Journal of the Neurological Sciences, 1990, 98:21-36 Elsevier

21

JNS 03352

Expression of various isoforms of neural cell adhesive molecules and their highly polysialylated counterparts in diseased human muscles D. Figarella-Branger ~, J. Nedelec 2, j. F. Pellissier 1, j. Boucraut ~, N. Bianco ~ and G. Rougon 2 1Laboratoire d'Anatomie Pathologique et de Neuropathologie, Facult~ de M~decine, Marseille (France) and 2URA 202 CNRS, Institut de Chimie Biologique, Marseille (France) (Received 5 December, 1989) (Revised, received 12 March, 1990) (Accepted 12 March, 1990)

SUMMARY

Antibodies directed against neural cell adhesive molecules (NCAM) and in particular a monoclonal antibody recognizing polysialylated isoforms, were used to characterize the expression of these molecules in normal and diseased human muscles. Normal subjects as well as patients with inflammatory, dystrophic and denervating diseases were examined. By immunohistochemistry the main observations were (1) satellite cells expressed the non-sialylated form of NCAMs, (2)regenerative fibers strikingly expressed NCAMs and their sialylated isoforms both on membranes and in the cytoplasm; (3) in denervated muscles, fibers in atrophic groups and some fibers in acute denervation expressed NCAMs on their membrane but not the highly sialylated form; (4) finally, some fibers in myotonic dystrophy and fibers with rimmed vacuoles also expressed NCAMs. Biochemical approaches, using enzymes such as endoglycosidase N and phosphatidylinositol phospholipase C combined with immunoblot analysis allowed visualization of the nature of the expressed isoforms. We have shown that non activated cells, i.e. satellite cells and denervated fibers do not express polysialylated NCAMs. This post-translational modification may be only observed in activated or regenerating fibers. This would parallel the sequence of NCAM expression occurring in normal myogenic pathways.

Correspondence to: Dr. G. Rougon, URA 202 CNRS, Institut de Chimie Biologique, 3 Place Victor Hugo, F-13003 Marseille Cedex, France. 0022-510X/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

22 Key words: Neural cell adhesive molecules; Glycosylation; Sialylation; Diseased human muscle; Development

INTRODUCTION Neural cell adhesive molecules (NCAMs) are known to play a pivotal role in regulating cell-cell interactions in the brain and other tissues (for review, see Edelman 1986a,b; Walsh 1988). They have been implicated in a wide range of morphogenetic and developmental events and, in particular, muscle innervation and synaptogenesis (Rutishauser et al. 1983; Covault et al. 1986; Tosney et al. 1986; Rieger et al. 1988). During these events the adhesive properties and function of NCAMs are modulated by the differential expression of various NCAM polypeptides. Indeed, NCAMs have an unusually large number of features which make it possible to modulate their expression and function at a variety of levels. The NCAM gene is a complex transcriptional unit and is a member of the immunoglobulin superfamily of recognition molecules (Walsh 1988). During myogenesis, developmentally regulated alterations in NCAM RNA splicing have been observed (Covault et al. 1986). It is clear that in human muscle, as it had already been shown in mouse and chick, these mechanisms probably contribute towards the generation ofNCAM polypeptide diversity. Besides alternative splicing and differential polyadenylation site selection, NCAMs are subject to a variety of post-translational modifications including glycosylation (Rougon et al. 1982; Finne et al. 1983), phosphorylation (Gennarini et al. 1984), and sulfation (Hoffman et al. 1982). In striated muscle, NCAM related polypeptide chains of 155, 145 and 125 kDa have been found. Recent studies have indicated that the 155 and 125 kDa isoforms found in mouse myotube cultures are released by phosphatidylinositol phospholipase C, whereas the NCAM 145 kDa isoform present in myoblasts is phosphorylated and resistant to the action of the enzyme (Moore et al. 1987). Thus, these forms are likely to be nontransmembrane and transmembrane forms, respectively. Data obtained by Rieger et al. (1985) on mouse diaphragm muscle showed that in immunoblot experiments NCAMs from developing embryos migrated as a polydisperse (140-200 kDa) form. This:heterogeneity, as in embryonic neural tissues, is confirmed largely by the presence of sialic acid. This had been confirmed by Moore's data (Moore et al. 1987) which showed that NCAMs expressed by C2 and C8 mouse muscle cell lines were sensitive to neuraminidase treatment. In vivo NCAMs are widely distributed on developing myotubes, but disappear as development proceeds. In adult muscle fibers, NCAMs are concentrated at the neuromuscular junction (NMJ) and on muscle satellite cells (Covault et al. 1986; Rieger et al. 1985; Moore and Walsh 1985). Following denervation or pharmacologically induced paralysis, NCAMs are reexpressed extrasynapticaUy (Covault et al. 1986; Moore and Walsh 1985). Thus, it seems that the levels of NCAMs parallel the susceptibility of muscle to innervation: it is high in embryonic, denervated, and paralysed muscles, and low in normal and reinnervated adult muscle.

