- Email: [email protected]
Distribution of anionic sites and glycoconjugates on the endothelial surfaces of the developing blood-brain barrier
Developmental Brain Research, 29 (1986) 69-79 Elsevier
69
BRD 50440
Distribution of Anionic Sites and Glycoconjugates on the Endothelial Surfaces of the Developing Blood-Brain Barrier A.W. VORBRODT, A.S. LOSSINSKY, D.H. DOBROGOWSKA and H.M. WISNIEWSKI
Department of Pathological Neurobiology, New York State Office of Mental Retardation and Developmental Disabilities, Institute for Basic Research in Developmental Disabilities, Staten Island, NY 10314 (U.S.A.) (Accepted March 18th, 1986)
Key words: anionic site - - lectin receptor - - blood-brain barrier - - brain development - - barrier maturation
The distribution of anionic sites detected in vitro with cationized ferritin and lectin-binding sites on the endothelial cell (EC) surface of brain micro-blood vessels was studied by electron microscopy. Gold-labeled lectins and glycoproteins and Lowicryl K4M-embedded brain samples obtained from mouse embryos (19th day), and from 1-, 5-, 12-, 24- and 48-day-old and adult mice were used. It was shown that the functional maturation of the blood-brain barrier (BBB) occurring in the mouse after birth'between the 12th and 24th day of life is accompanied by a disappearance of vesicular transport in capillaries and by the formation of a uniform, thin, negatively charged layer on the surface of the EC. Concomitantly the binding of lectins specific for fl-D-galactosyl (RCA) and sialyl (LFA and WGA) residues become progressively more intense and uniform on both luminal and abluminal fronts of the EC. The concentration of HPA-binding sites on the abluminal side of the EC and in the basement membrane increases. Similarlythe binding of Con A becomes more intense on abluminal than on luminal front of the EC. These observations suggest that extensive remodeling of anionic sites and surface glycoprotein layer and also the elaboration of ECs polarity occur during BBB maturation.
INTRODUCTION O u r previous studies indicate that the functional maturation of the b l o o d - b r a i n barrier (BBB), as evidenced by a 'closure' of brain m i c r o - b l o o d vessels (MBVs) to intravenously injected horseradish peroxidase ( H R P ) takes place between the 12th and 24th day of life in the mouse 34'37'40. This m a t u r a t i o n coincides with the a p p e a r a n c e of alkaline phosphatase activity in the p l a s m a l e m m a of the endothelial cells
(ECs). The restricted p e r m e a b i l i t y of brain M B V s representing BBB type of microvasculature in adult animals is due not only to a structural peculiarity3'12,23 but also to some biochemical and physicochemical properties of brain endothelia 2'12'2t'22'36. One of the factors p r e s u m a b l y restricting the m o v e m e n t of molecules across the vessel wall is the negative charge of the luminal surface of the E C 5-7.
In fenestrated endothelium, negatively charged molecules (anionic sites) are distributed unevenly, suggesting the existence of differentiated microdomains representing various endocytic and transport (transcytotic) organelles like p l a s m a l e m m a l vesicles, transendothelial channels, fenestrae, etc. 2°'2s. The study on the chemical nature of these anionic sites indicates that they are contributed by various glycoproteins and proteoglycans 2°'29,3°. The presence of a relatively uniform layer of anionic sites 15A6,37'38most p r o b a b l y contributed by various glycoconjugates 17,18,39, was also observed on the surface of brain endothelia. A l l of the a b o v e - m e n t i o n e d observations p r o m p t ed us to study the localization of anionic sites and glycoconjugates recognized by various lectins in brain endothelia during d e v e l o p m e n t as well as the functional maturation of the BBB. A s our experimental model, we e m p l o y e d the mouse in various stages of
Correspondence: A.W. Vorbrodt, Department of Pathological Neurobiology, New York State Office of Mental Retardation and Developmental Disabilities, Institute for Basic Research in Developmental Disabilities, 1050 Forest Hill Road, Staten Island, NY 10314, U.S.A. 0165-3806/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
FABLE [ Gold-labeled lectins and glycoproteins used in this study Lectin
Sugar specificity
Ricinus commums agglutinin 120 (RCA-1 J Limax flavus agglutinin (LFA) Wheat germ agglutinin (WGA~ Helix pomana agglutinin (HPA) Concanavalin A (Con A J
/LD-galactosc
Glycoprotein used in two-step method
Final concentrations of specific inhibitor ~. 1 M l)-galaclo~e
N-acetyl and N-glycolylneuraminic acid N-acetyl-glucosamine and N-acetylneuraminic acid ~z-N-acetyl-D-galactosamine terminal~ (t-l~-mannose and a-D-glucose
Fetuin (Sigma, type IVI
O. t M N-acct~lneuraminic acid
Ovomucoid (Sigma, type II1)
I~.1 M N-acetylglucosamine. 0.1 M N-acetvlneuraminic acid (I.2 M N-acctyt-~-galactosaminc
HRP, acidic isoenzyme (Sigma, type VII)
0,2 M c~-mcm,d-D-mannosidc
dylate buffer, p H 7.3. T i m e of p e r f u s i o n was 15 rain
d e v e l o p m e n t and m a t u r a t i o n similar to o u r p r e v i o u s study 34"37'4°. W e are i n t e r e s t e d if the m a t u r a n o n of
f o l l o w e d by i m m e r s i o n f i x a n o n up to 2 h in i c e - c o l d
B B B is p a r a l l e l e d by the a p p e a r a n c e of n e g a t i v e l y
fixative. E m b r y o n i c brains w e r e fixed by i m m e r s i o n
c h a r g e d m o l e c u l e s and specific g l y c o c o n j u g a t e s c o n -
in this fixative w i t h o u t p r i o r p e r f u s i o n . A f t e r fixation
tributing to t h e m a i n t e n a n c e of this charge.
brains w e r e r e m o v e d , c o r o n a l l y s e c t i o n e d , and small samples f r o m f r o n t o p a r i e t a l cortex w e r e s e l e c t e d for sectioning with Sorvall T C - 2 tissue s e c t i o n e r ( D u -
MATERIALS AND METHODS
P o n t ) o r for e m b e d d i n g in L o w i c r y l K 4 M . A f t e r fixation was c o m p l e t e d , the s p e c i m e n s w e r e
Animals All
experiments
BALBc/J
mice.
were
Brain
conducted
samples were
on
or
w a s h e d and i m m e r s e d for 4 h in ice-cold 0.1 M NH~CI
from
in 0.1 M c a c o d y l a t e buffer, p H 7.3. to b l o c k free al-
IM
taken
m o u s e e m b r y o s (19 days), n e w b o r n m i c e (1st day of
d e h y d e groups. This step was f o l l o w e d by o v e r n i g h t
life) o r 5-. 12-. 24- and 48-day-old animals. F o r c o m -
w a s h i n g in the s a m e buffer.
parison, adult animals ( 6 - 1 0 m o n t h s old) w e r e also Demonstration o f anionic sttes
used. Under
Nembutal anesthesia, animals (newborn.
T h e sections (35 ktm thick) o b t a i n e d f r o m t h e tissue
infants o r adult) w e r e fixed by i n t r a c a r d i a c p e r f u s i o n
s e c t i o n e r w e r e i n c u b a t e d for 15 rain at r o o m t e m p e r -
with a m i x t u r e of freshly p r e p a r e d 2 % p a r a f o r m a l d e -
ature (20 °CI in the s o l u t i o n of c a t i o n i z e d ferritin
h y d e and 1% or 0.1% g l u t a r a l d e h y d e in 0.1 M caco-
( C F ) in 0.1 M c a c o d y l a t e buffer, p H 7.3 (1 m g / m l ) ac-
Fig. 1 A segment of mechanically detached ECs from the brain capillary of a 19-day-old mouse fetus is shown. Both fronts of the EC become accessible and are decorated with CF. On the luminal surface, the plasmalemma proper is labeled with a continuous, thick layer of CF (arrows), whereas limiting membranes of plasmalemmal vesicles are not labeled (curved arrowsL The abluminal front of the EC is heavily labeled with clustered CF (arrowheads). E. endothelial cell: J. tight ]unction between ECs: L. vessel lumen: M. basement membrane; R. red blood cell: S. smooth muscle cell x75.000. Fig. 2. Labeling of the luminal surface of the EC of 1-day-old mouse brain MBV is shown. The plasmatemma proper is decorated with a thick, continuous layer of CF. Numerous plasmalemmal pits and vesicles are not decorated (curved arrows), x 87.000. Fig. 3. Heavy labeling with CF of the luminal surface of the EC of brain MBVs of 5-day-old mouse is shown. The limiting membranes of numerous plasmalemmal pits and vesicles are not labeled (curved arrows), x75,000. Fig. 4. Distribution of anionic sites on the luminal surface of brain capillary EC of 12-day-old mouse is shown, The labeling of the plasmalemma proper with CF is uniform and continuous (arrows). No plasmalemmal pits and vesicles are present. ×75,000. Fig. 5. A segment of the wall of brain arteriole of 12-day-old mouse is shown. The decoration with CF of the luminal surface (arrow) of the EC is less regular than in capillary (compare Fig. 4l. The limiting membrane of a plasmalemmal vesicle (curved arrow) is almost free of the label, x 87.000. Fig. 6. A segment of the wall of a brain arteriole of 24-day-old mouse is shown. The luminal plasmalemma proper is continuously but slightly irregularly labeled (arrows), whereas plasmalemmal pits and vesicles (curved arrow) are not labeled with CF. x75,000.
