Inclusion body myositis: peripheral nerve involvement. Combined morphological and electrophysiological studies on peripheral nerves

The occurrence of neuropathy in 5 cases of inclusion body myositis (IBM) was studied. The intramuscular nerve branches showed variable ultrastructural changes in all cases. The observed changes were loss of axons, wallerian degeneration and axon terminal atrophy. EMG with concentric needles showed myopathic motor unit potentials in all cases. A reduced interference pattern due to loss of motor units was found in all cases. All but one patient showed fibrillation potentials in several muscle groups. Motor nerve conduction velocities and F-wave latencies were pathological in two of the cases, in which there were motor unit potentials with increased amplitude and duration as evidence of reinnervation. Sural nerve biopsy in one of these cases revealed slight neuropathy of the axonal type. These findings support the concept of peripheral nerve involvement in many cases of IBM.

Prolonged ongoing discharges of sensory nerves as recorded in isolated nerves in the rat

Whether or not injury to a mammalian sensory nerve produces prolonged discharges is a controversial issue. Because of this controversy and its potential ramifications in both experimental and clinical conditions, we examined discharges in sectioned sensory nerves of the rat. In anesthetized rats, either a dorsal root or a saphenous nerve was isolated by sectioning both proximally and distally. Multi-unit recordings from these isolated nerves showed low levels of prolonged ongoing discharges often lasting for at least 2 hr. Furthermore, results from short-term (1-2 days) survival surgeries indicated that prolonged ongoing discharges could last for days. Sectioning the ventral root produced discharges for only a short period. Various pieces of evidence suggested that the sources of impulse generation are multiple, occurring in the middle of an uninjured axon at a site away from the injury, as well as at the injured site. There is circumstantial evidence which suggests that these prolonged discharges are produced in physiological conditions or at least under normal experimental conditions.

An autopsy case of peroneal muscular atrophy with rigidity and tremor. Ultrastructural and systematic morphometrical studies on peripheral nerves

An autopsy case of hereditary peroneal muscular atrophy (PMA) with rigidity and static tremor is presented. The patient developed slowly progressive distal muscular atrophy of the legs at the age of 15 years. By the age of 52 years, PMA became marked associated with pes cavus, and tremor and rigidity of the extremities were noted. Motor and sensory conduction velocities gradually depressed and lost near the end of his life. At autopsy, the major neuropathological abnormalities involved the peripheral nervous systems, and were characterized by axonal atrophy and loss of myelinated fibers. These changes involved both the proximal and distal nerves, being more severely affected in the distal. The pathological changes in other regions of the nervous systems were mainly confined to the spinal cord, dorsal ganglia and spinal nerve roots, and pigmented neurons in the brain stem. Morphometrically, the total fascicular area was much smaller than in control, but the total number of myelinated fibers greatly outnumbered that of control 75,200 to 48,200 at the proximal sciatic nerve and then gradually decreased towards the periphery; however, even in the distal sural nerve, the total number of myelinated fibers exceeded that of control (6820 to 5469). Thus, the density of myelinated fibers were much higher, being 1.5 to 2 times greater, than in control. Its abrupt decline at the distal nerve might account for neurogenic atrophy of the distal musculature. Unmyelinated fibers were slightly increased in density and not atrophic. This case is unique in its clinicopathology and does not belong to any subtypes of PMA including "neuronal plus".

Expression of NF-L in both neuronal and nonneuronal cells of transgenic mice: increased neurofilament density in axons without affecting caliber

We have generated transgenic mice containing additional copies of the murine NF-L gene in order to examine the consequences of neurofilament-L overexpression on axonal morphology. Founder mice were constructed to carry a transgene in which the presumptive 5' promoter sequences of NF-L were replaced with the strong murine sarcoma virus long terminal repeat promoter. The transgenes were expressed prominently in several tissues, including skeletal muscle and kidney where NF-L accumulated to approximately 2% of cell protein. This was not accompanied by an overt phenotype, except that expression in lens led to cataract formation. In the brains of these animals, transgene RNA levels exceeded the endogenous NF-L RNAs by up to 20-fold, although no additional protein accumulated, indicating posttranscriptional regulation of NF-L expression. However, in peripheral neurons transgene RNA was approximately fourfold higher than endogenous NF-L mRNAs, and a corresponding increase in NF-L subunits was found in axons arising from these neurons. Myelinated nerve fibers of transgenic animals contained increased numbers of NFs, assembled predominantly of NF-L. This was reflected in an increase in the density of axonal NFs; axonal caliber was not affected.

