA-induced loss of muscle strength. Decreasedforce production in skeletal muscle can be caused byreduced Ca2� release from the SR. However, the forcedeficits in CIA soleus muscles were accompanied by anincreased, rather than decreased, tetanic [Ca2�]i. Therate of tetanic [Ca2�]i decay was not different betweenCIA and control muscles, which indicates that theincreased tetanic [Ca2�]i in CIA muscles was due to aneffect on SR Ca2� release, rather than uptake. It isconceivable that increased production of NO results inan increased amplitude of Ca2� transients, given thatNO has been shown to regulate SR Ca2� release channelfunction by nitrosylation of cysteine residues of thisprotein (39).

When reduced SR Ca2� release is excluded as apossible mechanism of decreased muscle force, thedecrease in force in CIA muscles could, in principle, beattributed to the reduced ability of cross-bridges togenerate force and/or decreased myofibrillar Ca2� sen-sitivity (40). To distinguish between these 2 possibilities,force–[Ca2�]i curves were constructed. Ca50, which de-scribes myofibrillar Ca2� sensitivity (41), was not differ-ent between the 2 groups. In contrast, Pmax, whichrepresents the maximum Ca2�-activated force, wasmarkedly depressed in CIA muscles. Thus, the forcedepression in CIA muscles was caused by a reduction inthe maximum force–generating capacity of the cross-bridges. This interpretation is further reinforced by ourobservation that the reduction in maximum tetanic forceproduction in CIA muscles could not be overcome byapplication of caffeine, which is well known to substan-tially increase tetanic [Ca2�]i (42).

Over the past decade, accumulating evidenceindicates that redox stress is one of several mechanismsinvolved in skeletal muscle dysfunction (10). Further-more, decreased force production and the deleteriouseffects of peroxynitrite on muscle function have beenassociated with many pathologic conditions (8,43). Ex-posure of skinned muscle fiber to peroxynitrite donorshas been shown to cause a marked decline in maximalCa2�-activated force (19). Our observations of increasedtotal 3-NT content, which is a widely used marker ofprotein modifications induced by peroxynitrite-derived

radicals (15), are consistent with these previous findingsand suggest that acceleration of peroxynitrite formationcould play a crucial role in the development of myofi-brillar dysfunction in the CIA soleus muscle.

In addition to the nitration of tyrosine residues,the other protein modifications observed in this studymight also be induced by an increased rate of peroxyni-trite production (15). Cysteine modification is moststrongly implicated in contractile dysfunction of skeletalmuscle, in that it alters protein structure and the avail-ability of regulatory sites (37). We observed increasedSNO-Cys content in both MyHC and troponin I in CIAsoleus muscles. The oxidation or chemical modificationof 2 highly reactive cysteines of myosin S1 is a well-established mechanism for the strong inhibition of acto-myosin ATPase (mATPase) activity (44). Furthermore,experiments on the purified troponin subunit showedthat the troponin complex reconstituted with oxidizedtroponin I had markedly reduced mATPase activity,which could be restored by the reducing agent dithio-threitol (45).

Slowing of twitch contractile function was also aprominent feature in the CIA soleus muscles. V0 reflectsthe cycling rate of the cross-bridge, which is determinedprimarily by the mATPase activity. Thus, a decreased V0

indicates a reduction in mATPase activity in CIA soleusmuscle. Experiments on skinned muscle fibers showedthat an exogenous NO donor had inhibitory effects onmATPase activity and V0 (46). Taken together, thesefindings suggest that the decreased mATPase activityobserved in our study accounts for the impaired cross-bridge kinetics affecting the generation of force in CIAsoleus muscles.

Three types of NOS and several different spliceisoforms have been identified in skeletal muscle (16).Constitutively expressed nNOS and eNOS are activatedby the interaction with Ca2�/calmodulin, whereas iNOSis Ca2�/calmodulin-insensitive. Although the NOS activ-ity in skeletal muscle is normally dominated by nNOS,increased levels of NO in inflammatory states are tradi-tionally thought to result from increased expression ofiNOS, which is transcriptionally up-regulated by pro-inflammatory cytokines (16). Our results showing nodetectable iNOS in either group but, instead, a markedincrease in nNOS expression in CIA soleus muscles arenot surprising, because it was clearly demonstrated, inrat gastrocnemius muscle, that iNOS protein expressionpeaked 12 hours after endotoxin injection and disap-peared within 24 hours, whereas nNOS expression in-creased progressively over 24 hours (43). Interestingly, ithas been demonstrated that nNOS, rather than iNOS, is

PEROXYNITRITE AND MUSCLE WEAKNESS IN MURINE ARTHRITIS 3287

primarily responsible for 3-NT formation in the quadri-ceps femoris muscles of patients with chronic obstructivepulmonary disease (47). Thus, these findings suggest animportant pathologic role of sustained elevation of thenNOS level in skeletal muscle.