71 cording to the procedure of Thiirauf et al. 33. For control purposes, some sections were incubated under similar conditions in a solution of native ferritin (NF). Incubated sections were washed and fixed for 1 h
in 1% OsO 4 buffered with 0.1 M cacodylate buffer pH 7.3, washed again, en bloc stained with 0.5% uranyl acetate (pH 5.0) for 1 h at room temperature in complete darkness, dehydrated in ethanol and embedded in Spurr low viscosity resin.
The demonstratton of"lectin-binding sites Tissue samples after fixation (see above) were cut
in a two-step labeling m e t h o d (final concentrations of inhibitory sugars as shown in Table i J.
into approximately 1-mm 3 blocks, dehydrated in ~ce-
All ultrathin sections obtained from Spurr or Lowi-
cold ethanol and e m b e d d e d in hydrophilic resin Low-
cryl K4M e m b e d d e d s p e o m e n s were observed in a
icryl K4M. After polymerization at low temperature
Hitachi H U - 1 1 E or Philips 420 electron microscope.
under an ultraviolet lamp, specimens were cut with a diamond knife on a Sorvall MT 5000 ( D u P o n t ) mi-
RESULTS
crotome and picked up on 75- or 100-mesh formvarcarbon-coated nickel grids.
Distribution of anionic sites
Preparation of colloidal gold and lectin-gold complexes
luminal surface of ECs is covered with a relatively
Particles of colloidal gold with a mean diameter of 14 nm were prepared according to Frens 8. The preparation of lectin or g l y c o p r o t e m - g o l d complexes was performed according to the procedure of Roth 24-26 similar to what was previously described 39.
Staining procedure The incubation with lectins was performed at room temperature (20 °C) in a moist chamber. Initially, all grids with attached sections were placed on a drop of phosphate buffer-saline (PBS) for 5 min. Direct or in-
In brain MBVs of 19-day-old mouse embryos, the thick layer decorated with CF. The decoration of the plasmalemma proper of the E C is frequently not uniform. i.e., the thickness of the E C labeling varies from one blood vessel to another. Occasionally, clusters of CF can be noted, especially on the protrusions of the luminal p l a s m a l e m m a and in the vicinity of cell junctions. At this time of d e v e l o p m e n t , tight junctions between ECs are already well developed (Fig. 1). N u m e r o u s plasmalemmal pits and vesicles opened onto the vascular l u m e n are less or n o t deco-
tory sugar. Interaction of g l y c o p r o t e i n - g o l d complexes with
rated. In some areas of brain tissue where the ECs are mechanically detached from the adjacent astrocytic process, the ligand has access to the abluminal front of the EC which is heavily decorated with clustered CF (Fig. 1) indicating that the anionic sites are present on both fronts of the EC. In n e w b o r n animals (first day of life) the pattern of CF labeling is essentially similar to that described above. The luminal surface of the E C is covered by a continuous layer, heavily labeled with CF. The limit-
tissue structures was also controlled by omission of the first step, i.e.. incubation with non-labeled lectins
ing m e m b r a n e s of n u m e r o u s ptasmalemmal pits and vesicles are not decorated with CF. The b a s e m e n t
direct labeling was performed exactly as described previously39.
Control incubation Sections were p r e m c u b a t e d for 30 m m in a solution of appropriate inhibitory sugar (0.1 or 0.2 M), followed by incubation in a mixture of lectin and inhibi-
Figs. 7-11 represent ultrathin sections of mouse brain cortex embedded in Lowicryl K4M and incubated with lectin-gold or glycoprotein-gold complex for the detection of various glycoconjugates. Abbreviations as in Fig. 1 Fig. 7. Localization of RCA-binding sites (fl-D-galactosylresidues) in capillary walt of 19-day-old mouse embryo is shown, Both luminal (arrows) and abluminal fronts of the EC are irregularly labeled. Some labeling is also present in the cytoplasmic compartments of the EC, in large cytoplasmic vacuoles fcurved arrow) and in a hardly visible BM. x45.000. Fig. 8. RCA-binding sites in the endothelium of 1-day-old mouse are shown. The luminal surface of the EC is more regularly labeled than in embryonic MBVs, although occasionally some clustering of gold grains (arrows) can be noted. The abluminal front of the EC facing the hardly visible BM is decorated with dusters of gold complexes with a certain degree of order. ×66,000, Fig. 9. Localization of RCA-bindingsites in the wall of brain capillary of 12-day-old mouse is shown. Both the luminal and abluminal (arrowheads) fronts of the EC are labeled with numerous gold particles. The labeling of the BM is weak and irregular, x66.000. Fig. 10. The distribution of LFA-bindingsites (sialic acid residues) in the wall of a capillary of an embrYonic mouse brain is shown. There are only a few gold particles on the luminal (arrows) surface of the EC. in the EC cytoplasm and in the area of the BM. ×66.000. Fig. 11. The distribution of LFA-bindingsites in the wall of an arteriole in a 5-day-old mouse brain is shown. There are numerous gold particles located on the luminal front of the EC (arrows) and in subendothelial BM. Some labeling is also present on the surface of a smooth muscle cell. ×45.000.
73 membrane is hardly visible, similar to that in the embryonic vessels, whereas tight junctions between ECs are well developed and the intercellular rim is not penetrated by the CF (Fig, 2). In 5-day-old mice, the ECs of MBVs are still char-
acterized by the presence of numerous plasmalemmal pits and vesicles the limiting membranes of which are not labeled with CF (Fig. 3). On the contrary, the plasmalemma proper is continuously decorated with a relatively thick layer of CF.
74 In 12-day-old animals, only small fractions of MBVs is decorated similar to that of the 5-day-old mice. In the majority of capillaries, the luminal surface of ECs is covered with a continuous layer, uniformly labeled with CF (Fig. 4}. This layer of CF is evidently thinner than observed in earlier stages of development. The plasmalemmal pits and vesicles are almost completely absent from the ECs of capillaries. They are noticeable, however, only in the larger MBVs, presumably arterioles (Fig. 5). The limitmg membranes of these pits and vesicles show either a reduced or a complete absence of CF decoration. In older (24 and 48 days) and in adult animals, the decoration of the luminal plasmalemma of capillary ECs is essentially similar to those observed in most 12-day-old mice. The plasmalemmal vesicles and pits are only occasionally present, and the plasmalemma proper is continuously and uniformly decorated with a thin layer of CF. Only in arterioles, the labeling of the luminal surface of the EC is less regular than in capillaries and venules. The luminal and abluminal pits and plasmalemmal vesicles are located on both fronts of the EC and their limiting membranes are not labeled with CF (Fig. 6). The plasmalemma proper is decorated with unevenly distributed CF particles. occasionally forming clusters irregularly scattered on the luminal surface of the EC No labeling, or only accidentally scattered single ferritin grains are observed in brain sections incubated in the solution of NF.
Distribution of lectin binding sites RCA ([3-D-galactosyl residues J The labeling of both fronts of the EC of brain MBVs at various stages of development was most intense with this lectin. In embryonic brain, the distribution of RCA binding sites on the luminal surface of the EC is irregular There are short segments of the luminal plasmalemma proper completely deprived of the label, intermitted with segments heavily labeled. The protrusions of the cell surface are always labeled with several gold particles. Also the limiting membranes of cytoplasmic vesicles or vacuoles, which are barely visible in Lowicryl-embedded tissue sections, are labeled (Fig. 7). The abluminal front of the EC is also irregularly labeled with single or clustered gold particles. Some labeling is also scattered throughout the cytoplasm of the EC. At this stage of development, the basement membrane is hardly visible and only few gold grains are scattered in this structure (Fig. " M). In newborn mice. the labeling of both fronts and of the cytoplasm of the EC is slightly more intense than in the embryonic brain. The distribution of luminal labeling is more regular, and the intensity of the labeling of the abluminal side of the EC is also stronger. The gold particles are clustered and these clusters are distributed with some degree of regularity on the abluminal front of the EC (Fig. 8). The basement
Figs. 12-17 represent ultrathin sections of mouse brain cortex embedded in Lowicryl K4M and incubated with lectin-gold or glycoprotein-gold complexes for the detection of various glycoconjugates. Abbreviations as in Fig. l. Fig. 12. The distribution of LFA-binding sites in the wall of a brain capillary of a 12-day-old mouse is shown, In this vessel the labeling on the luminal surface is more intense than on the abluminal surface. There are also numerous gold particles scattered throughout lhc cytoplasmic structures of the EC and in the BM. x 57.000. Fig. 13. The localization of LFA-binding sites in a brain arteriole of a 12-day-old mouse is shown. The labeling of the luminal surface ot the EC is less regular larrowsl than in capillaries (compare Fig. 12L whereas the subendothelial BM is more regularly decorated. x 45.000. Fig. 14. The labeling of the capillary wall of a 12-day-old mouse with W G A is shown. The luminal surface of the EC is heavily and regularly labeled with gold particles (arrows) whereas the abluminal side of the E C and the BM are much less labeled in this vessel. T h e r e are also gold particles present in cytoplasmic vacuoles or vesicles (curved arrows), x 57.000. Fig. 15. The distribution of HPA-binding sites (terminal N-acetyl-D-galactosaminyl residues) in the wall of an embryonic brain capillary is shown. The gold particles are more numerous on the abluminal front of the EC (arrowheads) and in the area of the primitive BM than on the luminal surface (arrows) of the EC. x45,000. Fig. 16. HPA-binding sites in the wall of brain MBVs of a l-day-old mouse are shown. The majority of gold particles is concentrated on the abluminal side of the EC (arrowheads). Only single gold grains are scattered throughout the cell cytoplasm and on the luminal surface of the EC. x66.000. Fig. 17. The localization of HPA-binding sites in the watl of a brain capillary of 12-day-old mouse is shown. The main site of the labeling is the BM and abluminal front of the E C larrowheads) x66.000.
