Identification of the major proteins that promote neuronal process outgrowth on Schwann cells in vitro.

Schwann cells have a unique role in regulating the growth of axons during regeneration and presumably during development. Here we show that Schwann cells are the best substrate yet identified for promoting process growth in vitro by peripheral motor neurons. To determine the molecular interactions responsible for Schwann cell regulation of axon growth, we have examined the effects of specific antibodies on process growth in vitro, and have identified three glycoproteins that play major roles. These are the Ca2+-independent cell adhesion molecule (CAM), L1/Ng-CAM; the Ca2+-dependent CAM, N-cadherin; and members of the integrin extracellular matrix receptor superfamily. Two other CAMs present on neurons and/or Schwann cells-N-CAM and myelin-associated glycoprotein-do not appear to be important in regulating process growth. Our results imply that neuronal growth cones use integrin-class extracellular matrix receptors and at least two CAMs--N-cadherin and L1/Ng-CAM-for growth on Schwann cells in vitro and establish each of these glycoproteins as a strong candidate for regulating axon growth and guidance in vivo.

Autoradiographic localization of acetylcholine receptors in the Schwann cell membrane of the squid nerve fiber.

Intact and slit nerve fibers of the squid Sepioteuthis sepioidea were incubated in a 50-nM solution of [125I] alpha-bungarotoxin in artificial seawater, in the absence and in the presence of D-tubocurarine (10(-4) M). The distribution of the radioactive label was then determined by electron microscope autoradiography. It was found that, in the fibers exposed solely to the radioactive toxin, the label was located mainly at the axon-Schwann cell boundary in the intact nerve fibers or at the axonal edge of the Schwann cell layer in the axon-free nerve fiber sheaths. Label was also present in those regions of the Schwann cell layer rich in intercellular channels. No signs of radioactivity were observed in the nerve fibers exposed to the labeled toxin in the presence of D-tubocurarine. These results indicate that the acetycholine receptors previously found in the Schwann cell plasma membrane are mainly located over the cell surfaces facing the neighboring axon and the adjacent Schwann cells. These findings represent a further advance in the understanding of the relationship between the axon and its satellite Schwann cell.

Presence of the myelin-associated glycoprotein correlates with alterations in the periodicity of peripheral myelin.

The myelin-associated glycoprotein (MAG) is an integral membrane protein (congruent to 100,000 mol wt) which is a minor component of purified peripheral nervus system (PNS) myelin. In the present study, MAG was localized immunocytochemically in 1-micrometer thick Epon sections of 7-d and adult rat peripheral nerves, and its localization was compared to that of the major structural protein (Po) of PNS myelin. To determine more precisely the localization of MAG, immunostained areas in 1 micrometer sections were traced on electron micrographs of identical areas from adjacently cut thin sections.l MAG was localized in periaxonal membranes. Schmidt-Lantermann incisures, paranodal membranes, and the outer mesaxon of PNS myelin sheaths. Compact regions of PNS myelin did not react with MAG antiserum. The results demonstrate MAG's presence in "'semi-compact" Schwann cell or myelin membranes that have a gap of 12-14 nm between extracellular leaflets and a spacing of 5 nm or more between cytoplasmic leaflets. In compact regions of the myelin sheath which do not contain MAG, the cytoplasmic leaflets are "fused" and form the major dense line, whereas the extracellular leaflets are separated by a 2.0 nm gap appearing as paired minor dense lines. Thus, it is proposed that MAG plays a role in maintaining the periaxonal space, Schmidt-Lantermann incisures, paranodal myelin loops, and outer mesaxon by preventing "complete" compaction of Schwann cell and myelin membranes. The presence of MAG in these locations also suggests that MAG may serve a function in regulating myelination in the PNS.

Immunocytochemical localization of P0 protein in Golgi complex membranes and myelin of developing rat Schwann cells.