Little is known about the changes in antioxidantdefenses in the skeletal muscles of patients with RA.Our results suggest that these are reduced, given that theexpression of SOD2 was significantly decreased in thesoleus muscles of mice with CIA. The precise mecha-nism for the decrease in SOD2 content is unclear.However, it is conceivable that reduced expression andactivity of SOD2 may represent a peroxynitrite-dependent nitration of tyrosine residues of this enzyme,given that protein nitration has been shown to increasethe susceptibility of breakdown by the proteaosome(48). Our finding suggests that, regardless of the under-lying mechanism, decreased levels of SOD2 in CIAsoleus muscles will lead to an increased steady-stateconcentration of superoxide in mitochondria, forming apositive loop for enhancing peroxynitrite formation inthe presence of excessive NO production.

The present study results in female DBA/1 miceindicate that arthritis causes intrinsic contractile defectsin the soleus muscle; specifically, the ability of cross-bridges to generate force is impaired. The increasedlevels of nNOS and tyrosine nitration in myofibrillarproteins suggest that peroxynitrite-derived radicals areresponsible, at least in part, for this impairment. Aprevious study showed that treatment with a metallopor-phyrin, which has peroxynitrite-scavenging activity, ame-liorates septic contractile dysfunction in the diaphragmsof rats (49). Interestingly, it has been reported thatperoxynitrite scavenging markedly reduces both arthritisincidence and severity in murine CIA (50), but furtherstudies are required to reveal whether this treatmentalso improves muscle function.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising itcritically for important intellectual content, and all authors approvedthe final version to be published. Dr. Yamada had full access to all ofthe data in the study and takes responsibility for the integrity of thedata and the accuracy of the data analysis.Study conception and design. Yamada, Place, Kosterina, Ostberg,Zhang, Grundtman, Erlandsson-Harris, Lundberg, Glenmark, Bruton,Westerblad.Acquisition of data. Yamada, Place, Kosterina, Ostberg, Zhang,Grundtman, Bruton.Analysis and interpretation of data. Yamada, Place, Kosterina,Erlandsson-Harris, Lundberg, Glenmark, Bruton, Westerblad.

REFERENCES

1. Sokka T, Hakkinen A, Kautiainen H, Maillefert JF, Tolaza S,Hansen J, et al. Physical inactivity in patients with rheumatoidarthritis: data from twenty-one countries in a cross-sectional,international study. Arthritis Rheum 2008;59:42–50.

2. Lundberg IE, Nader GA. Molecular effects of exercise in patientswith inflammatory rheumatic disease. Nat Clin Pract Rheumatol2008;4:597–604.

3. Stenstrom CH, Minor MA. Evidence for the benefit of aerobic andstrengthening exercise in rheumatoid arthritis. Arthritis Rheum2003;49:428–34.

4. Helliwell PS, Jackson S. Relationship between weakness andmuscle wasting in rheumatoid arthritis. Ann Rheum Dis 1994;53:726–8.

5. Dalla Libera L, Ravara B, Gobbo V, Danieli Betto D, GerminarioE, Angelini A, et al. Skeletal muscle myofibrillar protein oxidationin heart failure and the protective effect of Carvedilol. J Mol CellCardiol 2005;38:803–7.

6. Szentesi P, Bekedam MA, van Beek-Harmsen BJ, van der LaarseWJ, Zaremba R, Boonstra A, et al. Depression of force productionand ATPase activity in different types of human skeletal musclefibers from patients with chronic heart failure. J Appl Physiol2005;99:2189–95.

7. Coirault C, Guellich A, Barbry T, Samuel JL, Riou B, Lecarpen-tier Y. Oxidative stress of myosin contributes to skeletal muscledysfunction in rats with chronic heart failure. Am J Physiol2007;292:H1009–17.

8. Lanone S, Mebazaa A, Heymes C, Henin D, Poderoso JJ, Panis Y,et al. Muscular contractile failure in septic patients: role of theinducible nitric oxide synthase pathway. Am J Respir Crit CareMed 2000;162:2308–15.