23 We have used antibodies recognizing different NCAM isoforms to study the cellular distribution and molecular forms of NCAMs in a series of 45 muscle biopsies from patients with a variety of neuromuscular syndromes. The aim of the present study was to examine the expression of the post-translational modification glycosylation of NCAMs and especially to search whether high sialylation was expressed in particular diseases. If so then its detection could be used to further improve the diagnosis of degenerative or regenerative syndromes that are not yet well characterized.

MATERIALS AND METHODS

Materials Muscle biopsies were selected from patients with a variety of clinical neuromuscular syndromes. Muscle specimens were flash-frozen at - 1 5 0 ° C in isopentane chilled by liquid nitrogen and stored at - 80 ° C. The diagnoses of patients were as follows: normal muscle control showing no microscopic abnormality, 6 (adults 4, infants 2); muscle from patients with various inflammatory myopathies, 12 (sclerodermia 1, polymyositis 2, adult dermatomyositis 4, childhood dermatomyositis 2, inclusion body myositis3); muscle from patients with muscular dystrophies, 12 (Duchenne dystrophy 2, facioscapulohumeral dystrophy 2, myotonic dystrophy 4 and oculopharyngeal muscular dystrophy 4); muscle with evidence of denervation, 13 (amyotrophic lateral sclerosis (ALS) with acute onset 3, chronic spinal muscular atrophy3, various chronic neuropathies 3, post poliomyelitic syndrome 1, and Werdnig-Hoffman disease 3). Two muscle specimens from patients with acute rhabdomyolysis were also tested. The diagnosis of muscle disease was established on the basis of the usual criteria pertaining to the clinical picture as well as the appropriate laboratory data, including electrophysiological studies, serum creatine kinase activity and microscopic evaluation of muscle biopsies. Reagents Two different anti-NCAM antibodies were used in the present study. These were an anti-NCAM site directed polyclonal antibody recognizing the NHz-terminal domain. This sequence had been shown to be shared by all NCAM isoforms (Rougon and Marshak 1986). The second was a mouse monoclonal antibody (anti-MenB) recognizing high polymeric forms of alpha 2.8 linked sialic acid expressed by "embryonic" forms of NCAMs (Rougon et al. 1986). Occasionally the mouse monoclonal HNK-1 antibody was also used and was produced from a hybridoma cell line obtained from the ATCC (Rockville, MD, U.S.A.). Antibodies produced in ascites fluids were used at a dilution of 1 : 2000 for HNK-1 and anti-MenB, and 1 : 5000 for the site directed polyclonal antibody. Phosphatidylinositol phospholipase C was purified in our laboratory from the Bacterium thuringiensis. Endoneuraminidase N was prepared as the soluble virion free form from bacteriophage KIF and was a kind gift from G. Alcaraz (CIML, Marseille).

24 Muscle extracts and immunoblot detection Muscle extracts were prepared as previously described (Rougon et al. 1982). Briefly, tissues were homogenized in 10 mM Tris-HC1, pH 7.8, 2 ~o Nonidet P-40, 1 mM EDTA, and 10 #M leupeptin, 1/~M alpha2-macroglobulin, 0.1 mM phenylmethylsulfonylchloride as protease inhibitors. The homogenates were centrifuged at 100000 × g for 1 h at 4 °C; the supernatant was collected and protein concentration determined. The protein concentration was adjusted to 10 mg/ml and samples were boiled for 3 min in reducing electrophoresis sample buffer. Aliquots containing equivalent amounts of proteins were then resolved on 7.5 °.o SDS-polyacrylamide gels, transferred to nitrocellulose sheats (NitroScren West NEN) and sequentially reacted with primary antibody at the dilution mentioned above, then in the case of anti-MenB antibody, with anti-isotype antibody rabbit monoclonal antimouse IgM at 1/~g/ml and finally with 12~I-labelled protein A (0.5 × 106/ml). Polyclonal anti-NCAM was directly revealed by incubation with protein A. Bound antibody was visualized by autoradiography. Enzymatic treatments For phosphatidylinositol phospholipase C treatment, tissue homogenates were prepared without detergent at a concentration of 10 mg/ml. They were incubated with or without 0.05 IU/ml enzyme for 1 h at 37 °C then soluble and particulate materials were separated by centrifugation (100000 × g, 1 h, 4 °C) as described by Jean et al. (1988). Treatment with endoneuraminidase N was conducted on homogenates in the presence of 2 ~o Nonidet P40 with 0.2 unit/rag protein for 1 h at 20 ° C (Vimr et al. 1984). In some instances some tissue sections were incubated for 3 h at 20 °C with endoneuraminidase 0.5 unit/ml in phosphate-buffered saline (PBS). Immunocytochemical procedure Five serial 4-#m thick cryostat sections were prepared from each block of tissue. The serial sections were stained with, or reacted for, the following: (1) hematoxylin-eosin (HE), (2) anti-NCAM site directed polyclonal, (3) anti-MenB monoclonal, and (4 and 5) controls. In 12 cases the anti-HNK1 antibody was also applied. Immunoperoxidase staining was performed using avidin biotin peroxidase complex (ABC kits Vector laboratories Inc., Institut Pasteur, Paris, France) as previously described (Hsu et al. 1981). Peroxidase was localized using AEC (3-aminoethylcarbazole) substrate. Control incubations included substitution of PB S for primary antibody and an irrelevant mouse IgM. The expression of the immunoreactive proteins was correlated with pathological changes in muscle fibers as revealed by HE.