P0 protein, the dominant protein in peripheral nervous system myelin, was studied immunocytochemically in both developing and mature Schwann cells. Trigeminal and sciatic nerves from newborn, 7-d, and adult rats were processed for transmission electron microscopy. Alternating 1-micrometer-thick Epon sections were stained with paraphenylenediamine (PD) or with P0 antiserum according to the peroxidase-antiperoxidase method. To localize P0 in Schwann cell cytoplasm and myelin membranes, the distribution of immunostaining observed in 1-micrometer sections was mapped on electron micrographs of identical areas found in adjacent thin sections. The first P0 staining was observed around axons and/or in cytoplasm of Schwann cells that had established a 1:1 relationship with axons. In newborn nerves, staining of newly formed myelin sheaths was detected more readily with P0 antiserum than with PD. Myelin sheaths with as few as three lamellae could be identified with the light microscope. Very thin sheaths often stained less intensely and part of their circumference frequently was unstained. Schmidt-Lanterman clefts found in more mature sheaths also were unstained. As myelination progressed, intensely stained myelin rings became much more numerous and, in adult nerves, all sheaths were intensely and uniformly stained. Particulate P0 staining also was observed in juxtanuclear areas of Schwann cell cytoplasm. It was most prominent during development, then decreased, but still was detected in adult nerves. The cytoplasmic areas stained by P0 antiserum were rich in Golgi complex membranes.

Antibodies to neurofilament, glial filament, and fibroblast intermediate filament proteins bind to different cell types of the nervous system.

Antisera were raised to the 210,000-dalton and the 49,000-dalton proteins of a fraction enriched in intermediate (10 nm) filaments from human brain. Proteins of the filament preparation were separated by SDS-polyacrylamide gel electrophoresis and used for immunization and subsequent analysis of the reactions of the sera by rocket immunoelectrophoresis. Anti-210,000-dalton serum precipitated proteins of molecular weights 210,000, 160,000, and 68,000, and, thus, reacted with all the neurofilament triplet components. Anti-49,000-dalton serum did not react with the triplet proteins but precipitated the 49,000-dalton protein. By immunofluorescence on tissue sections, anti-210,000-dalton serum bound to neuronal axons in sciatic nerve and cerebellum. In dissociated cell cultures, rat dorsal root ganglion cells and their processes bound the serum, whereas nonneuronal cells did not. Some cultured cerebellar neurons were also positive, whereas astrocytes were not. At the ultrastructural level, anti-210,000-dalton serum bound to intermediate filaments inside axonal processes. Anti-49,000-dalton serum bound to astrocytes in sections of the cerebellum, and cultured astrocytes had filaments that stained, whereas other cell types did not. In sciatic nerve sections, elements stained with this serum, but cultured cells from newborn sciatic nerve were negative. An antiserum against the 58,000-dalton protein of the cytoskeleton of NIL-8 fibroblasts strongly stained sciatic nerve sections, binding to Schwann cells but not to axons or to myelin. In cerebellar sections, astrocytes were positive, as were blood vessels and cells in the pia. In cell cultures, anti-58,000-dalton serum stained filaments inside Schwann cells, fibroblasts, and astrocytes, but neurons were negative. Cells in the cultures and tissue sections of the nervous system failed to react with antiserum to the 58,000-dalton protein of skin intermediate filaments. In these studies, astrocytes in vivo and in culture were the only cells which had antigens related to two classes of intermediate filaments.

Absence of filipin-sterol complexes from the membranes of active zones and acetylcholine receptor aggregates at frog neuromuscular junctions.

The polyene antibiotic filipin reacts specifically with membrane cholesterol and produces distinctive membrane lesions. We treated frog cutaneous and sartorius muscles with 0.04% filipin in a glutaraldehyde solution with or without prefixation with glutaraldehyde. Freeze-fracture of these muscles revealed numerous 19 to 38-nm protuberances and depressions (filipin-sterol complexes) in most areas of muscle, axon, and Schwann cell membranes. In the presynaptic membrane, however, these filipin-sterol complexes were absent from active zones consisting of ridges bordered with double rows of particles. In the postsynaptic membrane, filipin-sterol complexes were also virtually absent from the areas occupied by aggregates of large particles representing acetylcholine receptors. These results suggest that the membrane regions of active zones and acetylcholine receptor aggregates have a low cholesterol content.