75 membrane is still hardly visible and the labeling of this structure is irregular (Fig. 8, M). An essentially similar distribution of RCA binding sites is observed in brain MBVs of 5°day-old animals.
In the majority of capillaries and venules of 12-dayold mouse brains, the distribution of RCA binding sites becomes progressively more regular on both fronts of the EC (Fig. 9). Similar, almost continuous
l.
I.
7~
labeling of both fronts of the EC is present in older (24 and 48 days) and in adult animals. The basement membrane in these animals becomes well visible, but the labeling is rather scanty and irregular. Some differences in the distribution of R C A binding sites in ECs of capillaries, venules and arterioles were already described elsewhere 39. No labeling was observed after incubation of sections in a control medium containing D-galactose.
LFA (siafic acid residues) The labeling of MBVs in embryonic brain is rather scanty and irregular. There are only few sparsely scattered gold grains on the luminal front and in the cytoplasm of the EC (Fig. 10). Single or clustered gold particles are also occasionally present on the abluminal side of the EC and in a thin basement membrane. The labeling with LFA progressively increases after birth. In newborn mice, the general pattern of the distribution of gold particles is similar, although slightly more intense than in the embryos. In 5-day-old animals, the LFA binding sites are more numerous and more regularly distributed on the luminal and abluminal fronts of the EC, and they are also relatively numerous in the basement membrane. A similar distribution of labeling is also observed in arterioles, where smooth muscle cells are also labeled (Fig. 11). In the majority of brain capillaries of 12-day-old mice the labeling of the luminal surface of the EC is rather intense, although particular gold grains are scattered in irregular spaces (Fig. 12). The labeling of the abluminal front of the EC is less regular, although relatively numerous gold particles are scattered in the basement membrane. Some intracytoplasmic structures of ECs are frequently strongly labeled. The labeling of the luminal surface of the ECs of arterioles is less regular than in capillaries. The labeling with LFA of the subendotheliai basement membrane in arterioles is evidently more intense than in the capillary (Fig. 13). An essentially similar distribution pattern of LFAbinding sites is observed in older (24 and 48 days) and in adult animals. The preincubation of sections with N-acetylneuraminic acid almost completely inhibits the labeling of the brain sections.
WGA (N-acetyl-D-glucosarniny[ and sialyl residues1 The distribution of WGA-binding sites in embryonic, newborn and 5-day-old mice is essentially similar to that described previously As a rule, the labeling of the luminal front of the EC ~,~more intense with this lectin than with LFA. This difference l~ especially well visible in some brain MBVs of 12-dayold mice. The labeling with gold particles is relatively intense on the luminal front of the EC in some arterioles, where the cytoplasmic vesicle s or vacuoles are also labeled [Fig. t4). tress regular labeling with this lectin is present on the abluminal front of the EC and in the basement membrane. The distribution pattern of labeling with W G A is generally less regular than with LFA. In some ve~ sels, labeling appears in the basement membrane. whereas in others there is almost no labeling. [ncubation of seclions in a control medium containing N-acetvl-neuraminic acid only partially inhibits the labeling. After addition of N-acetylglucosamine, the inhibition is stronger but not complete. HPA tterrninal N-acetyl-O-galactosaminyl residues ~ In embryonic brain MBVs. these residues are present both on the luminal and abluminal fronts ofIhe EC and in a hardly recognizable basement membrane (Fig. 15). Some label is also present in elongated structures presumably representing perivascular astrocytic processes. In newborn animals, the abluminal front of the EC and the area of the basement membrane becomes more intensely labeled, whereas the luminal part of the EC remains almost unlabeled [Fig. 16). The disappearance of the labeling from the luminal front of the EC and increase of the labeling of the abluminal plasmalemma and especially of the basement membrane is a very characteristic feature of MBVs in 5-day-old animals. In 12-day-old animals the distribution pattern ot HPA-binding sites is similar to those observed in older and adult animals. On the luminal surface and in the cytoplasm of the EC. there are only a few solitary, irregularly scattered gold grains. On the contrary, the abluminal front of the EC of MBVs and the basement membranes are strongly labeled (Fig. 17). After incubation of sections in a control medium containing N-acetyl-D-galactosamine. the labeling completely disappears.
77
Con A (a-mannosyl and a-glucosyl residues) The pattern of the distribution of Con A binding sites in MBVs during development is rather inconsistent. Con A-binding sites are present in the MBVs of embryonic brain on both the luminal and abluminal surfaces of the ECs and also are scattered throughout the cytoplasm and nuclei of these cells. No noticeable differences in the intensity of the labeling can be found on both fronts of the EC. However, after birth, beginning from the 5th day of life there is diminution of the labeling on the luminal surface of the EC. Finally, in the majority of brain MBVs of 12-day-old animals, as well as in older and adult animals, the labeling of the abluminal front of the EC becomes evidently more intense than of the luminal surface. This characteristic distribution of Con A binding sites in the endothelia of brain MBVs of adult mice was already described 39. The incubation of sections in control medium containing a-methyl-D-mannoside only partially inhibits the labeling.
DISCUSSION Our present observations indicate that a negatively charged surface layer appears on the luminal front of brain MBVs endothelia before the functional maturation of the BBB. This surface layer, which is relatively thick and occasionally clustered in embryonic and newborn brains (up to the 5th day of life), becomes thinner and continuously uniform in the 12and 24-day-old mice. Occasionally observed labeling with CF of the abluminal plasmalemma suggests that the layer of anionic sites is present on both fronts of the EC. In embryonic and newborn mice brains, the tight junctions between ECs of blood vessels are well developed in contradistinction to the basement membrane which seems to be fully assembled after birth during the first weeks of life. The most characteristic features of the immature MBVs is the presence of numerous plasmalemmal pits and vesicles, frequently associated with the luminal or abluminal plasma membrane (PM) of the EC. We have previously observed that these structures are involved in transport of intravenously injected HRP across the vessel wall 13'14. Their limiting mem-
branes, in contrast to the plasmalemma proper, are not decorated with CF. The lack of negative charge makes them similar to the plasmalemmal vesicles of mouse visceral fenestrated endothelia, presumably engaged in transendothelial transport of various soluble macromolecules (solutes) including negatively charged serum proteins 2°,31. Although our knowledge on the biochemical character of anionic sites on the surface of ECs of brain microvasculature is rather limited, one can assume that they are contributed by various glycoproteins and sulfated proteoglycans, similarly like in other types of vasculature 2°,29'3°. According to Nag, only a part of anionic groups in brain endothelia is contributed by sialyl residues although their localization is similar to the distribution of binding sites for W G A 17'18. According to our previous 39 and present observations, both sialic (neuraminic) acid recognized by WGA and LFA, and also fl-D-galactosyl residues recognized by RCA, are present on both sides of the EC, and most probably contribute to the maintenance of the negative surface charge. The binding intensity of these lectins increases and their distribution becomes more regular and uniform, parallel to the maturation of the BBB. Because the binding sites of the above-mentioned lectins do not form such a densely packed layer on the surface of ECs of embryonic and newborn animal brains as do anionic sites labeled with CF, one can assume that some other glycoconjugates and also other molecules and residues (lipids and proteins), as suggested by Burry and Wood 4, contribute to the maintenance of the negative charge on the surface of the ECs. Our observation indicates also that in the course of the development and maturation of the BBB, the multiplication and redistribution of some glycoconjugates (for example fl-D-galactosyi and sialyl residues) occurs on both surfaces of the EC. We have observed also a gradual increase in the labeling of the basement membrane (BM) with HPA. It suggests that the components of the BM are especially rich in terminal N-acetyl-D-galactosaminyl residues (see Figs. 15-17), which together with sialyl residues (Figs. 11, 13) are relatively highly concentrated in this area. If the BM can be considered as a charge barrier or a selective filter, as postulated by some authors TM, the function and composition of this barrier at least concerning the presence of some glycoconjugates,
