303Biol Res 36: 303-312, 2003

INTRODUCTION

Ciliary neurotrophic factor (CNTF), acytokine expressed by glial cells, acts as asurvival factor for motor and sensoryneurons and that has also been described ashaving myotrophic properties (Helgren etal. 1994, Sleeman et al. 2000, Huang et al2002). The receptor for CNTF, present inneurons and skeletal muscles, is anheterotrimer formed by a specific CNTFreceptor subunit (CNTFRα), the beta sub-unit of the leukemia inhibitory factor (LIFR-ß) and glycoprotein gp130.

After the success of CNTF in pre-clinicaltrials this cytokin has been used in patientswith neurodegenerative diseases, such asamyotrophic lateral sclerosis (ALS) and isbeing considered for use in Huntington’sdisease (HD) to prevent or cure neuronaldegeneration and their associated motorand mental problems (ALS Study Group

Ciliary Neurotrophic Factor (CNTF) affects the excitableand contractile properties of innervated skeletal muscles

BEATRIZ U. RAMIREZ1, LEIRA RETAMAL2 and CECILIA VERGARA 2

1Facultad de Ciencias Médicas, USACH2Departamento de Biología, Facultad de Ciencias, U. de Chile

ABSTRACT

The well-established trophic role of CNTF upon neurons led to performing clinical trials in patients ofneurodegenerative diseases. However, trials were suspended due to side effects such as severe weight loss,hyperalgesia, coughing, muscle cramps and pain. So far it is not known how CNTF triggers the problems related toskeletal muscle cramps and pain. CNTF has also been described as a myotrophic factor for denervated skeletalmuscles, but the possibility that it affects innervated muscles has also been considered. Since a myotrophic factorcould be a valuable tool for treatment of several muscle diseases, we studied the effects of low doses of CNTFdelivered systemically by an osmotic pump, over the electrical and mechanical properties of innervated anddenervated fast and slow muscles.CNTF induced spontaneous electrical discharges and slowed twitches in innervated muscles, but did not preventthe changes induced by denervation. We postulate that the spontaneous discharges induced by CNTF in innervatedmuscles may be the cause of the cramps, coughing, and muscle ache reported by patients. At low doses, CNTFdoes not exert its myotrophic role over denervated muscles but clearly affects the excitable and contractileproperties of innervated muscles.

Key terms: CNTF, contraction and relaxation time, denervation, EMG, fibrillation, skeletal muscle.

1996, Bachoud-Levi 2000, Kordower1999).

The init ial cl inical tr ials with ALSpatients used peripheral subcutaneousadministration of CNTF, but patientsreported severe side effects that includedanorexia, skeletal muscle weight loss,hyperalgesia, coughing that was not relatedto lung problems, cramps and muscle pain(ALS Study Group 1996, Windisch 1998).As a way to circumvent these side effects,later studies have used intrathecally-delivered CNTF. With this applicationmethod one suspects that the peripheralcirculating levels of CNTF would be quitelow and patients would not suffer fromanorexia or skeletal muscle weight loss,although they did report muscle crampsand pain (Penn et al. 1997).

On the other hand, in the peripheralnervous system, CNTF is found in highquantities in Schwann cells around the

Corresponding author: Cecilia Vergara. Biology Department, Faculty of Sciences, University of Chile. Casilla 653, Santiago,Chile. Telephone: (56-2) 678-7313 - Fax: (56-2) 271-2893 - e-mail: cvergara@uchile.cl

Received: December 17, 2002. In revised form: May 16, 2003. Accepted: September 22, 2003

304

sciat ic nerve and upon denervat ion,circulating CNTF increases and also thespecific CNTF receptor binding alfa sub-unit (CNTFR-α) increases in skeletalmuscle (Weis et al. 1998).

Denervation induces many changes in themuscle cells. Among the most obvious areatrophy, the appearance of fibrillatorypotentials, slowing of twitch kinetics andaltered expression levels of many proteins,included CNTFRα. Helgren et al. (1994)proposed that CNTF has a myotrophic rolebecause they found that this cytokineattenuates the skeletal muscle wastinginduced by transection of the sciatic nerveand also spares the slowing of the twitch(Helgren et al. 1994). Nevertheless, Martinet al. (1996), using similar doses of CNTF,reported that CNTF did not protectdenervated muscles, but caused wasting ofnormally-innervated fast and slow muscles.

