Vascular permeability (problem of the blood-brain barrier) in the pineal organ of the rainbow trout,Salmo gairdneri

Summary The problem of the blood-brain barrier in the pineal organ of the rainbow trout, Salmo gairdneri, was investigated following intraperitoneal or intracardial injections of several tracers and dyes with different molecular weights. As demonstrated at the light-microscopic level, repeated injections of trypan blue or horseradish peroxidase (HRP) resulted in an accumulation of these substances in the pineal epithelium (parenchyma). By use of the electron microscope, HRP was found in electron-dense bodies, probably lysosomes, in (i) the endothelial cells and perivascular macrophages 4 h after intraperitoneal injection, (ii) the supporting cells and intrapineal or luminal macrophages 8 h after injection, and (iii) the receptor cells 24 h after injection of the tracer. Ferritin particles penetrated the fenestrated endothelium of pineal capillaries. They were confined to vesicles, vacuoles and the smooth endoplasmic reticulum of the supporting cells as well as to the synaptic vesicles and the smooth endoplasmic reticulum of the pineal photoreceptors. The intercellular passage of tannic acid mixed with the fixative was blocked at the luminal junctional complex separating the pineal lumen from the basal portion of the pineal epithelium. The passive intercellular transport of substances with high molecular weight from the bloodstream to the cerebrospinal-fluid compartment is thus prevented. However, no blood-brain barrier exists for exogenously administered proteins, which are rapidly taken up by pineal cells and actively transported in a transcellular manner.The findings on the blood-brain barrier of the pineal organ of the rainbow trout are discussed with particular reference to the endocrine capacity of pineal sensory organs.Fellow of the Alexander von Humboldt Foundation, Federal Republic of Germany.     

An ultrastructural study of the blood/air barrier in the guinea-pig

Horseradish peroxidase (HRP) was intravenously injected into guinea-pigs to ultrastructurally examine the permeability of the blood/air barrier. Adults were given 300 mg/kg of the tracer in a small volume of saline, anesthetized and sacrificed at intervals by either intratracheal filling or right ventricular perfusion with 3% glutaraldehyde. The reaction product had passed through endothelial clefts and accumulated in the interstitium as early as 1.5 min after injection. This same degree of penetration occurred with either fixation method used. Tight junctions between pneumocytes prevented passage of the reaction product into alveoli. Pinocytotic vesicles were numerous in both endothelial and epithelial cells, but did not significantly contribute to tracer transport. Ten minutes post-injection was selected as optimal for this model since the highest concentration of tracer was found in the tissues at this time.     

Tubular invaginations in cerebral endothelium and their relation to smooth-surfaced cisternae in hagfish (Myxine glutinosa)

Summary The organization of vesicular profiles in the endothelium of cerebral capillaries of the hagfish, Myxine glutinosa, has been reinvestigated. Judged from random thin sections the endothelial cells contain numerous vesicles and tubules, in contrast to brain endothelia of most other vertebrates. However, three-dimensional reconstructions based on ultrathin serial sections (thickness 18 nm) showed that the profiles represent a system of irregular tubular invaginations of the cell membrane, comparable to the vesicular invaginations demonstrated in extracerebral capillary endothelia of frogs and rats. In addition, smooth-surfaced cisternae were present in close relation to the invaginations. The function of endothelial invaginations is unknown. They do not transport macromolecules, because the blood-brain barrier is practically impermeable to proteins. However, since the system of the invaginations and smooth-surfaced cisternae is structurally similar to the system of caveolae and sarcoplasmic reticulum in smooth muscle cells, a common function seems likely. It is proposed that endothelial invaginations and smooth-surfaced cisternae are involved in regulation of cytosolic Ca⁺⁺-concentration.     