25 RESULTS

Cellular distribution of NCAMs and theirpolysialylated counterparts in normal and diseased muscles Normal muscle W h e n b i o p s i e s f r o m n o r m a l subjects (adult or i n f a n t ) were e x a m i n e d , different types o f N C A M positive cells c o u l d be identified. T h e y were (1) satellite cells (Fig. lc), (2) s o m e u n i d e n t i f i e d m o n o n u c l e a t e cells w h i c h a c c o r d i n g to their m o r p h o l o g y were

Fig. 1. NCAM and its polysialylated isoforms in normal adult muscle. Motor end-plate area stained with anti-NCAM (a) and anti-MenB (b); both pre- and post-synaptic structures are labelled with anti-NCAM ( x 235). (c) Satellite cell positive with anti-NCAM ( x 630). (d + e) section of an intramuscular nerve. NCAM is expressed on Sehwann cells or amyelinated fibers. Some of them are also positive with anti-MenB (e). Note that large myelinated axons are not labelled ( x 235).

26 likely to be fibroblasts or pericytes, (3) s o m e structures in i n t r a m u s c u l a r nerve such as u n m y e l i n a t e d a x o n s or S c h w a n n cells ( m y e l i n a t e d a x o n s were never stained) (Fig. l d), (4) n e r v e t e r m i n a l s a n d the i m m e d i a t e p r e t e r m i n a l p o r t i o n o f axons or t e r m i n a l s associated with S c h w a n n cells a n d p o s t - s y n a p t i c areas o f the N M J (Fig. la). O n l y studies p e r f o r m e d at the electron m i c r o s c o p i c level allow o n e to define the n a t u r e of the

Fig. 2. NCAM and its polysialylated isoforms in inflammatory myopathies. Polymyositis with numerous regenerative fibers labelled with anti-NCAM both on their membrane and intraeellularly (a). 45% are also reactive with anti-MenB (b) (x 155). Perifaseicular fibers in adult dermatomyositis are positive with anti-NCAM (c) and 80% with anti-MenB (d)( x 65). Fibers with rimmed vacuoles in inclusion body myositis are strongly positive with both anti-NCAM (e) and anti-MenB (f). Note that the vacuole is itself strongly labelled (arrow) ( x 235). All biopsies shown are from adult patients.

27 labelled compartment. Mature myofibers were never labelled with anti-NCAM antibody. With respect to anti-MenB reactivity, only some Schwann cells or unmyelinated axons were stained (Fig. le), satellite cells were always negative. Slightly positive anti-MenB staining was seen in zones corresponding to NMJs (Fig. lb). Interestingly, NCAM-positive staining sometimes associated with anti-MenB was observed on some spindle fibers as well as part of the nerve contacting them.

