fusion between myoblasts and adult muscle fibers...

Development 109, 139-148(1990)Printed in Great Britain © The Company of Biologists Limited 1990

139

Fusion between myoblasts and adult muscle fibers promotes remodeling of

fibers into myotubes in vitro

TIMOTHY J. HINTERBERGER* and KATE F. BARALD

Department of Anatomy and Cell Biology, Program in Neuroscience, Program in Cell and Molecular Biology, University of MichiganMedical School, Ann Arbor, Ml 48109-0616, USA

•Present address: Department of Biological Sciences, Purdue University, W. Lafayette, IN 47907, USA

Summary

Muscle satellite cells are residual embryonic myoblastprecursors responsible for muscle growth and regener-ation. In order to examine the role of satellite cells in theinitial events of muscle regeneration, we placed indi-vidual mature rat muscle fibers in vitro along with theirsatellite cells. When the satellite cells were allowed toproliferate, they produced populations of myoblasts thatfused together to form myotubes on the laminin sub-strate. These myoblasts and myotubes also fused withthe adult fibers. When they did so, the fibers lost theiradult morphology, and by 8 days in vitro, essentially allof them were remodeled into structures resemblingembryonic myotubes. However, when proliferatingsatellite cells were eliminated by exposure to cytosine

arabinoside (araC), the vast majority of fibers retainedtheir adult shape. Addition of C2C12 cells (a myoblastline derived from adult mouse satellite cells) to araC-treated fiber cultures resulted in their fusion with the ratmuscle fibers and restored the ability of the fibers toremodel, whereas addition of either a fibroblast cell lineor a transformed, non-fusing variant of C2C12 cells, oraddition of conditioned medium from C2C12 cells, failedto do so. These results imply that myoblast fusion isresponsible for triggering adult fiber remodeling in vitro.

Key words: satellite cells, myoblast, muscle fiber, myotube,fusion, rat muscle.

Introduction

Injured vertebrate skeletal muscle regenerates from itssatellite cells, a reserve population of myoblast precur-sors located between the sarcolemma and the basallamina of muscle fibers (Mauro, 1961; review, Cam-pion, 1984). Additional evidence implicates satellitecells in the normal growth process in adult muscle(Campion, 1984), a process that apparently does notrequire an injury 'trigger'. Satellite cells proliferate inresponse to muscle injury as well as in the enlargementof already-existing adult muscle fibers. In an injuryresponse, the resulting myoblasts fuse with each otherto form new immature fibers called myotubes (reviews:Allbrook, 1981; Carlson and Faulkner, 1983). In thenormal growth of mature fibers without injury, satellitecells fuse with the preexisting fibers (Moss and Leblond,1971). Mature fibers that survive an injury may alsoparticipate in the regeneration process. Multinucleatedsprouts, resembling embryonic myotubes, are oftenseen at the ends of mature fibers in histological sectionsof injured mammalian skeletal muscle (e.g. Clark, 1946;Walton and Adams, 1956; Roth and Oron, 1985). Gay

and Hunt (1954) dissected individual sprouted fibersfrom partially transected rat muscles to demonstrate thecontinuity between sprouts and mature fibers, laterconfirmed by electron microscopy (Shafiq, 1970; Ali,1979; Roth and Oron, 1985). Based on examination offibers at various times after injury, Gay and Hunt (1954)concluded that sprouts from fibers on both sides of asmall cut eventually fuse together, completely bridgingthe site of the lesion.

The means by which muscle fibers sprout has notbeen determined. A sprout could be the result of one ormore satellite-cell myoblasts fusing onto the end of amature fiber. Myotubes arising from myoblast fusionsecondarily fuse with mature fibers in injured bat webmuscle (Church, 1970). Sprouting was not seen at cutinjuries in rats following vinblastine injection, indi-cating that sprouting may involve cell division (Rothand Oron, 1985).

Another possibility is that some sprouts originateentirely from mature fibers by 'budding' (Shafiq, 1970;Ali, 1979). This phenomenon would be difficult todistinguish from myoblast fusion in histological sectionsof injured whole muscles, but it was clearly demon-

140 T. J. Hinterberger and K. F. Barald

strated by individual adult rat muscle fibers maintainedin cell culture. Bischoff (1980, 1986, 1988) used theflexor digitorum brevis muscle, a small, multipennatemuscle particularly well suited for dissociation andculturing due to the shortness of its fibers (0.4-1 mm).Collagenase digestion made it possible to separate thefibers from the connective tissue but left the fiber basallamina and underlying satellite cells attached (Bischoff,1979). Cultured mature fibers actively participated insprouting, with small pseudopodial outgrowths appear-ing after about a week and filling with myonuclei. Manyfibers eventually underwent complete dedifferentiationinto structures resembling embryonic myotubes (Bis-choff, 1980). When Bischoff (1980, 1986) treated somecultures with the antimitotic drug cytosine arabinoside(araC) to interfere with satellite cell proliferation, heindicated that araC had no effect on the behavior ofmature fibers but presented no specific data.