The complementary membranes forming the blood-brain barrier

Brain capillary endothelial cells form the blood-brain barrier. They are connected by extensive tight junctions, and are polarized into luminal (blood-facing) and abluminal (brain-facing) plasma membrane domains. The polar distribution of transport proteins allows for active regulation of brain extracellular fluid. Experiments on isolated membrane vesicles from capillary endothelial cells of bovine brain demonstrated the polar arrangement of amino acid and glucose transporters, and the utility of such arrangements have been proposed. For instance, passive carriers for glutamine and glutamate have been found only in the luminal membrane of blood-brain barrier cells, while Na-dependent secondary active transporters are at the abluminal membrane. This organization could promote the net removal of nitrogen-rich amino acids from brain, and account for the low level of glutamate penetration into the central nervous system. Furthermore, the presence of a gamma-glutamyl cycle at the luminal membrane and Na-dependent amino acid transporters at the abluminal membrane may serve to modulate movement of amino acids from blood-to-brain. Passive carriers facilitate amino acid transport into brain. However, activation of the gamma-glutamyl cycle by increased plasma amino acids is expected to generate oxoproline within the blood-brain barrier. Oxoproline stimulates secondary active amino acid transporters (Systems A and B(o)+) at the abluminal membrane, thereby reducing net influx of amino acids to brain. Finally, passive glucose transporters are present in both the luminal and abluminal membranes of the blood-brain barrier. Interestingly, a high affinity Na-dependent glucose carrier has been described only in the abluminal membrane. This raises the question whether glucose entry may be regulated to some extent. Immunoblotting studies suggest more than one type of passive glucose transporter exist in the blood-brain barrier, each with an asymmetrical distribution. In conclusion, it is now clear that the blood-brain barrier participates in the active regulation of brain extracellular fluid, and that the diverse functions of each plasma membrane domain contributes to these regulatory functions.     

Analog of vertebrate anionic sites in blood-brain interface of larval Drosophila

The blood-brain barrier ensures brain function in vertebrates and in some invertebrates by maintaining ionic integrity of the extraneuronal bathing fluid. Recent studies have demonstrated that anionic sites on the luminal surface of vascular endothelial cells collaborate with tight junctions to effect this barrier in vertebrates. We characterize these two analogous barrier factors for the first time on Drosophila larva by an electron-dense tracer and cationic gold labeling. Ionic lanthanum entered into but not through the extracellular channels between perineurial cells. Tracer is ultimately excluded from neurons in the ventral ganglion mainly by an extensive series of (pleated sheet) septate junctions between perineurial cells. Continuous junctions, a variant of the septate junction, were not as efficient as the pleated sheet variety in blocking tracer. An anionic domain now is demonstrated in Drosophila central nervous system through the use of cationic colloidal gold in LR White embedment. Anionic domains are specifically stationed in the neural lamella and not noted in the other cell levels of the blood-brain interface. It is proposed that in the central nervous system of the Drosophila larva the array of septate junctions between perineurial cells is the physical barrier, while the anionic domains in neural lamella are a charge-selective barrier for cations. All of these results are discussed relative to analogous characteristics of the vertebrate blood-brain barrier.     

THE INFLUENCE OF INTRAVASCULAR FLUID VOLUME ON THE PERMEABILITY OF NEWBORN AND ADULT MOUSE LUNGS TO ULTRASTRUCTURAL PROTEIN TRACERS

The permeability of the alveolar-capillary membrane of newborn and adult mice to horseradish peroxidase (HRP) and catalase was studied by means of ultrastructural cytochemistry, and the permeability to ferritin was studied by electron microscopy. The influence of varying volumes of intravenously injected fluid on the rate of leakage of the tracers from pulmonary capillaries was examined. The tracers were injected intravenously and the mice were sacrificed at timed intervals. Experiments on newborn mice with intranasally instilled HRP were also done. The tissues were fixed in formaldehyde-glutaraldehyde fixative. Chopped sections were incubated in Graham and Karnovsky's medium for peroxidase and in a modification of this medium for catalase. Tissues were postfixed in OsO₄ and processed for electron microscopy. In both newborn and adult mice, the ready passage of peroxidase through endothelial clefts was dependent on the injection of the tracer in large volumes of saline. When the tracer was injected in small volumes of saline, its passage through endothelial clefts was greatly reduced. Endothelial junctions of newborn mice were somewhat more permeable to HRP than those of adult mice. In all animals, alveolar epithelial junctions were impermeable to HRP. Catalase and ferritin did not pass through endothelial junctions. Intranasally instilled HRP in newborn mice was taken up by pinocytotic vesicles and tubules of flat alveolar cells.     