We have attempted to understand the basisof the controversial reports concerning themyotrophic effects of CNTF on rat muscleand the cause of side effects reported bypat ients , even at low p lasmat icconcentrations of CNTF.

To this aim we used rats as a model tostudy the effects of chronic systemicapplication of low doses of CNTF over: a)skeletal muscle contractile and electricalproper t ies in both, innervated anddenervated preparations, and b) lower legmuscle weight. Because a possible cross-talk between the signal transduct ionpathways mediated by cAMP and theCNTF-mediated s ignal ing has beendescribed in some cell types (McLennan,1996), we also tested the effect of forskolin,an activator of adenylate cyclase.

We found that at low doses, CNTF inducesthe appearance of spontaneous electricaldischarges in innervated muscles. It alsoinduced a slowing of the twitch, only ininnervated muscles, and did not protect themuscles from the denervation-inducedlengthening of the twitch. The spontaneouselectrical activity could account for the crampsor pain described by patients treated with thiscytokine. Forskolin also affected the kineticparameters of the twitch only for innervatedmuscles. Doses of CNTF that inducedspontaneous potentials did not affect the

weight of either innervated or denervatedmuscles.

METHODS

Male Sprague-Dawley rats aged 40 dayswere anaesthetized with an i.p. injection ofa ketamine (80 mg/Kg) and xylazinehydrochlor ide (12 mg/Kg) mix ture.Animals were divided into 1 control and 4experimental groups. Control animals(controls) received the osmotic pumploaded with saline. Animals in the firstexperimental group (unilateral denervation)were subjected to unilateral sciatic nervesection at the ischiatic notch height, (3-5mm of the nerve removed). Animals in thesecond and third groups were unilaterallydenervated as described and received apump loaded with 0.5 µg/ml rrCNTF (R&DSystems) dissolved in phosphate bufferedsaline containing 0.1 % bovine serumalbumin (unilateral denervation + CNTF)or loaded wi th 100 µM forskol in(Calbiochem) (unilateral denervation +forskolin). In the fourth experimental group,normal animals, whose sciatic nerves wereuntouched, received a pump loaded withCNTF (normals + CNTF). The pump (Alzet2002) was implanted in a pocket under theskin in the back of the animals and delivered0.5 µl/hr. The drug or saline was distributedthrough the systemic circulation to thetarget tissues because no catheter wasattached to the pump. The stability andbioactivity of CNTF delivered through thesepumps has been established previously bythe work of Guillet et al. 1999.

All rats received (i.p.) 0.1 ml of 0.5%Baytril (Bayer) to prevent infections. Theanimals were left to recover for 10 dayswith free access to food and water. Eachanimal was weighed before the operationas well as 10 days after, before theelect romyogram (EMG) and tens ionrecordings were performed.

Electromyogram (EMG) recording

The t ibial is anterior (TA) EMG wasrecorded in anaesthetized animals with a

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

305RAMIREZ ET AL. Biol Res 36, 2003, 303-312

concentric electromyography electrode(Nicolet 019-721300, gauge 27) insertedinto the muscle belly close to the knee orclose to the ankle. The electrode wasconnected into a Power Lab 4s dataacquisition system. Electrical activity wasrecorded at a sampling rate of 1 KHz. Ahigh pass filter was set at 10 Hz and a lowpass filter at 500 Hz. The input amplifierwas connected to a loudspeaker in order tohear the discharges. Occasionally the EMGfrom the extensor digitorum longus (EDL),soleus or gastrocnemius muscles was alsorecorded. The intensity of the electricalactivity was quantified as the standarddeviation of 3 s of recording, measured inµV. The baseline recorded in a restingmuscle at high amplification shows a verynarrow range of potential (low noise), andtherefore a small standard deviation. Thepresence of variable amplitude potentialsover the basel ine introduces a largevariability and therefore a higher standarddeviat ion. The stat ist ical analysis todetermine whether record ings f romdifferent experimental conditions weresignificantly different or not was done at afixed time after inserting the recordingelectrode.