7~q changes during the m a t u r a t i o n of brain microvasculature. As we previously discussed, the localization of some glycoconjugates recognized by H P A or Con A , suggests that they are not directly related to the maintenance of the negative charge on the surface of brain endothelia 39. These residues however, together with other glycoconjugates, not presently detected, can play an important role as receptors for a variety of circulating substances including various exogenous glycoproteins, regulatory molecules, hormones, etc. ~1"27. Thus, these residues can represent important components of the surface of ECs participating in transport mechanisms across the brain MBVs. The uneven distribution of some glycoconjugates recognized by H P A and Con A , on both fronts of the EC and changes n o t e d in their localizations are further indications of the existence of the previously postulated and discussed polarity of brain ECs~:~', which seems to be e l a b o r a t e d during postnatal maturation of the function of BBB. Several observations indicate, that after the damage of BBB function in adult animals in various pathological conditions, the c o m m o n feature of leaking vessels is a disruption or d i s a p p e a r a n c e of the endothelial surface negative charge 9"1516"3x. On the other hand, the neutralization or alteration of surface anionic sites by such cationic ligands like C F 32 or protamine sulfate 19'35 leads to an increased p e r m e a b i l i t y of endothelia of MBVs and to the o p e n i n g of the BBB. In the i m m a t u r e mouse brain, however, we have observed that the leaking MBVs do not lose their luminal surface negative charge, but they are characterized by the presence of n u m e r o u s plasmalemmal pits and vesicles. In this respect, they resemREFERENCES 1 Betz, A.L., Firth, J.A. and Goldstein, G.W., Polarity o! the blood-brain barrier: distribution of enzymes between the luminal and antiluminal membranes of brain capillary endothelial cells, Brain Res., 192 (1980) 17-28. 2 Bradbury, M., The Concept of a Blood-Brain Barrier, Wiley, New York, 1979, pp. 137-213. 3 Brightman, M.W., Morphology of blood-brain interfaces, Exp. Eye Res., Suppl. 25 (1977) 1-25. 4 Burry, R.W. and Wood, J.G., Contributions of lipids and proteins to the surface charge of membranes, J. Cell Biol., 82 (1979) 726-741. 5 Danom D. and Skutelsky, E., Endothelial surface charge
ble non-BBB type MBVs with extenswe v e ~ c u l a r transport. Thus, the mechanism of l e a k a g e of immature brain MBVs evidently differs in this respect from that observed in adult animals caused bv the damage of the BBB. O u r present observations show mr the first ttme the distribution of lectin binding sites within the entire section of the wall of brain M B V s including the luminal and abluminal fronts of the EC. b a s e m e n t m e m b r a n e and smooth muscle cells However. p o o r visibility of the PM and other c ~ t o m e m b r a n e s in Lowicryl K 4 M - e m b e d d e d tissue samples and labeling of various glycoconjugates located inside the EC cytoplasm, which are exposed in thin section, m a k e the precise recognition of binding sites located exclusively on the luminal or abluminal PM relatively difficult. F u r t h e r m o r e , it appears {tom the d a t a presented by H o r i s b e r g e r and Rosset:". that one gold granule may cover many receptor s~tes (up to t500 with wheat germ agglutinin on the surface of red blood cell). Fhus. one g o l d - l e c t i n grain located m the area of luminal or abluminal PM can also occasionally label the binding sites prescnl inside the cell body, t.c.. m some juxtaposed
ACKNOWLEDGEMENT The authors wish to express their appreciatton to Mrs. Patricia C o d o n e r for excellent secretarial assistance. This study is s u p p o r t e d in p a r t by G r a n t 1727t05 from N I N C D S . and tts possible relationship to thrombogenesls, Ann. N. Y Acad. Sci.. 275 t19761 47-63. Danon. D.. Laver-Rudich, Z. and Skutelsky, E.. Surface charge and flow properties of endothelial membranes in aging rats. Mech. Ageing Dev.. 14 (19807 145-153. Dermietzel. R., Thiirauf. N. and Kalweit. P.. Surface charges associated with fenestrated brain capillaries, J. Ultrastruct. Res.. 84 (1983) 111-119 8 Frens. G.. Controlled nucleation for the regulauon of the particle size .n monodisperse gold suspensions. Nature Phys. Sci.. 24l (1973)20-22. 9 Hart. M.N. and Schelper. R.L.. Endothelial surface charge alterations associated with blood-brain barrier disruption, J. Neuropathol. Exp. Neurol., 43 1to84)350.
79 10 Horisberger, M. and Rosset, J., Colloidal gold, a useful marker for transmission and scanning electron microscopy, J. Histochem. Cytochem., 25 (1977) 295-305. 11 Hughes, R.C., Membrane glycoproteins. A review of structure and function, Butterworths, London, 1976, pp. 1-5. 12 Joo, F., The blood-brain barrier in vitro: ten years of research on microvessels isolated from the brain, Neurochem. Int., 7 (1985) pp. 1-25. 13 Lossinsky, A.S., Vorbrodt, A.W. and Wisniewski, H.M., Ultracytochemical studies of transendothelial transport in cerebral microblood vessels during development, J. Cell Biol., 99 (1984) 283a. 14 Lossinsky, A.S., Vorbrodt, A.W. and Wisniewski, H.M., Characterization of endothelial cell transport in the developing mouse blood-brain barrier, Dev. Neurosci., in press. 15 Nag, S. and Arseneau, R., Alteration of endothelial surface charge in hypertension, J. Cereb. Blood Flow Metab., 3 (1983) $624-$625. 16 Nag, S., Cerebral endothelial surface charge in hypertension, Acta Neuropathol., 63 (1984) 276-281. 17 Nag, S., Ultrastructural localization of lectin receptors on cerebral endothelium, Acta Neuropathol., 66 (1985) 105-110. 18 Nag, S., Ultrastructural localization of monosaccharide residues on cerebral endothelium, Lab. Invest., 52 (1985) 553-558. 19 Nagy, Z., Peters, M. and Hfittner, I., Endothelial surface charge: blood-brain barrier opening to horseradish peroxidase induced by the polycation protamine sulfate, Acta Neuropathol., Suppl. 7 (1981) 7-9. 20 Palade, G.E., Simionescu, M. and Simionescu, N., Differentiated microdomains in the vascular endothelium. In M.L. Nossel and M.J. Vogel (Eds.), Pathobiology of the Endothelial Cell, Academic Press, New York, 1982, pp. 23-33. 21 Pardridge, W.M. and Oldendorf, W.H., Transport of metabolic substrates through the blood-brain barrier, J. Neurochem., 28 (1977) 5-12. 22 Rapoport, S.I., Blood-brain barrier in physiology and medicine, Raven, New York, 1976, pp. 3-75. 23 Reese, T.S. and Karnovsky, M.J., Fine structural localization of a blood-brain barrier to exogenous peroxidase, J. Cell Biol., 34 (1967) 207-217. 24 Roth, J., Application of lectin-gold complexes for electron microscopic localization of glycoconjugates on thin sections, J. Histochem. Cytochem., 32 (1983) 987-999. 25 Roth, J., Cytochemical localization of terminal N-acetyl-Dgalactosamine residues in cellular compartments of intestinal goblet cells: implications for the topology of O-glycosylation, J. CellBiol., 98 (1984) 399-406. 26 Roth, J., Lucocq, J. and Charest, P.M., Light and electron microscopic demonstration of sialic acid residues with the lectin from Limax flavus: a cytochemical affinity technique with the use of fetuin-gold complexes, J. Histochem. Cytochem., 32 (1984) 1167-1176. 27 Schulte, B.A. and Spicer, S.S., Histochemical methods for
characterizing secretory and cell surface sialoglycoconjugates, J. Histochem, Cytochem., 33 (1985) 427-438. 28 Simionescu, N., Simionescu, M. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium. I. Preferential distribution of anionic sites, J. Cell Biol., 90 (1981) 605-613. 29 Simionescu, M., Simionescu, N., Silbert, J.E. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium. II. Partial characterization of their anionic sites, J. Cell Biol., 90 (1981) 614-621. 30 Simionescu, M., Simionescu, N. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium: distribution of lectin receptors, J. Cell Biol., 94 (1982) 406-413. 31 Simionescu, M., Simionescu, N. and Palade, G.E., Preferential distribution of anionic sites on the basement membrane and the abluminal aspect of the endothelium in fenestrated capillaries, J. Cell Biol., 95 (1982) 425-434. 32 Skutelsky, E. and Danon, D., Redistribution of surface anionic sites on the luminal front of blood vessel endothelium after interaction with polycationic ligand, J. Cell Biol., 72 (1976) 232-241. 33 Thiirauf, N., Dermietzel, R. and Kalweit, P., Surface charges associated with fenestrated brain capillaries. I. In vitro labeling of anionic sites, J. Ultrastruct. Res., 84 (1983) 103-110. 34 Vorbrodt, A.W., Cytochemical characterization of endothelial cell surface during development of blood-brain barrier (BBB). Abstracts of Vllth lnternatl. Congr. Histochem. Cytochem., Helsinki, Finland, 1984, p. 471. 35 Vorbrodt, A.W., Lassmann, H., Wisniewski, H.M. and Lossinsky, A.S., Ultracytochemical studies of the bloodmeningeal barrier (BMB) in rat spinal cord, Acta Neuropathol., 55 (1981) 113-123. 36 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Enzyme cytochemistry of blood-brain barrier (BBB) disturbances, Acta Neuropathol., Suppl. VII (1983) 43-57. 37 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Cytochemical and ultrastructural characterization of developing blood-brain barrier (BBB), J. Neuropathol. Exp. Neurol., 43 (1984) 350. 38 Vorbrodt, A.W., Lossinsky, A.S., Wisniewski, H.M., Suzuki, R., Yamaguchi, T., Masaoka, M. and Klatzo, I., Ultrastructural observations on the transvascular route of protein removal in vasogenic brain edema, Acta Neuropathol., 66 (1985) 265-273. 39 Vorbrodt, A.W., Dobrogowska, D.H., Lossinsky, A.S. and Wisniewski, H.M., Ultrastructural localization of lectin receptors on the luminal and abluminal aspects of brain micro-blood vessels, J. Histochem. Cytochem., 34 (1986) 251-261. 40 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Localization of alkaline phosphatase activity in endothelia of developing and mature mouse blood-brain barrier, Dev. Neurosci., in press.
