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.     

Innervation of the vasa nervorum: changes in human diabetics.

Transperineurial and epineurial vessels are innervated by plexuses of unmyelinated axons. Human sural nerve biopsies were examined ultrastructurally and immunocytochemically with an antibody which recognizes a neuronal and neuroendocrine protein, PGP 9.5, to characterize perivascular axons of these plexuses. Diabetics exhibited a greater degree of abnormal innervation of the vasa nervorum than nondiabetics with and without neuropathy. Abnormal innervation included: a reduction in the percentage of vessels exhibiting perivascular axons and a concomitant increase in the percentage of vessels having denervated Schwann cell units, particularly around vessels confined to perineurial compartments, and remaining axons in nerves from diabetics exhibited fewer varicosities. Denervated arterioles of diabetics also displayed structural changes indicating injury. The arteriolar structural defects and loss of neurogenic control of neural blood flow may lead to or aggravate endoneurial ischemia or hypoxia. The patchy, focal endoneurial fiber loss that is prominent in proximal nerves and associated with the distal myelinated fiber loss of some diabetic patients may be due in part to perivascular denervation of the vasa nervorum.     

Analysis of unsaturated fatty acids of endoneurium and perineurium from normal and degenerating rat sciatic nerve. Morphological correlations

The major portion of the endoneurial lipids is found in myelin. Since perineurial cells differ morphologically from endoneurial cell components, we attempted to determine whether these morphological differences also extended to a difference in fatty acid (FA) composition. Under normal circumstances, unsaturated FAs are more abundant than saturated ones (55-60% of total FAs) in endoneurium and perineurium. A characteristic biochemical difference between these two structures lies in the distribution of linoleic acid (C18:2(n-6)) which represents 20% of total FAs in perineurium and only 2% in endoneurium. Wallerian degeneration takes place after injection of pure glycerol into the endoneurium. This is followed by regeneration characterized by a proliferation of perineurial cells infiltrating the center of the nerve fascicule forming microcompartments. The changes in linoleic acid content reflect these morphological changes. A marked increase in linoleic acid is detected in the endoneurial fraction in parallel with the observed infiltration of perineurial cells into the nerve fascicule.     

Permeability of blood nerve barriers in the diabetic rat

An albumin-Evans blue conjugate has been used as a fluorescent tracer to demonstrate the increased permeability of endoneurial capillaries and perineurial sheath of the sciatic nerve of the alloxan-diabetic rat. The significance of the extravasation of protein into the endoneurial space is discussed in relation to the altered dynamics of the endoneurial microcirculation. It is suggested that tissue hypoxia produced in this way may be a cause of the segmental demyelination which occurs in these nerves.     

PERMEABILITY OF THE PERINEURIUM OF SMALL NERVE FASCICLES: AN ULTRASTRUCTURAL STUDY USING FERRITIN IN RATS

The permeability of the perineurial cells was studied with ferritin as a tracer. It was injected beneath the muscle fascia of the lateral aspect of the fore limb of rats. After various intervals the animals were killed and the distribution of ferritin in the nerves of the underlying muscle was observed. Two minutes after injection, ferritin was found surrounding nerve fascicles of various sizes. Some medium sized fascicles containing ten to twenty myelinated nerve fibres, were bounded by a single layer of perineurial cells. In the endoneurium of these fascicles only very occasional ferritin particles were observed. Ferritin was, however, found in perineurial micropinocytotic vesicles which communicated with the epineurial extracellular compartment and also in some vesicles which were entirely intracellular and in a few vesicles which had an opening towards the endoneurium. Occasionally two or more vesicles had fused with one another. Fifteen minutes after the injection of ferritin, numerous ferritin particles were seen in the endoneurium of all nerve fascicles covered by a single layer of perineurial cells. The vesicles of these cells then frequently contained numerous ferritin particles. Sixty minutes after injection of ferritin the tracer was observed within the endoneurium of fascicles of all sizes, including those with a multilayered perineurium. In the latter, ferritin was seen between all the perineurial cell layers as well as in micropinocytotic vesicles of the perineurial cells. The results of this study strongly indicate that the perineurium of small nerve fasicles is permeable to ferritin and that this permeability is due to pinocytosis in the perineurial cells.     