The denervation of the muscles waschecked before recording by the lack ofresponse upon central stimulation of thenerve.

Tension recordings

Tension was recorded in vivo at roomtemperature (~20ºC) after isolating thedistal tendon of the slow soleus and the fastEDL. The leg was fixed to a specially-designed table to hold the rat leg with twofixtures, one inserted at the knee and one atthe ankle. Each tendon was tied to a semi-isometric force transducer (UFI, model1030), taking care that the thread was at aright angle to the transducer. Basal musclelength was adjusted for maximal tension.The transducer was connected to theacquisition system (Power Lab 4s). Datawere recorded at 1 KHz. Twitch kineticparameters were determined after directsupramaximal stimulation of the muscle.

After recordings, gastrocnemius, soleus,EDL and TA muscles were dissected andfrozen in liquid nitrogen into pre-weightedEppendorf tubes. The weight of each musclewas determined later.

Statistical analysis

Results are expressed as mean ± SEM. Themeasurements were made on the number ofanimals indicated in the correspondingfigure, table or text. Statistical significanceof the results was tested by the Student’s“ t” test.

RESULTS

Effects of CNTF and forskolin over theexcitability of the nerve-muscle system. Inthe presence of CNTF, innervated musclesdeveloped spontaneous potentials.

The EMG of control animals was veryreproducible, there was just a small baselinenoise and no spontaneous activity afterinsertional activity had vanished. The EMGrecorded from the normal + CNTF treatedrats showed spontaneous potentials thatdiffered from fibrillation in sound andfrequency. In most muscles we recordedsmall (20 µV) isolated spikes, but frequentlywe recorded bursts of spontaneouspotent ia ls wi th large sp ikes ofapproximately 200 µV, as shown in Figure1. The amplitude of these spontaneouspotentials varied among different normal +CNTF treated muscles, but because theyoccurred over a stable base line, they werereadily distinguishable from normal EMGs.Similar activity was recorded for theinnervated muscles of un i la tera l lydenervated + CNTF treated animals.Therefore, data from muscles of normal +CNTF treated animals and from theinnervated muscles of un i la tera l lydenervated + CNTF treated animals werepooled for statist ical analysis of themembrane potentials after CNTF treatment.Because statistical analysis of the intensityof the membrane potentials was done overthree seconds of recording at a fixed time

306

after inserting the electrode, the evaluatedper iod for each condi t ion was notnecessarily the one with the highest activityfor each recording. Normal musclesdisplayed a low basal activity, accounted

for only by noise. As shown in Table I, thespontaneous bursts induced by CNTF resultin a doubling of the value for standarddeviation in the innervated TA muscles, ascompared with normal (untreated) muscles.Innervated EDL developed s imi larspontaneous electrical discharges afterCNTF treatment (data not shown). Thespontaneous potentials developed in CNTF-exposed innervated muscles were intenseenough to trigger spontaneous contractions.Figure 2 shows a spontaneous twitch thatappeared after a twitch induced by electricalstimulation of a normal + CNTF treatedEDL muscle. The spontaneous twitch wassmaller than the evoked one, indicatingthat only a fraction of the total number ofmuscle fibers were contracting at the sametime. The occurrence of spontaneoustwitches is an extremely rare phenomena ina normal neuromuscular preparation, andwe never recorded one in any of the controlanimals.

EMG recordings from innervated musclesof forskolin- treated animals were identicalto the controls.

Neither CNTF nor forskolin affect thef ibr i l latory potent ia ls in denervatedmuscles

Figure 3 shows EMG recordings fromcontrol, denervated and denervated + CNTFmuscles. The EMG was recorded in two

Figure 1. CNTF induces spontaneous discharges in innervatedmuscles. EMG recordings from the innervated Tibialis Ante-rior of a normal control rat and from an animal that wasexposed to CNTF for 10 days. The spontaneous activityrecorded from the innervated CNTF-treated muscle is differentfrom fibrillation and shows a burst of high amplitude spikesand a few low amplitude potentials.