69
BRD 50440
Distribution of Anionic Sites and Glycoconjugates on the Endothelial Surfaces of the Developing Blood-Brain Barrier A.W. VORBRODT, A.S. LOSSINSKY, D.H. DOBROGOWSKA and H.M. WISNIEWSKI
Department of Pathological Neurobiology, New York State Office of Mental Retardation and Developmental Disabilities, Institute for Basic Research in Developmental Disabilities, Staten Island, NY 10314 (U.S.A.) (Accepted March 18th, 1986)
Key words: anionic site - - lectin receptor - - blood-brain barrier - - brain development - - barrier maturation
The distribution of anionic sites detected in vitro with cationized ferritin and lectin-binding sites on the endothelial cell (EC) surface of brain micro-blood vessels was studied by electron microscopy. Gold-labeled lectins and glycoproteins and Lowicryl K4M-embedded brain samples obtained from mouse embryos (19th day), and from 1-, 5-, 12-, 24- and 48-day-old and adult mice were used. It was shown that the functional maturation of the blood-brain barrier (BBB) occurring in the mouse after birth'between the 12th and 24th day of life is accompanied by a disappearance of vesicular transport in capillaries and by the formation of a uniform, thin, negatively charged layer on the surface of the EC. Concomitantly the binding of lectins specific for fl-D-galactosyl (RCA) and sialyl (LFA and WGA) residues become progressively more intense and uniform on both luminal and abluminal fronts of the EC. The concentration of HPA-binding sites on the abluminal side of the EC and in the basement membrane increases. Similarlythe binding of Con A becomes more intense on abluminal than on luminal front of the EC. These observations suggest that extensive remodeling of anionic sites and surface glycoprotein layer and also the elaboration of ECs polarity occur during BBB maturation.
INTRODUCTION O u r previous studies indicate that the functional maturation of the b l o o d - b r a i n barrier (BBB), as evidenced by a 'closure' of brain m i c r o - b l o o d vessels (MBVs) to intravenously injected horseradish peroxidase ( H R P ) takes place between the 12th and 24th day of life in the mouse 34'37'40. This m a t u r a t i o n coincides with the a p p e a r a n c e of alkaline phosphatase activity in the p l a s m a l e m m a of the endothelial cells
(ECs). The restricted p e r m e a b i l i t y of brain M B V s representing BBB type of microvasculature in adult animals is due not only to a structural peculiarity3'12,23 but also to some biochemical and physicochemical properties of brain endothelia 2'12'2t'22'36. One of the factors p r e s u m a b l y restricting the m o v e m e n t of molecules across the vessel wall is the negative charge of the luminal surface of the E C 5-7.
In fenestrated endothelium, negatively charged molecules (anionic sites) are distributed unevenly, suggesting the existence of differentiated microdomains representing various endocytic and transport (transcytotic) organelles like p l a s m a l e m m a l vesicles, transendothelial channels, fenestrae, etc. 2°'2s. The study on the chemical nature of these anionic sites indicates that they are contributed by various glycoproteins and proteoglycans 2°'29,3°. The presence of a relatively uniform layer of anionic sites 15A6,37'38most p r o b a b l y contributed by various glycoconjugates 17,18,39, was also observed on the surface of brain endothelia. A l l of the a b o v e - m e n t i o n e d observations p r o m p t ed us to study the localization of anionic sites and glycoconjugates recognized by various lectins in brain endothelia during d e v e l o p m e n t as well as the functional maturation of the BBB. A s our experimental model, we e m p l o y e d the mouse in various stages of
Correspondence: A.W. Vorbrodt, Department of Pathological Neurobiology, New York State Office of Mental Retardation and Developmental Disabilities, Institute for Basic Research in Developmental Disabilities, 1050 Forest Hill Road, Staten Island, NY 10314, U.S.A. 0165-3806/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
FABLE [ Gold-labeled lectins and glycoproteins used in this study Lectin
Sugar specificity
Ricinus commums agglutinin 120 (RCA-1 J Limax flavus agglutinin (LFA) Wheat germ agglutinin (WGA~ Helix pomana agglutinin (HPA) Concanavalin A (Con A J
/LD-galactosc
Glycoprotein used in two-step method
Final concentrations of specific inhibitor ~. 1 M l)-galaclo~e
N-acetyl and N-glycolylneuraminic acid N-acetyl-glucosamine and N-acetylneuraminic acid ~z-N-acetyl-D-galactosamine terminal~ (t-l~-mannose and a-D-glucose
Fetuin (Sigma, type IVI
O. t M N-acct~lneuraminic acid
Ovomucoid (Sigma, type II1)
I~.1 M N-acetylglucosamine. 0.1 M N-acetvlneuraminic acid (I.2 M N-acctyt-~-galactosaminc
HRP, acidic isoenzyme (Sigma, type VII)
0,2 M c~-mcm,d-D-mannosidc
dylate buffer, p H 7.3. T i m e of p e r f u s i o n was 15 rain
d e v e l o p m e n t and m a t u r a t i o n similar to o u r p r e v i o u s study 34"37'4°. W e are i n t e r e s t e d if the m a t u r a n o n of
f o l l o w e d by i m m e r s i o n f i x a n o n up to 2 h in i c e - c o l d
B B B is p a r a l l e l e d by the a p p e a r a n c e of n e g a t i v e l y
fixative. E m b r y o n i c brains w e r e fixed by i m m e r s i o n
c h a r g e d m o l e c u l e s and specific g l y c o c o n j u g a t e s c o n -
in this fixative w i t h o u t p r i o r p e r f u s i o n . A f t e r fixation
tributing to t h e m a i n t e n a n c e of this charge.
brains w e r e r e m o v e d , c o r o n a l l y s e c t i o n e d , and small samples f r o m f r o n t o p a r i e t a l cortex w e r e s e l e c t e d for sectioning with Sorvall T C - 2 tissue s e c t i o n e r ( D u -
MATERIALS AND METHODS
P o n t ) o r for e m b e d d i n g in L o w i c r y l K 4 M . A f t e r fixation was c o m p l e t e d , the s p e c i m e n s w e r e
Animals All
experiments
BALBc/J
mice.
were
Brain
conducted
samples were
on
or
w a s h e d and i m m e r s e d for 4 h in ice-cold 0.1 M NH~CI
from
in 0.1 M c a c o d y l a t e buffer, p H 7.3. to b l o c k free al-
IM
taken
m o u s e e m b r y o s (19 days), n e w b o r n m i c e (1st day of
d e h y d e groups. This step was f o l l o w e d by o v e r n i g h t
life) o r 5-. 12-. 24- and 48-day-old animals. F o r c o m -
w a s h i n g in the s a m e buffer.
parison, adult animals ( 6 - 1 0 m o n t h s old) w e r e also Demonstration o f anionic sttes
used. Under
Nembutal anesthesia, animals (newborn.
T h e sections (35 ktm thick) o b t a i n e d f r o m t h e tissue
infants o r adult) w e r e fixed by i n t r a c a r d i a c p e r f u s i o n
s e c t i o n e r w e r e i n c u b a t e d for 15 rain at r o o m t e m p e r -
with a m i x t u r e of freshly p r e p a r e d 2 % p a r a f o r m a l d e -
ature (20 °CI in the s o l u t i o n of c a t i o n i z e d ferritin
h y d e and 1% or 0.1% g l u t a r a l d e h y d e in 0.1 M caco-
( C F ) in 0.1 M c a c o d y l a t e buffer, p H 7.3 (1 m g / m l ) ac-
Fig. 1 A segment of mechanically detached ECs from the brain capillary of a 19-day-old mouse fetus is shown. Both fronts of the EC become accessible and are decorated with CF. On the luminal surface, the plasmalemma proper is labeled with a continuous, thick layer of CF (arrows), whereas limiting membranes of plasmalemmal vesicles are not labeled (curved arrowsL The abluminal front of the EC is heavily labeled with clustered CF (arrowheads). E. endothelial cell: J. tight ]unction between ECs: L. vessel lumen: M. basement membrane; R. red blood cell: S. smooth muscle cell x75.000. Fig. 2. Labeling of the luminal surface of the EC of 1-day-old mouse brain MBV is shown. The plasmatemma proper is decorated with a thick, continuous layer of CF. Numerous plasmalemmal pits and vesicles are not decorated (curved arrows), x 87.000. Fig. 3. Heavy labeling with CF of the luminal surface of the EC of brain MBVs of 5-day-old mouse is shown. The limiting membranes of numerous plasmalemmal pits and vesicles are not labeled (curved arrows), x75,000. Fig. 4. Distribution of anionic sites on the luminal surface of brain capillary EC of 12-day-old mouse is shown, The labeling of the plasmalemma proper with CF is uniform and continuous (arrows). No plasmalemmal pits and vesicles are present. ×75,000. Fig. 5. A segment of the wall of brain arteriole of 12-day-old mouse is shown. The decoration with CF of the luminal surface (arrow) of the EC is less regular than in capillary (compare Fig. 4l. The limiting membrane of a plasmalemmal vesicle (curved arrow) is almost free of the label, x 87.000. Fig. 6. A segment of the wall of a brain arteriole of 24-day-old mouse is shown. The luminal plasmalemma proper is continuously but slightly irregularly labeled (arrows), whereas plasmalemmal pits and vesicles (curved arrow) are not labeled with CF. x75,000.
