Acta Physiol Scand 1991, 142, 443456

ADONIS

000 16772910014Ox

Elect romyography and morphology during regeneration of muscle injury in rats T. HURME"?, M. L E H T O f , B. FALCKS, H. T A I N I O f and H. K A L I M O " Departments of Pathology", Paavo Nurmi Center? and Neurophysiologys, University of Turku and University Central Hospital of Turku, Department of SurgeryI, University Central Hospital of Tampere, Finland. B., TAINIO, H. & KALIMO, H. 1991. Electromyography HURME, T., LEHTO,M., FALCK, and morphology during regeneration of muscle injury in rats. A m Physzol Scund 142, 443456. Received 28 October 1990, accepted 20 March 1991. ISSN 0001-6772. Departments of Pathology Surgery, Paavo Nurmi Center and Neurophysiologj-, University of Turku and University Central Hospital of Turku, Department of Surgery, University Central Hospital of Tampere, Finland. Healing of the partially ruptured rat gastrocnemius muscle was studied correlating electromyographical findings with morphological changes. Fibrillation potentials and positive sharp waves were seen both proximal and distal to the injury 7 days after the injury and disappeared in the proximal part by day 14 and in the distal part by day 21. Late components of motor action potential were observed from day 14 onwards. Denervation was mainly myogenic, i.e. caused either by rupture of myofibres, whence the abjunctional stump lost its contact with the neuromuscular junction on the adjunctional stump, or by necrosis of the segment of the ruptured myofibre lying underneath the neuromuscular junction. Lesser extent of denervation was neurogenic, i.e. caused by damage to intramuscular nerve fibres. The reinnervation occurs either by regeneration of the necrotized myofibres, by regeneration of the severed nerves, or by collateral innervation of new neuromuscular junctions in the abjunctional stumps. The present study indicates that electromyography may be useful in the diagnosis and follow-up of skeletal muscle injuries.

Key ~ o r d :s Denervation, electromyography, neuromuscular junctions, regeneration.

Rupture of a muscle is commonly caused either by external violence, contusion, or by internal strain by forceful uncoordinated muscle contraction. An experimental model for studying regeneration following muscle contusion injury was developed by Jarvinen & Sorvari (1975). I n that model the lesion is produced in rat gastrocnemius muscle by a strike with a spring loaded blunt hammer, which ruptures both muscle fibres and the supporting connective tissue (Lehto et al. 1986). Following the trauma, the ends of the ruptured myofibres are retracted and the gap formed at the site of the injury is at first filled with a haematoma and later on with Correspondence : Tim0 Hurme, Department of Pathology, University of Turku, Kiinamyllynkatu 10, SF-20520 Turku. Finland

proliferating cells which synthesize extracellular matrix components (Lehto et al. 1985). Regeneration of the transsected muscle fibres occurs at both stumps and finally the scar separating the stumps is traversed by newly formed fibres (Jarvinen 1975). During the healing process the scar diminishes in size and the function of the muscle is gradually restored. I n this model the contusion destroys the muscle in such a region that intramuscular nerve branches may also be damaged. O n the other hand, since a muscle fibre receives its neural input at one location only, at the neuromuscular junction (NMJ), the part of fibre distal to the disruption loses its connection with the NMJ and thereby its neural input. Thus, the muscle fibres in this model of contusion muscle injury may be denervated in two different ways.

443

444

T . Hzirme et

a1

Previous studies have s h o w that intact innermtion is necessar!. for regeneration of muscle cells beyond the myotube phase (Schick &Jerusalem 1973, Carlson &- Faulkner 1983). Thcrcfore, regeneration of the damaged intramuscular nerve twigs may significantly affect the healing myofibres after contusion injur!-. I n the present stud!- we investigated the electrophysiological events in correlation with the morphologically detectable damage of nerve and muscle fibres a n d their subsequent regeneration in the a h \ - e mentioned model of muscle contusion