Endocytosis in the uterine luminal and glandular epithelial cells of mice during early pregnancy

We have localized horseradish peroxidase (HRP) in the mouse uterus after intravenous administration on days 1 and 5 of pregnancy in an effort to understand how serum proteins reach the uterine lumen. Direct movement of HRP into uterine and glandular lumina was blocked by the epithelial tight junctions on both days. In luminal and glandular epithelial cells at both times, HRP was localized in endocytic vesicles along the basolateral membranes, multivesicular bodies (mvb), elongated dense bodies below the nucleus (bdb), and many small vesicles near the apical surface of the cells. The uptake of HRP was most extensive in the luminal epithelium on day 1: the number of tracer-containing apical vesicles and bdb was largest, and there were also clusters of vesicles containing the tracer above the nucleus. Acid phosphatase was localized on day 1 in mvb and bdb in both cell types, indicating that these structures are lysosomes. It appeared that HRP followed two pathways after basolateral endocytosis by the epithelial cells: it was transported to the apical region of the cells, where it was present in small vesicles that may release their contents into the uterine or glandular lumina, or it was transported to lysosomes. To investigate whether macromolecules may be transported from the uterine lumen to the stroma, we also studied endocytosis at the apical pole of luminal epithelial cells after intraluminal injection of HRP. There was no detectable uptake of HRP from the lumen on day 1, and no tracer was detected in the intercellular spaces or basement membrane region. On day 5, a large amount of HRP was taken up from the lumen into apical endocytic vesicles, mvb, and dense bodies, but tracer was not present in the Golgi apparatus, lateral intercellular spaces, or the basement membrane region at the times studied. These observations indicate that there was no transport of luminal macromolecules to the uterine stroma on day 1, while the possibility of transport on day 5 requires further study.     

Transport processes through the blood-brain barrier

The existence of the blood-brain barrier is due to tight junctions between endothelial cells preventing the passage of liquid and solute material at the capillary level. Substances can thus pass across the blood-brain barrier if they are lipophilic or if they have transport systems in the membranes of endothelial cells. The luminal membrane brings metabolites needed for the brain function, the abluminal one plays an important part in removing substances from brain, this can happen against a concentration gradient and thus needs energy. Ions are transported differently by the 2 membranes. Sodium and chloride have carriers and potassium is transported very actively by the sodium-potassium ATPase of the abluminal membrane. Blood-brain glucose influx is very important and happens by carrier transport at the 2 membranes. Efflux seems to use the same transport system as the influx. Transport of ketone bodies seems to happen only from blood to brain, the carriers being reversibly used for brain-blood transport of pyruvic and lactic acid. Amino-acid transport is very different on the luminal and abluminal membranes. On the luminal membrane there are 2 transport systems, one for basic amino acids, the other one, the L system, for neutral amino-acids. All neutral amino-acids are transported through the abluminal membrane by the L, A and ASC systems. There exists a system of transport for basic amino-acids, and a very active one for acid amino-acids. Some systems for the transport of hormones, vitamins and for some peptides exist also at the blood-brain barrier which thus plays a very important role in the regulation of brain metabolism.     

Formation of a barrier by brain microvessel endothelial cells in culture

Endothelial cells (EC) isolated from bovine brain microvessels produce a continuous monolayer when grown in primary culture. The EC are joined together by tight junctions and contain few pinocytotic vesicles. Horseradish peroxidase (HRP) is unable to penetrate this in vitro barrier system. Exposure of the cells to 1.6 M arabinose produces a reversible separation of the tight junctions with penetration of HRP across the monolayer in a pattern similar to that observed in animals after infusion of hyperosmotic solutions into the carotid artery. The behavior of brain microvascular cells in culture suggest that they retain properties important to the formation of the blood-brain barrier.     

Opening of blood-brain barrier in Triturus cristatus carnifex by hyperosmolar mannitol solutions

The permeability of the newt cerebral capillaries to lanthanum ion has been studied after perfusion with mannitol solutions of increasing molarity. In the control specimens lanthanum deposits were limited to the luminal side of the capillaries and tracer did not spread to the pericapillary spaces due to the tight junctions. Treatment with hypertonic solutions of mannitol (0.25M, 0.5M, 1M) caused opening of the blood brain barrier with a progressive increase in lanthanum between the endothelial cell edges, in the basal lamina and in the extracellular spaces of the nervous parenchyma in relation to the molarity of the mannitol solution. The spread of lanthanum is probably due to opening of the tight junctions between the endothelial cells, since pinocytotic vesicles labelled with tracer were not evident.     