The dynamics of macrophage recruitment after nerve transection

In the present study rat sciatic nerves (n = 60) were transected; in half of the animals the nerve was allowed to regenerate freely, in the other half the regeneration was prevented by suturing beside the point of transection. Macrophages were stained with ED-1 antibody and counted (number/mm²) in both the epi- and endoneurium 3, 7, 14, 48 and 56 days post transection. Macrophages were observed first in the epineurium; the local density of macrophages was considerably higher in the epineurium than in the endoneurium during the first few days. The number of macrophages in the epineurium was maximal at 3 days (1000–2000/mm²), and thereafter it declined sharply. In the endoneurium macrophages were most abundant after 2 weeks (1000/mm²), after which their number declined steadily. A migration of epineurial macrophages appeared to take place through the perineurium from epineurial areas containing a high concentration of macrophages. Initially an endoneurial accumulation of macrophages was noted in the subperineurial area. These findings suggest an alternative route for macrophages into the endoneurial space. No statistical difference was observed between the regenerating and non-regenerating experimental groups. The present study indicates that regenerating axons do not have an effect on the number of macrophages in either the epineurium or the endoneurium. Received: 21 March 1996 / Revised, accepted: 25 September 1996     

Giant axonal neuropathy: report of two siblings with endocrinological and histological studies

Giant axonal neuropathy in two siblings was reported. The fact that two cases are found in the same family supports this disorder is genetically determined and recessively inherited. These two cases, similar to the cases reported in literature, had chronic peripheral neuropathy and CNS symptoms, and also petit mal absence and mental retardation in elder sister (case 1) and precocious puberty in younger sister (case 2). Sural nerve biopsies in both cases disclosed axonal swellings or giant axons filled with aggregated neurofilaments, and that aggregated intermediate-sized filaments were found within cytoplasm of Schwann cells, endothelial cells of intra and extra-neurial capillaries and of extra-neurial arterioles, perineurial cells and endoneurial fibroblasts. Skin biopsies in both cases disclosed that aggregated intermediate-sized filaments were also found within cytoplasm of fibroblasts, Langerhans' cells, melanocytes and endothelial cells of capillaries, lymphatic vessels and arterioles. The diagnosis of giant axonal neuropathy can be made only by the findings in skin biopsy.     

Perineurial cells filled with collagen in ‘atypical’ Cogan’s syndrome

Cogan’s syndrome is a rare clinical entity characterized by non-infectious interstitial keratitis with vestibuloauditory dysfunction. The clinical course is extremely variable. In the majority of patients, there appears to be an underlying systemic process, often a “vasculitis”. We were able to study for the first time a sural nerve biopsy of a 38-year-old female with clinically suggested Cogan’s syndrome associated with a severe multiplex type of neuropathy. There were unusual cells in or below the perineurium and along perineurial extensions into the endoneurium which were usually associated with blood vessels and which have thus far not been described in association with any type of peripheral neuropathy. The unusual cells were identified as perineurial cells because (1) they were frequently associated with the perineurium and its endoneurial extensions; (2) they were immunoreactive for antibodies against epithelial membrane antigen (EMA) but did not react with antibodies against protein S100, GFAP, and CD 68; and (3) they showed focally accumulated pinocytotic vesicles and hemidesmosomes. Some of these cells were clearly immunoreactive with antibodies against collagen VI. Electron microscopic examination revealed numerous intracellular bundles of collagen fibers which were surrounded by an amorphous basal lamina-like material, indicating that they were located within intracellular projections of the surface membrane. The number of myelinated and unmyelinated nerve fibers was severely reduced corresponding to the clinical manifestation of the neuropathy and to the atrophy, especially of the distal arm and leg muscles. It is concluded that the changes were caused by a special type of autoimmune reaction involving blood vessels and perineurial cells of peripheral nerves.     