M -41'E R I AL S AND M E T H O D S Forty-two male Wistar rats (12-16 weeks old) were used in this stud!. -411 animals were housed in individual cages and fed with commercial pellets and m t e r (id hhltzim. .4 standard contusion injur!- was induced to the left calf under light ether anaesthesia using a blunt spring loaded hammer. The gastrocnemius muscle was partially ruptured without tearing the overlying skin (Jarvinen & Sorvari 1975). The injury as about 1 cm distal to the point at which the ncr\es to the gastrocnemius, soleus and piantaris muscles branch from the tibia1 nerve. The da!- of the traumatization was designated dak- 0. At each of the folloaing time points, 2, 5 , 7, 10, 14, 21, 35, 42, and .56 d a y after the trauma, two rats were examined neuroph!-siologicall~ whereafter they were sacrificed for morphologic analysis. Two untreated rats seri-ed as controls for neuroph logical examinations, 3 rats were used as controls the morphological studies and further 5 untraumatized animals were used to study the innervation pattern of gastrocnemius muscle by dissecting the muscle under a microscope. The experiments were approved by the ethics committee of Turliu University. Electromyographic (EMG)recordings of the gastrocnemius muscle were done with a concentric needle electrode (Dantec 13L23, Dantec Electronics, Skovlunde, Denmark), with external diameter of 0.35 mm, the area of the recording surface was 0.04 mm'. The signals were amplified and displa?-ed with a standard commercial EXIG unit (Dantec Neuromatic 2000). The gastrocnemius muscle was carefull!- explored under ether anaesthesia with the concentric needle electrode both distal and proximal to the muscle injur) (Fig. 1 C). The number of fibrillation potentials and positive sharp waves in 20 insertions was observed. If fibrillation potentials and positive sharp waves were observed in 2 4 sites the finding was scored as mild, 5 f 5 sites with fibrillation potentials and positive sharp \laves was scored as moderate and 16 or more sites with tihrillation potentials and positive sharp waves as ahundant. The needle electrode was left in

place and the sciatic nerve was stimulated in the thigh using a supramaximal stimulation intensity with monopolar needle electrodes (Dantec 13L49). The motor action potential (hL4P) was recorded with the needle electrode. T h e latency from the stimulus to the first and last components following the main MAP were measured. Filter settings used for all recordings were 20 Hz-5 kHz. For histology and immunohistochemistry gastrocnemius muscles were fixed at constant length with phosphate buffered paraformaldehyde. For paraffin embedding muscle samples were obtained sagittally from the midline of the injury site and transversely from levels distal and proximal to the trauma site. Routine 5 ,um sections were stained with the van Gieson or Herovici technique. For the demonstration of the nerve injury and muscle regeneration the sections were stained histologically with Luxol Fast Blue-cresyl violet stain and immunohistochemically with monoclonal mouse anti-desmin antibody (Zymed Laboratories Inc., San Francisco, USA). T h e bound antibodies were visualized by using the biotin-avidinperoxidase method (ABC kits, Vector Laboratories, Burlingame, USA) with diaminobenzidine as chromogen. For enzyme histochemistry specimens from both distal and proximal parts of the injured gastrocnemius muscles were obtained on days 2, 5 , 7, 14 and 21 immediatelj- after death and snap frozcn in freon 22 cooled \I-ith liquid nitrogen. Serial sections of 5 ,um were cut on a cryostat and reacted for acetylcholinesterase (AChE) (Karnovsky & Root 1964). For semithin epon-sections a part of the samples were refixed Lvith glutaraldehyde, postosmicated, dehydrated and embedded in epon. T h e sections were stained kvith toluidinc blue.

RESULTS Jnnerzation pattern of gastrocnemius T h e r e was considerable variation i n t h e innervation pattern between rats though a certain basic pattern seemed to prevail. Both the medial and lateral heads of the gastrocnemius muscle had 3 main branches which ran on the deep surfaces of t h e muscle bellies a n d entered the muscle at 3 different levels (Fig. 1). After entering the muscle the nerves divided into multiple branches. T h e larger intramuscular branches ran parallel t o the myofibres deep between fascicles and the small twigs with u p to 10 axons ran mainly towards the muscle surface perpendicular to t h e long axis of the muscle fibres. On the basis of AChE-stainings the NMJs in both heads wcre located i n 3 obliquely oriented planes (Fig. 2 A ) a t different levels

Electromyography in muscle injury

~

/"\

A

445

m

B

C

Fig. 1. (A, B) Schematic representation of two variants in the innervation pattern of the gastrocnemius muscle. The tibia1 nerve send 3 main branches to both the medial and lateral head. These ran somewhat varying courses on the deep surface of the muscle to finally enter the muscle belly at 3 different levels, after which they divided into multiple intramuscular branches. (C) The site of injury and the points of insertion of the EMG electrodes.

corresponding to the main branches to each head. Because of this orientation the NMJs in sagittal sections usually ran in 3 slightly anteriorly descending rows (Fig. 2). In transverse sections NMJs are present in 2-3 radically oriented rows (Fig. 2B). The relationship of the trauma to the NMJs is depicted in Figure 3.