Studies of adhesion molecules mediating interactions between cells of peripheral nervous system indicate a major role for L1 in mediating sensory neuron growth on Schwann cells in culture.

The involvement of the adhesion molecules L1, N-CAM, and J1 in adhesion and neurite outgrowth in the peripheral nervous system was investigated. We prepared Schwann cells and fibroblasts (from sciatic nerves) and neurons (from dorsal root ganglia) from 1-d mice. These cells were allowed to interact with each other in a short-term adhesion assay. We also measured outgrowth of dorsal root ganglion neurons on Schwann cell and fibroblast monolayers. Schwann cells (which express L1, N-CAM, and J1) adhered most strongly to dorsal root ganglion neurons by an L1-dependent mechanism and less by N-CAM and J1. Schwann cell-Schwann cell adhesion was mediated by L1 and N-CAM, but not J1. Adhesion of fibroblasts (which express N-CAM, but not L1 or J1) to neurons or Schwann cells was mediated by L1 and N-CAM and not J1. However, inhibition by L1 and N-CAM antibodies was found to be less pronounced with fibroblasts than with Schwann cells. N-CAM was also strongly involved in fibroblast-fibroblast adhesion. Neurite outgrowth was most extensive on Schwann cells and less on fibroblasts. A difference in extent of neurite elongation was seen between small- (10-20 microns) and large- (20-35 microns) diameter neurons, with the larger neurons tending to exhibit longer neurites. Fab fragments of polyclonal L1, N-CAM, and J1 antibodies exerted slightly different inhibitory effects on neurite outgrowth, depending on whether the neurites were derived from small or large neurons. L1 antibodies interfered most strikingly with neurite outgrowth on Schwann cells (inhibition of 88% for small and 76% for large neurons), while no inhibition was detectable on fibroblasts. Similarly, although to a smaller extent than L1, N-CAM appeared to be involved in neurite outgrowth on Schwann cells and not on fibroblasts. Antibodies to J1 only showed a very small effect on neurite outgrowth of large neurons on Schwann cells. These observations show for the first time that identified adhesion molecules are potent mediators of glia-dependent neurite formation and attribute to L1 a predominant role in neurite outgrowth on Schwann cells which may be instrumental in regeneration.

Ultrastructural localization of phospholipid synthesis in the rat trigeminal nerve during myelination.

A method for the ultrastructural localization of acyltransferase enzymes involved in phospholipid metabolism has been applied to the developing rat trigeminal nerve. Determination of acyltransferase levels in the nerve indicated that a peak of activity occurs at the 8th day after birth with gradual declines of activity up to 15 days. Morphological surveys and determinations of cholesterol levels suggested that heavy myelin formation occurs in the nerve during this latter period. Fixed nerves incubated in a medium for localization of acyltransferases indicated deposition of reaction product associated with Golgi cisternae, intracellular smooth vesicles, and the plasma membrane of the Schwann cell in the incipient stages of myelin formation. Golgi-derived vesicles appeared to move toward the Schwann cell surface and fuse with the plasma membrane. Activity continued to be detectable in the plasma membrane of the internal mesaxon as long as cytoplasm was evident and mature myelin membrane was not yet formed. Cells in which myelin formation appeared advanced showed little or no enzyme marker. Consistent with cytochemical observations were biochemical determinations of acyltransferases which showed high levels of the enzymes in microsomes, while no activity could be detected in the myelin fraction. Acyltransferase reaction product was also observed in the Golgi apparatus of ganglion cell bodies, axoplasmic smooth vesicles, and the axolemma. Localization of acyltransferase enzymes in Schwann cells, ganglion cell bodies, and axons during development of the nerve is discussed in relation to membrane biogenesis in the nervous system.

Distribution of N-CAM in synaptic and extrasynaptic portions of developing and adult skeletal muscle.