Investigating mechanical behaviour at a core-sheath interface in peripheral nerve

As peripheral nerves bend and stretch, internal elements need to move in relation to each other. However, the way in which intraneural components interact is poorly understood. Previous work identified a distinct core and sheath in the rat sciatic nerve and provides a useful model with which to investigate this interaction. Here we have focused on identifying the mechanical and anatomical characteristics of the interface between core and sheath. Nerve samples, 15 and 20 mm long, of rat sciatic nerves were harvested and placed in a purpose-built jig, and a tensile testing machine was used to pull core from sheath. Mechanical tests of specimens in which core had been previously pulled from sheath by 25% of its initial length achieved a mean pull-out force approximately six times smaller than that achieved using intact controls. These results are consistent with the proposal that core-sheath interactions involve physical connections rather than a viscous fluid interface. Anatomical features of this interface were characterised using transmission electron microscopy. It appeared that sheath was derived from epineurium and most of the perineurium, whilst core consisted of endoneurium and a small proportion of the perineurium: the plane of cleavage appeared to involve the innermost perineurial cell layer.     

Perineurial window: demyelination in nonherniated endoneurium with reduced nerve blood flow.

The perineurial window, created by surgical incision of the perineurial sheath allowing its contents to herniate into the epineurial space, provides an experimental model of primary demyelination, the cause of which is unclear. Because the injury is localized and involves distortion of tissue at the lesion site, ischemia is suspected as a cause of demyelination. To study the mechanism of demyelination in the perineurial window model, we measured nerve blood flow (NBF) with a laser Doppler flowmeter before and after perineurial rupture in rat sciatic nerve and assessed the spatial distribution of demyelinated fibers, particularly in the nonherniated portion of the endoneurium. Nerve blood flow at the site of the perineurial window was reduced significantly with an average level of NBF approximately 50% of presurgical values 10 minutes, 60 minutes and 6 hours after surgery. By light microscopic examination, most nerve fibers that herniated through the perineurial window underwent demyelination by 7 days. In addition, focal lesions of subperineurial demyelination were found in the nonherniated endoneurium in the adjacent subperineurial region and proximally and distally to the perineurial window. Endoneurial vessels adjacent to the perineurial incision appeared to be compressed. We suggest that ischemia contributes to the process of demyelination in the perineurial window model.     

A study of the perineurial diffusion barrier of a peripheral ganglion

Summary The perineurial diffusion barrier to horseradish peroxidase (HRP) and ferritin was investigated in superior cervical ganglia of rats and mice. The ganglion was surrounded by a delicate epineurium and 2–5 perineurial lamellae joined by zonulae occludentes and desmosomes.Following local application of tracers the animals were killed after 5, 30, and 60 min and the distribution of HRP and ferritin was studied by light and electron microscopy. The inner layers of the ganglionic perineurium prevented diffusion of both HRP and ferritin into the endoneurium. Owing to the fewness of the perineurial lamellae investing the ganglion. HRP had often extended to the innermost lamella 60 min after application.HRP and ferritin were present in vesicles of ganglionic perineurial cells. There was no passage of tracers via intercellular junctions.     

Pathology of local anesthetic-induced nerve injury

Summary Nerve fiber injury and endoneurial edema were induced by the injection of the local anesthetic 2-chloroprocaine, tetracaine, procaine, etidocaine or mepivacaine into the soft tissue and fascia surrounding the sciatic nerve of Sprague-Dawley rats. Light microscopy demonstrated that the perineurial barrier was not mechanically damaged by the surgical procedure but, at 48 h post-injection, perineurial permeability was increased. Previous observations of leakage of horseradish peroxidase and the present report of neutrophils and eosinophils in the endoneurium indicate a disruption of blood-nerve barrier systems. Endoneurial edema was observed in the subperineurial, interstitial and perivascular regions. Axonal degeneration and demyelination occurred; the latter associated with accumulation of large lipid droplets in Schwann cells. Degranulation of mast cells, proliferation of fibroblasts and macrophage activity were noteworthy in affected areas. The findings are remarkable in that this is the first model of endoneurial edema by a neurotoxin which penetrates the perineurium, disrupting barrier system and inducing nerve fiber injury.Supported in part by USPHS NS18715, NS14162 and NS07078 and the Veteran's Administration Research Service     

Is part of the molecular basis of the perineurial barrier function the lack of endogenous carbohydrate-binding proteins?