Of course, fibers might sprout chiefly by one means invivo and by another in vitro. However, the potential ofboth phenomena to occur together has been shown inculture, where myoblasts can fuse with a fiber alreadyundergoing dedifferentiation (Bischoff, 1988). The nu-clei within this muscle fiber would then originate fromtwo different sources, preexisting myonuclei and satel-lite cell myoblasts.

Leaving aside the question of the relationship be-tween muscle fiber behavior in vitro and muscle regen-eration in vivo, the present cell culture study addressesthe role of satellite cells in muscle fiber sprouting anddedifferentiation. We used araC to eliminate prolifer-ating cells (including fibroblasts and any other residualcell types as well as satellite-cell myoblasts) from someof the cultures. Since myonuclei do not replicate(Stockdale and Holtzer, 1961; Shafiq et al 1968; Bisch-off and Holtzer, 1969; Pullman and Yeoh, 1978),myotubes and mature fibers are unaffected by araCtreatment. We found that essentially all fibers inuntreated cultures lost their adult morphology andacquired characteristics of embryonic myotubes.[3H]thymidine labeling showed that proliferating myo-blasts fused extensively with the fibers. In the absenceof numerous myoblasts in cultures treated with araC,however, very few fibers underwent these changes.Replacing the native, rat myoblasts with mouse satel-lite-cell-derived C2C12 myoblasts, which also fusedwith the muscle fibers, greatly restored the capacity offibers to dedifferentiate, and myotubes containingmingled rat and mouse nuclei were produced. Replac-ing satellite cells with fibroblasts or with non-fusingtransformed myoblasts failed to restore this behavior.

These results indicate that fusion between satellite-cell myoblasts (or myotubes) and mature fibers pro-moted fiber dedifferentiation. To avoid implying aspecific pattern of changes in gene expression in themature fibers, and to emphasize that separation ofmononucleated cells from mature fibers (see Reznik,1976) was not observed, we prefer to call this phenom-enon 'remodeling' rather than dedifferentiation. Abrief account of this work has been published (Hinter-berger and Barald, 1988).

Materials and methods

Preparation of rat muscle fiber culturesFibers were isolated from flexor digitorum brevis muscles ofWistar rats 4-12 months old through a modification of theprocedures of Bekoff and Betz (1977), Jay and Barald (1985),and Bischoff (1986). After the animals were killed by CO2asphyxiation, the muscles were aseptically removed andincubated with 0.3% collagenase (Sigma, type I, 1830 col-lagenase units mg"1) in 5 ml of Dulbecco's PBS (Gibco) in atube mounted on a tilted rotator at 37°C for 3h. They werethen transferred to D-PBS plus 5% horse serum (M.A.Bioproducts), and bundles of fibers were carefully loosenedfrom their tendons with fine forceps. To do this, the tips ofclosed forceps were inserted and spread between the softenedtendons and the rows of muscle fibers. As much of the tendonsand nerves as possible were removed, and the fiber bundleswere triturated through a wide-mouthed pipet followed by aPasteur pipet to yield individual fibers. These were collectedby sedimentation through a standing tube of D-PBS plus 5 %horse serum. Remaining fragments of tendons and nerves aswell as undissociated fiber bundles were picked out of thefibers under a dissecting microscope, and sedimentation wasrepeated. Typically ten to twenty thousand viable musclefibers were obtained from each animal.

Fibers were plated on 35 mm tissue culture dishes (Falcon)that had been coated with polylysine and laminin (Sigma) asdescribed by Foster et al. (1987) but with the laminin solutionleft on the dishes for 5-10 days at 4°C before rinsing withmedium. Approximately 500-1000 fibers were placed in1-1.5 ml of culture medium per dish. The medium consistedof Dulbecco's modified Eagle's medium (D-MEM, Gibco),10% horse serum (M.A. Bioproducts), 5% chick embryoextract (Nishi and Berg, 1979), 2mM glutamine, 1% anti-biotic-antimycotic solution, and 12//gml~l gentamycin(Gibco or Sigma). AraC (Sigma) was added to individualdishes from 100 x stock solution to give IOJIM. Cultures weremaintained at 37°C in 5 % CO2. Medium was changed on thecultures after 4 days by careful pipeting under a dissectingmicroscope, in order to avoid detaching the fibers. For fibercounting, cultures were stained with 0.1% trypan blue(Gibco) in D-PBS, rinsed in D-PBS, and observed with phasecontrast at lOOx. Only fibers excluding the dye were counted.