Previous studies of denervated and cultured muscle have shown that the expression of the neural cell adhesion molecule (N-CAM) in muscle is regulated by the muscle's state of innervation and that N-CAM might mediate some developmentally important nerve-muscle interactions. As a first step in learning whether N-CAM might regulate or be regulated by nerve-muscle interactions during normal development, we have used light and electron microscopic immunohistochemical methods to study its distribution in embryonic, perinatal, and adult rat muscle. In embryonic muscle, N-CAM is uniformly present on the surface of myotubes and in intramuscular nerves; N-CAM is also present on myoblasts, both in vivo and in cultures of embryonic muscle. N-CAM is lost from the nerves as myelination proceeds, and from myotubes as they mature. The loss of N-CAM from extrasynaptic portions of the myotube is a complex process, comprising a rapid rearrangement as secondary myotubes form, a phase of decline late in embryogenesis, a transient reappearance perinatally, and a more gradual disappearance during the first two postnatal weeks. Throughout embryonic and perinatal life, N-CAM is present at similar levels in synaptic and extrasynaptic regions of the myotube surface. However, N-CAM becomes concentrated in synaptic regions postnatally: it is present in postsynaptic and perisynaptic areas of the muscle fiber, both on the surface and intracellularly (in T-tubules), but undetectable in portions of muscle fibers distant from synapses. In addition, N-CAM is present on the surfaces of motor nerve terminals and of Schwann cells that cap nerve terminals, but absent from myelinated portions of motor axons and from myelinating Schwann cells. Thus, in the adult, N-CAM is present in synaptic but not extrasynaptic portions of all three cell types that comprise the neuromuscular junction. The times and places at which N-CAM appears are consistent with its playing several distinct roles in myogenesis, synaptogenesis, and synaptic maintenance, including alignment of secondary along primary myotubes, early interactions of axons with myotubes, and adhesion of Schwann cells to nerve terminals.

Distribution and role in regeneration of N-CAM in the basal laminae of muscle and Schwann cells.

The neural cell adhesion molecule (N-CAM) is a membrane glycoprotein involved in neuron-neuron and neuron-muscle adhesion. It can be synthesized in various forms by both nerve and muscle and it becomes concentrated at the motor endplate. Biochemical analysis of a frog muscle extract enriched in basal lamina revealed the presence of a polydisperse, polysialylated form of N-CAM with an average Mr of approximately 160,000 as determined by SDS-PAGE, which was converted to a form of 125,000 Mr by treatment with neuraminidase. To define further the role of N-CAM in neuromuscular junction organization, we studied the distribution of N-CAM in an in vivo preparation of frog basal lamina sheaths obtained by inducing the degeneration of both nerve and muscle fibers. Immunoreactive material could be readily detected by anti-N-CAM antibodies in such basal lamina sheaths. Ultrastructural analysis using immunogold techniques revealed N-CAM in close association with the basal lamina sheaths, present in dense accumulation at places that presumably correspond to synaptic regions. N-CAM epitopes were also associated with collagen fibrils in the extracellular matrix. The ability of anti-N-CAM antibodies to perturb nerve regeneration and reinnervation of the remaining basal lamina sheaths was then examined. In control animals, myelinating Schwann cells wrapped around the regenerated axon and reinnervation occurred only at the old synaptic areas; new contacts between nerve and basal lamina had a terminal Schwann cell capping the nerve terminal. In the presence of anti-N-CAM antibodies, three major abnormalities were observed in the regeneration and reinnervation processes: (a) regenerated axons in nerve trunks that had grown back into the old Schwann cell basal lamina were rarely associated with myelinating Schwann cell processes, (b) ectopic synapses were often present, and (c) many of the axon terminals lacked a terminal Schwann cell capping the nerve-basal lamina contact area. These results suggest that N-CAM may play an important role not only in the determination of synaptic areas but also in Schwann cell-axon interactions during nerve regeneration.

Basement membrane proteins, interstitial collagens, and fibronectin in neurofibroma.