TABLE I

CNTF increases electrical activity only in innervated muscles

The intensity of electromyographic activity of Tibialis Anterior muscles was quantified as the standard deviation of3 seconds of recording. After CNTF treatment this parameter doubled for innervated muscles, due to the presence of

sporadic bursts superimposed over basal activity. The fibrillatory activity of denervated muscle results in a highvalue of standard deviation, as compared with normal innervated muscles, due to the big variability of the

fibrillatory potentials. Nevertheless, this value is not statistically different for denervated only and denervated +CNTF muscles, therefore, we can not suggest that CNTF affects fibrillatory activity in denervated muscles.

Condition µV P

Normal innervated TA 3.32 ± 0.33 (8)Innervated + CNTF 6.44 ± 0.47 (8)Denervated 13.67 ± 2.15 (16)Denervated + CNTF 17.96 ± 2.14 (16)

Data are mean ± SEM. In brackets, the number of independent records measuredn.s: non-significant

< 0.001

n.s.

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

307

different spots in every muscle (near theknee or the ankle) but no difference wasdetected between the two records in eachmuscle. As expected, denervation inducedthe appearance of fibrillatory potentials.F ibr i l la tory act iv i ty cons is ted ofspontaneous potent ia ls o f var iab leamplitude and frequency displayed duringthe whole recording period, and thereforewas quite different from the short burststhat occur in innervated muscles treatedwith CNTF. The frequency and intensity offibrillatory potentials were quite variableamong d i f ferent an imals for bothdenervated or denervated + CNTF muscles.As shown in Table I , there was nostatistically-significant difference betweenthe intensity of the electrical activity ofonly denervated and denervated + CNTFtreated muscles. The same lack of effect onf ibr i l latory act iv i ty was observed inforskolin-treated animals (not shown).

Effects of CNTF and forskolin on the twitchkinetics for innervated or denervated EDLand soleus muscles

We evaluated the effects of CNTF andforskolin upon the twitch characteristics inthe unilateral denervated animals treatedwith either drug, using the muscles from

animals that had received unilateral sectionof the sciatic nerve, but no drug treatmentas controls. The twitch parameters wereevaluated in fast (EDL) and slow (soleus)muscles in both legs. As expected, incontrol animals, all denervated muscles hadslower contraction and half-relaxation timesthan the corresponding contralateralinnervated muscles (Fig 4, front and backcolumns at the left in each panel). Figure4A shows that in the presence of CNTF orforskolin, contraction time is significantlyincreased for the innervated soleus (frontmiddle and right columns). Nevertheless,neither drug caused a significant diminutionin contraction time for the denervatedmuscle (back middle and right columns).For the EDL (Fig 4C), the effects of thesedrugs are very similar, except that in thiscase CNTF potent iates the effect of

Figure 2. CNTF induces spontaneous contraction. Isometrictension recorded from an innervated EDL muscle previouslyexposed to CNTF for 10 days. The record shows a twitchinduced by electrical stimulation followed by a spontaneouscontraction.

Figure 3. CNTF does not alter fibrillatory activity indenervated muscles. EMG recordings from the distal portionof the Tibialis Anterior. The upper record is from an innervatedmuscle, the second from a 10-day denervated muscle, and thethird from a 10-day denervated-CNTF exposed muscle.Denervated muscles show typical fibrillatory potentials.

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

308

denervation with respect to the slowing ofcontraction time (back left and middlecolumns). For the innervated EDL, CNTFincreases contraction time from 36.6 ± 0.8to 40.8 ± 2.2 and forskolin to 41.3 ± 0.5(mean ± SEM). Even though both drugsincrease contract ion t ime, the h ighdispersion of the data for CNTF-treatedmuscles makes the difference insignificant(Fig 4 C front middle column). With respectto half-relaxation times, CNTF again causeda slowing of this parameter for innervatedsoleus and EDL muscles, but we did notfind that either drug caused a diminution ofthe slowing caused by denervation (Fig 4Band D). For the EDL, CNTF also intensifiedthe effect of denervation (Fig 4D, left andmiddle columns at the back). In summary,our studies showed that at low doses, CNTFdid not counteract the denervation-inducedslowing of the contractile characteristicsof the twitch, but did induce a slowing ofthe twitch in innervated muscles (Fig 5)and also intensify the effect of denervationfor the EDL. Because innervated musclesbecame slower, the difference between

denervated and innervated musclesdecreased.