Cocultures of mouse myoblasts with rat muscle fibersC2C12 myoblasts, a subclone (Blau et al. 1983) of cellsoriginally derived from adult mouse muscle satellite cells(Yaffe and Saxel, 1977), were from a stock provided by Dr H.Blau, Stanford University. Non-fusing, ras-transformedC2C12 cells (Olson et al. 1987), designated C41, were pro-vided by Dr E. Olson, University of Texas, Houston. Both ofthese cell lines were grown in D-MEM with 20% fetal calfserum (Flow), 0.5% chick embryo extract, and O.lmgmF1

kanamycin (Sigma). Balb/c 3T3 fibroblasts, obtained from DrM. Imperiale, University of Michigan, were grown inD-MEM plus 5-10 % horse serum and 1 % antibiotic-antimy-cotic solution. After aliquots of these cells were added to 4day muscle fiber cultures, the cocultures were maintained infiber culture medium, which was replaced after 2 days. Foridentification of mouse cell nuclei, cocultures were stainedwith bisbenzimide (Hoechst 33258, Sigma) according to theprocedure described by Blau et al. (1983). Bisbenzimide-stained cultures were observed with Leitz filter combination D(excitation 355—425 nm, transmission >460nm). Mouse nu-clei appear speckled due to the presence of A-T rich satellite

Myoblasts promote adult muscle fiber remodelling 141

DNA, while rat nuclei appear uniformly pale (Hardeman etal. 1988).

[3H]thymidine labeling[3H]thymidine was added to individual culture dishes to a finalconcentration of 10 nCi ml"'. To demonstrate fusion betweensatellite-cell myoblasts and muscle fibers in cultures nottreated with araC, [3H]thymidine was present continuouslyfrom the start of culturing. To determine whether any cellsproliferated and fused with fibers beginning after the end ofaraC treatment, the label was added only after removal of thedrug on day 4. For autoradiography, cultures were fixed inmethanol:formalin 9:1, rinsed in distilled water, and airdried. They were coated with Kodak NTB2 emulsion bypipeting several drops onto the dishes, tilting them to coverthe surface, and pipeting out the excess. After drying, theywere exposed at 4°C for 2 weeks.

Results

Muscle fiber remodeling in the presence of satellite cellsImmediately after dissociation and usually for the first 2days in culture, the muscle fibers retained their matureappearance, displaying prominent striations, peripheralmyonuclei, and jagged myotendinous junctions. Mono-nucleated cells, presumably satellite cells, were at-tached to the surface of most fibers (Fig. 1). Thesatellite cells, as well as fibroblasts from remainingconnective tissue fragments, began to migrate onto thelaminin substrate after 1 day. These cells proliferatedrapidly, and by 4 days, satellite-cell myoblasts hadbegun to fuse to form small myotubes (Figs'2, 3).

Also by 4 days, the appearance of the mature musclefibers began to change to resemble embryonic myo-tubes. Vesicles appeared within the fibers, and portionsof some fibers flattened onto the substrate yet remainedstriated. In many cases, mononucleated cells madecontact with muscle fibers as though preparing to fusewith them (Fig. 3A). Myotubes that had formed fromsatellite cells fused with the fibers as well (Figs 2, 3B).

Typically by 6-7 days, nearly all the muscle fiberswere remodeled to various extents. Many entirely losttheir mature appearance and fused into the network ofmyotubes forming from satellite cells. Myotubes de-rived from mature fibers were usually marked byclusters of large vesicles (Fig. 2) not stainable for lipidor for glycogen (data not shown). At the same time,some fibers had only a few vesicles or small myotube-like sprouts or small regions of flattening. Many fiberstwitched extensively after 5-6 days in vitro. Thisappeared to contribute to adhesion failure, which led todetachment and rounding up of some remodeled fibers.Muscle fiber remodeling thus generated a wide range ofmorphologies, from flattened sheets to thin tubes torounded 'myoballs,' with many fibers being partly oneshape and partly another (Fig. 2). By day 8,97.5±0.75% (mean of 3 experiments±s.E.M.) of theviable muscle fibers in untreated, control cultures hadbecome remodeled (Fig. 4). In all experiments, fiberswere scored as remodeled if they were at least partiallyflattened onto the substrate, if they had sprouts contain-ing one or more nuclei, or if they had rounded up andaccumulated vesicles. Fiber contraction without vesicleformation, or the appearance of vesicles in an otherwiseunaltered fiber were not considered sufficient evidenceof remodeling.

In order to substantiate the assumption, based onobservations of living cultures, that myoblasts fusedwith mature fibers, [3H]thymidine labeling was used totrack the nuclei of proliferating cells. In cultures ex-posed continuously to lOnCiml"1 [3H]thymidine for8-10 days, muscle fibers remodeled as usual; followingautoradiographic processing, these were all found tocontain labeled nuclei (Fig. 5). At higher concen-trations of [3H]thymidine, cell proliferation and myo-genesis were inhibited (data not shown), in agreementwith the results of Levis et al. (1970) with embryonic ratmyoblasts.