Effects of colchicine on central and peripheral nerve terminals of the locust Schistocerca gregaria

The treatment of locusts (Schistocerca gregaria) with colchicine leads to various changes in the ultrastructure of both central and peripheral nerve terminals. These changes include a marked reduction in synaptic vesicle density, which is further enhanced following stimulation. There are also increased numbers of coated pits and membraneous cisternae together with enlarged terminal areas and disruption of mitochondrial ultrastructure. The capacity of isolated central nerve terminals (synaptosomes) to take up the radiolabelled transmitter precursor [3H]-choline is enhanced following in vivo treatment of locusts with colchicine, whilst in vitro treatment leads to a reduction in choline uptake. The results are discussed with regard to the postulated effects of colchicine on axoplasmic transport.

Anti-macrophage CR3 antibody blocks myelin phagocytosis by macrophages in vitro

Myelin phagocytosis in Wallerian degeneration of peripheral nerves depends on invasion of nerves by non-resident macrophages. The present study was done to clarify the role of the macrophage complement receptor type 3 (CR3) in myelin removal. Myelin phagocytic capacity of invading macrophages was abolished by treatment of cultured nerves and macrophages with anti-CR3 antibody or by serum complement depletion with cobra venom factor. This indicates that myelin phagocytosis is mediated by the macrophage CR3.

Peripheral neuropathy in xeroderma pigmentosum

The pathology of the peripheral nervous system (PNS) in 2 autopsied cases of group A xeroderma pigmentosum (De Sanctis Cacchione syndrome) are presented. Motor nerves including those of the oculomotor systems were severely affected, but involvement of the sensory system was even more marked. Minor hypertrophic changes were present in the distal portions of the peripheral nerve trunks, but there was no appreciable difference in the density of myelinated nerve fibres between proximal and distal levels. Morphometric data including teased fibre analyses and g ratio scattergrams suggest that the underlying pathogenetic mechanism is that of a neuronopathy. Unmyelinated axons were also severely depleted. Review of the previous literature revealed that the pathological changes of the PNS in group A xeroderma pigmentosum are thought to be slowly progressive, which is also suggested by the severe and widespread sclerotic changes of the CNS in the present 2 cases.

Receptors for immunoglobulin G demonstrated on human peripheral nerve fibres by electron microscopy

The localization of class III of receptors for the Fc (fragment-crystallizable) region of immunoglobulin G (IgG) (Fc gamma RIII) on human peripheral nerve fibres was investigated by indirect immunoperoxidase staining of frozen nerve sections with the monoclonal antibody anti-Leu-IIb. Transmission electron microscopy revealed that Fc gamma RIII are localized to the entire surface membrane of the Schwann cell. Myelin and axons were not stained. Fc gamma RIII on human Schwann cells may be of significance in binding immune complexes, in phagocytosis, in release of lymphokines as well as cytotoxic and inflammatory mediators and in local immunoregulation.

Chronic inflammatory demyelinating polyneuropathy in childhood: ultrastructural features of peripheral nerve biopsies in four cases

Peripheral nerve biopsies (PNB) from four children suffering from subacute or chronic inflammatory demyelinating polyneuropathy were studied by electron microscopy. Remyelinating features with onion bulb formations, inflammatory cell infiltrates and active demyelinating lesions were strongly suggestive of the disease. In the first case, a second PNB, performed after 7 months of severe subacute course, showed a striking evolution of the lesions. In the second case and in spite of severe neurological symptoms, the PNB was almost normal, suggesting that inflammatory lesions were mainly located in the proximal parts of the nerve. No signs of active demyelination could be seen in the third case but onion bulb formations and inflammatory cell infiltrates were present. In the fourth case, characteristic lesions of active demyelination were associated with a history of familial polyneuropathy; this association suggests an auto-immune process in certain kindreds with hereditary motor and sensory neuropathy.

