Perivascular Channels Research Articles

Abstract TP452: Systemic Inflammation Reverses Stroke-induced Changes in Astrocytic Aquaporin-4 Expression

Introduction: Aquaporin-4 (AQP4) is a perivascular water channel that plays a key role in regulating the blood-brain barrier (BBB). Stroke acutely reduces expression and polarization of AQP4 on astrocytic endfeet, and these vascular changes might persist chronically and evolve with systemic inflammation. Reduced polarization of AQP4 after traumatic brain injury (TBI) leads to impaired perivascular clearance of debris and cognitive decline in mice, and if similar changes occur after stroke, then they may also contribute to post-stroke dementia. Hypothesis: Stroke will cause lasting changes in vascular organization, reducing astrocytic AQP4 expression and polarization in the area of injury. Methods: Thirty male Balb/cJ mice were subjected to distal middle cerebral artery ligation, allowed to survive for 28 days, and sacrificed 24 hours after injection with lipopolysaccharide (LPS, 0.33 mg/kg i.p.) or saline to model systemic inflammation. Three sequential coronal sections (20 μm) near the site of infarct (0.25 mm from Bregma) were stained with lectin, AQP4, and glial fibrillary acidic protein (GFAP) to evaluate BBB organization. Vessel density was evaluated with AngioTool. AQP4 expression was evaluated by Western blot. Results: Stroke increased vascular density in the injured hemisphere (11 ± 0.8% vs 7.6 ± 0.8% in sham mice, p < 0.05). Decreased AQP4 expression and polarization near the site of the infarct was observed, which were reversed with LPS (Figure 1). LPS increased AQP4 protein levels in stroke (2.5 ± 0.37 fold-increase, p < 0.05), but not sham mice. Conclusions: Stroke caused a chronic increase in vascular density, with peri-infarct vessels displaying decreased expression and changed polarity of AQP4. However, with systemic inflammation, this is reversed, potentially contributing to neuroinflammation and worsened cognitive outcomes in stroke patients who encounter infections via increased permeability of these abnormal vessels.

Commentary: Structural and functional features of central nervous system lymphatic vessels.

Commentary: Structural and functional features of central nervous system lymphatic vessels.

Mechanisms of astrocytic K(+) clearance and swelling under high extracellular K(+) concentrations.

In response to the elevation of extracellular K(+) concentration ([K(+)]out), astrocytes clear excessive K(+) to maintain conditions necessary for neural activity. K(+) clearance in astrocytes occurs via two processes: K(+) uptake and K(+) spatial buffering. High [K(+)]out also induces swelling in astrocytes, leading to edema and cell death in the brain. Despite the importance of astrocytic K(+) clearance and swelling, the underlying mechanisms remain unclear. Here, we report results from a simulation analysis of astrocytic K(+) clearance and swelling. Astrocyte models were constructed by incorporating various mechanisms such as intra/extracellular ion concentrations of Na(+), K(+), and Cl(-), cell volume, and models of Na,K-ATPase, Na-K-Cl cotransporter (NKCC), K-Cl cotransporter, inwardly-rectifying K(+) (KIR) channel, passive Cl(-) current, and aquaporin channel. The simulated response of astrocyte models under the uniform distribution of high [K(+)]out revealed significant contributions of NKCC and Na,K-ATPase to increases of intracellular K(+) and Cl(-) concentrations, and swelling. Moreover, we found that, under the non-uniform distribution of high [K(+)]out, KIR channels localized at synaptic clefts absorbed excess K(+) by depolarizing the equivalent potential of K(+) (E K) above membrane potential, while K(+) released through perivascular KIR channels was enhanced by hyperpolarizing E K and depolarizing membrane potential. Further analysis of simulated drug effects revealed that astrocyte swelling was modulated by blocking each of the ion channels and transporters. Our simulation analysis revealed controversial mechanisms of astrocytic K(+) clearance and swelling resulting from complex interactions among ion channels and transporters.