9. Drew JS, Farkas GA, Pearson RD, Rochester DF. Effects of achronic wasting infection on skeletal muscle size and contractileproperties. J Appl Physiol 1988;64:460–5.

10. Supinski GS, Callahan LA. Free radical-mediated skeletal muscledysfunction in inflammatory conditions. J Appl Physiol 2007;102:2056–63.

11. Ward CW, Reiken S, Marks AR, Marty I, Vassort G, LacampagneA. Defects in ryanodine receptor calcium release in skeletalmuscle from post-myocardial infarct rats. FASEB J 2003;17:1517–9.

12. Wehrens XH, Lehnart SE, Reiken S, van der Nagel R, Morales R,Sun J, et al. Enhancing calstabin binding to ryanodine receptorsimproves cardiac and skeletal muscle function in heart failure.Proc Natl Acad Sci U S A 2005;102:9607–12.

13. Supinski G, Nethery D, Nosek TM, Callahan LA, Stofan D,DiMarco A. Endotoxin administration alters the force vs. pCarelationship of skeletal muscle fibers. Am J Physiol 2000;278:R891–6.

14. Yamada T, Mishima T, Sakamoto M, Sugiyama M, Matsunaga S,Wada M. Oxidation of myosin heavy chain and reduction in forceproduction in hyperthyroid rat soleus. J Appl Physiol 2006;100:1520–6.

15. Szabo C, Ischiropoulos H, Radi R. Peroxynitrite: biochemistry,pathophysiology and development of therapeutics. Nat Rev DrugDiscov 2007;6:662–80.

16. Pacher P, Beckman JS, Liaudet L. Nitric oxide and peroxynitrite inhealth and disease. Physiol Rev 2007;87:315–424.

17. Sandhu JK, Robertson S, Birnboim HC, Goldstein R. Distributionof protein nitrotyrosine in synovial tissues of patients with rheu-matoid arthritis and osteoarthritis. J Rheumatol 2003;30:1173–81.

18. Cuzzocrea S, McDonald MC, Mota-Filipe H, Mazzon E, Costan-tino G, Britti D, et al. Beneficial effects of tempol, a membrane-permeable radical scavenger, in a rodent model of collagen-induced arthritis. Arthritis Rheum 2000;43:320–8.

19. Supinski G, Stofan D, Callahan LA, Nethery D, Nosek TM,

3288 YAMADA ET AL

DiMarco A. Peroxynitrite induces contractile dysfunction and lipidperoxidation in the diaphragm. J Appl Physiol 1999;87:783–91.

20. Williams RO. Collagen-induced arthritis as a model for rheuma-toid arthritis. Methods Mol Med 2004;98:207–16.

21. Campbell IK, Hamilton JA, Wicks IP. Collagen-induced arthritisin C57BL/6 (H-2b) mice: new insights into an important diseasemodel of rheumatoid arthritis. Eur J Immunol 2000;30:1568–75.

22. Marinova-Mutafchieva L, Williams RO, Mason LJ, Mauri C,Feldmann M, Maini RN. Dynamics of proinflammatory cytokineexpression in the joints of mice with collagen-induced arthritis(CIA). Clin Exp Immunol 1997;107:507–12.

23. Feldmann M, Brennan FM, Williams RO, Woody JN, Maini RN.The transfer of a laboratory based hypothesis to a clinically usefultherapy: the development of anti-TNF therapy of rheumatoidarthritis. Best Pract Res Clin Rheumatol 2004;18:59–80.

24. Hardin BJ, Campbell KS, Smith JD, Arbogast S, Smith J, MoylanJS, et al. TNF-� acts via TNFR1 and muscle-derived oxidants todepress myofibrillar force in murine skeletal muscle. J ApplPhysiol 2008;104:694–9.

25. Ekdahl C. Muscle function in rheumatoid arthritis: assessment andtraining. Scand J Rheumatol Suppl 1990;86:9–61.

26. Hakkinen A, Kautiainen H, Hannonen P, Ylinen J, Makinen H,Sokka T. Muscle strength, pain, and disease activity explainindividual subdimensions of the Health Assessment Questionnairedisability index, especially in women with rheumatoid arthritis.Ann Rheum Dis 2006;65:30–4.

27. Kokkola R, Li J, Sundberg E, Aveberger AC, Palmblad K, YangH, et al. Successful treatment of collagen-induced arthritis in miceand rats by targeting extracellular high mobility group box chro-mosomal protein 1 activity. Arthritis Rheum 2003;48:2052–8.