Fig. 1. Muscle fibers 1 day in vitro with apparent satellite cells (arrows) still attached. Phase contrast. Scale line=0.1mm.


A B

DAY 4

DAY 5

DAY 6

DAY 7

DAY 8

I I !

Fig. 2. Four representative adult muscle fibers (a-d) photographed at daily intervals from 4 or 5 days to 8 days in vitro.Arrowheads of graduated sizes follow single fibers throughout the time course (in a,b,d). Transition of arrowheads fromsmall to large indicate that the fiber has remodeled. Note small myotubes (m), derived from satellite cells [see examples in adays 6 and 7), and vesicles (*s in c and d days 7 and 8] in remodeling fibers. Stars [see d days 7 and 8] indicate partiallydetached fibers forming 'myoballs.' The dark corners seen on some frames are from orientation marks drawn on the dishes.Phase contrast. Scale line=0.5mm.

Effect of araC treatmentProliferating cells were eliminated by exposure to 10 /JMaraC during the first 4 days in vitro. In these cultures,the extent of muscle fiber remodeling observed at 8 dayswas drastically reduced; only 2.5±0.45% (3 exper-iments) of the viable fibers in these cultures becameremodeled to resemble myotubes (Figs 4, 6). Satellitecells were not found on the surface of araC-treatedfibers that retained their mature appearance, althoughretention of satellite cells on a small fraction of suchfibers cannot be ruled out. Some myoblasts and fibro-blasts initiated mitosis after the removal of araC. To testwhether myoblast proliferation and fusion were respon-sible for the remodeling of those araC-treated fibersthat did so, araC was maintained in some cultures forthe entire 8-day culture period and replenished after 4days. With this protocol, 1.5±0.40% (2 experiments)of the fibers still remodeled. To examine furtherwhether satellite cell proliferation following the re-

moval of araC and subsequent myoblast fusion withmuscle fibers might lead to fiber remodeling, 4 culturesthat had been treated with araC for 4 days were exposedto [3H]thymidine for 4 days in the absence of the drug.None of the 53 fibers that remodeled in these culturescontained any labeled nuclei (Fig. 7), indicating thatmyoblasts had not divided and then fused with themduring the labeling period.

Effects of myoblast addition to araC-treated culturesAs a test of the hypothesis that myoblast fusion canpromote muscle fiber remodeling, exogenous myoblastswere added to fibers whose native, satellite-cell myo-blasts had been largely eliminated by araC treatment.In cultures that received 105 mouse C2C12 cells afteraraC medium was replaced with regular medium on day4, muscle fiber remodeling was largely restored (Fig. 4).Within 1 day after their addition, the myoblasts werefusing with muscle fibers, and a few of the fibers began


100

Fig. 3. Details of contacts between adult muscle fibers (af)grown without araC and other cells at 4 days in vitro. Notesmall myotubes (mt), points of contact or fusion betweenfibers and other cells (arrows), and a dead fiber fragment(*). Phase contrast. Scale line=0.1mm.

spreading onto the laminin substrate. Often by 2-3 daysof coculturing, many C2C12 myotubes had formed andwere fusing with the rapidly remodeling fibers. In 4-5day cocultures (i.e. 8-9 days in vitro for the fibers), anetwork of myotubes, consisting of both remodeled ratmuscle fibers and C2C12 mouse myotubes, covered thedishes. Distinguishing between remodeled rat musclefibers and newly formed C2C12 myotubes was difficultwith phase contrast microscopy, but we estimate that atleast 95 % of the fibers remodeled. Fusion between theC2C12 cells and the muscle fibers was confirmed withbisbenzimide staining, which showed nuclei from bothspecies mingled in the same myotubes (Fig. 8).

To control for the possibility that simply the presenceof many proliferating cells, rather than fusion betweenthe myoblasts and muscle fibers, was in fact responsiblefor fiber remodeling, the same number of non-fusingC41 mouse myoblasts was added to some araC-treatedcultures in place of C2C12 myoblasts following removalof araC on day 4. Precise fiber counts were impossibledue to the overgrowth by C41 cells, but clearly fewerthan 5 % of the fibers flattened onto the substrate,formed sprouts, or acquired large vacuoles as fibers did

40

20

Untreated AraC-treated

AraC plusC2C12»

AraCplusC41s

AraCplus3T3s

AraCplusC2C12conditioned

medium

Fig. 4. Extents of muscle fiber remodeling at 8-9 days invitro following various treatments. Each column representscounts of at least 1000 fibers from 2-3 experiments.Columns without S.E.M. bars represent estimated counts(see text). All araC treatments shown here were for thefirst 4 days; ceH lines or conditioned medium were added onday 4.