Occurrence and co-localization of amyloid β-protein and apolipoprotein E in perivascular drainage channels of wild-type and APP-transgenic mice

The deposition of the amyloid β-protein (Aβ) is a hallmark of Alzheimer's disease (AD). One reason for Aβ-accumulation and deposition in the brain may be an altered drainage along perivascular channels. Extracellular fluid is drained from the brain towards the cervical lymph nodes via perivascular channels. The perivascular space around cerebral arteries is the morphological correlative of these drainage channels. Here, we show that Aβ is immunohistochemically detectable within the perivascular space of 25 months old wild-type and amyloid precursor protein (APP)-transgenic mice harboring the Swedish double mutation driven by a neuron specific promoter. Only small amounts of Aβ can be detected immunohistochemically in the perivascular space of wild-type mice. Cerebrovascular and parenchymal Aβ-deposits were absent. In APP-transgenic mice, large amounts of Aβ were found in the perivascular drainage channels accompanied with cerebrovascular and parenchymal Aβ-deposition. The apolipoprotein E (apoE) immunostaining within the perivascular channels did not vary between wild-type and APP-transgenic mice. Almost 100% of the area that represents the perivascular space was stained with an antibody directed against apoE. Here, Aβ co-localized with apoE indicating an involvement of apoE in the perivascular clearance of Aβ. Fibrillar congophilic amyloid was not seen in wild-type mice. In APP-transgenic animals, congophilic fibrillar amyloid material was seen in the wall of cerebral blood vessels but not in the perivascular space. In conclusion, our results suggest that non-fibrillar forms of Aβ are drained along perivascular channels and that apoE is presumably involved in this clearance mechanism. Overloading such a clearance mechanism in APP-transgenic mice appears to result in insufficient Aβ-clearance, increased Aβ-levels in the brain and the perivascular drainage channels, and finally in Aβ-deposition. In so doing, our results strengthen the hypothesis that an alteration of perivascular drainage supports Aβ-deposition and the development of AD.

P1-029: Occurrence of amyloid β-protein and co-localization with apolipoprotein E in perivascular drainage channels of wild-type and APP-transgenic mice

P1-029: Occurrence of amyloid β-protein and co-localization with apolipoprotein E in perivascular drainage channels of wild-type and APP-transgenic mice

A 3’-UTR polymorphism in the oxidized LDL receptor 1 gene increases Aβ40 load as cerebral amyloid angiopathy in Alzheimer’s disease

It is presently unclear whether polymorphic variations in the oxidized low-density lipoprotein receptor 1 (OLR1), or low-density lipoprotein receptor-related protein 1 (LRP1), genes act as risk factors for Alzheimer's disease (AD). In the present study, we have investigated the extent of amyloid beta protein (Abeta) deposition as cerebral amyloid angiopathy (CAA) or senile plaques (SP) in relationship to OLR1 +1071 and +1073 polymorphisms and LRP1 C766T polymorphism in patients with AD There was an increased Abeta40 load as CAA, but not as SP, in frontal cortex of AD patients carrying OLR1+1073 CC genotype, compared to those with CT, TT or CT+TT genotypes, but only in those individuals without apolipoprotein (APOE) epsilon4 allele. No differences in total Abeta or Abeta42 load as CAA or SP between OLR1+1073 genotypes was seen, nor were there any differences between OLR1+1071 and LRP1 genotypes for any measure of Abeta. Present data suggests that homozygosity for the C allele for OLR1+1073 polymorphism, selectively in individuals without APOE epsilon4 allele, may impair clearance of Abeta, and particularly Abeta40, from the brain across the blood-brain barrier, leading to its 'diversion' into perivascular drainage channels, thereby increasing the severity of CAA in such persons.