The sugar part of cellular glycoconjugates and specific endogenous sugar receptors, i.e., lectins, can establish a system of biological recognition based on protein-carbohydrate interactions. An assortment of labelled (neo)glycoproteins, carrying different types of sugar moieties, is synthesized to localize respective sugar receptors. With these tools, the histochemical patterns of endogenous carbohydrate-binding receptors of the epi-, peri-, and endoneurium were analyzed in human sural and accessory nerves and in swine sciatic nerve. This approach is complementary to the application of plant lectins, focusing on endogenous carbohydrate-binding proteins (lectins). In contrast to the epi- and endoneurium, which bound certain types of carbohydrates, such endogenous sugar receptors were histochemically not detectable in the perineurial cells. Moreover, no histochemical reaction was present in the "connective tissue septa" localized in the endoneurium in which the endoneurial vessels were embedded. This common property supplies evidence that these septa are composed of perineurial cells. They may represent a barrier in addition to the capillary endothelium. Our observations suggest histogenetical differences between the cell populations of epi- and endoneurium vs. perineurium. This significant difference in the ability to bind carbohydrate residues, conjugated to a carrier protein, is contradictory to the assumption that perineurial cells and fibroblasts are functional variants of the same cell type. The histochemical patterns of endogenous carbohydrate-binding receptors found in human and swine nerves were similar but not identical, with exception of the perineurium, reflecting phylogenetic differences in the expression of sugar-binding proteins. The absence of specific sugar receptors in perineurial cells, however, seems to be a more general phenomenon.(ABSTRACT TRUNCATED AT 250 WORDS)     

Fine structural and immunohistochemical identification of perineurial cells connecting proximal and distal stumps of transected peripheral nerves at early stages of regeneration in silicone tubes

Perineurial cells are specialized connective tissue cells that form a barrier between endoneurium and epineurium in normal nerves. In the present study, the formation of the perineurium after transection of rat sciatic nerves was investigated. The cord bridging the gap between proximal and distal stumps through silicone tubes was studied 3, 7, 12, 18, and 21 days after surgery using electron microscopy and antibodies against epithelial membrane antigen (EMA), a marker for perineurial cells that has thus far not been applied to the study of differentiating cells in nerve tubulation systems. Initially, a thin cord consisting of fibrin bridged the gap between the stumps. At 7 days, longitudinal cells had migrated from both stumps toward the center of the tubes on the surface of the fibrin cord. These cells were immunoreactive with anti-EMA. At 12 days, ultrastructural features of perineurial cells (desmosomes, tight junctions, actin filaments with dense bodies, tonofilaments) were prominent in these cells. Subsequently, the gap was bridged through the perineurial tube by endothelial cells, pericytes, fibroblasts, Schwann cells, and axons. At 21 days, a single large nerve fascicle ensheathed by a mature perineurium was found between the stumps. Thus, the first cells to connect proximal and distal stumps in the investigated nerve regeneration silicon chamber system are perineurial cells. Through the tube formed by these cells, blood vessels and nerve fibers bridge the gap. Therefore, establishment of a perineurial connection between nerve stumps appears to be important in the sequence of events during nerve regeneration.The results of this study were presented in part at the Post Graduate Boerhaave Course: Brachial Plexus Injury, Leyden, March 25 and 26, 1993 [17] and at the 38th Annual Meeting of the Deutsche Gesellschaft für Neuropathologie and Neuroanatomie, Berlin, October 6–9, 1993 [27]     

Immunofluorescence studies in a case of rheumatoid neuropathy.

Findings in a sural nerve biopsy from a patient with rheumatoid neuropathy are reported. Inflammatory changes in different stages were observed in epineurial arterioles. Arterioles with fibrinoid necrosis contained IgG, IgM, rheumatoid factor and complement. The same substances were found at the inner site of the perineurial sheaths. It is suggested that these proteins leaked out of endoneurial vessels and became trapped in the perineurial sheaths. With anti-herpes simplex virus serum fluorescence was seen in monuclear cells of infiltrated vessel walls. This virus could not be isolated in tissue culture.     

Inner perineurial cell vulnerability in ischemia

The perineurium of peripheral nerve plays important roles in anatomical organization of fiber groups, in endoneurial fluid homeostasis and in maintenance of tensile strength but little is known about the functional and structural alterations of the perineurium with injury. Large arteries of supply to lower limb of Sprague-Dawley rats were ligated to study the structural reactions of perineurium at 36 h and at 7 days after induction of ischemic injury. Lipid droplets were found to be an early reactive change to ischemia in multiple cell types including perineurial, endothelial and Schwann cells. In peripheral nerve levels showing early myelinated fiber injury the inner perineurial sheath was widened and was undergoing degeneration. The inner layers of perineurial cells showed swelling, organelle disruption and membrane dissolution while outer layers remained intact. Inner perineurial cell degeneration is a prominent early feature of ischemic injury and may be an important mechanism of altered endoneurial homeostasis, fiber function and structure.     