Diseased muscles These were classified according to clinical diagnosis as inflammatory,dystrophic and denervated muscles. Inflammatory myopathies. The most obvious feature concerning inflammatory myopathies was that, when regenerative fibers were observed on the basis of HE staining, all of them strongly expressed NCAM, with both a cytoplasmic and membranous localization. In contrast, necrotic or degenerative fibers did not. In all cases in which HNK1 antibody staining was performed, regenerative fibers never stained positive. When corresponding serial sections were probed for polysialic units expression, positive staining was detected, although around 50~o cells were positive when compared with NCAM staining. Like with NCAM, cytoplasmic and membranous staining was strong as shown in Fig. 2a,b for polymyositis. It is noteworthy that all anti-MenB positive cells were also NCAM-positive. This agrees well with the fact that the molecules bearing polysialosyl units are likely to be NCAMs. At early stages of regeneration both antibody staining were also detected intracellularly with a punctate distribution. In dermatomyositis, when perifascicular atrophy occurred, the perifascicular fibers were strongly labelled with anti-NCAM antibody and 80~ of them were also reactive with anti-MenB antibody (Fig. 2c,d). The same findings were observed in sclerodermia, although in this case, only few atrophic perifascicular fibers expressed MenB epitopes. In inclusion body myositis, fibers with large rimmed vacuoles were stained with both antibodies. The vacuole was itself strongly labelled (Fig. 2e,f, arrow). In this disease, numerous fibers, especially hypertrophic fibers, also slightly expressed NCAM on their surface. It should also be mentioned that in inflammatory myopathies, some fibers that exhibited no detectable alterations with HE, also weakly expressed NCAM. Dystrophic muscles. Four different pathologies classified in this group were observed, namely Duchenne dystrophy, facioscapulohumeral dystrophy (FSH), myotonic dystrophy and oeulopharyngeal muscular dystrophy (OPMD). All regenerative fibers based on HE staining appeared NCAM-positive; anti-MenB positive cells were also seen, but here again in lower numbers than those expressing NCAM except for Duchenne dystrophy in which polysialylated fibers were particularly numerous. Young regenerative fibers were the most positive fibers with a cytoplasmic, punctate distribution of both NCAM and sialylated polymers. This feature was striking in biopsies of patients with Duchenne dystrophy where groups of regenerating fibers

28

Fig. 3. NCAM and its polysialylated isoforms in dystrophic muscle. Numerous fibers with small diameter are anti-NCAM-positive (a) and to a lesser extent anti-MenB-positive (b) (arrows) in myotonic dystrophy ( x 325). (c + d) Sarcoplasmic masses (arrows) are strongly stained with both antibodies in myotonic dystrophy ( x 325). Groups of regenerative fibers in Duchenne's dystrophy at early stages are strongly positive with both anti-NCAM (e) and anti-MenB (f). Necrotic areas (stars) containing numerous macrophages are visible (× 325). Regenerative fibers in Duchenne's dystrophy observed at a latter stage of regeneration are anti-NCAM (g) and anti-MenB (h) positive (arrows). Note the punctate staining in the cytoplasm ( x 200).

29 occurred (Fig. 3e-h). In this disease large hypercontracted fibers never expressed NCAMs. In every case of muscle dystrophy observed, some segmented fibers were stained with anti-NCAM antibody. The intensity of labelling was often stronger at the level of segmentation. Sometimes only a portion of the segmented fiber was stained as if segregation of N C A M expression occurred. When small polygonal or rounded fibers were observed they were strongly NCAM positive but very few expressed the polysialylated form. The 2 biopsies from patients with F S H dystrophies showed severe myopathic changes. In addition to regenerative and small rounded fibers, numerous fibroblasts which also expressed N C A M were observed. In patients with oculopharyngeal muscular dystrophies, fibers showing linned vacuoles also expressed N C A M and, to a lesser extent, its polysialylated form. An interesting feature of muscles from myotonic dystrophic patients was the expression of NCAM and often sialylated polymers at the surface of many fibers with central nuclei. Staining was found predominantly in type I fibers although it was not limited exclusively to these structures (Fig. 3a,b). In addition, sarcoplasmic masses also strongly expressed both N C A M and sialylated forms (Fig. 3c,d). Denervated muscles. Groups of atrophic fibers and angular isolated fibers were observed in all adult patients. In addition, pathological changes of acute denervation were seen in three cases of ALS, i.e., necrosis, cytoplasmic bodies, numerous target fibers and no type grouping. In these cases, muscle biopsies were performed less than 6 months after the onset of the disease. In contrast, type grouping was obvious in muscle samples from patients with chronic spinal muscular atrophy, chronic neuropathies and post poliomyelitic syndrome. In all these cases muscle biopsy was performed 3 or more years after the onset of the disease. In all denervated muscles, most of fibers in atrophy groups were positive for N C A M but not for the anti-MenB antibody, with the exception of one patient with ALS in which highly polysialylated NCAMs were also observed. The distribution of staining differed from that observed in regenerative fibers in that it was more concentrated on the membrane and occasionally weak and diffuse in the cytoplasm (Fig. 4a,b). Angular isolated fibers were either negative or positive to both antibodies or positive only to anti-NCAM. In denervated muscle some hypertrophic or rounded fibers, and some normal fibers by HE also expressed NCAMs on their cytoplasmic membrane. All or only part of the surface membrane was stained. The number of such fibers was variable from case to case. It was high in patients with acute denervation, low in cases with chronic neuropathies and post-poliomyelitic syndrome. But, such fibers were not seen in chronic spinal muscular atrophy. In contrast, target fibers were always NCAM negative. The number of satellite cells increased in adult denervated muscles. In Werdnig-Hoffman disease, most of the round atrophic fibers showed membrane expression of NCAMs and were occasionally anti-MenB positive. Hypertrophic fibers were always negative (Fig. 4e,f). In these cases, the number of satellite cells was not increased. Finally, in 2 cases of rhabdomyolysis, regenerative fibers were seen at different