Reduced myelinated fiber size correlates with loss of axonal neurofilaments in peripheral nerve of chronically streptozotocin diabetic rats

Peripheral sensory nerve abnormalities were investigated in long-term streptozotocin diabetic rats using quantitative analysis. To determine whether the characteristic structural changes occur with a proximodistal gradient, three levels of the sensory peripheral nervous system were investigated: the postganglionic segment of the dorsal root, the midportion of the sciatic nerve, and the distal sural nerve. Reduction of myelinated fiber size due to reduced axonal caliber was the most characteristic change at both proximal and distal levels of the peripheral nerve. The relationship between axonal size and myelin spiral length indicated a more severe axonal atrophy in the distal portion. The axonal atrophy was related to a proportional loss of axonal neurofilaments at proximal levels, whereas in the distal sural nerve the loss of neurofilaments exceeded that which would be expected for axonal size. The universal reduction of axonal size in diabetic nerve may be accounted for by impaired supply of neurofilaments or reduced neurofilament synthesis. Such cytoskeletal defects may, in turn, lead to distal axonal degeneration or contribute to the susceptibility of diabetic nerve to various external noxi, including ischemia and hypoglycemia.

High density of nodes of Ranvier in the CNS-PNS transitional zone

Node of Ranvier density was examined at three levels along rat lumbar motoneuron axon bundles: where they lie in the central nervous system, in the peripheral nervous system and in the transitional zone (TZ) between these. Density was considerably and significantly greater in the TZ than in either of the other locations. It is possible that such densely packed nodes in the TZ could interact electrically with one another. Because of its structure and position and because it contains a relatively pure fibre population, the rat L4 ventral rootlet TZ lends itself readily to electrophysiological investigation of this possibility.

Immunopathological factors in peripheral nerve allograft rejection: quantification of lymphocyte invasion and major histocompatibility complex expression

The numbers of helper T and cytotoxic T lymphocytes and macrophages were quantified, and the expression of major histocompatibility complex (MHC) class I and class II molecules was examined in rat peripheral nerve allografts from 1 to 14 days after implantation, using the indirect immunoperoxidase method for light and electron microscopy. Two centimetre segments of peripheral nerve freshly obtained from inbred Dark Agouti strain rats were inserted in a gap created in n. fibularis or n. tibialis of young adult inbred Wistar strain rats, using fascicular nerve repair techniques under general anaesthesia. There was a gradual increase in the number of helper T and cytotoxic/suppressor T cells from day 2 with peak numbers of both types of T cells observed around day 7. The results suggest that the critical time for T cell proliferation is between day 6 and day 7 post-operatively. The number of macrophages increased over 10 days, with peak numbers being observed at day 10 post-operatively. This is in accord with the pattern of rejection observed in allografts of other tissue. Schwann cells were found to express MHC class I and class II molecules by day 2 post-operatively, which is well before there is any substantial T cell and macrophage infiltration. It may be that the donor Schwann cells act as antigen presenting cells, triggering the immune response and finally becoming a target of the rejection process.

The receptive part of the primary afferent axon is most vulnerable to systemic capsaicin in adult rats

The present study shows that systemic capsaicin in adult rats results in a significant loss of axons in the subepidermal nerve plexus of the posterior leg but no loss of axons in the sural nerves of these same animals. These data are interpreted as indicating that the receptive part of the peripheral sensory axon is destroyed but that the cell body and most of the peripheral axon remains intact. Thus we suggest that the receptive part of the peripheral sensory axon is the most vulnerable part of the primary afferent neuron to capsaicin in these animals. These findings may explain the observation that adult rats treated with systemic capsaicin are deficient in their responses to certain painful stimuli but usually do not show obvious signs of primary afferent neuron death. We also suggest that as the dose of capsaicin is increased the whole neuron dies. It remains to be determined if the peripheral damage reported here is related to the striking loss of primary afferent markers in the dorsal horn that is also seen after this treatment.