71 cording to the procedure of Thiirauf et al. 33. For control purposes, some sections were incubated under similar conditions in a solution of native ferritin (NF). Incubated sections were washed and fixed for 1 h
in 1% OsO 4 buffered with 0.1 M cacodylate buffer pH 7.3, washed again, en bloc stained with 0.5% uranyl acetate (pH 5.0) for 1 h at room temperature in complete darkness, dehydrated in ethanol and embedded in Spurr low viscosity resin.
The demonstratton of"lectin-binding sites Tissue samples after fixation (see above) were cut
in a two-step labeling m e t h o d (final concentrations of inhibitory sugars as shown in Table i J.
into approximately 1-mm 3 blocks, dehydrated in ~ce-
All ultrathin sections obtained from Spurr or Lowi-
cold ethanol and e m b e d d e d in hydrophilic resin Low-
cryl K4M e m b e d d e d s p e o m e n s were observed in a
icryl K4M. After polymerization at low temperature
Hitachi H U - 1 1 E or Philips 420 electron microscope.
under an ultraviolet lamp, specimens were cut with a diamond knife on a Sorvall MT 5000 ( D u P o n t ) mi-
RESULTS
crotome and picked up on 75- or 100-mesh formvarcarbon-coated nickel grids.
Distribution of anionic sites
Preparation of colloidal gold and lectin-gold complexes
luminal surface of ECs is covered with a relatively
Particles of colloidal gold with a mean diameter of 14 nm were prepared according to Frens 8. The preparation of lectin or g l y c o p r o t e m - g o l d complexes was performed according to the procedure of Roth 24-26 similar to what was previously described 39.
Staining procedure The incubation with lectins was performed at room temperature (20 °C) in a moist chamber. Initially, all grids with attached sections were placed on a drop of phosphate buffer-saline (PBS) for 5 min. Direct or in-
In brain MBVs of 19-day-old mouse embryos, the thick layer decorated with CF. The decoration of the plasmalemma proper of the E C is frequently not uniform. i.e., the thickness of the E C labeling varies from one blood vessel to another. Occasionally, clusters of CF can be noted, especially on the protrusions of the luminal p l a s m a l e m m a and in the vicinity of cell junctions. At this time of d e v e l o p m e n t , tight junctions between ECs are already well developed (Fig. 1). N u m e r o u s plasmalemmal pits and vesicles opened onto the vascular l u m e n are less or n o t deco-
tory sugar. Interaction of g l y c o p r o t e i n - g o l d complexes with
rated. In some areas of brain tissue where the ECs are mechanically detached from the adjacent astrocytic process, the ligand has access to the abluminal front of the EC which is heavily decorated with clustered CF (Fig. 1) indicating that the anionic sites are present on both fronts of the EC. In n e w b o r n animals (first day of life) the pattern of CF labeling is essentially similar to that described above. The luminal surface of the E C is covered by a continuous layer, heavily labeled with CF. The limit-
tissue structures was also controlled by omission of the first step, i.e.. incubation with non-labeled lectins
ing m e m b r a n e s of n u m e r o u s ptasmalemmal pits and vesicles are not decorated with CF. The b a s e m e n t
direct labeling was performed exactly as described previously39.
Control incubation Sections were p r e m c u b a t e d for 30 m m in a solution of appropriate inhibitory sugar (0.1 or 0.2 M), followed by incubation in a mixture of lectin and inhibi-
Figs. 7-11 represent ultrathin sections of mouse brain cortex embedded in Lowicryl K4M and incubated with lectin-gold or glycoprotein-gold complex for the detection of various glycoconjugates. Abbreviations as in Fig. 1 Fig. 7. Localization of RCA-binding sites (fl-D-galactosylresidues) in capillary walt of 19-day-old mouse embryo is shown, Both luminal (arrows) and abluminal fronts of the EC are irregularly labeled. Some labeling is also present in the cytoplasmic compartments of the EC, in large cytoplasmic vacuoles fcurved arrow) and in a hardly visible BM. x45.000. Fig. 8. RCA-binding sites in the endothelium of 1-day-old mouse are shown. The luminal surface of the EC is more regularly labeled than in embryonic MBVs, although occasionally some clustering of gold grains (arrows) can be noted. The abluminal front of the EC facing the hardly visible BM is decorated with dusters of gold complexes with a certain degree of order. ×66,000, Fig. 9. Localization of RCA-bindingsites in the wall of brain capillary of 12-day-old mouse is shown. Both the luminal and abluminal (arrowheads) fronts of the EC are labeled with numerous gold particles. The labeling of the BM is weak and irregular, x66.000. Fig. 10. The distribution of LFA-bindingsites (sialic acid residues) in the wall of a capillary of an embrYonic mouse brain is shown. There are only a few gold particles on the luminal (arrows) surface of the EC. in the EC cytoplasm and in the area of the BM. ×66.000. Fig. 11. The distribution of LFA-bindingsites in the wall of an arteriole in a 5-day-old mouse brain is shown. There are numerous gold particles located on the luminal front of the EC (arrows) and in subendothelial BM. Some labeling is also present on the surface of a smooth muscle cell. ×45.000.
73 membrane is hardly visible, similar to that in the embryonic vessels, whereas tight junctions between ECs are well developed and the intercellular rim is not penetrated by the CF (Fig, 2). In 5-day-old mice, the ECs of MBVs are still char-
acterized by the presence of numerous plasmalemmal pits and vesicles the limiting membranes of which are not labeled with CF (Fig. 3). On the contrary, the plasmalemma proper is continuously decorated with a relatively thick layer of CF.
74 In 12-day-old animals, only small fractions of MBVs is decorated similar to that of the 5-day-old mice. In the majority of capillaries, the luminal surface of ECs is covered with a continuous layer, uniformly labeled with CF (Fig. 4}. This layer of CF is evidently thinner than observed in earlier stages of development. The plasmalemmal pits and vesicles are almost completely absent from the ECs of capillaries. They are noticeable, however, only in the larger MBVs, presumably arterioles (Fig. 5). The limitmg membranes of these pits and vesicles show either a reduced or a complete absence of CF decoration. In older (24 and 48 days) and in adult animals, the decoration of the luminal plasmalemma of capillary ECs is essentially similar to those observed in most 12-day-old mice. The plasmalemmal vesicles and pits are only occasionally present, and the plasmalemma proper is continuously and uniformly decorated with a thin layer of CF. Only in arterioles, the labeling of the luminal surface of the EC is less regular than in capillaries and venules. The luminal and abluminal pits and plasmalemmal vesicles are located on both fronts of the EC and their limiting membranes are not labeled with CF (Fig. 6). The plasmalemma proper is decorated with unevenly distributed CF particles. occasionally forming clusters irregularly scattered on the luminal surface of the EC No labeling, or only accidentally scattered single ferritin grains are observed in brain sections incubated in the solution of NF.
Distribution of lectin binding sites RCA ([3-D-galactosyl residues J The labeling of both fronts of the EC of brain MBVs at various stages of development was most intense with this lectin. In embryonic brain, the distribution of RCA binding sites on the luminal surface of the EC is irregular There are short segments of the luminal plasmalemma proper completely deprived of the label, intermitted with segments heavily labeled. The protrusions of the cell surface are always labeled with several gold particles. Also the limiting membranes of cytoplasmic vesicles or vacuoles, which are barely visible in Lowicryl-embedded tissue sections, are labeled (Fig. 7). The abluminal front of the EC is also irregularly labeled with single or clustered gold particles. Some labeling is also scattered throughout the cytoplasm of the EC. At this stage of development, the basement membrane is hardly visible and only few gold grains are scattered in this structure (Fig. " M). In newborn mice. the labeling of both fronts and of the cytoplasm of the EC is slightly more intense than in the embryonic brain. The distribution of luminal labeling is more regular, and the intensity of the labeling of the abluminal side of the EC is also stronger. The gold particles are clustered and these clusters are distributed with some degree of regularity on the abluminal front of the EC (Fig. 8). The basement
Figs. 12-17 represent ultrathin sections of mouse brain cortex embedded in Lowicryl K4M and incubated with lectin-gold or glycoprotein-gold complexes for the detection of various glycoconjugates. Abbreviations as in Fig. l. Fig. 12. The distribution of LFA-binding sites in the wall of a brain capillary of a 12-day-old mouse is shown, In this vessel the labeling on the luminal surface is more intense than on the abluminal surface. There are also numerous gold particles scattered throughout lhc cytoplasmic structures of the EC and in the BM. x 57.000. Fig. 13. The localization of LFA-binding sites in a brain arteriole of a 12-day-old mouse is shown. The labeling of the luminal surface ot the EC is less regular larrowsl than in capillaries (compare Fig. 12L whereas the subendothelial BM is more regularly decorated. x 45.000. Fig. 14. The labeling of the capillary wall of a 12-day-old mouse with W G A is shown. The luminal surface of the EC is heavily and regularly labeled with gold particles (arrows) whereas the abluminal side of the E C and the BM are much less labeled in this vessel. T h e r e are also gold particles present in cytoplasmic vacuoles or vesicles (curved arrows), x 57.000. Fig. 15. The distribution of HPA-binding sites (terminal N-acetyl-D-galactosaminyl residues) in the wall of an embryonic brain capillary is shown. The gold particles are more numerous on the abluminal front of the EC (arrowheads) and in the area of the primitive BM than on the luminal surface (arrows) of the EC. x45,000. Fig. 16. HPA-binding sites in the wall of brain MBVs of a l-day-old mouse are shown. The majority of gold particles is concentrated on the abluminal side of the EC (arrowheads). Only single gold grains are scattered throughout the cell cytoplasm and on the luminal surface of the EC. x66.000. Fig. 17. The localization of HPA-binding sites in the watl of a brain capillary of 12-day-old mouse is shown. The main site of the labeling is the BM and abluminal front of the E C larrowheads) x66.000.