Forskol in behaved as CNTF forcontraction time and did not affect half-relaxation time.

A low dose of CNTF does not affect the wetweight of normal or denervated skeletalmuscles

The effect of CNTF on the weight ofdenervated and innervated lower legmuscles (TA, EDL, so leus andgastrocnemius) was studied in the animalsthat received unilateral sciat ic nervesectioning after 10 days of treatment.Controls were animals that receivedunilateral sciatic nerve transection but noCNTF. As expected, in control animals allthe denervated muscles decreased in weight,as compared with the correspondingcontralateral innervated muscles. Thedecrease was approximately 40% for theTA, soleus and gastrocnemius but only~25% for the EDL, consistent with the

Figure 4. CNTF alters kinetic parameters of the twitch. Twitch kinetic parameters from control, CNTF or forskolin exposedanimals. In all animals one sciatic nerve was cut. The figure shows the average for 6-9 controls and 5-9 drug-treated musclesfor contraction and half-relaxation times. Statistically-significant differences were calculated separated for innervated ordenervated muscles (Student’s “t”-test; a = p< 0.01, b = p < 0.05). In each case drug treatment was compared to its corresponding(innervated or denervated) control. The standard error for each condition is too small to be appreciated at the scale shown.

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

309

slower time course of weight loss describedfor this muscle (Helgren et al., 1994). Asshown in Table II, CNTF exposure had noeffect on the weight of innervated ordenervated muscles. The d i f ferencebetween muscles in the control group andthe CNTF-t reated animals was notsignificant for the four different musclesstudied, whether innervated or denervated.

DISCUSSION

The most important finding of this study isthat low doses of CNTF evoke spontaneouspotentials in innervated skeletal muscles.Also, the kinetic parameters of the twitchof innervated muscles are altered by thiscytokine.

The suggested use of CNTF to preventneurodegeneration in HD and ALS (ALSStudy Group 1996, Kordower et al., 1999)

requires long-term application of this drug.Therefore, it is important to ascertain itseffects not only on degenerating neurons,but also on other target tissues such asdenervated and innervated muscles. Thespontaneous e lect r ica l act iv i ty o finnervated CNTF-exposed musclesreported here could be related to the legand back cramps reported by CNTF-treatedpatients (Penn et al., 1997). In our hands,application of CNTF provoked small musclecontractions. A similar situation could havebeen described as cramps by patients treatedwith this cytokine.

To our knowledge, this is the first reportshowing that CNTF can induce spontaneouselectrical discharges in innervated muscles.Although we do not know the mechanisms,this effect may result from: a) direct action ofCNTF on the muscle fibers, b) indirect actionmediated by the corresponding motoneuron,or c) indirect action of CNTF over blood borncells that could liberate other cytokinesresponsible for the observed effects.

Direct effects of any drug over the muscleshould be observed over denervated as wellas over innervated fibers. If only innervatedmuscles are affected the effect of the drugis probably nerve mediated.

Dif ferent ia l e f fects of CNTF overinnervated and denervated muscles

We suspect that CNTF may have both directand nerve-mediated effects on the twitchkinetic parameters. Direct effects on muscle

Figure 5. CNTF slows the twitch in innervated muscles.Isometric twitches evoked by stimulation of the soleus in anormal animal (control) and in a rat treated with CNTF for 10days. Despite both muscles being innervated, the twitch fromthe CNTF-exposed muscle presents a slower time course.

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

TABLE II

A low dose of CNTF does not affect weight for innervated or denervated muscles

Weight of lower leg muscles (Tibialis Anterior, Extensor Digitorm Longus, Gastrocnemius and Soleus) from 10 daysunilaterally denervated and unilaterally denervated + CNTF exposed rats.