Ultrastructural morphometric study of efferent nerve terminals on murine bone marrow stromal cells, and the recognition of a novel anatomical unit: the "neuro-reticular complex"

In order to extend our understanding of the role of nerve fibers in the structure and function of bone marrow stroma, we have examined nerve terminals, arterioles, and capillaries in femoral bone marrow tissues of 50 C57BL strain mice, using electron microscopy and morphometric methods. Within the adventitia of arterioles, a particular type of cell, termed periarterial adventitial (PAA) cell, is characterized by a thin veil-like cytoplasm which concentrically surrounds both nerves and arterioles. Nerve fibers containing both unmyelinated and myelinated axons are distributed mainly between the layers of PAA cells, but are found rarely on the sinus walls or within the hematopoietic parenchyma. Quantitatively, the efferent nerve terminals with many synaptic vesicles are distributed mainly beside arterial smooth muscle cells (Type I: 58.8%) or between the layers of PAA cells (Type III: 33.2%), and rarely in hematopoietic parenchyma (Type II: 5.3%) or on sinus walls (Type IV: 2.7%). In the case of Type II-IV nerve terminals, efferent (autonomic) nerves and bone marrow stromal cells which are connected by gap junctions (sinus adventitial reticular cells, intersinusoidal reticular cells, and PAA cells) appear to constitute a potential functional unit for signal conduction. We would like to propose a new term for this anatomical unit in marrow, the "neuro-reticular complex."

Glucose transporters at the blood-nerve barrier are associated with perineurial cells and endoneurial microvessels

The blood-nerve barrier consists of continuous layers of cells linked by tight junctions and includes the endothelial cells which line the endoneurial capillaries and the perineurial cells which surround fascicles of nerve fibers. A facilitated transport carrier protein allows D-glucose to penetrate the barrier. To determine the specific cellular location of the transport system, an antiserum to a synthetic peptide corresponding to the carboxyl-terminus of the glucose transporter protein was used for light and electron immunocytochemical analyses. Glucose transporters were abundant both in endoneurial capillaries and the perineurial sheath. In perineurium, transporters were located in the plasma membranes and cytoplasm of the perineurial cells. Approximately two-thirds of the transporters associated with perineurial cells were localized in the plasma membranes. Perineurial cells are thus similar to cerebral endothelial cells in that they lack a large intracellular pool of transporters which might be sensitive to hormonal regulation. The presence of hexose carriers in perineurium suggests that glucose transport from epineurium to endoneurium may play a significant role in the metabolism of peripheral nerve fibers. These results support the concept that the blood-nerve barrier serves as an important nutrient delivery system.

Ultrastructural changes in glial cells during regeneration of cockroach peripheral nerve

After peripheral nerve 5 in the cockroach Periplaneta americana was cut, changes occurring in the glial cells in the proximal stump were studied immediately after damage and during the process of nerve regeneration. During the first week haemocytes accumulated outside the nerve and morphologically similar granule-containing cells appeared inside the nerve. These cells were involved in phagocytic activity. Between the second and the fourth week, signs of regeneration were distinguishable; many small axonal sprouts were formed which were surrounded by glial processes, and the nerve stump increased in length. During this period the glial cells produced large amounts of extracellular material in which the bundles of axons and glia were embedded. The structural differences between glial and perineurial cells were lost during these stages of regeneration and there was no restriction to the penetration of the extracellular tracer lanthanum. After 8 weeks, reinnervation of the muscles had taken place, perineurial and glial cells were again distinguishable, and the perineurial cells were able to exclude lanthanum.

Elimination of intramedullary axon collaterals of cat spinal alpha-motoneurons following peripheral nerve injury

The motor nerve supplying the medial gastrocnemius (MG) muscle was transected in the popliteal fossa of adult cats. The proximal nerve stump was ligated to prevent reinnervation. Three, six or twelve weeks later, axotomized MG motoneurons were intracellularly labelled with horseradish peroxidase, and the morphology of their intramedullary axon collateral systems was investigated quantitatively. The results were compared with corresponding data obtained from normal MG motoneurons. The peripheral chronic axotomy induced a gradual decrease in the number of recurrent axon collaterals originating from the lesioned MG motoraxons within the spinal cord. At 12 weeks postoperatively, this decrease amounted to 40%. The elimination preferentially involved axon collaterals originating from juxta-somatic regions of the motoraxons. In the axon collateral trees persisting in the axotomized MG neurons the tree size, branching patterns and number of synaptic boutons were all normal. Thus, no signs of a gradual deterioration of individual axon collateral systems were observed at any postoperative stage studied. The results are discussed in relation to other retrograde degenerative and regenerative events induced by axotomy.