75 membrane is still hardly visible and the labeling of this structure is irregular (Fig. 8, M). An essentially similar distribution of RCA binding sites is observed in brain MBVs of 5°day-old animals.
In the majority of capillaries and venules of 12-dayold mouse brains, the distribution of RCA binding sites becomes progressively more regular on both fronts of the EC (Fig. 9). Similar, almost continuous
l.
I.
7~
labeling of both fronts of the EC is present in older (24 and 48 days) and in adult animals. The basement membrane in these animals becomes well visible, but the labeling is rather scanty and irregular. Some differences in the distribution of R C A binding sites in ECs of capillaries, venules and arterioles were already described elsewhere 39. No labeling was observed after incubation of sections in a control medium containing D-galactose.
LFA (siafic acid residues) The labeling of MBVs in embryonic brain is rather scanty and irregular. There are only few sparsely scattered gold grains on the luminal front and in the cytoplasm of the EC (Fig. 10). Single or clustered gold particles are also occasionally present on the abluminal side of the EC and in a thin basement membrane. The labeling with LFA progressively increases after birth. In newborn mice, the general pattern of the distribution of gold particles is similar, although slightly more intense than in the embryos. In 5-day-old animals, the LFA binding sites are more numerous and more regularly distributed on the luminal and abluminal fronts of the EC, and they are also relatively numerous in the basement membrane. A similar distribution of labeling is also observed in arterioles, where smooth muscle cells are also labeled (Fig. 11). In the majority of brain capillaries of 12-day-old mice the labeling of the luminal surface of the EC is rather intense, although particular gold grains are scattered in irregular spaces (Fig. 12). The labeling of the abluminal front of the EC is less regular, although relatively numerous gold particles are scattered in the basement membrane. Some intracytoplasmic structures of ECs are frequently strongly labeled. The labeling of the luminal surface of the ECs of arterioles is less regular than in capillaries. The labeling with LFA of the subendotheliai basement membrane in arterioles is evidently more intense than in the capillary (Fig. 13). An essentially similar distribution pattern of LFAbinding sites is observed in older (24 and 48 days) and in adult animals. The preincubation of sections with N-acetylneuraminic acid almost completely inhibits the labeling of the brain sections.
WGA (N-acetyl-D-glucosarniny[ and sialyl residues1 The distribution of WGA-binding sites in embryonic, newborn and 5-day-old mice is essentially similar to that described previously As a rule, the labeling of the luminal front of the EC ~,~more intense with this lectin than with LFA. This difference l~ especially well visible in some brain MBVs of 12-dayold mice. The labeling with gold particles is relatively intense on the luminal front of the EC in some arterioles, where the cytoplasmic vesicle s or vacuoles are also labeled [Fig. t4). tress regular labeling with this lectin is present on the abluminal front of the EC and in the basement membrane. The distribution pattern of labeling with W G A is generally less regular than with LFA. In some ve~ sels, labeling appears in the basement membrane. whereas in others there is almost no labeling. [ncubation of seclions in a control medium containing N-acetvl-neuraminic acid only partially inhibits the labeling. After addition of N-acetylglucosamine, the inhibition is stronger but not complete. HPA tterrninal N-acetyl-O-galactosaminyl residues ~ In embryonic brain MBVs. these residues are present both on the luminal and abluminal fronts ofIhe EC and in a hardly recognizable basement membrane (Fig. 15). Some label is also present in elongated structures presumably representing perivascular astrocytic processes. In newborn animals, the abluminal front of the EC and the area of the basement membrane becomes more intensely labeled, whereas the luminal part of the EC remains almost unlabeled [Fig. 16). The disappearance of the labeling from the luminal front of the EC and increase of the labeling of the abluminal plasmalemma and especially of the basement membrane is a very characteristic feature of MBVs in 5-day-old animals. In 12-day-old animals the distribution pattern ot HPA-binding sites is similar to those observed in older and adult animals. On the luminal surface and in the cytoplasm of the EC. there are only a few solitary, irregularly scattered gold grains. On the contrary, the abluminal front of the EC of MBVs and the basement membranes are strongly labeled (Fig. 17). After incubation of sections in a control medium containing N-acetyl-D-galactosamine. the labeling completely disappears.
77
Con A (a-mannosyl and a-glucosyl residues) The pattern of the distribution of Con A binding sites in MBVs during development is rather inconsistent. Con A-binding sites are present in the MBVs of embryonic brain on both the luminal and abluminal surfaces of the ECs and also are scattered throughout the cytoplasm and nuclei of these cells. No noticeable differences in the intensity of the labeling can be found on both fronts of the EC. However, after birth, beginning from the 5th day of life there is diminution of the labeling on the luminal surface of the EC. Finally, in the majority of brain MBVs of 12-day-old animals, as well as in older and adult animals, the labeling of the abluminal front of the EC becomes evidently more intense than of the luminal surface. This characteristic distribution of Con A binding sites in the endothelia of brain MBVs of adult mice was already described 39. The incubation of sections in control medium containing a-methyl-D-mannoside only partially inhibits the labeling.
DISCUSSION Our present observations indicate that a negatively charged surface layer appears on the luminal front of brain MBVs endothelia before the functional maturation of the BBB. This surface layer, which is relatively thick and occasionally clustered in embryonic and newborn brains (up to the 5th day of life), becomes thinner and continuously uniform in the 12and 24-day-old mice. Occasionally observed labeling with CF of the abluminal plasmalemma suggests that the layer of anionic sites is present on both fronts of the EC. In embryonic and newborn mice brains, the tight junctions between ECs of blood vessels are well developed in contradistinction to the basement membrane which seems to be fully assembled after birth during the first weeks of life. The most characteristic features of the immature MBVs is the presence of numerous plasmalemmal pits and vesicles, frequently associated with the luminal or abluminal plasma membrane (PM) of the EC. We have previously observed that these structures are involved in transport of intravenously injected HRP across the vessel wall 13'14. Their limiting mem-
branes, in contrast to the plasmalemma proper, are not decorated with CF. The lack of negative charge makes them similar to the plasmalemmal vesicles of mouse visceral fenestrated endothelia, presumably engaged in transendothelial transport of various soluble macromolecules (solutes) including negatively charged serum proteins 2°,31. Although our knowledge on the biochemical character of anionic sites on the surface of ECs of brain microvasculature is rather limited, one can assume that they are contributed by various glycoproteins and sulfated proteoglycans, similarly like in other types of vasculature 2°,29'3°. According to Nag, only a part of anionic groups in brain endothelia is contributed by sialyl residues although their localization is similar to the distribution of binding sites for W G A 17'18. According to our previous 39 and present observations, both sialic (neuraminic) acid recognized by WGA and LFA, and also fl-D-galactosyl residues recognized by RCA, are present on both sides of the EC, and most probably contribute to the maintenance of the negative surface charge. The binding intensity of these lectins increases and their distribution becomes more regular and uniform, parallel to the maturation of the BBB. Because the binding sites of the above-mentioned lectins do not form such a densely packed layer on the surface of ECs of embryonic and newborn animal brains as do anionic sites labeled with CF, one can assume that some other glycoconjugates and also other molecules and residues (lipids and proteins), as suggested by Burry and Wood 4, contribute to the maintenance of the negative charge on the surface of the ECs. Our observation indicates also that in the course of the development and maturation of the BBB, the multiplication and redistribution of some glycoconjugates (for example fl-D-galactosyi and sialyl residues) occurs on both surfaces of the EC. We have observed also a gradual increase in the labeling of the basement membrane (BM) with HPA. It suggests that the components of the BM are especially rich in terminal N-acetyl-D-galactosaminyl residues (see Figs. 15-17), which together with sialyl residues (Figs. 11, 13) are relatively highly concentrated in this area. If the BM can be considered as a charge barrier or a selective filter, as postulated by some authors TM, the function and composition of this barrier at least concerning the presence of some glycoconjugates,