30

Fig. 4. NCAM and its polysialylated isoforms in denervated muscles (a-f) and rhabdomyolysis (g + h). Large groups of atrophic fibers are strongly labelled with anti-NCAM (a) but not with anti-MenB antibody (b) in a patient with chronic spinal muscular atrophy. NCAM staining is both membranous and diffuse in the cytoplasm ( × 235). Dark aspect in (b) is due to hemalun counterstain of the nuclei. Acute denervation in an adult patient in which numerous fibers showed membranous stain with anti-NCAM (c) but not with anti-MenB (d). The fiber (arrow), positive with both antibodies, is very likely a regenerative fiber ( x 325). Small round fibers positive with anti-NCAM (e) in a biopsy from a patient with Werdnig-Hoffmann disease.

31 stages o f maturation. All regenerative fibers were stained with a n t i - N C A M and among these approximately 50 % were also anti-MenB positive (Fig. 4g,h). When such sections were incubated with endoneuraminidase prior the addition of anti-MenB antibody, all the immunoreactivity was lost, whereas the sections were still immunoreactive with a n t i - N C A M antibody (not shown).

Biochemical analysis of NCAM isoforms In order to analyse the various isoforms o f N C A M expressed we prepared muscle extracts from biopsy samples. The homogenates were probed using anti-MenB and a n t i - N C A M antibodies. Whatever the nature o f the sample examined, anti-MenB immunoreactivity was always undetectable. This was due to a decreased sensitivity of immunoblot detection when conducted with a monoclonal antibody. To ascertain the presence of polysialosyl units, we tested the effect o f endosialidase k n o w n to remove polysialosyl units on the mobility of NCAM-positive molecules on polyacrylamide gels (Vimr et al. 1984). This

ENDO_F -+

-

PI.PLC

I

+1

-

+

, 200

I 4200

,116

16

A

PSNP B

Fig. 5. Immunoblot analysis of NCAM immunoreactivity probed with polyclonal antibody. (A) Samples from patients exhibiting polymyositis lanes 1-2 and inclusion body myositis lanes 3-4, respectively, were treated (lanes 2 and 4), or not (lanes 1 and 3), with endosialidase before gel electrophorcsis. Note that enzyme treatment which digests polysialosyl-units, influences NCAM migration profile. (B) Polymyositis patient was treated (lanes 3-4) or not (lanes 1-2) with enzyme phosphatidylinositol phospholipase C (PI-PLC) then NCAM immunoreactivity was examined in particulate (lanes 1-3) and soluble fractions (lanes 2-4). Note that PI-PLC solubiliz¢ a part of NCAM (lane 4) previously associated with membrane. • 52% are also anti-MenB-positive, among these 14% are strongly labelled (f). Staining is essentially located to the cytoplasmic membrane. Hypertrophic fibers (stars) are always negative. (x 235). (g + h) Rhabdomyolysis in which fibers at different stages of regeneration are observed. Fibers strongly positive with both anti-NCAM and anti-MenB antibodies (arrows) as welt as 45% fibers positive only with anti-NCAM (stars) are seen. We postulate that when regeneration evolved fibers loose their expression of polysialosyl-units NCAM molecules ( x 155).

32 1

234567

200,--

116

,--

92.5 ,.-66.2"-

45 ,,-!Q

~ ii!!il¸i i ¸¸¸~il~! i!!ilii!i!!!iiiiii~:iiii~i!i!i!~ii ii

Fig. 6. Immunoblot analysis of anti-NCAM polyclonal immunoreactive isoforms in diseased and normal human muscle. Lane 1, polymyositis; lanes 2-3, inclusion body myositis; lane4, denervation; lane 5, Duchenne's muscular dystrophy; lane 6, normal adult; lane 7, normal infant patients. Equal quantities of proteins were run in each lane. Positions of molecular weight markers are shown.