Muscle weight (mg)

Condition TAden

TAin

EDLden

EDLin

Gasden

Gasin

Solden

Solin

Unil. den. 319 ± 13 529 ± 17 111 ± 4 148 ± 5 1029 ± 35 1694 ± 59 80 ± 6 144 ± 4

Unil. den + CNTF 305 ± 9 545 ± 17 120 ± 6 159 ± 6 1047 ± 36 1795 ± 71 84 ± 6 158 ± 7

Mean ± SEM from 21 controls and 12 CNTF treated animalsden: denervatedin: innervated

310

could be receptor-mediated or non-specific,as is the case for striatal neurons, which donot express NGF nor CNTF receptors butare protected from degeneration by bothtrophic factors through an unknownmechanism of action (Kordower et al., 1999).One indication that CNTF may exert a directaction on muscle fibers comes from its effecton the twitch kinetic parameters. The slowingof contraction and half-relaxation times indenervated EDL suggests a direct effect onthe muscle, since the nerve was sectioned atthe time the pump containing CNTF wasimplanted. Given that twitch time is alsolengthened by CNTF in innervated EDLmuscles, we can speculate that this effectalso has a direct component over the muscle.

A clear indication of CNTF acting throughthe nerve comes from our results on thesoleus twitch t ime because this drugsignificantly increased contraction andrelaxation times in innervated muscles anddid not a f fect denervated so leus(Communicated as preliminary results inRamirez et al. 2001).

With respect to muscle excitability, wedetected CNTF-induced sporadic burstsonly in innervated muscles. We cannotdisregard the possibility that spontaneous,irregular bursts induced by CNTF indenervated muscles could have beenmasked by fibrillation.

In our experiments, CNTF may have beenconveyed to the spinal motoneuron somasby retrograde axonal transport. Retrogradetransport of radiolabeled CNTF has beenobserved in peripheral nerves (Curtis et al.,1993, Mufson, 1999) , a l though i tsmechanism has not been established (Neetand Campenot , 2001) . Fur thermore,intramuscular injections of an adenoviralvector coding for a secretable form of CNTFreduced the degenerat ion of phrenicmyelinated nerve fibers in mouse mutantswith motoneuronal degeneration (Haase etal., 1999). It is possible then, that thespontaneous electrical activity observed ininnervated muscles could be due to an effectof CNTF over motoneurons that indirectlyaffected muscles. Stoop and Poo (1995)and Liou et al (1997) have shown that CNTFpotent iates the spontaneous and theimpulse-evoked release of neurotransmitter

at developing neuromuscular synapses inXenopus cell cultures.

Forskolin mimics the effect of CNTF forcontraction time only in innervated muscles.Since forskolin increases cAMP levels, themotoneuron mediated effects of CNTF oncontraction time could be cAMP-mediated.Because forskol in has no ef fect ondenervated muscles, we can speculate thatthe direct effects of CNTF over muscle arenot mediated by cAMP.

Duration of contraction is a function oftemperature, the coupling between thetranverse-tubule dihidropyridin receptors,ryanodin receptors, and the actomyosinATPase. On the other hand, relaxation timeis primarily controlled by the rate of Ca2+

uptake into the sarcoplasmic reticulum.Therefore, wi th the exempt ion oftemperature, any of these parameters couldhave been af fected by CNTF. Ourmeasurements were done at roomtemperature, and therefore the twitchkinetics were slowed with respect tomeasurements done at 37ºC.

Myotrophic role of CNTF on denervated orinnervated muscles

The reports on the protective effects ofCNTF on the denervation-induced atrophyand denervated muscle funct ionalperformance are contradictory. The doseand type of administration of this cytokine(Kordower et al., 1999), the way results arepresented, and the type of muscle studiedmay contribute in part of the discrepancies.Helgren et al. (1994) reported that CNTFattenuated the soleus weight-loss inducedin rats after sectioning the sciatic nerve. Incontrast, Martin et al. (1996) reported thatCNTF induced a dose-dependent weightloss of innervated gastrocnemius and thatthe weight of denervated-CNTF exposedmuscles was not different from the weightof just denervated muscles. BecauseHelgren et al. (1994) expressed their resultsas the ratio of the wet weight of denervated/innervated muscles that were or were notexposed to CNTF, it is not clear whetherthe effect of the drug was on the innervatedor on the denervated muscle.

