- Email: [email protected]
Glucose transporter immunoreactivity in the hypothalamus and area postrema
Brain Research Bullerin, Vol. 24. pp. 525-528 Q Pergamon Press pk. 1990. Printed in the U.S.A.
0361-9230/90
$3.00 + .OO
Glucose Transporter Immunoreactivity in the Hypothalamus and Area Postrema JOHN K. YOUNG Department of Anatomy, Howard University, 520 W Street, NW, Washington, D.C. 20059 AND CHUNG WANG Department of Biochemistry and Molecular Biology, Harvard University, Cambridge, MA 02138 Received 6 November
1989
YOUNG, J. K. AND C. WANG. Glucose transporter immunoreactivity in the hypothalamus and area postrema. BRAIN RES BULL 24(3) 525-528, 1990.-To study the glucose transporter (GT) protein in two glucose-sensitive areas of the rat brain, frozen coronal sections at the level of the median eminence (ME) and area postrema (AP) were stained immunocytochemically with an antibody raised against human erythrocyte glucose transporter. Immunoreactivity was mainly confined to blood vessels in most brain areas but was lacking in those of the ME and AF’,which also lack a normal blood-brain barrier. This suggests that glucose entry into these brain areas, unlike others, is not limited or regulated by capillary glucose transport systems. Tanycyte processes stained strongly for GT and for glycogen and thus may have an unusual glucose metabolism. Hypothalamus
Area postrema
Glucose transporter
Tanycytes
investigation of patterns of GT immunoreactivity in these specialized brain areas. This was the goal of this study.
SUBSTANTIAL advances in the understanding of glucose transport in various tissues has been made possible by the identification and sequencing of the glucose transport protein of human red blood cells, a protein first isolated by taking advantage of its ability to bind cytochalasin B in a manner specifically displaceable by glucose (18). Since then, numerous studies characterizing this protein in muscle, adipose, and brain tissue have appeared (2, 4, 5, 10, 19). Most studies of the transporter protein in brain have characterized it biochemically and reported a marked enrichment of glucose transporter (GT) protein in brain microvessel preparations (4,lO). Immunocytochemical studies of dog and human brains have also localized the GT protein to endothelial cells of blood vessels (7,14); thus far, the only analogous study of the rat brain has been a description of glucose-displaceable cytochalasin B binding in brain sections (23). A better understanding of glucose transport in two areas of the rat brain-neural tissue adjacent to the median eminence (ME) and area postrema (AP)-may be of particular value in understanding brain function. These two brain areas have a high blood-brain barrier permeability; also, elevations in blood glucose appear to influence the function of these areas and can cause localized cell death if blood glucose elevations are chronic (6, 13, 26). An antibody raised against the human erythrocyte glucose transporter was recently purified for biochemical and immunohistochemical study of the GT protein in muscle and in brain synaptosomes (2,25). This antibody was found to have a substantially greater affinity for rat brain microsomes than for microsomes from muscle and fat and thus would seem suitable for immunohistochemical
METHOD
Rabbit anibody against human erythrocyte GT protein was prepared and then affinity purified according to previously published methods (25). Afftnity purified antibody was then used, at a dilution of 1: 100, in a histochemical staining procedure described below. Brain tissue was obtained from five 3-month-old male SpragueDawley albino rats that were perfused through the heart with 4% paraformaldehyde in 0.2 M phosphate buffer (pH 7.3) under deep chloral hydrate anesthesia. Fixative was contained in a reservoir 166 cm above each rat, producing an estimated IV pressure of about 122 mm of mercury. Brains were removed, fixed overnight, and then transferred to 20% sucrose-formalin until sinking; then, 30 p frozen sections were cut and placed in phosphate buffered saline (PBS-8.5 g NaCl, 0.85 g NaaHPO,, and 0.54 g KH,PG, per 1 of distilled water). After washing for several hr in PBS, free-floating sections were processed through the following steps (24): 1) 30-min incubation in PBS + 0.3% Triton X-100 containing 3% normal goat serum, 2) 24-hr incubation (4 degrees C) in GT antibody, diluted 1:lOO in PBS + 2% normal goat serum + 0.3% Triton X-100, 3) three five-mm washes in PBS-O. 1% T&on, 4) 30-min incubation with biotinylated goat-anti-rabbit IgG (Vectastain ABC kit, Vector Laboratories, Burlingame, CA), 5) three
525
YOUNG AND WANG
FIG. 1, (A) Coronal section of hypothalamus, showing the distribution of immunoreactivity for glucose transporter protein. Immunoreactive blood vessels of various sizes are visible in all areas except the median eminence. Also, strongly staining, thin tanycyte processes can be seen (arrow). (B) Control section stained with nonimmune rabbit serum. (C) Section stained for glycogen using the periodic acid-Schiff method. Glycogen deposition in tanycyte processes is visible (arrow). Magnification bar= 100 IL.