Microscopical findings The structural details of the healing of the ruptured muscle have been described in detail elsewhere (Hurme et al. 1991). Here we describe the repair process to the extent that is needed for understanding the denervation phenomena. Throughout the repair process, three different zones could be defined (Fig. 3A). The central zone (CZ) contained at early stages a haematoma, and was later filled with granulation tissue. The cavity was surrounded by the regeneration zone (RZ), where at the early stages the necrosis of muscle fibres occurred. Outside the traumatized area there was a surviving zone (SZ) were myofibres remained viable. Two days following the traumatization the necrotized parts of the fibres in RZ were restricted by formation of a demarcation membrane that sealed the plasmalemma of the breached myofibre (Fig. 4). At this stage the necrotized part appeared as cylinders of basal

lamina filled with macrophages and lined by spindle shaped, strongly desmin positive myoblasts derived from activated satellite cells (Fig. 4). The central zone was filled with a haematoma. By day 5 regenerating myoblnsts had fused into large multinucleated anti-desmin positive myotubes, which by day 7 had assumed a definitely cross-striated appearance. These myotubes filled the preserved old basal lamina cylinders in RZ and their tips now began to grow through the increasingly dense connective tissue in the CZ. At 14 days a few of the basophilic regenerating muscle fibres had succeeded to extend across the entire gap between the surviving stumps (Fig. SA). Many of the regenerated fibres had a fairly interlaced orientation as they also had still on day 21, when zone between the surviving muscle fibre stumps had become occupied by thin regenerating fibres within the connective tissue. Thereafter the regenerated new myofibres began to increase in thickness though they remained smaller in diameter than those in the uninjured areas of the muscle (Fig. 5B). On days 28, 35,42 and 56 the area of the connective tissue decreased gradually, but was still discernible on day 56. At the site of the trauma only few damaged nerve twigs of size up to 10 axons were observed : myelin was disrupted and a few phagocytic cells were also present within the perineurial sheath (Fig. 6A). Fairly close to the lesion both laterally 18-2

446

T. Hurme et al.

Fig. 2. T h e neuromuscular junctions in the medial head of gastrocnemius muscle are seen (A) in the sagittal section in a row extending obliquely from deep-proximal to superficial-distal and (B) in transverse section from the deep part of the muscle in a radially oriented row. Note the longitudinall! running nerve branch (arrow). DE = deep, P = proximal, S = superficial, DI = distal. .icet!-lcholincstcrase stain, bars -1:200 p m , B : 100 lrm.

Electromyography in muscle injury

447

B Fig. 3. For legend see p. 448.

and deep to the CZ well preserved nerve branches were present, even in areas where the muscle fibres were separated by reactive inflammatory and connective tissue cells (Fig. 6B, C). Depending on the site and extent of the contusion, the necrosis extended in some fibres to the level of the NMJ-zone on day 2, the fibre underneath the endplate appeared necrotized or was composed of regenerating myoblasts (Fig.

7A). Besides, many intact fibres with NMJs were seen in transverse sections next to necrotic fibres, and longitudinal sections revealed that necrosis had not reached the level of the NMJ. 5 days following traumatization AChE positive NMJs were occasionally found on basal lamina bound myotubes both in the proximal and distal stumps (Fig. 7B). In some longitudinal sections majority of NMJs were present on basal lamina

448

C

T . Hurme et al.