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

311

Helgren et al. (1994) also concluded thatCNTF protects the so leus f rom thedenervation-induced slowing of the twitchby comparing the difference in the peakcontract ion t imes for innervated anddenervated muscles that were or were notexposed to CNTF. They reported that forthe CNTF treated group, the differencebetween innervated and denervated muscleswas not significant and considered this amanifestation of the CNTF’s protectiverole. We also find a smaller difference forthe CNTF treated group as compared tocontrols, but this is because CNTF causes aslowing of the contraction time for theinnervated soleus (see Fig 4A control andCNTF innervated and denervated) .Never the less, in our exper imenta lconditions we also find, as Helgren’s groupdescribes, the protective effect of CNTFover the denervation-induced atrophy forthe soleus, quantif ied as fiber cross-sectional area (Ramirez et al. 2000).

We did not find differences in muscleweight between the control and CNTF-treated muscles, whether innervated ordenervated. One reason for this may be thatwe used a lower dose than Helgren et al(1994) or Martin et al. (1996). Differentparameters may have d i f ferent ia lsensitivities in one organ, and there couldalso be differential sensitivities betweenorgans and between muscle fiber types.Spontaneous electrical activity has not beenrecorded by other authors after CNTFtreatment, but our studies showed that dosesthat induced spontaneous discharges ininnervated muscles did not modify the wetweight of the same muscles.

It is not easy to compare the effects ofCNTF under d i f ferent dosages andappl icat ion methods. The range forsubcutaneous applications was 1mg/Kgdaily in rats in the reports of Helgren et al.(1994) and Martin et al. (1996) and of 15-30 µg/Kg every 48 hours in humans (ALSStudy Group 1996) . For in t ra thecaldelivery, 0.2 mg/day every 48 hours havebeen given to human patients (Penn et al.1997). The dose used for the experimentsreported here is approximately equivalentto 24 ng/Kg/day. The reason for deliveringCNTF chronically with a pump rather than

by daily injections is to avoid animal stress.CNTF stability in osmotic pumps has beendemonstrated by Marques and Neto (1997).They reported that 1/4 of CNTF bio-activityis detectable af ter 7 days of pumpimplantation in an identical experimentaldesign to the one we used. Moreover, Guilletat al. (1999) detected CNTF in plasma after14 days of treatment monitored by Elisaand TF1 cells bioassays. Although we useda very low dose of CNTF, we foundnotorious effects on excitabi l i ty andcontract i le proper t ies of innervatedmuscles.

Because the use of neurotrophic factorsfor t reatment o f neurodegenerat ivedisorders seems promising, it is importantthat the specific side effects of the cytokineswould be characterized. Here we presentevidence that in addition to its known actionas a neurotrophic factor, CNTF affectsskeletal muscle performance in denervatedand innervated preparations. In the latter,the increased neuromuscular excitabilitymay be responsible for the cramps and paindescribed by patients.

ACKNOWLEDGEMENTS

We want to thank Dr. Daniel Wolff, Dr.Juan Bacigalupo, Dr. Patricio Sáez andRossana Marsh for their comments on themanuscript. Financed by Fondecyt 198-1053 and Dicyt 029743-RU.

REFERENCES

ALS-CNTF-Treatment Study Group (ACTSG) (1996) Adouble-blind placebo-controlled clinical trial ofsubcutaneous recombinant human ciliary neurotrophicfactor (rhCNTF) in amyotrophic lateral sclerosis.Neurol 46: 1244-1249

BACHOUD-LEVI A, DEGLON N, NGUYEN J, BLOCH J,BOURDET C, WINKEL L, REMY P, GODDARD M,LEFAUCHEUR JP, BRUGIERE P, BAUDIC S,CESARO P, PESCHANSKI M, AEBISCHER P (2000)Neuroprotective gene therapy for Huntington’s diseaseusing a polymer encapsulated BHK cell line engineeredto secrete human CNTF. Hum Gene Ther 11: 1723-1729

CURTIS R, ADRYAN KM, ZHU Y, HARKNESS PJ,LINDSAY RM, DISTEFANO PS (1993) Retrogradeaxonal transport of ciliary neurotrophic factor isincreased by peripheral nerve injury. Nature 365:253-255

RAMIREZ ET AL. Biol Res 36, 2003, 303-312

312

GUILLET C, AUGUSTE P, MAYO W, KREHER P,GASCAN H (1999) Ciliary neurotrophic factor is aregulator of muscular strength in aging. J Neurosci19: 1257-1262