Expression pattern of matrilins and other extracellular matrix proteins characterize distinct stages of cell differentiation during antler development

Deer antler regeneration is a uniquely intense and complex process, which involves chondrogenic and intramembranous ossification. Cell differentiation in the developing antler of red deer, Cervus elaphus, was characterized with extracellular matrix markers. Expression of the four matrilin genes was monitored by immunohistochemistry and in situ hybridization and compared to cartilage markers collagen II and cartilage link protein, the bone component collagen I, and the endothelial basement membrane constituent laminin. The mesenchyme layer at the very tip of the velvet antler was enriched in link protein, indicative of the role of hyaluronan in apical morphogenesis. Matrilin-2, formerly described as a component of hard and soft connective tissue matrices, was identified here also as a marker of cells with high differentiation potential: it is expressed predominantly by mesenchyme cells, prechondrocytes and preosteoblasts. In addition to matrilin-3, documented as a component of the bony extracellular matrix, expression of the other three matrilin genes was observed in osteoprogenitor cells and osteoblasts. A layer of presumed osteoprogenitor cells, which surrounded the perivascular channels, expressed all four matrilins and collagen I. As a consequence, all four matrilins, including matrilin-1, previously detected in the skeleton only in cartilage, were found associated to collagen I-rich structures in a thin layer bordering the columns of hypertrophic chondrocytes. Cells with similar morphology and expression pattern were identified in the periosteum. Altogether all cell types of the chondrogenic and osteogenic lineage that expressed the four matrilins were in a separate study [Faucheux, C., Nicholls, B.M., Allen, S., Danks, J.A, Horton, M.A., Price, J.S., 2004. Recapitulation of the parathyroid hormone-related peptide-Indian hedgehog pathway in the regenerating deer antler. Dev. Dyn. 231, 88-97] positive for parathyroid hormone-related peptide and its receptor.

High endothelial venules of the rat lymph node. A review and a question: is their activity antigen specific?

The high endothelial venules (HEVs) of the lymph nodes are sites for transvascular lymphocyte traffic. Due mostly to the wide scale of variations manifested by the HEVs and to frequently restricted conditions of observation, reports often differed on their morphological or functional features, which has led to opposing views on aspects of the functioning of HEVs. In the present review, we analyze previous reports and attempt to derive comprehensive proposals to reconcile variations in actual observations under diverse conditions. This analysis shows that the features typical of the HEV endothelial cells (HEV cells) are stimulated to emerge by antigens and the proper lymphocytes and mediators. The stimulation would implicate drained lymphocytes migrating in the perivascular channel, immediately cuffing an HEV's endothelium. A marked pleomorphism of HEV cells betrays the fact that they undergo individual stimulation and a somewhat heterogeneous activity. Other facts indicate that the subendothelial spaces of HEV cells are sites of interactions between drained lymphocytes, HEV cells, and recruited blood lymphocytes. Facts also reveal time- and site-related variations in the intensity of recruitment of blood lymphocytes by HEV cells and topographically related variations in the nature of the recruited cells. Analysis of some other observations, often ignored, lead to the conclusion that recruitment of lymphocytes by HEV cells for the sake of participating in local specific immune activities is antigen specific, despite the implication of homing receptors of lymphocytes.

High endothelial venules of the rat lymph node. A review and a question: Is their activity antigen specific?

Background The high endothelial venules (HEVs) of the lymph nodes are sites for transvascular lymphocyte traffic. Due mostly to the wide scale of variations manifested by the HEVs and to frequently restricted conditions of observation, reports often differed on their morphological or functional features, which has led to opposing views on aspects of the functioning of HEVs. Methods In the present review, we analyze previous reports and attempt to derive comprehensive proposals to reconcile variations in actual observations under diverse conditions. Results This analysis shows that the features typical of the HEV endothelial cells (HEV cells) are stimulated to emerge by antigens and the proper lymphocytes and mediators. The stimulation would implicate drained lymphocytes migrating in the perivascular channel, immediately cuffing an HEV's endothelium. A marked pleomorphism of HEV cells betrays the fact that they undergo individual stimulation and a somewhat heterogeneous activity. Other facts indicate that the subendothelial spaces of HEV cells are sites of interactions between drained lymphocytes, HEV cells, and recruited blood lymphocytes. Facts also reveal time- and site-related variations in the intensity of recruitment of blood lymphocytes by HEV cells and topographically related variations in the nature of the recruited cells. Conclusions Analysis of some other observations, often ignored, lead to the conclusion that recruitment of lymphocytes by HEV cells for the sake of participating in local specific immune activities is antigen specific, despite the implication of homing receptors of lymphocytes. © 1996 Wiley-Liss, Inc.