Fig. 5. Autoradiograph of a muscle fiber culture exposed to[3H]thymidine continuously for 8 days. Note mixing ofunlabeled, preexisting myonuclei (arrows) and labeled,newly added myonuclei in a remodeled fiber. Hematoxylinstained. Scale line=50mm.

in untreated cultures (Fig. 4). However, approximatelyhalf of the viable fibers in cocultures with C41 cellsbecame highly contracted, far more than in untreatedor araC-treated cultures (data not shown). In anothercontrol experiment, 105 Balb/c 3T3 mouse fibroblastswere added to araC-treated rat muscle fibers on day 4.This too resulted in fewer than 5 % of fibers remodelingafter 8-9 days in vitro (Fig. 4), although, again, manymore fibers were highly contracted than in untreated orin araC-treated cultures.

Another concern was that some diffusible factor fromthe C2C12 myoblasts, rather than their demonstrated


Fig. 6. Muscle fibers in cultures treated with araC on days0-4. (A) A fiber cultured for 6 days. (B) A fiber culturedfor 15 days. Note retention of striations and jaggedmyotendinous junctions at the ends of the fibers. Phasecontrast. Scale bars are 0.1mm.

fusion with the muscle fibers, could be promotingremodeling. Therefore conditioned medium fromC2C12 cultures, filtered and mixed 2:1 with freshmuscle fiber medium, was placed on muscle fibers after4 days araC treatment. The number of fibers thatremodeled in these cultures (4.05±0.60 %, 2 exper-iments) was not significantly different (/)>0.05, two-tailed f-test) from the number that did so in araC-treated cultures maintained in unconditioned C2C12medium (2.15±0.15%, 2 experiments).

Discussion

Satellite cells from adult rat muscle promoted a strikingremodeling of mature muscle fibers in this in vitrostudy. Over a period of about a week, most of the fibersmaintained on laminin substrates under conditions thatpermitted proliferation and fusion of their satellite cellscame to resemble embryonic myotubes. Identificationof preexisting myonuclei within the remodeled fiberssupports the idea that sprouted or remodeled regions offibers can form by 'budding' from mature regions(Shafiq, 1970; Ali, 1979; Bischoff, 1980, 1986). At thesame time, our results indicate that remodeling of

7A

B

Fig. 7. Autoradiographs of muscle fibers exposed to[3H]thymidine on days 4-8 following araC treatment ondays 0-4. Fibroblast nuclei are labeled, but myonuclei areall unlabeled. Hematoxylin stained. Scale lines=0.1mm.

muscle fibers may depend on satellite cell or myoblastfusion with the fibers.

Eliminating most of the mononucleated cells fromthe cultures by araC treatment greatly decreased thenumber of muscle fibers that remodeled into myotubesby 8 days in vitro. In this respect, our results appear todiffer from those of Bischoff (1980,1986), who reportedthat virtually all fibers eventually converted into myo-tube-like structures in both araC-treated and untreatedcultures. However, this may be a function of the time inculture, since Bischoff maintained fibers in vitro for 3weeks or longer. When the culture period for our araC-treated muscle fibers was extended beyond 2 weeks, wealso noted changes in the morphology of many morefibers, including shrinkage and constriction, centraliz-ation and clumping of nuclei, and loss of striations(unpublished observations). Since these events weredifferent from those observed in untreated cultures andtheir onset was substantially delayed, we consider themto represent a separate phenomenon, possibly dener-vation atrophy (Gutman and Zelena, 1962). Also,Bischoff (1980) exposed fibers to araC for only 2-3days, possibly allowing more satellite cells to surviveand then fuse with the fibers. Based on our results, wewould expect that changes in muscle fiber morphology


Fig. 8. Portion of a muscle fiber cocultured with 105 C2C12 mouse myoblasts for 4 days following araC treatment for 4 days.Rat myonuclei (arrows) are mixed with mouse myonuclei in the remodeled fiber. Bisbenzimide stained. Scale line=50mm.

should be significantly delayed and protracted, as wellas qualitatively different, in any thoroughly araC-treated cultures compared with untreated ones.

The finding that addition of exogenous myoblastsrestored normal remodeling ability to araC-treatedmuscle fibers demonstrated that the fibers themselveswere unaffected by araC. More importantly, it supportsthe hypothesis that satellite-cell myoblasts somehowpromoted fiber remodeling in the untreated cultures. Inthe presence of either the native satellite cells or theadded C212 myoblasts, fiber remodeling was associatedwith fusion between myoblasts and fibers. To controlfor the possibility that it was simply the presence ofproliferating cells rather than rnyoblast fusion withfibers that was responsible for promoting fiber remodel-ing, either non-fusing C41 myoblasts or 3T3 fibroblastswere added to araC-treated muscle fibers instead.Neither of these cell types significantly promoted fiberremodeling, although static contraction of the musclefibers was more prevalent in their presence. Anothercontrol procedure, the addition of conditioned mediumfrom growing C2C12 cultures to araC-treated musclefibers, also failed to restore remodeling ability to astatistically significant extent. Yet C2C12 conditionedmedium enhanced fiber remodeling slightly, suggesting

that diffusible factors from myoblasts may play somerole. The fiber contraction seen in the presence of C41or 3T3 cells also might well be caused by some factorreleased from those cells.