is shown in Fig. 5A for 2 patients exhibiting polymyositis (lanes 1 and 2) and inclusion body myositis (lanes 3 and 4), respectively. After endosialidase treatment, the NCAM immunoreactive product was a major desialo form migrating with an apparent molecular mass of approximately 140 kDa. NCAM polyclonal antibody (Fig. 6) revealed heterodisperse material migrating between 120 and 200 kDa (lanes 1-3). In some other cases, corresponding to denervation and Duchenne's dystrophy (lanes 4 and 5), the bands were more discrete and corresponded to an absence or a very low level of polysialylation. Lanes 6 and 7 correspond to the biopsies of normal adult and young subjects, respectively. NCAMs were undetectable in lane 6 in which no intramuscular nerve was observed. The heavy band at 55 kDa revealed in lane 3, represents heavy chains of immunoglobulin because they could be revealed by treatment with protein A alone. Generally, the intensity of the bands observed on immunoblots corresponded to the intensity of staining observed on sections by immunohistochemical treatment. Recent studies have indicated that the NCAM 155 and NCAM 125 kDa isoforms found in mouse myotube cultures are released by the bacterial enzyme phosphatidylinositol phospholipaseC, whereas the NCAM 145kDa isoform present in myoblasts is phosphorylated and resistant to this treatment (Moore et al. 1987). Using phosphatidylinositol phospholipase C we searched for the expression of phosphatidylinositol-sensitive forms in diseased human muscles. This is shown in Fig. 5B for a patient suffering from polymyositis in whom regenerative fibers were observed. Enzyme treatment liberated a soluble form which represented about 20 ~o of the total immunoreactive NCAMs. It migrated as a large band between 130 and 140 kDa. This migratory pattern may be due to the polysialylation of the liberated band(s).

33 DISCUSSION The present studies were performed to gain some insight into the regulation and description of various isoforms of NCAM expressed in both normal and diseased human muscle. Data published by Rutishauser et al. (1988) demonstrate that the presence of NCAM on a cell surface can have either a positive or negative effect on its overall interaction with other cells or substrates depending on the molecules' polysialic content. It has been previously described that NCAMs were reexpressed in some human muscle diseases and, in particular, that they were found associated with regenerating fibers (Walsh and Moore 1985; Cashman et al. 1987). However, no precise data on the nature of the isoforms or state of glycosylation have been reported. Because NCAMs with relatively low polysialic content were found predominantly in early development and adult brain tissues (Rothbard et al. 1982) we first examined satellite cells in normal muscle. It is clear from our data that in normal muscle, either from adults or children, the NCAM forms expressed by satellite cells were always anti-MenB negative. Thus there may be a parallel with the development of the nervous system in which precursor cells from neuroepithelium do not express polysialic acid until the neuronal phenotype is acquired (Rothbard et al. 1982; Rougon, G., unpublished data). It is tempting to speculate that there are signaling factor(s) activating satellite cells to proliferate and to differentiate along the myogenic pathway which would concomitantly turn on the expression of polysialic units by NCAMs. Reexpression of NCAMs was always observed in diseases in which regenerative fibers were identified by HE staining. This is in agreement with data from either animal experimental models (Sanes et al. 1986) or human biopsies (Schubert et al. 1989; Cashman et al. 1987; Lanier et al. 1989). Indeed, data from Schubert et al. (1989) using anti-Leul9 antibody described as recognizing a HNK-1 positive structure on lymphocyte killer cells are, in fact, related to NCAM as the 2 antigens have been recently shown to be similar (Lanier et al. 1989). However, we never observed any staining with anti HNK-1 antibody in regenerative fibers. Our observations allow one to distinguish between 2 populations among regenerating fibers: one was positive only with NCAM, and the other positive with anti-MenB and anti-NCAM antibodies. We propose that the expression of polysialosyl units is transient and parallel to normal muscle development, these units being heavily expressed during the early stages of regeneration. Interestingly, the staining with both anti-NCAM and anti-MenB was not restricted to the sarcolemma in regenerating fibers, as it is generally found in developing myofibers. This very likely reflects active biosynthesis. Further studies using electron microscopy are needed to establish the precise nature of the MenB positive intracellular compartments. Thus, in dermatomyositis, perifascicular fibers expressing both anti-NCAM and anti-MenB immunoreactivity are likely to be regenerative fibers. Contrary to the first observations of Walsh and Moore (1985) we have found the expression of NCAM in denervating diseases. This is in agreement with data reported by Sanes' group (Sanes et al. 1986; Cashman et al. 1987). Staining found on denervated