Fig. 3. ( A ) .4n overview of the lesion in the ruptured gastrocnemius muscle on day 2 demonstrating the different zones. CZ = central zone, RZ = regeneration zone, SZ = surviving zone, the border line between the two latter zones is indicated by the dashed line. Though the myfibres are breached fairly sharplj-, the contraction of the ruptured stumps results in formation of qide gap between them. The distribution of two distal rows of neuromuscular junctions is schematically depicted on this section with circles. Herovici stain, bar 1 mm. (B, C) A schematic presentation of the three different alternatives for denervation. (B) In myogenic denervation either the rupture of myofibres interrupts the contact of the abjunctional part with its NMJ (fibres 1 & 3 ) , or the myofibre is ruptured so close to the NMJ that the necrosis extends beyond the level of the NMJ (fibre 2). (C) In neurogenic denervation the axons are damaged by the contusion.

Fig. 4. T h e necrotized part of the myofibre is demarcated on day 2 from the intensely antidesmin positive surviving part of the breached myofibres. T h e preserved basal lamina forms a cylinder, which is lined by numerous myoblasts (two of which marked with an arrow) and which contains macrophages that have phagoc!-tosed the necrotized sarcoplasm. Anti-desmin haematoxylin counterstain, bar 50 pm.

+

Electromyography an muscle injury

Fig. 5. (A) The site of injury 14 days after trauma. The ruptured muscle fibres have regenerated within the preserved basal lamina cylinders and they continue their growth as myotubes across the connective tissue with an attempt to re-establish the continuity between the stumps. (B) 56 days after injury the scar tissue between the stumps has nearly disappeared. The diameter of the regenerated muscle fibres is smaller than that in the uninjured areas, and their orientation is somewhat irregular, interlaced. A : Anti-desmin haematoxylin counterstain. B : vanGieson stain, Bars 100fim.

+

449

450

T. Hurme et al.

Fig. 6. For legend see facing page. bound myotubes indicating that most of the fibres innervated by the distal branch had become denervated. On days 7 and 10 the findings in 4C:hF: stained section were fairly similar to those on day 5 , some YMJs were still found on basal lamina hound myotubes. After 14 and 21 days NIIJs were found in the proximal NMJ zones in similar radial lines as in control sections, and the same was true for some distal NMJ-zones. In some animals the trauma had involved the distal NhlJ-zone and a more regular pattern was observed, the radial line of the NMJs being interrupted by scar tissue.

E c1 G' findings Spontuneous uztiz.it.y. The amount of fibrillation potentials and positive sharp w-aves following the traumatization are presented in Table 1. Five days following the traumatization increased insertional activity and a few fibrillation potentials and positive sharp waves were seen.

Fibrillation potentials and positive sharp waves were observed 7 and 10 days after the injury in both proximal and distal part of the muscle (Fig. 8). On the 14th day the fibrillation potentials and positive sharp waves had disappeared from the proximal segment of the gastrocnemius muscle, but a few fibrillation potentials and positive sharp waves were still seen in the distal part. On day 21 no fibrillation potentials and positive sharp waves were observed any more, only some fasciculations were observed in both the distal and proximal segment of the gastrocnemius muscle. 1,ater on no abnormal spontaneous activity was seen. I,ute components following the muin M A P . TN-o,five and seven days following traumatization there were no late components following the main MAP (Table 2). Ten days following the traumatization small satellites following the main motor action potential were observed in the proximal segment of the gastrocnemius (Fig. 9.1). The latency of these small late components

Electromyography in muscle injury

Fig. 6. Semithin epon-sections from the trauma site on day 2. (A) A small nerve twig undergoing Wallerian degeneration with disrupted and phagocytosed myelin is seen in the trauma site next to necrotic myofibres (arrows). (B) Well preserved nerve branches (arrow) are encountered very close to the injured area and (C) even in an area where muscle fibres are separated by oedema inflammatory cells. Toluidine blue, bars, A: 25 pm, B and C: 50 ,urn.

Fig. 7. For legend see p. 452.