five-mm washes in PBS-0.1% Triton, 6) 30-min incubation in PBS containing 0.3% hydrogen peroxide to nullify sources of endogenous pseudoperoxidase, and 7) three more washes in PBS, followed by a 60-min incubation in ABC reagent (avidin-biotinperoxidase complex). After completion of these steps, sections were again washed and incubated for 15 min in a solution of 0.05% diaminobenzidine in Tris buffer (pH 7.2) containing 0.03% hydrogen peroxide to produce a brown precipitate over rabbit IgG binding sites. Finally, sections were dried onto gelatinized slides and coverslipped. Control sections were incubated with nonimmune rabbit serum diluted in 2% normal goat serum-PBS to examine nonspecific binding. Special features of these brain areas were also illustrated by other methods. Glycogen in the arcuate area was demonstrated by rapidly perfusing 3 rats with fixative, fixing brains overnight, immersing tissue blocks in acidic alcohol + formalin for 2 hr, and then preparing 30 p frozen sections for staining (9). Sections were stained for glycogen via the periodic acid-Schiff method after blockade of nonspecific aldehyde groups by immersion in alcoholic dimedone (9). Specific staining of glycogen was checked by comparison with sections that had been exposed to saliva for 1 hr at 40 degrees C. In order to more clearly visualize the appearance of blood vessels in the area postrema, three additional rats were perfused with 0.5% pamformaldehyde-1.5% glutaraldehyde in 0.2 M phosphate buffer, small blocks of medulla were dehydrated, and then were embedded in EM Bed 812 resin (Electron Microscopy Sciences, Fort Washington, PA). One micron thick sections of this tissue were cut using a Sorvall JV-4 microtome and dry glass knives and were stained with 0.1% toluidine blue in borate buffer. RESULTS lrnrnunoreactivity
for the GT protein
in most of the coronal
brain sections was restricted to a thin rim on the inner surfaces of blood vessels of all sizes (Figs. 1 and 2). This thin layer of reaction product was not wider in larger blood vessels than in capillaries, suggesting that the smooth muscle layers of these larger blood vessels did not contribute to this immunoreactivity. In the hypothalamus, intense staining of ependymal tanycyte processes traversing the arcuate nucleus was also observed (Fig. 1A). Examination of similar sections after periodic acid-Schiff staining revealed dense accumulations of PAS positive material in tanycyte processes (Fig. 1C) that presumably was glycogen, since prior exposure of sections to saliva prevented this staining reaction (not shown). Immunoreactive blood vessels were absent from two brain areas. Inspection of 10 sections in which the median eminence had not been tom during sectioning failed to find a single blood vessel with clear immunoreactivity, although numerous immunoreactive blood vessels were present in the nearby arcuate nucleus and irregular patches of immunoreactivity were present in the median eminence (Fig 1). Also, 21 sections through the area postrema showed a striking absence of immunoreactive vessels in this region (Fig. 2). Semi-thin sections of this area showed a marked perivascular edema around vessels within the AP itself, but a normal appearance of the neuropil around vessels just a short distance away from the AP (Fig. 2B and C). Inspection of other circumventricular organs (organum vasculosum of the lamina terminalis in the anterior hypothalamus and the subfornical organ) that are also known to have unusually permeable capillaries failed to reveal any distinctive patterns of immunoreactivity: immunoreactive blood vessels close to the ventricular margin were seen in both areas (not shown). In general, sections stained with nonimmune rabbit serum showed only a low level of light brown, nonspecific staining (Fig. 1B).
GLUCOSE TRANSPORTERS
527
IN THE HYPOTHALAMUS
FIG. 2. (A) Coronal section of the area postrema in the medulla, showing the distribution of immunoreactivity for glucose transporter protein. Blood vessels within the triangular-shaped area postrema proper are unstained. (B) Semi-thin (1 u) section of area uostrema embedded in EM Bed 812 resin. Blood vessels within the area-postrema proper show pericapillary edema (arrow). (C) Portion of B at a higher magnification, showing capillary with pericapillary edema (arrow). Note the more normal appearing capillary at the right, outside of the area postrema. Magnification bars= 100 p.