A scanning electron-microscopic study of the rat thymus with special reference to cell types and migration of lymphocytes into the general circulation.

The three-dimensional structure of the rat thymus was studied by combined scanning- and transmission electron microscopy. The thymus consists mainly of four types of cells: epithelial cells, lymphocytes, macrophages, and interdigitating cells (IDCs). The epithelial cells form a meshwork in the thymus parenchyma. Cortical epithelial cells are stellate in shape, while the medullary cells comprise two types: stellate and large vacuolated elements. A continuous single layer of epithelial cells separates the parenchyma from connective tissue formations of the capsule, septa and vessels. Surrounding the blood vessels, this epithelial sheath is continuous in the cortex, while it is partly interrupted in the medulla, suggesting that the blood-thymus barrier might function more completely in the cortex. Cortical lymphocytes are round and vary in size, whereas medullary lymphocytes are mainly small, although they vary considerably in surface morphology. Two types of large wandering cells, macrophages and IDCs, could be distinguished, as well as intermediate forms. IDCs sometimes embraced or contacted lymphocytes, suggesting their role in the differentiation of the latter cells. Perivascular channels were present around venules and some arterioles in the cortico-medullary region and in the medulla. A few lymphatic vessels were present in extended perivascular spaces. The present study suggests the possible existence of two routes of passage of lymphocytes into the general circulation. One is via the lymphatics, while the other is through the postcapillary venules into the blood circulation. Our SEM images give evidence that lymphocytes use an intracellular route, i.e., the endothelium of venules.

The immune system and the nervous system.

The immune system may interfere with brain function. The central nervous system may also influence the activity of the immune system. The central nervous system is functionally protected by the blood-brain barrier. The central nervous system is functionally protected by the blood-brain barrier. The endothelial cells of the brain capillaries are linked by tight junctions, resulting in an almost continuous interior wall which restricts the transfer of plasma proteins. The barrier function is modified by inflammatory meningeal lesions, stroke and epileptic seizures. Antigenic material may penetrate the barrier and enter the nerve tissue. The phagocytic cells in the central nervous system are mainly of haematogenous origin. The number of such cells in the brain is very low. There are also few lymphocytes under normal circumstances. These cells circulate from the blood, through the vessel walls and into the perivascular spaces, along the perivascular channels and to the CSF and back to the blood. This circulation may increase enormously during inflammatory conditions. In multiple sclerosis, the number of T-lymphocytes in the CSF is increased, corresponding to a preponderance of T-lymphocytes in the perivascular cell infiltrates in and around the lesions. Thus, the individual elements of the immune system are all present in the brain, which is only partially immunologically privileged. The mechanisms underlying the brain's immunological privilege may be of a non-immunological nature. As yet there are only few data which indicate that auto-immunity is a prominent feature in diseases of the human brain. The central nervous system also exerts a modulating influence upon the immune response. This may take place both by secretion of hormones and by a nervous/neurotransmitter influence upon the immune system.

Thymic cell migration in the subnodular spaces of draining lymph nodes of rats

Little is known about the pathway of possible lymphocyte traffic in the nodules (germinal centers) of the node. Observations reported here indicate the involvement of the subnodular spaces. These spaces, which constitute the inner limit of the nodules, are contiguous to the perivascular channels of the postcapillary venules which partially encircle each nodule. Twelve hours after a local transfer of labeled thymic cells, they were observed in the subnodular spaces and in the perivascular channels of the draining nodes. It is proposed that the spaces and the channels provide a pathway for the rapid migration of lymphocytes entering a node via the afferent lymph and, probably, carrying an immunogenic information. The pathway would permit these cells to transmit the information rapidly to the appropriate cell population(s) of a node draining a tissue undergoing an immunological process.