45 1

452

T. Hurme et al.

Fig. 7 . (.4)On day 2 one AChE positive NMJ (1) is seen on a necrotized myofibre (asterisk) euemplifying myogenic denervation. Next to this there is another NMJ (2) on a preserved ni!-ofbre. (R) On da!- 5 AChE positive NMJs are present on 3 basal lamina bound, basophilic myotubes, indicating regeneration of the necrotized myofibre to be reinnervated by the preserved XAlJs. -4: .~cetylcholinesterase ranGieson counterstain, B : Acetylcholinesterase haematoxylin counterstain. Bars, .I : 100 pm, R : 50 p m

+

+

T a b l e 1. Amount of fibrillation potentials and positive sharp waves following traumatization. inc.ins. ( = increased insertional actiI-it!-), fasc. ( = fasciculations), moder. ( = moderata). 1 3 ~ \ s Proximal site after Rat 2 trauma Rat I ~~

Distal site Rat 1

Rat 2 VO Inc ins

~

1

\O

->

Inc ins lloder Some

\lo Inc ins Inc in? \o

It

O \

VO

zi 28 35 42 56

Fasc

Fasc

\o Inc ins \loder Some Some rase

YO

Y O

Y O

NO

h0 No No

Y O

WO

YO

YO NO

KO

NO

Y O

N O

I

10

\llld Some Some Fasc

here 1-5.4-19.8 ms in the initial stage. On days 21-35 these late components had a maximum latenc! ranging from 1 6 1 3 ms. On da! 42 there Irere late components with verj long latenu, up t o 122 ms (Fig 9B)

DISCUSSION Neurophysiological findings obviously indicate that denervation of muscle fibres occur after contusion muscle injury in this model of ours. Following a nerve injury the membrane characteristics of denervated muscle fibres change. 'The changes are the result of active protein synthesis (Grampp et ul. 1972, Thesleff & Ward 1973) and tahe some time to develop. T h e resting membrane potential is reduced (Redfern & Thesleff 1971) and the membrane potential starts to oscillate spontaneously, as a result of which spontaneous action potentials, called fibrillation potentials and positive sharp waves--rhythmical or irregular-an be found (Purves & Sakman 1974). T h e first electrophysiological signs of the injury in the form of increased insertional activity were observed 5 days after the injury. Fibrillation potentials and positive sharp waves appeared both proximal and distal to the injury 7 and 10 days following the injury. Development of fibrillation potentials and positive sharp waves in the injured muscles corresponds well with

Electromyography in muscle injury

I-

453

+

,Oms

Fig. 8. Muscle rupture 10 days following the injury. Fibrillation potentials and positive sharp waves in the proximal part of the gastrocnemius muscle (low pass filter 5 kHz, high pass filter

20 Hz).

2. Latency (ms) of late components following the main motor action potential (MAP). The latency of the first/last component are given.

Table

Days after trauma 2 5 7 10 14 21 28 35 42 56

Proximal site

Distal site

Rat 1

Rat 2

Rat 1

Rat 2

-1-/-/-

-1-/-/-

-1-/-I-/-

-I-

12/21 12/29 10116 10143 11/22 8/18 9/18

9/20 Nolobs 8/20 10139 11/17 11/17 12/35

15/23 12/24 16/45 11/31 8/52 8/23

-/-

-/-/-

Nolobs 10120 11/18 9/19 101122 7/35

previous observations on the appearance of fibrillation potentials and positive sharp waves following nerve injury : Smith & Thesleff (1976) observed the first fibrillation potentials and positive sharp waves in the diaphragm 3 days following crush of the phrenic nerve and the number of fibrillation potentials and positive sharp waves reached maximum 10 days after the injury. Our previous study demonstrated that the wedge-shaped edge of the hammer ruptures the muscle fibres reasonable sharply in one plane only (Hurme et al. 1991). Since the main nerve trunks advance distally along the deep surface of

the gastrocnemius muscle fascia and the depth of the contusion is only about one third to one half of the thickness of gastrocnemius muscle, these main trunks are hardly disrupted by the injury. Besides, some nerves within the site of injury seemed to withstand the mechanical force they had been exposed to surprisingly well (Fig. 6B, C). On the other hand, most intramuscular branches run towards the muscle surface, i.e. more or less parallel to the direction of the strike of the hammer. Finally NMJs in gastrocnemius muscle exist in several oblique oriented planes proximally and distally to the contusion injury. Hence, only a limited number of intramuscular nerve branches and NMJs are directly damaged by the hammer within the site of injury. Only nerve branches that innervate the NMJ at or distal to the injury are likely to be affected. In conclusion, it is likely that fairly few muscle fibres are denervated neurogenically because of direct damage to nerve fibres (Fig. 6). The main denervation is obviously myogenic, i.e. caused by the rupture and necrosis of the myofibres. In mammalian limb muscles there is only one NMJ in each fibre. An injury that cuts the muscle fibre in two parts will leave one part, the adjunctional stump, innervated with the NMJ and the other part, the abjunctional stump, without innervation. Functionally the abjunctional part of the muscle fibre probably behaves as if it would have been denervated, and gives rise to denervation activity. Besides, if the NMJ is close to the injury the necrosis of the fibre may extend

4.54

T . Hzirmr et al.