7~q changes during the m a t u r a t i o n of brain microvasculature. As we previously discussed, the localization of some glycoconjugates recognized by H P A or Con A , suggests that they are not directly related to the maintenance of the negative charge on the surface of brain endothelia 39. These residues however, together with other glycoconjugates, not presently detected, can play an important role as receptors for a variety of circulating substances including various exogenous glycoproteins, regulatory molecules, hormones, etc. ~1"27. Thus, these residues can represent important components of the surface of ECs participating in transport mechanisms across the brain MBVs. The uneven distribution of some glycoconjugates recognized by H P A and Con A , on both fronts of the EC and changes n o t e d in their localizations are further indications of the existence of the previously postulated and discussed polarity of brain ECs~:~', which seems to be e l a b o r a t e d during postnatal maturation of the function of BBB. Several observations indicate, that after the damage of BBB function in adult animals in various pathological conditions, the c o m m o n feature of leaking vessels is a disruption or d i s a p p e a r a n c e of the endothelial surface negative charge 9"1516"3x. On the other hand, the neutralization or alteration of surface anionic sites by such cationic ligands like C F 32 or protamine sulfate 19'35 leads to an increased p e r m e a b i l i t y of endothelia of MBVs and to the o p e n i n g of the BBB. In the i m m a t u r e mouse brain, however, we have observed that the leaking MBVs do not lose their luminal surface negative charge, but they are characterized by the presence of n u m e r o u s plasmalemmal pits and vesicles. In this respect, they resemREFERENCES 1 Betz, A.L., Firth, J.A. and Goldstein, G.W., Polarity o! the blood-brain barrier: distribution of enzymes between the luminal and antiluminal membranes of brain capillary endothelial cells, Brain Res., 192 (1980) 17-28. 2 Bradbury, M., The Concept of a Blood-Brain Barrier, Wiley, New York, 1979, pp. 137-213. 3 Brightman, M.W., Morphology of blood-brain interfaces, Exp. Eye Res., Suppl. 25 (1977) 1-25. 4 Burry, R.W. and Wood, J.G., Contributions of lipids and proteins to the surface charge of membranes, J. Cell Biol., 82 (1979) 726-741. 5 Danom D. and Skutelsky, E., Endothelial surface charge
ble non-BBB type MBVs with extenswe v e ~ c u l a r transport. Thus, the mechanism of l e a k a g e of immature brain MBVs evidently differs in this respect from that observed in adult animals caused bv the damage of the BBB. O u r present observations show mr the first ttme the distribution of lectin binding sites within the entire section of the wall of brain M B V s including the luminal and abluminal fronts of the EC. b a s e m e n t m e m b r a n e and smooth muscle cells However. p o o r visibility of the PM and other c ~ t o m e m b r a n e s in Lowicryl K 4 M - e m b e d d e d tissue samples and labeling of various glycoconjugates located inside the EC cytoplasm, which are exposed in thin section, m a k e the precise recognition of binding sites located exclusively on the luminal or abluminal PM relatively difficult. F u r t h e r m o r e , it appears {tom the d a t a presented by H o r i s b e r g e r and Rosset:". that one gold granule may cover many receptor s~tes (up to t500 with wheat germ agglutinin on the surface of red blood cell). Fhus. one g o l d - l e c t i n grain located m the area of luminal or abluminal PM can also occasionally label the binding sites prescnl inside the cell body, t.c.. m some juxtaposed
ACKNOWLEDGEMENT The authors wish to express their appreciatton to Mrs. Patricia C o d o n e r for excellent secretarial assistance. This study is s u p p o r t e d in p a r t by G r a n t 1727t05 from N I N C D S . and tts possible relationship to thrombogenesls, Ann. N. Y Acad. Sci.. 275 t19761 47-63. Danon. D.. Laver-Rudich, Z. and Skutelsky, E.. Surface charge and flow properties of endothelial membranes in aging rats. Mech. Ageing Dev.. 14 (19807 145-153. Dermietzel. R., Thiirauf. N. and Kalweit. P.. Surface charges associated with fenestrated brain capillaries, J. Ultrastruct. Res.. 84 (1983) 111-119 8 Frens. G.. Controlled nucleation for the regulauon of the particle size .n monodisperse gold suspensions. Nature Phys. Sci.. 24l (1973)20-22. 9 Hart. M.N. and Schelper. R.L.. Endothelial surface charge alterations associated with blood-brain barrier disruption, J. Neuropathol. Exp. Neurol., 43 1to84)350.
79 10 Horisberger, M. and Rosset, J., Colloidal gold, a useful marker for transmission and scanning electron microscopy, J. Histochem. Cytochem., 25 (1977) 295-305. 11 Hughes, R.C., Membrane glycoproteins. A review of structure and function, Butterworths, London, 1976, pp. 1-5. 12 Joo, F., The blood-brain barrier in vitro: ten years of research on microvessels isolated from the brain, Neurochem. Int., 7 (1985) pp. 1-25. 13 Lossinsky, A.S., Vorbrodt, A.W. and Wisniewski, H.M., Ultracytochemical studies of transendothelial transport in cerebral microblood vessels during development, J. Cell Biol., 99 (1984) 283a. 14 Lossinsky, A.S., Vorbrodt, A.W. and Wisniewski, H.M., Characterization of endothelial cell transport in the developing mouse blood-brain barrier, Dev. Neurosci., in press. 15 Nag, S. and Arseneau, R., Alteration of endothelial surface charge in hypertension, J. Cereb. Blood Flow Metab., 3 (1983) $624-$625. 16 Nag, S., Cerebral endothelial surface charge in hypertension, Acta Neuropathol., 63 (1984) 276-281. 17 Nag, S., Ultrastructural localization of lectin receptors on cerebral endothelium, Acta Neuropathol., 66 (1985) 105-110. 18 Nag, S., Ultrastructural localization of monosaccharide residues on cerebral endothelium, Lab. Invest., 52 (1985) 553-558. 19 Nagy, Z., Peters, M. and Hfittner, I., Endothelial surface charge: blood-brain barrier opening to horseradish peroxidase induced by the polycation protamine sulfate, Acta Neuropathol., Suppl. 7 (1981) 7-9. 20 Palade, G.E., Simionescu, M. and Simionescu, N., Differentiated microdomains in the vascular endothelium. In M.L. Nossel and M.J. Vogel (Eds.), Pathobiology of the Endothelial Cell, Academic Press, New York, 1982, pp. 23-33. 21 Pardridge, W.M. and Oldendorf, W.H., Transport of metabolic substrates through the blood-brain barrier, J. Neurochem., 28 (1977) 5-12. 22 Rapoport, S.I., Blood-brain barrier in physiology and medicine, Raven, New York, 1976, pp. 3-75. 23 Reese, T.S. and Karnovsky, M.J., Fine structural localization of a blood-brain barrier to exogenous peroxidase, J. Cell Biol., 34 (1967) 207-217. 24 Roth, J., Application of lectin-gold complexes for electron microscopic localization of glycoconjugates on thin sections, J. Histochem. Cytochem., 32 (1983) 987-999. 25 Roth, J., Cytochemical localization of terminal N-acetyl-Dgalactosamine residues in cellular compartments of intestinal goblet cells: implications for the topology of O-glycosylation, J. CellBiol., 98 (1984) 399-406. 26 Roth, J., Lucocq, J. and Charest, P.M., Light and electron microscopic demonstration of sialic acid residues with the lectin from Limax flavus: a cytochemical affinity technique with the use of fetuin-gold complexes, J. Histochem. Cytochem., 32 (1984) 1167-1176. 27 Schulte, B.A. and Spicer, S.S., Histochemical methods for
characterizing secretory and cell surface sialoglycoconjugates, J. Histochem, Cytochem., 33 (1985) 427-438. 28 Simionescu, N., Simionescu, M. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium. I. Preferential distribution of anionic sites, J. Cell Biol., 90 (1981) 605-613. 29 Simionescu, M., Simionescu, N., Silbert, J.E. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium. II. Partial characterization of their anionic sites, J. Cell Biol., 90 (1981) 614-621. 30 Simionescu, M., Simionescu, N. and Palade, G.E., Differentiated microdomains on the luminal surface of the capillary endothelium: distribution of lectin receptors, J. Cell Biol., 94 (1982) 406-413. 31 Simionescu, M., Simionescu, N. and Palade, G.E., Preferential distribution of anionic sites on the basement membrane and the abluminal aspect of the endothelium in fenestrated capillaries, J. Cell Biol., 95 (1982) 425-434. 32 Skutelsky, E. and Danon, D., Redistribution of surface anionic sites on the luminal front of blood vessel endothelium after interaction with polycationic ligand, J. Cell Biol., 72 (1976) 232-241. 33 Thiirauf, N., Dermietzel, R. and Kalweit, P., Surface charges associated with fenestrated brain capillaries. I. In vitro labeling of anionic sites, J. Ultrastruct. Res., 84 (1983) 103-110. 34 Vorbrodt, A.W., Cytochemical characterization of endothelial cell surface during development of blood-brain barrier (BBB). Abstracts of Vllth lnternatl. Congr. Histochem. Cytochem., Helsinki, Finland, 1984, p. 471. 35 Vorbrodt, A.W., Lassmann, H., Wisniewski, H.M. and Lossinsky, A.S., Ultracytochemical studies of the bloodmeningeal barrier (BMB) in rat spinal cord, Acta Neuropathol., 55 (1981) 113-123. 36 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Enzyme cytochemistry of blood-brain barrier (BBB) disturbances, Acta Neuropathol., Suppl. VII (1983) 43-57. 37 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Cytochemical and ultrastructural characterization of developing blood-brain barrier (BBB), J. Neuropathol. Exp. Neurol., 43 (1984) 350. 38 Vorbrodt, A.W., Lossinsky, A.S., Wisniewski, H.M., Suzuki, R., Yamaguchi, T., Masaoka, M. and Klatzo, I., Ultrastructural observations on the transvascular route of protein removal in vasogenic brain edema, Acta Neuropathol., 66 (1985) 265-273. 39 Vorbrodt, A.W., Dobrogowska, D.H., Lossinsky, A.S. and Wisniewski, H.M., Ultrastructural localization of lectin receptors on the luminal and abluminal aspects of brain micro-blood vessels, J. Histochem. Cytochem., 34 (1986) 251-261. 40 Vorbrodt, A.W., Lossinsky, A.S. and Wisniewski, H.M., Localization of alkaline phosphatase activity in endothelia of developing and mature mouse blood-brain barrier, Dev. Neurosci., in press.