Our results raise fundamental questions about themaintenance of the differentiated state by adult musclefibers, both in vitro and in vivo. Why did muscle fiberdissociation and culturing lead to their fusion withmyoblasts, and why did fusion in vitro lead to suchcomplete fiber remodeling? Fusion between satellite-cell myoblasts and mature muscle fibers in a variety ofsituations in vivo is not associated with fiber remodel-ing. Satellite cells proliferate and fuse with fibers duringnormal growth (Moss and Leblond, 1971), and fusion isaccelerated during muscle hypertrophy (Schiaffino etal.1976) or when satellite cells are propagated in vitro andthen grafted back into a muscle (Lipton and Schultz,1979), without noted loss of adult fiber morphology.The sprouting of fibers, accumulation of vesicles, fusionbetween myoblasts and fibers, and fusion betweenremodeled fibers that we observed in vitro is character-istically seen in vivo only following muscle injury(Clark, 1946; Gay and Hunt, 1954; Walton and Adams,1956; Roth and Oron, 1985). Myoblast fusion may turnout to be a necessary but insufficient condition for the


remodeling of mature fibers, with denervation or dis-ruption of normal tissue architecture also beingrequired. Growth factors present in the serum andembryo extract that were used in vitro, or factorsreleased from proliferating myoblasts and fibroblasts invivo, may help to trigger remodeling. Our conclusionthat myoblast fusion promotes cultured muscle fiberremodeling still remains to be explicitly tested ininjured muscle in vivo, although the observation thatvinblastine injection reduced fiber sprouting in rats(Roth and Oron, 1985) is supportive.

While these results demonstrate a major role formyoblast fusion in promoting morphological remodel-ing of muscle fibers in vitro, the mechanisms of remo-deling and its control remain unknown. Myoblasts andmyotubes arising from satellite cells certainly addednuclei and cytoplasm to the mature fibers. Yet the fibersalso appeared to be active participants in the remodel-ing process, as they rapidly (sometimes less than 2 days)and often completely lost their normal morphology.This participation presumably involves structural geneproducts not normally expressed in mature fibers, andone possibility is that these genes might be expressedonly by the newly added myoblast nuclei. We havefound that the embryonic rat muscle cell surface antigenH36 (Kaufman et al. 1985) appears to become moreabundant on remodeling fibers when native satellite-cellmyoblasts are present than when C2C12 cells are added(unpublished observations), suggesting that the myo-blast nuclei may be responsible for H36 expression.Alternatively, factors in the myoblast cytoplasm, orfactors expressed by their nuclei after fusion, couldinfluence gene expression by preexisting myonuclei.Such factors might be similar or identical to those thattrigger muscle gene expression by non-muscle nuclei inmixed heterokaryons (review, Blau et al. 1989). Geneproducts characteristic of muscle fiber remodeling mayof course be expressed both by preexisting and by newlyadded myonuclei. Studies using additional species-specific antibodies or nucleic acid probes for these geneproducts in cocultures of muscle fibers and myoblastsfrom different animals will distinguish among thesepossibilities. Even without reagents that are species-specific, it may be possible to visualize products ex-pressed by individual newly added and preexistingmyonuclei in the remodeled fibers, just as mRNAs(Harris et al. 1989; Fontaine and Changeux, 1989) andmembrane proteins (Pavlath et al. 1989; Ralston andHall, 1989) have been shown to be localized aroundindividual nuclei in cultured myotubes. This is the firstreport of fusion between adult individual rat musclefibers and a mouse satellite cell line. This ntovel culturesystem presents muscle biologists with a new modelsystem for studies of gene expression, extinction, regen-eration and possible recapitulation (or not) of molecu-lar events in embryonic muscle development.

Although the present results provide strong evidencethat myoblast fusion promoted rapid muscle fiberremodeling within the culture period examined, they donot address the question whether myoblast fusion is anabsolute requirement for fiber remodeling. The obser-

vation that 2-3% of fibers in araC-treated cultureswithout added myoblasts underwent remodeling, andthe observation that myonuclei of remodeled fibersexposed to [3H]thymidine following araC treatmentwere unlabeled, might suggest that some fibers remo-deled in the absence of myoblast fusion. However,these results are not inconsistent with an absoluterequirement for myoblast fusion in fiber remodeling,since satellite cells may have divided once during theseveral hours that elapsed between muscle dissectionand addition of araC to the final cultures, therebyescape elimination, and subsequently fuse with thefibers. Satellite cell mitosis has been observed within30min after explantation of rat muscles into organculture (Maltin, 1987), and many satellite cells can fusefollowing only one round of mitosis in injured musclesin vivo (Grounds and McGeachie, 1987,1989). Further-more, it is entirely possible that some satellite cellsfused with fibers in the presence of araC, not dividingeven once. Approaches that facilitate identification ofnon-dividing satellite cells on araC-treated musclefibers, and then correlate fusion between these cells andtheir fibers with subsequent fiber remodeling, will beneeded to determine whether any mature fibers can infact initiate remodeling independently of myoblastfusion.