The development of Meissner corpuscles in primate digital skin

Digital skin from fetal and neonatal primates was examined using light and electron microscopic techniques to document the development of the Meissner corpuscle. Generation of the receptor was initiated early in the third trimester by fine neurites which left the superficial dermal nerve plexus, ascended the dermal papillae and entered the basal epidermis. As maturation of the Meissner corpuscle continued, the axons were confined to the apex of the dermal papilla, where they were oriented parallel to the surface to the surface of the skin and terminated among cytoplasmic extensions of presumptive lamellar cells. During late fetal life the complexity of the lamellar cell stacking increased and the lamellar cell bodies were found solely at the base of the receptor. Numerous axon terminals were evident between the cytoplasmic lamellae. The appearance of the neonatal Meissner corpuscle was indistinguishable from that of the adult, indicating that the complete cycle of development is concluded before birth.

Ipsilateral and cross-over elongation of the motor nerve by nerve grafting: an experimental study in sheep

In difficult reconstructions, ipsilateral or cross-over nerve grafting is sometimes necessary to achieve reinnervation and motor function. This experimental study in sheep was to answer the question of limitation of elongation of a motor nerve by grafting, the question of the optimal time for suturing the nerve graft to the muscle nerve, and the question of the successful application of this surgical technique in extremities. In 18 sheep, the vastus nerve was elongated by a saphenous nerve graft as long as possible up to 30 cm (step 1). In 10 animals the nerve graft was applied ipsilaterally, and in 8 animals it was used as a cross-over nerve graft to the contralateral limb. The time between nerve grafting and connection of the distal end of the nerve graft to the freshly cut rectus nerve supplying the rectus femoris muscle (step 2) was variable: 0, 3, 6, 9, and 12 months. In all animals, the final experiments (step 3) were performed 6 months after the last operation (step 2). Muscle force measurements in the rectus femoris muscle and quantitative analysis of the number and diameter of myelinated nerve fibers in cross sections of the nerve biopsies at different levels showed that elongation of a motor nerve by nerve grafting is principally not limited. The functional results were rather inhomogeneous and therefore unpredictable (ipsilateral group: maximum tetanic tension = 27 to 172 N; cross-over group: 0 to 227.5 N). Nevertheless, crossover nerve grafting is recommended for selected cases even in extremities. There was no correlation between the time interval between the two operations and the functional or morphologic results, although better functional results were obtained when the distal nerve suture (step 2) was performed some months after nerve grafting (step 1). A clear correlation was found only between the number of regenerated axons in the rectus nerve behind the second suture line and the muscle function.

In-vivo quantitative imaging of catecholaminergic nerve terminals in brain and peripheral organs using positron emission tomography (PET)

Positron emission tomography (PET), has made possible quantitative imaging of the origins of positron-emitting isotopes, such as 11C and 18F, in intact animals and in humans. Lack of absolute specificity of enzymes, storage mechanisms and transporters allows 11C-or 18F-labelled "false transmitters" to be formed, stored and released from nerve terminals. Discussed are the assumptions, limitations, and advantages of 18F-6-fluoroDOPA, 18F-6-fluorodopamine, 18F-6-fluorometaraminol, and 11C-N-methyl-metaraminol (m-hydroxyephedrine) for PET imaging of dopaminergic nerve terminals in brain or sympathetic innervation in peripheral organs.

Non-specific cholinesterase activity of the developing peripheral nerves and its possible function in cells in intimate contact with growing axons of chick embryo

The results presented here demonstrate non-specific cholinesterase (nChE) activity in the developing peripheral nerves of chick embryos at stages 25-26 according to Hamburger and Hamilton (1951, J. Morphol. 88, 49-92). Under the light microscope the use of simultaneous staining for nChE activity and silver proteinate impregnation revealed the axons to be surrounded by cells exhibiting nChE activity in the main nerve trunks and in the growing tips of nerves. Nerve branches arising from the main nerve trunks contained cells with positive reaction for nChE activity, too. Electron-dense particles of the reaction product indicating nChE activity were found in the rough endoplasmic reticulum and in the perinuclear envelope of cells in close contact with growing nerve fibers and their growth cones. The same distribution of nChE activity was found in cells which were located near to nerve fasciculi but without direct contact with axons. Surprisingly, the cells in close contact with axons and their growth cones exhibited the end product of nChE activity in the outer part of their plasma membrane. The cells enveloping axons within the nerve trunks were apparently Schwann cells, while those around the growth cones at nerve tips could be identified as Schwann cells and/or mesenchymal cells of the hindlimb. The nChE reaction product was also detected in the axolemma of nerve fibers and their growth cones. The distribution of nChE activity in the developing peripheral nerves of chick embryos suggests that these molecules may influence the process of axonal elongation and locomotion. Several possible mechanisms of nChE action on growing axons can be presumed: (i) intracellular Ca2+ level regulation; (ii) providing an adhesive substrate; and (iii) butyrate production affecting the cell metabolism and the distribution of neurotubules and neurofilaments. It is also assumed that nChE molecules are involved in the interactions of nerve fibers with Schwann cells and/or mesenchymal cells as well as in interneuronal interactions.