28. Lannergren J. The force-velocity relation of isolated twitch andslow muscle fibres of Xenopus laevis. J Physiol 1978;283:501–21.

29. Bruton J, Tavi P, Aydin J, Westerblad H, Lannergren J. Mitochon-drial and myoplasmic [Ca2�] in single fibres from mouse limbmuscles during repeated tetanic contractions. J Physiol 2003;551:179–90.

30. Andrade FH, Reid MB, Allen DG, Westerblad H. Effect ofhydrogen peroxide and dithiothreitol on contractile function ofsingle skeletal muscle fibres from the mouse. J Physiol 1998;509:565–75.

31. Fagan JM, Sleczka BG, Sohar I. Quantitation of oxidative damageto tissue proteins. Int J Biochem Cell Biol 1999;31:751–7.

32. Bradford MM. A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle ofprotein-dye binding. Anal Biochem 1976;72:248–54.

33. Edman KA. The velocity of unloaded shortening and its relation tosarcomere length and isometric force in vertebrate muscle fibres.J Physiol 1979;291:143–59.

34. Hennig R, Lomo T. Firing patterns of motor units in normal rats.Nature 1985;314:164–6.

35. Allen DG, Westerblad H. The effects of caffeine on intracellularcalcium, force and the rate of relaxation of mouse skeletal muscle.J Physiol 1995;487:331–42.

36. Ischiropoulos H, al-Mehdi AB. Peroxynitrite-mediated oxidativeprotein modifications. FEBS Lett 1995;364:279–82.

37. Ferreira LF, Reid MB. Muscle-derived ROS and thiol regulationin muscle fatigue. J Appl Physiol 2008;104:853–60.

38. Alvarez B, Radi R. Peroxynitrite reactivity with amino acids andproteins. Amino Acids 2003;25:295–311.

39. Stamler JS, Sun QA, Hess DT. A SNO storm in skeletal muscle.Cell 2008;133:33–5.

40. Allen DG, Lannergren J, Westerblad H. Muscle cell functionduring prolonged activity: cellular mechanisms of fatigue. ExpPhysiol 1995;80:497–527.

41. Westerblad H, Allen DG. Changes of myoplasmic calcium con-centration during fatigue in single mouse muscle fibers. J GenPhysiol 1991;98:615–35.

42. Fryer MW, Neering IR. Actions of caffeine on fast- and slow-twitch muscles of the rat. J Physiol 1989;416:435–54.

43. El-Dwairi Q, Comtois A, Guo Y, Hussain SN. Endotoxin-inducedskeletal muscle contractile dysfunction: contribution of nitric oxidesynthases. Am J Physiol 1998;274:C770–9.

44. Tiago T, Simao S, Aureliano M, Martin-Romero FJ, Gutierrez-Merino C. Inhibition of skeletal muscle S1-myosin ATPase byperoxynitrite. Biochemistry 2006;45:3794–804.

45. Horwitz J, Bullard B, Mercola D. Interaction of troponin subunits:the interaction between the inhibitory and tropomyosin-bindingsubunits. J Biol Chem 1979;254:350–5.

46. Perkins WJ, Han YS, Sieck GC. Skeletal muscle force andactomyosin ATPase activity reduced by nitric oxide donor. J ApplPhysiol 1997;83:1326–32.

47. Barreiro E, Gea J, Corominas JM, Hussain SN. Nitric oxidesynthases and protein oxidation in the quadriceps femoris ofpatients with chronic obstructive pulmonary disease. Am J RespirCell Mol Biol 2003;29:771–8.

48. Souza JM, Choi I, Chen Q, Weisse M, Daikhin E, Yudkoff M,et al. Proteolytic degradation of tyrosine nitrated proteins. ArchBiochem Biophys 2000;380:360–6.

49. Nin N, Cassina A, Boggia J, Alfonso E, Botti H, Peluffo G, et al.Septic diaphragmatic dysfunction is prevented by Mn(III)porphy-rin therapy and inducible nitric oxide synthase inhibition. IntensiveCare Med 2004;30:2271–8.

50. Mabley JG, Liaudet L, Pacher P, Southan GJ, Groves JT, SalzmanAL, et al. Part II: beneficial effects of the peroxynitrite decompo-sition catalyst FP15 in murine models of arthritis and colitis. MolMed 2002;8:581–90.

PEROXYNITRITE AND MUSCLE WEAKNESS IN MURINE ARTHRITIS 3289