Localization of adenylate cyclase in the neurointermediate lobe of the rat pituitary

The ultrastructural localization of adenylate cyclase was accomplished cytochemically in the neurointermediate lobe of Sprague-Dawley rats. The main concentration of the reaction product was found on the plasmalemma of neurosecretory nerve fibers, their terminals and plasma membranes of pituicytes. Positive reaction for adenylate cyclase was found less regularly in endothelial cells, pericapillary spaces and processes of the basal lamin. The septum between the pars nervosa and the pars intermedia showed heavy deposits of the reaction product, especially around the neurosecretory nerve fibers but also around other types of nerve fibers. Reaction for adenylate cyclase was not seen in the cells of the pars intermedia. When the substrate (ATP) was omitted, no reaction product was found. These findings support the suggestion of an involvement of cyclic AMP in the release mechanism of neurohypophysial of an involvement of cyclic AMP in the release mechanism of neurohypophysial hormones from the neurosecretory nerve terminals, and possibly also their transfer into blood vessels and perivascular channels.

Canine Lung Allograft Lymphatic Alterations

Lymphatic obstruction has not been emphasized as a feature of lung allograft rejection. However, accumulation of fluid and cellular infiltrate, aggravated by lymph stasis, results in impaired lung function. In this study, lung specimens were recovered at varying times up to 133 days after either reimplantation (7 dogs) or allografting (29 dogs). Azathioprine and prednisone were administered to 17 allograft recipients. The presence of abnormally dilated perivascular, peribronchiolar, and subpleural lymphatic channels was a consistent histological finding, most striking in specimens recovered from untreated allograft recipient dogs. Attenuated lymphatic alterations were noted in immunosuppressed allograft recipients. In these animals the pulmonary lymphatics seemed to be ineffectual in clearing the allograft of the accumulating cellular infiltrates and fluid during rejection.

The glio-vascular system of cephalopods

The branches of the cerebral arteries run to the centre of each lobe of the brain and from there radiate outwards. The arteries are lined by endothelial cells, surrounded by pericytes. Outside these are large extracellular spaces containing collagen. These spaces continue as a system of ‘glio-vascular’ channels among the tissues. These channels contain collagen and other extracellular material and nuclei belonging probably to muscle cells and fibroblasts. This system permeates the neuropil and in the cell layers provides wrappings for the perikarya. The channels and extracellular material form tunnels of ‘trophospongium’ within the neuronal cytoplasm. Glial fingers also penetrate into these channels but are not well seen by light microscopy. The system of spaces among the tissues communicates with an elaborate set of branching ‘lymphoid' channels. These collect into veins either in the membrane around the brain or at the centre of the optic lobe. The veins discharge to the pharyngo-ophthalmic vein and the brain is thus provided with a venous system partly isolated from the main haemocoele. The true neuroglia cells are of two types. The protoplasmic glia cells have much-branched processes, especially in the neuropil. They have abundant end-feet attached to the outside of the perivascular channels and to the glio-vascular strands. They occur in the cell layers and their processes probably penetrate the neurons. The fibrous glia have very long processes, mostly smooth and with little branching. The processes run in bundles accompanying the main tracts of nerve fibres in the neuropil and extending into the cell layers.

Histological and ultrastructural study of the "corticomedullary perivenular lymphoid sheaths" of the mouse thymus gland

The present study is concerned with morphological investigations on some special vascular structures of the mouse thymus pertinent to the problem of «cellular traffic» in this organ. Thymuses from 20 normal mice of C3Hf/Gs and C57BL strains, ranging in age from 30 to 50 days, were examined by both light and electron microscopy. At the cortico-medullary junction wide post-capillary venules are present, running parallel to the border between cortex and medulla. They receive at right angles narrow capillaries coming especially from the cortex. The wall of these venules is formed of flattened endothelium, its basement membrane and epithelial reticular cells arranged as adventitial cells. The basement membrane of the endothelium is more or less regularly split in two layers (the inner «vascular» layer and the outer «parenchymal» layer) thus delineating a perivascular space containing one or more rows of lymphoid cells. This space, which has on the whole the character of a cylindrical perivascular channel, accompanies the cortico-medullary venules, although not necessarily along their entire course, and it is recognizable even around the cortical capillaries draining into them. The two layers of the basement membrane are lined on the inner side by the laminar cytoplasm of epithelial reticular cells, often interconnected by desmosomes. The ultrastructural characteristics of the perivascular space do not support the hypothesis of its lymphatic nature, as maintained by other authors. Lymphocytes were found in the process of migrating through either the «parenchymal» or «vascular» layer of the basement membrane. Although the direction of migration cannot be determined by static pictures, some morphological data seem to suggest that lymphocytes, at the level of the cortico-medullary venules, move from the thymic parenchyma towards the perivascular space and from there into the circulating blood stream. It is suggested that the perivascular apparatus be termed «cortico-medullary perivenular lymphoid sheath».