Our thanks to Drs H. Blau, E. Olson, and M. Imperiale forsupplying cell lines and to Dr B. Carlson for donating animals.This work was supported by an N1H Postdoctoral TrainingGrant fellowship to T.J.H. and by NIH, MDA, and U. of M.Rackham Faculty grants to K.F.B.

References

ALI, M. A. (1979). Myotube formation in skeletal muscleregeneration. J. Anat. (Cambridge) 128, 553-562.

ALLBROOK, D. (1981). Skeletal muscle regeneration. Muscle Nerve4, 234-245.

BEKOFF, A. AND BETZ, W. (1977). Properties of isolated adult ratmuscle fibres maintained in tissue culture. J. Physiol. (London)271, 537-547.

BISCHOFF, R. (1979). Tissue culture studies on the origin ofmyogenic cells during muscle regeneration in the rat. In MuscleRegeneration (ed. A Mauro), pp. 13-29. New York: RavenPress.

BISCHOFF, R. (1980). Plasticity of the myofiber-satellite cellcomplex in culture. In Plasticity of Mitscle (ed. D. Pette), pp.119-129. Berlin: Walter de Gruyter and Co.

BISCHOFF, R. (1986). Proliferation of muscle satellite cells on intactmyofibers in culture. Devi Biol. 115, 129-139.

BISCHOFF, R. (1988). Mature myofibers suppress proliferation ofattached satellite cells. J. Cell Biol. 107, 687a.

BISCHOFF, R. AND HOLTZER, H. (1969). Mitosis and the process ofdifferentiation of myogenic cells in vitro. J. Cell Biol. 41,188-200.

BLAU, H. M., CHIU, C.-P. AND WEBSTER, C. (1983). Cytoplasmic

activation of human nuclear genes in stable heterokaryons. Cell32, 1171-1180.

BLAU, H. M., PAVLATH, G. K., CHIU, C.-P., HARDEMAN, E.,MILLER, S., BLAKELY, B., PAN, L., PETERSON, C. AND SCHAFER,B. W. (1989). Heterokaryons: A model system for studying cellcommittment and differentiation. In Cellular and MolecularBiology of Muscle Development (ed. L. H. Kedes and F. E.Stockdale), pp. 851-861. New York: Alan R. Liss.

CAMPION, D. R. (1984). The muscle satellite cell: a review. Int.Rev. Cytol. 87,225-251.


CARLSON, B. M. AND FAULKNER, J. A. (1983). The regeneration ofskeletal muscle fibers following injury: A review. Med. Sci.Sports Excr. 15, 187-198.

CHURCH, J. C. T. (1970). Cell quantitation in regenerating bat webmuscle. In Regeneration of Striated Muscle and Myogenesis (ed.A. Mauro, S. A. Shafiq, and A. T. Milhorat), pp. 101-117.Amsterdam: Excerpta Medica.

CLARK, W. E. L. (1946). An experimental study of theregeneration of mammalian striped muscle. J. Anal. (Cambridge)80, 24-36.

FONTAINE, B. AND CHANGEUX, J.-P. (1989). Localization of nicotinicacetylcholine receptor alpha-subunit transcripts duringmyogenesis and motor endplate development in the chick. J. CellBiol. 108, 1025-1037.

FOSTER, R. F., THOMPSON, J. M. AND KAUFMAN, S. J. (1987). A

laminin substrate promotes in rat skeletal muscle cultures:analysis of replication and development using antidesmin andanti-BrdUrd monoclonal antibodies. Devi Biol. 122, 11-20.

GAY, A. J., JR AND HUNT, T. E. (1954). Reuniting of skeletalmuscle fibers after transection. Anat. Rec. 120, 853-864.

GROUNDS, M. D. AND MCGEACHIE, J. K. (1987). A model ofmyogenesis in vivo, derived from detailed autoradiographicstudies of regenerated skeletal muscle, challenges the concept ofquantal mitosis. Cell Tissue Res. 250, 563-569.

GROUNDS, M. D. AND MCGEACHIE, J. K. (1989). Myogenic cells ofregenerating adult chicken muscle can fuse into myotubes after asingle cell division in vivo. Expl Cell Res. 180, 429-439.

GUTMAN, E. AND ZELENA, J. (1962). Morphological changes in thedenervated muscle. In The Denervated Muscle (ed. E. Gutman),pp. 57-102. Prague: Czechoslovak Academy of Sciences.