The distribution and nature of extracellular matrix proteins in neurofibroma tissue was studied by indirect immunofluorescence, immunoelectron microscopy, immunoblotting, and rotary shadowing. The most striking feature was an extensive network of basement membranes localized mainly around Schwann cells and small blood vessels. The major components, collagen IV, laminin, and nidogen, were mainly deposited in the lamina densa. Some laminin and nidogen could be extracted with 0.5 M NaCl and were shown by electrophoresis to have the characteristic chain and fragment patterns described previously for these proteins isolated from the mouse Engelbreth-Holm-Swarm (EHS) sarcoma. Fragments of collagen IV and collagen VI were solubilized by limited proteolytic digestion and identified after rotary shadowing. The more remote interstitial regions of the tumor contained cross-striated collagen fibrils which were composed of collagen III (diameter, 20-30 nm) or collagen I (diameter, 40-50 nm). Collagen fibrils thicker than 80 nm were not found. The interstitial regions also contained collagen VI as a fine filamentous network near cells and between collagen fibrils. Deposits of fibronectin were rather small and showed a scattered distribution. The data indicate that Schwann cells contribute considerably to matrix production in neurofibroma which may therefore be a suitable model for studying basement membranes of neuroectodermal origin.

Distribution of albumin-like immunoreactivity in rat sciatic nerve after nerve crush injury.

To understand better the role of local factors in the response of peripheral nerve to crush injury, we studied the distribution of albumin-like immunoreactivity (A-LI) in the rat sciatic nerve from one day to eight weeks (wk) after a crushing injury; we used electron microscopic immunocytochemistry. In the nerve distal to the crush degenerating axons demonstrated intra-axonal A-LI, and by one wk most of the Schwann cells also showed A-LI. As regenerating sprouts entered the distal nerve, those Schwann cells in contact with sprouts lost their A-LI, while those cells not in contact with axons retained immunoreactivity up to eight wk after injury. Proximal to the nerve crush many axons showed intra-axonal A-LI from one to two wk after injury, despite appearing normal ultrastructurally. This immunoreactivity diminished as the distance from the crush site increased. Many Schwann cells proximal to the crush also showed A-LI from one to four wk after injury. These findings suggest that an albumin-like protein may play a role in the response of Schwann cells and axons to injury.

Neuroimmunological electron microscopy with microwave-accelerated fixation.

Immunoelectron microscopy is an important tool used to determine the precise location of immune complexes. Standard concentrations of glutaraldehyde destroy these complexes. This paper describes a method in which the period of glutaraldehyde fixation is shortened by concomitant microwave treatment. Using 1.25% glutaraldehyde and microwave fixation ideal preservation and demonstration of MHC class I antigen on Schwann cells was obtained by the peroxidase method.

Glial fibrillary acidic polypeptides in peripheral glia. Molecular weight, heterogeneity and distribution.

Glial fibrillary acidic (GFA) polypeptides are present in major categories of rat peripheral glia including non-myelin-forming Schwann cells, enteric glia and some satellite cells. They can be detected both immunochemically and immunohistochemically. The immunoreactivity is associated with a polypeptide which has an MW of 49 000, indistinguishable from that of glial fibrillary acidic protein (GFAP) from rat brain. In spite of this, the GFA polypeptides found in the peripheral nervous system and central nervous system are not identical since they can be distinguished both immunohistochemically and immunochemically by a monoclonal GFAP antibody which recognizes GFAP in astrocytes and some enteric glia, but not GFAP in non-myelin-forming Schwann cells, satellite cells and many enteric glia. GFA-related molecules can also be detected in human Schwann cells by immunofluorescence. The results suggest, however, that the glial filament polypeptides of peripheral glia and astrocytes are less closely related in the human than in the rat. The glial distribution of GFAP is closely paralleled by 2 cell surface proteins, Ran-2 and A5E3 antigen. Although GFAP, Ran-2 and A5E3 are individually expressed by diverse cell types, the phenotype GFAP+, Ran-2+, A5E3+ defines a narrow group including only non-myelin-forming Schwann cells, enteric glia and astrocytes. These observations suggest that the non-myelin-forming cells of the central and peripheral nervous system may share some common functions.