.-.--. . .-. . . -. . . -.~,....... ?

beyond the level of the NMJ in which case both stumps of the ruptured fibre are denervated (Figs. .3B & 7.4). 'I'hc fibrillation potentials and positive sharp

uat-es had disappeared in the proximal part of the muscle by day 14 and in the distal part by day 21. T h e disappearance of fibrillation potentials and positive sharp waves indicate that

Electromyography in muscle injury muscle fibres have become reinnervated. In nerve lesions following crush of the nerve in mouse just proximal to its entrance into the muscle, signs of reinnervation are seen 9 days following the injury (Sellin et al. 1980). Following the proximal crush injury of the rat sciatic nerve, signs of muscle reinnervation have been seen 14 days following the injury (Warsawski et al. 1975). In a proximal injury the nerve must grow from the lesion to the muscle, which explains why in our muscle injury the first signs of reinnervation are seen somewhat earlier. Reinnervation was also followed in our experiment by late components of MAP. In an intact muscle that MAP has a triphasic shape without late components. A few late components following the MAP were seen in the proximal part of the muscle 10 days after the injury. On day 14 late components were seen in the proximal and distal part of the muscle. The reinnervation of the denervated and regenerated muscle fibres may occur in several ways depending on the way the axon or the muscle fibre has been injured. If the adjunctional fibre stump loses its innervation because of the myonecrosis (fibre 2 in Fig. 3 B), it is restored by the regeneration of the muscle fibre underneath the NMJ, which is preserved as shown by AChE staining (Figs. 7A, B) and which thus should have retained its neural connection (Burden et al. 1979). If axons or their terminals were transsected (neurogenic lesion, Fig. 3 C) the reinnervation takes place as ' normally ' after axotomy. It has been demonstrated that the regenerating axon terminals, either form the original or collateral axons, usually reoccupy the old deserted NMJs which remain viable (Letinsky et al. 1976, McMahan et al. 1978). In our model this applies to the adjunctional stumps, whereas in the abjunctional stumps new NMJs must be formed somewhere along the muscle fibre to be innervated by axons nearby. One possibility is that the abjunctional parts of the muscle fibres fuse with innervated adjunctional stumps. Still another possibility is the formation of myomuscular junctions and it could be speculated that ephaptic myomuscular transmission across such junction is possible. Ultrastructural evidence for such myomuscular junctions in our model has not yet been found, but neurophysiologically existence of such ephaptic transmission is suggested by the complex repetitive discharges that are seen in several

455

neuromuscular disorders (Trontelj & Stilberg 1983). Whether this ephaptic transmission could be functionally significant is not known. The width of the injury was about 2 mm which might explain the slight delay in the disappearance of fibrillations and positive sharp waves from the distal part of the injury. The newly formed terminal axon branches to endplates conduct slowly, which gave rise to a temporal dispersion of the MAP. The latency of the late components was around 40-45 ms up to day 28. After this there was a slight increase in the maximum latency, even up to 122 ms. The slight increase in the maximum latency of the late components following the MAP 42 days after the injury may be due to innervation of the newly formed muscle fibres. The increase in temporal dispersion of the MAP at this stage may be due to an additional delay in the conduction along the thin muscle fibres. The reorganization and regeneration of the injury were not complete even at 56 days following the injury. There were still numerous late components and the organization of the regenerating muscle fibres were complex and interlaced. Remodelling and resorption of the fibrous scar has also not completely occurred and thickened endomysial and perimysial structures can be observed within the site of injury. Morphological observations are in concordance with electromyographic findings, indicating that neurophysiological examinations may be useful in the diagnosis and follow-up of muscle injuries. The technical assistance of Mss Liisa Lempiainen, Paula Merilahti and Toini Vieno, and the skilled photography by Jaakko Liippo are gratefully acknowledged. This study was supported by grant from Research Council for Physical Education and Sport, Ministry of Education, Finland.