HAASE G, PETTMANN B, BORDET T, VILLA P, VIGNEE, SCHMALBRUCH H (1999) Therapeutical benefitof ciliary neurotrophic factor in progressive motorneuronopathy depends on the route of delivery. AnnNeurol 45: 296-304

HELGREN ME, SQUINTO SP, DAVIS HL, PARRY DJ,BOULTON TG, HECK CS, ZHU Y, YANCOPOULUSGD, LINDSAY RM, DISTEFANO PS (1994) Trophiceffect of ciliary neurotrophic factor on denervatedskeletal muscle. Cell 76: 493-504

HUANG S, WANG F, HONG G, WAN S, KANG H (2002)Protective effects of ciliary neurotrophic factor ondenervated skeletal muscle. J Huazhong Univ SciTechnolog Med Sci 22: 148-51

KORDOWER JH, ISACSON O, EMERICH DF (1999)Cellular delivery of trophic factors for the treatmentof Huntington’s disease: is neuroprotection possible?Exp Neurol 159: 4-20

LIOU JC, YANG RS, FU WM (1997) Regulation of quantalsecretion by neurotrophic factors at developingmotoneurons in Xenopus cell cultures. J Physiol503:129-139

MCLENNAN AJ, GASKIN AA, VINSON EN, MARTINEZLC (1996) Ciliary neurotrophic factor receptor αmRNA in NB41A3 neuroblastoma cells: regulationby cAMP. Eur J Pharmacol 295 : 103-108

MARQUES MJ, NETO HS (1997) Ciliary neurotrophicfactor stimulates in vivo myotube formation in mice.Neurosci Letters 234: 43-46

MARTIN D, MERKEL E, TUCKER KK, MCMANAMANJL, ALBERT D, RELTON J, RUSSEL DA (1996)Cachectic effect of ciliary neurotrophic factor oninnervated skeletal muscle. Am J Physiol 271: R1422-1428

MUFSON EJ, KROIN JS, SENDERA TJ, SOBREVIELAT (1999) Distribution of retrograde transport of trophicfactors in the central nervous system: functionalimplications for the treatment of neurodegenerativediseases. Prog Neurobiol 57: 451-484

NEET KE, CAMPENOT RB (2001) Receptor binding,in terna l iza t ion , and re t rograde t ranspor t o fneurotrophic factors. Cell Mol Life Sci 58: 1021-1035

PENN RD, JEFFREY MD, KROIN S, YORK MM,CEDARBAUM JM (1997) In t ra theca l c i l ia ryneurotrophic factor del ivery for t reatment ofamyotrophic la tera l sc leros is (phase I t r ia l ) .Neurosurgery 40: 94-100

RAMIREZ BU, BAHAMONDE MI, TRONCOSO V,VERGARA C (2000) Ciliary neurotrophic factoraffects the expression of the apamin sensitive calciumactivated potassium channel in skeletal muscle. ContCient Tecnol USACH 123: 31-39

RAMIREZ BU, RETAMAL L, VERGARA C (2001)Trophic capacity of adult motoneurons is modified byCNTF. Biol Res 34 (3-4) R134, Abstract 251

SLEEMAN MW, ANDERSON KD, LAMBERT PD,YANCOPOULOS GD, WIEGAND S J (2000). Theciliary neurotrophic factor and its receptor, CNTFRalpha. Pharm Acta Helv 74: 265-272

STOOP R, POO M (1995) Potentiation of transmitterrelease by Ciliary Neurotrophic Factor requiressomatic signaling. Sci 267: 695-699

WEIS J, LIE DC, RAGOSS U, ZUCHNER SL, SCHRODERJM, KARPATI G, FARRUGGELLA T, STAHL N,YANCOPOULOS GD, DISTEFANO PS (1998)Increased expression of CNTF receptor alpha indenervated human skeletal muscle. J NeuropatholExp Neurol. 57: 850-857

WINDISCH M, GSCHANES A, HUTTER-PAIER B (1998)Neurotrophic activities and therapeutic experiencewith a brain derived peptide preparation. J NeuralTransm Suppl 53: 289-298

RAMIREZ ET AL. Biol Res 36, 2003, 303-312