Fiber counts at multiple sites along the rat ventral root after neonatal peripheral neurectomy or dorsal rhizotomy

We hypothesized that the afferent fibers in the ventral root of the rat are the third branches of dorsal root ganglion cells; these afferent processes in the ventral root are of varying length and end bluntly along the length of the root. In the case of an injury at either the central or the peripheral processes of the dorsal root ganglion cells in the neonatal stage, these fibers sprout at the blunt endings along the length of the ventral root. We cut either the sciatic nerve or the dorsal root on one side in neonatal rats. After the rats were fully grown, the number of both myelinated and unmyelinated fibers was counted in electron photomicrographs at multiple sites along the length of the ventral root. We observed a greatly increased number of unmyelinated fibers in the ventral root after the sciatic nerve had been cut at the neonatal stage. The magnitude of increase was more at the distal than at the proximal portion of the ventral root, suggesting that added fibers originated from the distal side. Neonatal dorsal rhizotomy, however, did not produce the same result. These results are consistent with our hypothesis that peripheral nerve injury at the neonatal stage triggers sprouting of the third branches of the dorsal root ganglion cells which end bluntly along the length of the ventral root in the normal animal.

[Age-related characteristics of the myeloarchitectonics of the peripheral nerves]

By means of macromicroscopy and microscopy myeloarchitectonics of morphologically and functionally different somatic and visceral nerves (branches: mandibular nerve, cervical spinal nerves, inferior laryngeal nerve, hepatic plexus) have been studied in 11 age groups. During the prenatal period of ontogenesis asynchronism of myelogenesis is stated in various muscle branches of the nerves, dependent on formation of function in corresponding muscles and muscle groups. As demonstrate investigations on peculiarities of myelogenesis course in the somatic and visceral nerves studied, during the period of pre- and postnatal ontogenesis, its dynamics embraces three stages of myelogenesis, determined by G. B. Stovichek for visceral nerves: productive myelogenesis, stages of stabilization and involution. The stage of productive myelogenesis in the somatic nerves studied lasts up to the end of the adolescent period. Two phases are determined in it: the first lasts up to 2-3 years; the second--up to the end of the adolescent period and is characterized with a complete formation of the myelin fibers spectrum. In the visceral nerves studied increase of general amount of myelin fibers and their differentiation are completed simultaneously during the adolescent age. The stabilization stage of myeloarchitectonics of the nerves studied corresponds to the mature age (I and II periods) and the involution stage--to the elderly (and old) age.

Cytomembranous inclusions in the peripheral nerves in AIDS

We report finding tubuloreticular inclusions (TRI) in the endothelial cells of endo- and epineurial vessels in the sural nerve of 11 patients with AIDS. Six patients had a painful peripheral neuropathy, one a non-painful sensory neuropathy, one an acute inflammatory demyelinating polyradiculoneuropathy and one a thalidomide-related neuropathy. Two patients had no clinical evidence of neuropathy. The TRI are not specific to one neuropathy and are unlikely to contribute to the pathogenesis of peripheral nerve syndromes in AIDS.

Neuroectodermal differentiation in "extraskeletal Ewing's sarcoma"

A small round cell tumor of soft tissue arising in the retroperitoneum of an 18-year-old woman is reported. The light and electron microscopic features of the tumor were mostly indistinguishable from Ewing's sarcoma of the bone and in some parts from differentiating neuroblastoma with mature neurons and Schwann cells. An immunohistochemical study using anti-NSE antibody showed many widespread NSE-positive cells lying singly and, more often, in clusters including the undifferentiated areas. The tumor was not anatomically related to the paravertebral sympathetic ganglia or adrenal glands. The present findings strongly suggest that certain extraskeletal Ewing's sarcomas are very primitive neuroectodermal tumors.