REFERENCES BURDEN, S.J., SARGENT, P.B. & MCMAHAN, L.J. 1979. Acetylcholine receptors in regenerating muscle accumulate at original synaptic sites in the absence of the nerve. 3 Cell Bid 82, 412415. CARLSON, B.M. & FAULKNER, J.A. 1983. The regeneration of skeletal muscle fibers following injurj-: a review. Med Sci Sports Exert 15, 187-198. GRAMPP,W., HARIS,J.B. & THESLEFF, S. 1972. Inhibition of denervation changes in skeletal muscle by blockers of protein synthesis. f Physiol 221, 743-754.

456

T. Hurme et al.

HLRME, T., KALIMO, H., LEHTO, M. & JARVINEN, M. 1991. Healing of skeletal muscle injury. An ultrastructural and immunohistochemical study. .&lea' Sci Sports Exerc 23 ( 7 ) . JARVINEN, M.& SORVARI, T. 1975. IIealing of a crush injury in rat striated muscle 1. Description and testing a new method of inducing a standard injury to the calf muscles. .'lcta Path Microbiol Scand Sect A 83, 259-265. JARVIXEN, M . 1975. Healing of a crush injury in rat striated muscle 2. %. histological study of the effect of early mobilization and immobilization on the repair process. Acto Path itlicrobiol Scand Srcr A 83, 269-282. KARNOVSKY, M.J. & ROOT, L. 1964. A 'directcoloring ' thiocholine method for cholinesterases. f Ifisrochem Cjjtochem 12, 219-221. I A H T O , M., DUANE, V.C. & RESTALL, D. 1985. Collagen and fibronectin in the healing skeletal muscle injury. An immunohistochemical study of the effects of physical activity in the repair of the injured gastrocnemius muscle in the rat. f Bone j'oint Surg 66B, 820-828. , $1. & NELIMARKKA, 0. 1986. Scar formation after skeletal muscle injury. A histological and autoradiographical study in rats. .-lrch Orthop Trauma Surg 104, 366370. LETISSKY, hl.K., FISCHBECK, K.H. & MCMAHAS, C.J. 1976. Precision of reinnervation of original postsynaptic sites in muscle after a nerve crush. 3 ,Veurocytol 5, 691-718. bfC%lAf*.4S, C.J., SANES,J.R. & AMARSH.4LL,L.M.

1978. Cholinesterase is isolated with the basal lamina at the neuromuscular junction. Nature 271, 172-174. D. & SAKMAN, B. 1974. Membrane properties PURVES, underlying spontaneous activity of denervated muscle fibers. 3 Physiol 239, 125-153. REDFERN, P. & THESLEFF, S. 1971. Action potential generation in denervated rat skeletal muscle. I1 T h e action of tetrodotoxin. ,4cta Physiol Scand 82, 70-78. SCHICK, G. & JERUSALEM, F. 1973. Ultrastrukture Befunden in der fruchen Regenerationphase des denervierten Rattenmuskels. Beitr Pathol 148, 127-140. SELLIN, L.C., LIBELIUS, R., LUNDQUIST, I., T~GERUD, S. & THESLEFF, S. 1980. Membrane and biochemical alterations after denervation and during reinnervation of mouse skeletal muscle. Acta Physiol Scand 110, 181-186. SMITH, J.V.& THESLEFF, S. 1976. Spontaneous activity in denervated mouse diaphragm muscle. f Ph,ysiol 257, 171-186. THESLEFF, S. & WARD,M.R. 1975. Studies on the mechanism of fibrillation potentials in denervated muscle. 3 Physiol 244, 313-323. TRONTELJ, J. & STALBERG, E. 1983. Bizarre repetitive discharges recorded with single fiber EMG. f Nrurol Neurosurg Ps,ychiatry 46, 3 10-3 16. WARSAWSKI, M., TELERMAN-TOPPET, N., DURDU, J., GRAF,A. & COERS,C. 1975. The early stages of neuromuscular regeneration after crushing the sciatic nerve in the rat. J Neurol Sci 24, 21-32.