Fine structure and permeability of capillaries in the stria vascularis and spiral ligament of the inner ear of the guinea pig

Summary The blood capillaries in the stria vascularis and the spiral ligament of guinea pigs were studied by electron microscopy with freeze-fracture and thin section methods, including tracer experiments with horseradish peroxidase (HRP) and microperoxidase (MP). The endothelial cells of the capillaries of both tissues are connected by tight junctions, and contain about the same number of micropinocytotic vesicles. In cases of intravascular administration before fixation, both of the tracers stained the perivascular space and almost all endothelial vesicles in the stria vascularis. On the other hand, the perivascular space and many vesicles in the spiral ligament were unstained. The endothelial tight junctions in the stria vascularis prevented the penetration of HRP, but sometimes allowed the penetration of MP. Those of the spiral ligament were impermeable to both tracers. In cases of tracer administration after fixation, leakage spots of HRP from capillaries were sparsely located all over the stria vascularis. Transendothelial channels and isolated fenestrae formed by micropinocytotic vesicles were detected. It is concluded that the capillaries of the stria vascularis are similar to the muscle capillaries and to the capillaries of the elasmobranch brain, whereas those in the spiral ligament are similar to the brain capillaries of higher vertebrates.     

Neutral amino acid transport by the blood-brain barrier. Membrane vesicle studies.

Endothelial cell membranes, the site of the blood-brain barrier, were obtained from the capillaries of cow brain. The luminal and abluminal membranes were separated by centrifugation on a discontinuous Ficoll gradient. Electron microscopy revealed that the membrane preparations consisted almost entirely of sealed vesicles. The release of latent enzyme activity showed that both membrane preparations were primarily right side out. Radiolabeled L-phenylalanine uptake by luminal vesicles was proportional to membrane protein concentration, with less than 10% binding. Transport was by a high affinity carrier (Km 11.8 +/- 0.1 microM, asymptotic standard error) that showed little or no stereospecificity, and was independent of Na+ or H+ gradients. Transport was inhibited by L-tryptophan, L-leucine, 2-aminobicyclo[2,2,1]heptane-2-carboxylate and D-phenylalanine, but not by N-(methylamino)-isobutyrate. Abluminal membranes showed an additional component in which a Na+ gradient accelerated the transport of both phenylalanine and N-(methylamino)-isobutyrate. These studies demonstrate the utility of membrane vesicles as a model to characterize the transport properties of the distinct membranes of the polar endothelial cells that form the blood-brain barrier.     

Receptor-mediated transcytosis of follicle-stimulating hormone through the rat testicular microvasculature

Accumulated data suggest that endothelial cells express specific receptors for several peptide and (glyco)protein hormones that may transport hormones across the cell to be delivered to the interstitial fluid and tissue target cells. Surprisingly, very little information is available on the actual endothelial organelles involved in this cellular process. In the present study the transfer of follicle-stimulating hormone (FSH) through the endothelial barrier of rat testes was examined by analysing the binding and transport of gold-tagged recombinant human (rh)FSH under various conditions using electron microscopy. At 4 degrees C the probe bound specifically to the luminal surface of the endothelial cells without internalization. The use of 125I-rhFSH, which allows precise quantitation of the binding, confirmed the specificity of hormone interaction with the testicular microvasculature. At 37 degrees C the hormone was internalized via coated pits and vesicles into an extensive subluminal tubulo-vesicular compartment and was transported across the endothelium via a system of tubules and vesicles. Moreover, monoclonal antibodies against the FSH receptor ectodomain coupled to colloidal gold followed the same route. In contrast, a non specific, fluid-phase uptake via caveolae was observed for a major plasma protein - rat serum albumin and a fluid-phase tracer - peroxidase. These results suggest that FSH transcytosis across the testicular endothelial barrier is receptor-mediated and involves luminal uptake via coated pits/vesicles, sorting at the level of luminal early endosomes, and transcellular transport through transcytotic tubulo-vesicular organelles. Similar receptor-mediated pathways are likely to be involved in the physiological functioning of a number of other protein and peptide hormones that must translocate specifically from blood to the target cells.     