A subset of Schwann cells in peripheral nerves contain a 50-kDa protein antigenically related to astrocyte intermediate filaments.

Antisera raised to the astrocyte intermediate filament structural protein stained elements in the peripheral nerves of several species. These elements were not associated with myelinated nerve fibers, were more common in splenic and vagus nerves than in the sciatic nerve, and persisted after nerve transection. In teased nerve preparations antigen-positive cells appeared to be the Schwann cells that surround small diameter, unmyelinated axons. Absorption of the antiserum with purified rat spinal cord 50-kDa protein or with bovine splenic nerve cytoskeletal extract blocked the reaction with CNS astrocyte processes or with PNS nerve fibers. Immunoblots of cytoskeletal preparations of bovine splenic nerve or rat sciatic nerve showed that the antigen from peripheral nerves comigrated at 50 kDa with antigen from bovine or rat spinal cord or cultured rat astrocytes. The CNS and PNS 50-kDa proteins from bovine tissues were subjected to limited digestion with Staphylococcus aureus protease V8. After separation on SDS-gels, antigenic peptides were detected by immunoblotting. The pattern of antigenic peptides for the CNS and PNS proteins were identical. We conclude that Schwann cells associated with nonmyelinated axons contain a cytoskeletal protein that is the same size and has the same peptide map as the major structural protein of astrocyte intermediate filaments.

Schwann cells of the olfactory nerves contain glial fibrillary acidic protein and resemble astrocytes.

Two antisera to glial fibrillary acidic protein from human brain and an antiserum to a 49 k dalton glial filament protein from human brain detected a cross-reacting antigen in the Schwann cells of the olfactory and vomeronasal nerves. The antigen was demonstrated at light- and electron-microscope levels. It was found throughout the cytoplasm and in association with cytoplasmic filaments of olfactory nerve Schwann cells in intact tissue and in Schwann cells grown in vitro. This observation, together with observations on the ultrastructure of olfactory nerve Schwann cells, relates them to central astroglia and to glial cells of the myenteric plexus, rather than to Schwann cells of other peripheral nerves. The unusual properties of olfactory nerve Schwann cells are of interest in relation to the regenerative abilities of the olfactory nerves.

Localization of immunoreactive epidermal growth factor receptors in human nervous system.

Epidermal growth factor is a well-defined peptide which stimulates cell growth and elicits cell responses in a variety of tissues by binding to specific receptors, EGF-R. A specific antiserum against the EGF receptor, which has previously been used to characterize EGF-R in human skin, fibroblasts, and smooth muscle, was used to survey the distribution of EGF-R in human nervous system. Portions of formalin-fixed, paraffin-embedded autopsy specimens were examined by use of immunohistochemical staining (PAP technique) with EGF-R antiserum. Many types of nerve cells, e.g., cerebral cortical pyramidal cells, hippocampal pyramidal cells, Purkinje cells, anterior horn cells, and dorsal root ganglion neurons, contained immunoreactive EGF-R. However, immunoreactive EGF-R were not detected in astrocytes, oligodendrogliocytes, and other small neurons such as granule cells. Intense immunostaining for EGF-R was also detected in ependymal cells from choroidal and extrachoroidal locations. Although immunoreactive EGF-R is widely distributed in human nervous system, the functional role of EGF and its receptor in the nervous system remains unknown.

Use of lead aspartate block staining in quantitative EM autoradiography of phospholipids: application to myelinating peripheral nerve.