SOME OBSERVATIONS ON THE ULTRASTRUCTURE OF THE ADENOHYPOPHYSIS OF THE RABBIT.

SUMMARY The ultrastructure of the adenohypophysis of the rabbit is described preliminary to reporting changes after experimental procedures. Fixation by perfusing with gluteraldehyde enabled selected regions of the gland to be removed with accuracy. Separate descriptions of the pars distalis proper, zona tuberalis, pars tuberalis and pars intermedia are therefore included. In the pars distalis proper four types of granular cell were recognized although their function cannot be accurately determined. For convenience, therefore, they have been designated 1, 2, 3 and 4. In addition a fifth type of cell (type 5) is described which is also present in the other areas. This cell, as well as having possible phagocytic functions, appears to be concerned in the formation of a perivascular channel. Two types of cell are recognized in the zona tuberalis, which are similar in appearance to the 3 and 4 cells of the pars distalis, although not necessarily identical in function. The characteristic cells of the pars tuberalis are rich in cytoplasmic RNA and contain large numbers of intracellular fibrils. It is suggested that the ribosomes are concerned in the synthesis of a sedentary protein which may take the form of the microfibrils. The pars intermedia contains a predominant cell type with large granules of varying density. The relationship of these granules to the specific hormone is discussed.

Cytologic Features and Cellular Migration in the Cortex and Medulla of Thymus in the Young Adult Rat

Abstract The cells of the cortex and medulla of thymus and their mitoses were described in 10-week old (200 Gm.) male rats. The four main cell types found in cortex (reticular cells, large, medium and small lymphocytes were related to each other by transitional cell types. The presence of numerous mitoses of the four cell types indicates rapid cell production. Since the thymus of 10-week old rats is not growing and, therefore, each cell population must be in a steady state, the mitotic production of new cells of a given type must be balanced by transformation into cells of another type or by emigration out of the cortex. Evidence is presented in support of the transformation of reticular cells into large lymphocytes; and of these into lesser and lesser sized lymphocytes. As for small lymphocytes, the evidence indicates that they migrate form cortex to medulla. The medulla contains numerous small lymphocytes, some reticular cells, and rare large and medium lymphocytes. The lymphocytes of the medulla (presumed to have migrated from the cortex) often show nuclear processes which are attributed to ameboid motion. The medulla of the rat thymus contains many blood vessels, most of which are enclosed within "perivascular channels." Diapedesis of lymphocytes, chiefly small ones, is frequently seen across the walls of both the perivascular channels and the blood vessels themselves. Furthermore, higher counts of lymphocytes in venous than in arterial blood of thymus indicate that these cells directly enter the blood circulation. In conclusion, cells of the lymphocytic series are produced by mitosis in the cortex of the thymus. The evidence indicates that lymphocytes arising in this region migrate into the medulla. Thence, these cells pass into perivascular channels and into the enclosed blood vessels to reach the circulation.