34 fibers could be easily distinguished from that seen in regenerative fibers in that it was less intense and more confined to the cytoplasmic membrane. In cases in which intracellular staining was seen, it was always rather diffuse in the cytoplasm and not punctated. The expression of the low polysialylated form of NCAM on membranes seems to be one of the best criteria of acute denervation. This hypothesis is further supported by data from the 3 cases of acute ALS in which these fibers were particularly numerous. Angular isolated fibers, which morphologically could be either early denervated fibers or regenerative fibers, were occasionally anti-MenB positive. Based on our observations, we consider this property a characteristic of regeneration rather than denervation. As already reported by several groups (Schubert et al. 1989; Cashman et al. 1987), when reinnervation successfully occurred, as confirmed by observation of type grouping, NCAM expression was lost from the fibers. In Werdnig-Hoffmann disease, almost all small round fibers were positive for NCAM and, in some instances, for anti-MenB antibodies. This raises the question whether to decide if these fibers are "true" denervated fibers or "normal" fetal fibers that have never been innervated, especially for the ones bearing polysialic units. As a further support of this hypothesis, there is no evidence of satellite cell activation in Werdnig-Hoffmann muscle which is in contrast to what is reported from experimental denervation studies (Murray and Robbins 1982). Thus, although our observations are similar concerning NCAM, our conclusions differ from those ofWalsh et al. (1987) who favor an unstable innervation of previously innervated myofibers. In Steinert's disease, NCAM was expressed on the fiber surfaces reminiscent of that observed in denervation cases. Similar findings were previously reported by Walsh et al. (1988). Although membrane activity mediated by nerve appears to repress the expression of the NCAM gene, NCAM expression can be reinduced in skeletal muscle by hypothyroidism (Thompson et al. 1987). Clinical signs of Steinert's disease are often associated with hormonal disorders (Harper 1979), which perhaps may, besides denervation, participate in the abnormal expression of NCAMs on muscle fibers. In the same way, we cannot exclude that other unidentified parameters could play a role on NCAM expression. Such unknown influences might occur in diseases like inclusion body myositis and oculopharyngeal muscular dystrophy. Preliminary biochemical examination of biopsy samples by immunoblot showed that, in most cases studied, NCAMs could be revealed without the need to concentrate the NCAM positive material as was necessary in the Cashman et al. studies (1987). This is generally the case when site directed polyclonal antibody is used. With our experimental conditions anti-MenB immunoreactivity was undetectable. However, the presence of alpha-2-8 linked neuraminic acid units on NCAM molecules was confirmed by pretreatment of samples with endosialidase known to cleave off such structures, followed by comparison of positions of migration of the NCAM positive bands. As described with immunohistochemical observations, the polysialic units appeared to be more abundant in diseases exhibiting active regeneration. The use of another enzyme, phosphatidylinositol phospholipase C, showed that a portion of NCAM was releasal01e as a soluble form. Combining the use of endosialidase and this enzyme is needed to precisely determine the desialylated size of phosphatidylinositol anchored chains. In

35 normal myotubes, 2 of them have been reported with molecular weights of 125 and 155 kDa, respectively (Moore et al. 1987). In conclusion, this study further confLrrns that NCAM is a complex transcriptional unit and that expression of some isoforms may be restricted at some stages of development in human muscles. Our observations lead us to propose that non-activated cells such as satellite cells or denervated cells do not express polysialylated NCAMs. This post-translational modification would only occur in activated or regenerating fibers.

ACKNOWLEDGEMENTS

We thank F. Br6s and F. Fabre for typing the manuscript and C. Guinot for technical assistance. This work was supported in part by grants from ARC No. 6895 and AFM (association fran~aise contre les myopathies) to G. Rougon.

REFERENCES Cashman, N., J. Covault, R. Wollman and J. Sanes (1987) Neural cell adhesion molecule in normal, denervated and myopathic human muscle. Ann. Neurol., 21: 481-489. Covault, J., J.P. Merlie, C. Goridis and J.R. Sanes (1986) Molecular forms of N.CAM and its RNA in developing and denervated skeletal muscle. J. Cell. Biol., 102: 731-739. Edelman, G.M. (1986a) Cell adhesion molecules in neural histogenesis. Aanu. Rev. Physiol., 48:417-430. Edelman, G. M. (1986b) Cell adhesion molecules in the regulation of animal form and tissue pattern. Annu. Rev. Cell. Biol., 2: 81-116. Finne, J., U. Finne, H. Deagostini-Bazin and C. Goridis (1983) Occurrence of alpha 2.8 linked polysialosyl units in a neural cell adhesion molecule. Biochem. Biophys. Res. Commun., 112: 482-487. Harper, P. S. (1979) Endocrine abnormalities in myotonic dystrophy. In: Myotonic Dystrophy, vol. 9, Major Problems in Neurology, Saunders Co., Philadelphia, PA, pp. 115-138. Hoffman, S., B. Sarkin, P. White, R. Brakenbury, R. Mailhammer, U. Rutishauser, B. Cunningham and G.M. Edelman (1982) Chemical characterization of a neural cell adhesion molecule purified from embryonic brain membranes. J. Biol. Chem., 257: 7720-7729. Hsu, S.M., L. Raine, and H. Fanger (1981) Use of avidin biotin peroxidase complex (ABC) in immunoperoxidase techniques. A comparison between ABC and unlabelled antibody (PAP) procedures. J. Histochem. Cytochem., 29: 577-580. Jean, F., P. Malapert, (3. Rougon and J. Barbet (1988) Cell membrane but not circulating carcinoembryonic antigen is linked to a phosphatidylinositol containing hydrophobic domain. Biochem. Biophys. Res. Comm., 155: 794-800. Lanier, L., R. Testi, J. Bindl and J. Phillips (1989) Identity of Leu-19 leucocyte differentiation antigen and neural cell adhesion molecule. J. Exp. Med., 169: 2233-2238. Moore, S.E. and F. S. Walsh (1985) Specific regulation of N.CAM/D2CAM cell adhesion molecule during skeletal muscle development. EMBO J., 4: 623-630. Moore, S.E., J. Thompson, J. Kirkness, G. Diekson and F.S. Waish (1987) Skeletal muscle neural cell adhesion molecule (N.CAM): changes in protein and in RNA species during myogenis of muscle cell lines. J. Cell Biol., 105: 1377-1386. Murray, M. A. and N. Robbins (1982) Cell proliferation in denervated muscle: identity and origin of dividing cells. Neuroscience, 7: 1823-1833. Rieger, F., M. Grumet and G.M. Edelman (1985) N.CAM at the vertebrate neuromuscular junction. J. Cell Biol., 101: 285-293. Rieger, F., M. Nicolet, M. Pin¢on-Raymond, M. Murawsky, G. Levi and G. M. Edelman (1988) Distribution and role in regeneration of N.CAM in the basal laminae of muscle and Schwann cells. J. Cell Biol., 107: 707-719.