HARDEMAN, E., MINTY, A . , BENSON-VOSMAN, P., KEDES, L. AND

BLAU, H. M. (1988). In vitro system for characterizing clonalvariation and tissue-specific gene regulatory factors based uponfunction. J. Cell Biol. 106, 1027-1034.

HARRIS, D. A., FALLS, D. L. AND FISCHBACH, G. (1989).

Differential activation of myotube nuclei following exposure toan acetylcholine receptor-inducing factor. Nature 337, 173-176.

HINTERBEROER, T. AND BARALD, K. F. (1988). Formation of hybridmyotubes in vitro from adult rat muscle fibers and mousemyoblasts. J. Cell Biol. 107, 687a.

JAY, J. C. AND BARALD, K. F. (1985). Improved culture ofindividual muscle fibres with and without spinal cord explants ina collagen gel. J. Neurosci. Methods 15, 229-234.

KAUFMAN, S. J., FOSTER, R. F., HAYE, K. R. AND FAIMAN, L. E.

(1985). Expression of a developmentally regulated antigen on thesurface of skeletal and cardiac muscle cells. J. Cell Biol. 100,1977-1987.

KONIGSBERG, U . R . , L.IPTON, B . H . AND KONIGSBERG, I. R. ( 1 9 7 5 ) .The regenerative response of single mature muscle fibers isolatedin vitro. Devi Biol. 45, 260-275.

LEVIS, A. G., PICCINNI, E., MARTINUCCI, G. AND ZANUSO, G.

(1970). Inhibiting action of tritiated thymidine on thedifferentiation of rat myoblasts in culture. Riv. Biol. 63, 392-410.

UPTON, B. H. AND SCHULTZ, E. (1979). Developmental fate ofskeletal muscle satellite cells. Science 205, 1292-1294.

MALTIN, C. A. (1987). Observations of satellite cells in rat skeletalmuscle incubated in vitro. Cell Tiss. Res. 245, 177.

MAURO, A. (1961). Satellite cells of skeletal muscle fibers. J.Biophys. Biochem. Cytol. 9, 493-495.

Moss, F. P. AND LEBLOND, C. P. (1971). Satellite cells are a sourceof nuclei in muscles of growing rats. Anat. Rec. 170, 421-436.

NISHI, R. AND BERG, D. (1977). Dissociated ciliary ganglionneurons in vitro: survival and synapse formation. Proc. natn.Acad. Sci. USA 74, 5171-5175.

OLSON, E. N., SPIZZ, G. AND TAINSKY, M. A. (1987). The

oncogenic forms of N-ras or H-ras prevent skeletal myoblastdifferentiation. Molec. Cell. Biol. 7, 2104-2111.

PAVLATH, G. K., RICH, K., WEBSTER, S. G. AND BLAU, H. (1989).

Localization of muscle gene products in nuclear domains. Nature,Lond. 337, 570-573.

PULLMAN, W. E. AND YEOH, G. C. T. (1978). The role ofmyonuclei in muscle regenration: An in vitro study. J. Physiol.(London) 96, 245-252.

RALSTON, E. AND HALL, Z. W. (1989). Transfer of a proteinencoded by a single nucleus to nearby nuclei in multinucleatedmyotubes. Science 244, 1066-1069.

REZNIK, M. (1976). Origin of the myogenic cell in the adult striatedmuscle of mammals. Differentiation 7, 65-73.

ROTH, D. AND ORON, U. (1985). Repair mechanisms involved inmuscle regeneration following partial excision of the ratgastrocnemius muscle. Expl Cell Biol. 53, 107-114.

SCHIAFFINO, S., PIEROBON BORMIOU, S. AND ALOISI, M. (1976). The

fate of newly formed satellite cells during compensatory musclehypertrophy. Virchows Arch. B cell Path. 21, 113-118.

SHAFIQ, S. A. (1970). Satellite cells and fiber nuclei in muscleregeneration. In Regeneration of Striated Muscle and Myogenesis(ed. A. Mauro, S. A. Shafiq, and A. T. Milhorat), pp. 122-132.Amsterdam: Excerpta Medica.

SHAFIQ, S. A., GORYCKI, M. A. AND MAURO, A. (1968). Mitosis

during postnatal growth in skeletal and cardiac muscle of the rat.J. Anat. (Cambridge) 103, 135-141.

STOCKDALE, F. E. AND HOLTZER, H. (1961). DNA synthesis andmyogenesis. Expl Cell Res. 24, 508-520.

WALTON, J. N. AND ADAMS, R. D. (1956). The response of thenormal, the denervated and the dystrophic muscle-cell to injury.J. Bad. Path. 72, 273-298.

YAFFE, D. AND SAXEL, O. (1977). Serial passaging anddifferentiation of myogenic cells isolated from dystrophic musclecells. Nature, Lond. 270, 725-727.

(Accepted 6 February 1990)

fusion between myoblasts and adult muscle fibers...

Documents