For quantitation of electron microscope (EM) autoradiographs, micrographs must contain clear images which are relatively free of heavy metal precipitates. Satisfactory contrast is usually obtained by staining individual ultra-thin sections with lead citrate. It was recently reported that sequential block staining of tissue with ferrocyanide-reduced osmium tetroxide and lead aspartate produced excellent contrast for EM autoradiography, with sections relatively free of lead precipitate. This protocol avoids the manipulation involved in staining individual ultra-thin sections. We have adapted this method to quantitative EM autoradiographic studies, primarily of phospholipid metabolism in peripheral nerve. We show that block staining with lead aspartate provides: (a) ultrastructural contrast of routinely high quality for myelinated peripheral nerve; (b) high (greater than 98%) retention of glycero-labeled lipid during dehydration and embedment; and (c) a distribution of de novo tritiated glycerol-labeled lipid in ultra-thin sections that is quantitatively identical to the distribution recorded for samples stained by the more laborious post-embedment method. During a 2-hr labeling period in vivo, tritiated glycerol is incorporated into phosphatidylcholine (44%), phosphatidylethanolamine (22%), other phospholipids (16%), and neutral lipids (15%). The analysis of grain distribution in developing sciatic nerve labeled for 2 hr with tritiated glycerol demonstrates that myelinating Schwann cells play the major role in synthesis of endoneurial lipids. Lipid synthesis in myelinated fibers is localized in perinuclear regions of Schwann cell cytoplasm. These regions lie external to compact myelin. Unmyelinated fibers and other endoneurial cells independently incorporate glycerol into lipids.

Gastrointestinal stromal tumors. An immunohistochemical study of cellular differentiation.

Forty-five benign and 11 malignant gastrointestinal stromal tumors (GIST) were immunohistochemically studied for the presence of desmin, muscle actins (MA) and S-100 protein. To facilitate the analysis, the tumors were divided into four groups by light microscopy: (1) typical leiomyomas comparable to peripheral leiomyomas (n = 9); (2) cellular spindle cell tumors (n = 29); (3) round cell tumors ("leiomyoblastomas" n = 7); and (4) sarcomas (n = 11). The typical leiomyomas were desmin- and MA-positive throughout, and showed well-differentiated smooth muscle cells by electron microscopy, similar to the normal gastric smooth muscle cells. All esophageal leiomyomas belonged to this group. Nineteen of 29 of the Group 2 tumors showed desmin positivity and 20 of 29 showed MA positivity, but usually only in less than 10% of the tumor cells, and in many instances it was very difficult to determine whether the positive cells were real tumor cells or entrapped muscle cells. Only 5 of 29 of Group 2 tumors showed widespread desmin positivity and 11 of 29 showed similar MA positivity. Of round cell tumors, only 1 of 7 showed desmin-positive cells and 3 of 7 MA-positive cells. None of the sarcomas showed desmin, while MA positivity was found in 6 of 11 cases, often in a large number of tumor cells. Seven tumors showed a significant number of S-100 positive tumor cells, but four of these also showed a high number of desmin- and MA-positive cells, suggesting that these tumors represented complex proliferations of muscle and Schwann cell elements. Two purely S-100 positive benign probably Schwann cell-like tumors were found, both in the small bowel. Small number of S-100 positive cells were commonly found in GIST, and these probably represented entrapped Schwann cells, because many tumors showed simultaneous proliferation of non-neoplastic nerves.

Schwann cells cultured from adult rats contain a cytoskeletal protein related to astrocyte filaments.

Cells that bound antibody to the astrocyte intermediate filament protein were cultured from adult rat sciatic nerve. The antigen was intracellular, finely filamentous, and formed perinuclear caps in response to colchicine, all properties of intermediate filaments. Cytoskeletal proteins of these cultures were separated by SDS-gel electrophoresis, transferred to nitrocellulose paper, and shown to bind the glial-specific antiserum to a protein of 50,000 daltons. All the cells that bound this serum had a Schwann cell surface antigen, Ran-1, whereas fibroblasts from the nerve had Thy-1 surface antigen and did not contain the astrocyte filament antigen. These results prove that some Schwann cells from adult nerve, in contrast to fibroblasts or immature Schwann cells, have an intermediate filament protein that shares antigenic determinants with, or may be identical to, the astrocyte filament protein.