36 Rothbard, J. B., R. Brackenbury, B. Cunningham and G. Edelman (1982) Differences in the carbohydrate structures of neural cell adhesion molecules from adult and embryonic chicken brains. J. Biol. Chem., 257:11064-11069. Rougon, G. and D. Marshak (1986) Structural and immunological characterization of the amino terminal domain of mammalian neural cell adhesion molecules. J. Biol. Chem., 261: 3396-3401. Rougon, G., M. Deagostini-Bazin, M. Hirn and C. Goridis (1982) Tissue and developmental stage specific forms of a neural cell surface antigen linked to differences in glycosylation of a common polypeptide, EMBO J., 1: 1239-1244. Rougon, G., C. Dubois, N. Buckley, J. Magnani and W. Zollinger (1986) A monoclonal antibody against meningococcus group B polysaccharides distinguishes embryonic from adult N.CAM. J. CellBiol., 103: 2429-2437. Rutishauser, U., M. Grumet and G.M. Edelman (1983) N.CAM mediates initial interaction between spinal cord neurons and muscle cells in culture. J. Cell Biol., 97: 145-152. Rutishauser, U., A. Achenson, A. Hall, D. Mann and J. Sunshine (1988) The neural cell adhesion molecule (N.CAM) as a regulator of cell-cell interactions. Science, 240: 53-57. Sanes, J., M. Schachner and J. Covautt (1986) Expression of several adhesive molecules (N.CAM, LI, J 1, NILE, uvomorulin, laminin, fibronectin and a heparan sulfate proteoglycan) in embryonic, adult, and denervated adult skeletal muscle. J. Cell Biol., 102: 420-431. Schubert, W., C. Zimmermann, M. Cramer and A. Sparzinski-Povitz (1989) Lymphocyte antigen Leu-19 as a molecular marker of regeneration in human skeletal muscle. Proc. Natl. Acad. Sci. USA, 86: 307-311. Thompson, J., S. Moore and F. Walsh (1987) Thyroid hormones regulate expression of the neural cell adhesion molecule in adult skeletal muscle. F E B S Lett., 219: 135-138. Tosney, KW., M. Watanabe, L. Landmasser and U. Rutishauser (1986) The distribution of NCAM in the chick hindlimb during axon outgrowth and synaptogenesis. Dev. Biol., 114: 437-452. Vimr, E. R., R. D. McCoy, H. F. Vollger, C. Wilkison and F. A. Troy (1984) Use of prokaryotic derived probes to identify polysialic acid in neonatal neuronal membranes. Proc. Natl. Acad. Sci. USA, 81:1971-1975. Walsh, F.S. (1988) The N.CAM gene is a complex transcriptional unit. Neurochem. Int., 12: 262-267. Walsh, F. and S. Moore (1985) Expression of cell adhesion molecule N.CAM in diseases of adult human skeletal muscle. Neurosci. Lett., 59: 73-78. Walsh, F., S. Moore and B. Lake (1987) Cell adhesion molecule N.CAM is expressed by denervated myofibers in Werdnig-Hoffman and Kugelberg-Welander type spinal muscular atrophies. J. Neurol. Neurosurg. Psychiat., 50: 539-542. Walsh, F. S., S. Moore and J. Dickson (1988) Expression of membrane antigens in myotonic dystrophy. J. Neurol. Psychiat. , 51: 136-138.