Astrocyte-mediated induction of alkaline phosphatase activity in human umbilical cord vein endothelium: an in vitro model

The blood-brain barrier, localized in the endothelium of the cerebral capillaries, is characterized by the existence of tight junctions, a low mitochondrial density, a low number of vesicles and a high activity of certain enzymes like alkaline phosphatase and gamma-glutamyl transpeptidase. Astroglial cells secrete a product that induces brain microvessel endothelial cells to differentiate into endothelial cells with blood-brain barrier properties. If rat astrocytes were grown together with human umbilical cord vein endothelial cells in a co-culture system in which there is no cellular contact between both cell types, alkaline phosphatase activity was induced in the endothelial cells after three days of co-culturing. If the endothelial cells were cultured in astrocyte conditioned medium, alkaline phosphatase activity was also induced, and preliminary results showed that formation of tight junctions occurred after five days. These observations support the hypothesis that astrocytes induce the differentiation of non-blood-brain barrier endothelial cells into endothelial cells with blood-brain barrier properties, in this study based on alkaline phosphatase-activity induction and induction of tight junction formation. These inductive processes are produced by a soluble factor released by the astrocytes.     

A blood-brain barrier without tight junctions in the fly central nervous system in the early postembryonic stage

Summary The anatomical basis of the vertebrate blood-brain barrier is a series of tight junctions between endothelial cells of capillaries in the central nervous system. Over two decades ago, tight junctions were also proposed as the basis of the blood-brain barrier in insects. Currently there is a growing understanding that septate junctions might possess barrier properties in various invertebrate epithelial cells. We now examine these two views by studying the blood-brain barrier properties of the early postembryonic larva of a dipteran fly (Delia platura) by transmission electron microscopy. Newly hatched larvae possess a functioning blood-brain barrier that excludes the extracellular tracer, ionic lanthanum. This barrier is intact throughout the second instar stage as well. The ultrastructural correlate of this barrier is a series of extensive septate junctions that pervade the intercellular space between adjacent perineurial cells. No tight junctions were located in either nerve, glial or perineurial cell layers. We suggest that the overall barrier might involve septate junctions within extensive, meandering intercellular clefts.     

Transtubular transport of proteins in rabbit proximal tubules

The purpose of the present experiments was to study possible different pathways of intracellular transport of proteins after luminal and basolateral uptake in isolated rabbit proximal tubules. Tubules were exposed to cationized ferritin (CF) in the perfusion fluid and horseradish peroxidase (HRP) in the bath simultaneously or to HRP in the bath alone for 30 min. The peritubular fluid (bath) and perfusion fluid were then exchanged and the tubules either fixed immediately or allowed to function during chase-periods for 10, 20, 30, or 60 min before fixation to follow the migration of the proteins through the cells. The proteins were to a large extent found separated in different vacuoles and lysosomes at all time periods studied, indicating separate pathways after uptake via the luminal and basolateral membranes respectively. About 0.5% of the CF taken up by the cells was transported through the cells and became located in the intercellular spaces. HRP was transported from the peritubular fluid to the apical cytoplasm of the tubules indicated by a gradual accumulation of small HRP-containing vesicles, first in the basal part of the cells and then in the apical cytoplasm. In tubules perfused with both CF and HRP in the perfusate, the CF and HRP were found together in apical vacuoles and lysosomes. After perfusion with HRP alone, this tracer was found in similar large vacuoles and lysosomes in the apical cytoplasm, in contrast to the small HRP-filled vacuoles seen after uptake from the bath.     