PERIVASCULAR SPACES OF THE MAMMALIAN CENTRAL NERVOUS SYSTEM

SummaryThe advances in the knowledge of the histology of the central nervous system which have been made possible by improved staining techniques have complicated and confused the anatomical conception of the perivascular and perineuronal spaces. As a first step towards the resolution of this confusion, a clear picture must be formed of the elements which intervene between the tunica media of the cerebral and spinal blood vessels and the neurons.The blood vessels in the central nervous system possess little or no adventitial coat, and this element is replaced by a reticular perivascular sheath continuous with the pia‐arachnoid envelope of the brain and spinal cord. The question as to whether this reticular sheath extends to cover the capillaries remains to some extent unanswered, though the probability is that it does not do so. No clearly defined tissue layer serves to separate the outer wall of the reticular perivascular sheath from the neuron and its processes. In fixed preparations a felted network, formed by the perivascular feet of the neuroglia, which resembles a membrane, is seen, but we do not accept the view that this structure is part of the normal anatomy of the central nervous system. There remains an element on which little stress has previously been laid: this is the ground substance of the central nervous system, a tissue which is probably a muco‐polysaccharide in constitution, and which by virtue of its physical and chemical properties may well be of considerable importance in relation to the physiology of the central nervous system.The views of the earlier authorities on the anatomy of the perivascular and perineuronal spaces are difficult to reconcile with modern knowledge of the histology of the tissues concerned, and are of interest chiefly because of the influence which they have had on descriptions of these structures which are still accepted to‐day. After the first description of the perivascular space by Pestalozzi, important landmarks in the subsequent history of the spaces were: the adoption of the term ‘Virchow‐Robin space’ for the perivascular space; the description by His of a second space external to this; the discovery by Obersteiner that the perineuronal spaces were continuous with the space of His, and the confusion by later workers of the space of His with that of Virchow‐Robin, so that the perineuronal spaces were described as communicating through a perivascular canalicular system with the subarachnoid space.The most influential figure in moulding the modern conception of the perivascular and perineuronal spaces has been Weed. It appears that he accepted the notion, prevailing at the time at which his researches were carried out, of a complete canalicular system of spaces opening into the subarachnoid space; and it was unfortunate that the limitations of his Prussian blue technique were such as to lead him to support this misconception.Schaltenbrand & Bailey and Patek did much to resolve the confusion produced by the existence of the two systems of spaces, the true perivascular space of Virchow‐Robin and the artifact space of His or Held. Schaltenbrand & Bailey regarded the combined outer wall of the reticular perivascular sheath and glial membrane (their ‘Piaglialmembran’) as separating the two systems of spaces. Patek, who used an improved version of Wee?s technique, described one true perivascular space, the Virchow‐Robin space, with which the perineuronal spaces did not communicate, and three artifact spaces, with one of which, that between the glial membrane and the brain substance, the perineuronal spaces were in communication.The consideration of the views of previous workers, taken in conjunction with our own observations, leads us to believe that there are two systems of perivascular spaces: (1) the true perivascular spaces bounded by the layers of the reticular perivascular sheath; and (2) a great system of artifact spaces produced by shrinkage in the preparation of histological material and extending from the perineuronal spaces through the artifact perivascular spaces to the epispinal spaces of His.Such a conception of the histology of the perivascular and perineuronal spaces must, if accepted, alter some modern views on certain problems in connexion with the anatomy and physiology of the central nervous system. We suggest that the perivascular spaces serve neither for the production nor the absorption of the cerebro‐spinal fluid, but, being channels in which there is no established flow in one direction, act as cushions between the expansile vessels and the nerve cells. We also believe that our investigations show that the anatomical arrangements are such that the perivascular channels can play no significant part in the metabolism of the neuron.Whilst the perivascular and perineuronal spaces have in all probability little relevance to the problems of the blood‐brain barrier and the anatomy of the synapse, it is evident that the importance of the ground substance in relation to these problems has not been adequately appreciated.

FORCED SPINAL DRAINAGE IN ITS RELATION TO INFECTIONS OF THE CENTRAL NERVOUS SYSTEM

In 1919, Weed and McKibbon 1 demonstrated that the intravenious injection of a hypotonic solution increased the intracranial pressure. In 1923 and 1924, Weed 2 found that the intravenous injection of a hypotonic solution in animals caused a hydrosis of the central nervous system. The essential changes were a hydrosis of the cells in the choroid plexus, a collection of fluid in the perineural spaces, a dilatation of the perivascular channels that lead into the subarachnoid space and an increased intracranial pressure. In 1928, Kubie 3 demonstrated that if a hypotonic solution is injected into the blood stream of animals there is a marked increase in intracranial pressure and a striking hydration of the central nervous system. If, however, a needle is introduced into the subarachnoid space and the increased fluid is allowed to drain, there is no rise in intracranial pressure and no hydrosis of the central nervous system