Microangiopathy in human diabetic neuropathy

Summary Morphological change of endoneurial and perineurial vessels accompanied severe loss of myelinated axons in peripheral nerves of each of 17 patients with diabetic neuropathy. Vascular mural thickening averaged 18.9±9.9 m² in diabetic capillaries (n=11) vs. 6.9±4.1 m² in controls (n=7). Electron microscopy revealed vigorous endothelial proliferation as well as thickening and reduplication of basal lamina in each instance. Particular attention was paid to vessels which penetrate the perineurium en route to the endoneurial intertitium, since they provide a major portion of the endoneurial blood supply. Luminal narrowing and mural thickening of these vessels was compounded by basal laminar thickening of the perineurium. Fenestrated endoneurial capillary endothelium was noted in one case. Both demyelination and axonal degeneration were observed with intra-axonal glycogen accumulation in some axons. Morphometric analysis revealed extensive myelinated nerve fiber loss in diabetic nerves. These morphological findings emphasize the impact of diabetic microangiopathy on specialized endothelium and suggest that local anatomic factors in the perineurial sheath render the nerve vulnerable to chronic ischemia.Supported in part by the National Institute for Communicative Disorders and Stroke NS-14162 and by the Veterans Administration Research Service     

Endoneurial fibrosis following nerve transection

Summary Indirect immunofluorescent techniques with antibodies to type I, III, and V collagens and fibronectin were used to study rat sciatic nerve tributaries after transection with intact contralateral nerves as controls.Codistribution of type I and III collagens characterized the epineurium of normal nerve. In the perineurium, however, type I collagen was absent, but type III and V collagens and fibronectin were detected. Type I and III collagens were codistributed in the endoneurial stroma where a homogeneous staining by antibodies to fibronectin was also observed.During the 4-week observation period after transection the perineurium reacted by slight thickening which was most clearly demonstrated by staining with antibodies to fibronectin and to type V collagen. A widening of the type I-negative cleft also occurred. Endoneurial, type V collagen-positive cuffs around the nerve fibers became disorganized, and a concomitant increase of the stroma containing type I and III collagens and fibronectin was observed.The codistribution of the fibrous collagen types appeared similar in normal epineurium and endoneurium. Type V collagen was locatd in the perineurium and in endoneurial cuffs surrounding the nerve fibers. The present data indicate that collagen accumulation takes place in the perineurium and endoneurium of transected nerve. The cell type responsible for the synthesis of the connective tissue material is discussed.Financially supported by a grant (to V. S.) from the Research and Science Foundation of Lääke Oy and by institutional grants from the Turku University Foundation and the Sigrid Jusélius Foundation     

Role of axon-deprived Schwann cells in perineurial regeneration in the rat sciatic nerve

The role of Schwann cells (SC) in perineurial regeneration after nerve injury has not yet been resolved. It was hypothesized that SC alone are able to induce at least partial morphological restoration of the destroyed orthotopic perineureum (PN). To test the hypothesis, a permanently denervated segment of the rat sciatic nerve was made acellular by freeze-thawing, except in its most proximal part where non-neuronal cells were left intact. Restoration of the frozen segment by these cells was examined by electron microscopy and immunohistochemistry of the SC marker, S-100 protein, 4 and 8 weeks after injury. The PN regenerated from undifferentiated fibroblast-like cells. In the presence of migrant SC without axons, regenerated cells in the place of the former PN were stacked in several layers and, in accordance with the hypothesis, partially expressed typical features of the perineurial cells (PC): pinocytotic vesicles, short fragments of basal lamina and tight junctions. Migrant SC induced formation of pseudo-minifascicles even in the epineurium. In these, SC organized the adjacent fibroblasts into a multilayered circular sheath, and induced their partial differentiation towards perineurial cells. Further experiments demonstrated that regenerating axons are required for complete morphological differentiation of the regenerated perineurial cells either in the orthotopic PN or in minifascicles.