The ultrastructural basis of transcapillary exchanges

A brief survey is given of current views correlating the ultrastructural and permeability characteristics of capillaries. Observations based on the use of peroxidase (mol wt 40,000), as an in vivo, and colloidal lanthanum, as an in vitro, ultrastructural tracer, are presented. In capillaries with "continuous" endothelium, the endothelial intercellular junctions are thought to be permeable to the tracers, and are regarded as maculae occludentes rather than zonulae occludentes, with a gap of about 40 A in width between the maculae. Some evidence for vesicular transport is also presented. It is inferred that the cell junctions are the morphological equivalent of the small-pore system, and the vesicles the equivalent of the large-pore system. Peroxidase does not apparently cross brain capillaries: the endothelial cell junctions are regarded as zonulae occludentes, and vesicles do not appear to transport across the endothelium. This is regarded as the morphological equivalent of the blood-brain barrier for relatively large molecules. The tracers appear to permeate the fenestrae of fenestrated capillaries, and the high permeability of these capillaries to large molecules is attributed to the fenestrae. Capillaries with discontinuous endothelium readily allow passage of the tracers through the intercellular gaps. A continuous basement membrane may act as a relatively coarse filter for large molecules. In general, the morphology of capillaries correlates well with physiological observations.     

Uptake of horseradish peroxidase from CSF into the choroid plexus of the rat,with special reference to transepithelial transport

Summary Protein uptake from cerebral ventricles into the epithelium of the choroid plexus, and transport across the epithelium were studied ultrastructurally in rats. Horseradish peroxidase (HRP, MW 40,000) was used as protein tracer. Steady-state ventriculo-cisternal perfusion with subatmospheric pressure (-10cm of water) in the ventricular system was applied. HRP dissolved in artificial CSF was perfused from the lateral ventricles to cisterna magna for various times, and ventriculo-cisternal perfusion, vascular perfusion or immersion fixation with a formaldehyde-glutaraldehyde solution was performed.Coated micropinocytic vesicles containing HRP were seen both connected with the apical, lateral and basal epithelial surface and within the cells. Heavily HRP-labeled vesicles were often fused with the lining membrane of slightly labeled or unlabeled intercellular spaces. Since the apical tight junctions of the epithelium never appeared open or never contained HRP in the spaces between the fusion points, and since the intercellular spaces between adjacent epithelial cells below the junctions only infrequently contained tracer after 5 min, by increasing amounts after 15–60 min of HRP perfusion, a vesicular transport of HRP from the apical epithelial surface to the intercellular spaces, bypassing the tight junctions, is suggested.In addition to the transepithelial transport, micropinocytic vesicles also transported HRP to the lysosomal apparatus of the epithelial cells. With increasing length of exposure to HRP, a sequence of HRP-labeled structures could be evaluated, from slightly labeled apical vacuoles and multivesicular bodies to very heavily labeled dense bodies.     

Mechanism of action of collagenase on the blood-brain barrier permeability. Increase of endothelial cell pinocytotic activity as shown with horse-radish peroxidase as a tracer.

The ultrastructural mechanism of the protease induced blood-brain barrier permeability-increase was studied with horse-radish peroxidase as a tracer. After intravenous injection of collagenase or pronase, a significantly increased number of pinocytotic vesicles was found in brain capillary endothelial cells. alpha-Chymotrypsine did not exert such an action.     

Transport of pinocytic vesicles in the eye of a snail,Helix aspersa

The heads of small adult snails, Helix aspersa, were injected with horseradish peroxidase (HRP) for one to five hours before extirpating the eyes and preparing them cytochemically for electron microscopy. There was internalization of tracer by pinocytic vesicles (pinosomes) at the bases of types-I and -II sensory cells, ganglion cells and, in lesser amounts, by pigmented supportive cells. Vesicles and vacuoles filled with HRP were transported in two directions: lensward as far distad as the ends of the cells (retrograde) and brainward down the optic nerve (anterograde). We believe that the numerous reacted vacuoles in the cell somata are formed by fusion of vesicles, tubules and C-shaped organelles filled with tracer; we present evidence that they become secondary lysosomes. Sensory cell type II possesses more HRP-reacted vacuoles distally than the other retinal cells. Other vesicles are also described. There was no uptake of tracer by the distal ends of the retinal cells following injection HRP into the hemolymph. The swelling of the optic nerve, immediately behind the eye, contains more HRP-filled pinosomes and vacuoles than does the nerve below the dilatation. The significance of endocytosis and transport of pinosomes within the eye and down the optic nerve is discussed.