DISCUSSION
The basic finding of this study-that the GT protein bound by this antibody is mainly detected in blood vessels of rat brain tissue and is absent from blood vessels of the ME and AP-is in accord with biochemical and histochemical studies (4,5, 10) and with the absence of GT immunoreactivity in AP vessels in the human brain reported by Kalaria et al. (14). The absence of GT immunoreactivity in blood vessels of the median eminence and area postrema
is likely related to the high permeability of these vessels: a specialized glucose transport system across these vessels would seem unnecessary (8). Other authors have also reported that rat AP vessels fail to stain for another protein associated with blood-brain barrier function in brain capillaries (22). In contrast to these two circumventricular organs, capillaries near the organum vasculosum lamina terminalis and subfomical organ did appear immunoreactive; this may reflect a functional difference from the AP and MB, or may partly be due to the absence of a clear anatomical border around the most permeable areas of these latter organs that made it difficult to pinpoint a specific area that would contain specialized capillaries. Some apparent variation in capillary characteristics between different circumventricular organs is supported by staining of pineal, but not AP capillaries, for GT immunoreactivity in the human brain (14). Another indication of a specialized function of capillaries in the area postrema was the marked pericapillaty edema found around AP vessels. This is probably not a simple artifact of excessive perfusion pressure, since nearby capillaries had a normal appearance. The ultrastructure of this phenomenon was studied by Sandoz, who concluded that the pericapillary edema was the result of both highly permeable capillaries and a relatively impermeable perivascular glial sheath that surrounded them and promoted perivascular fluid accumulation (20). According to this data, glucose entry into the area postrema would thus be regulated by glia and not by capillary endothelia. This interpretation would be consistent with data indicating an important influence of glia upon the reactivity of tissue near the area postrema to a toxic form of glucose, goldthioglucose (27,28). Staining of tanycyte processes for the GT protein, in addition to glycogen deposition in tanycytes, a phenomenon known for some time (21), suggests that tanycytes may have a specialized glucose metabolism. The functional significance of this is uncertain. Tanycytes have a number of unusual features, including positive staining for iron and for proteins associated with GABA, dopamine, and glutamate uptake and metabolism (1, 15, 17, 26). Tanycyte processes show a close anatomical relationship with neural processes containing dopamine or LHRH and may influence the excitability of these processes (12,17). Uptake of glucose by tanycytes, causing a change in tanycyte function, may thus be one way in which glucose influences LHRH secretion (3). A more precise localization of GT immunoreactivity in semi-thin and thin sections would provide more information about brain glucose transport; however, we and others have found that exposure of tissues to glutaraldehyde or osmium abolishes GT antigenicity, making adequate preservation of ultrastructure problematic. The only study successfully overcoming this problem has reported that 89% of GT immunoreactivity is associated with luminal and abluminal plasma membrane of endothelial cells (7). The reasons for the absence of GT immunoreactivity in the neuropil in this study, as well as in other studies, are not clear. Neurons are known to have a lowered content of GT protein compared to most cells, and unlike glia or other cell types, fail to increase glucose uptake in response to insulin (11,16). Altered amounts, molecular configuration (glycosylation), and antigenicity of neural GT protein may thus underlie the lack of neuropil immunoreactivity (11,19). Other cytochemical/biochemical approaches may be required to investigate the basis of neuronal glucose uptake in the hypothalamus and elsewhere in the brain.
ACKNOWLEDGEMENT Supported ton, D.C.
by a grant from the Sugar
Association,
Inc.,
Washing-
528
YOUNG
AND WANG
REFERENCES 1. Alho, H.; Fremeau, R. T., Jr.; Tiedge, H.; Wilcox, J.; Bovolin, P.; Brosius, J.; Roberts, J. L.; Costa, E. Diazepam binding inhibitor gene expression: location in brain and peripheral tissue of rat. Proc. Natl. Acad. Sci. USA 85:7018-7022; 1988. 2. Bhattacharyya, M. V.; Brodsy, J. L. Characterization of the glucose transporter from rat brain synaptosomes. Biochem. Biophys. Res. Commun. 155:685-691; 1988. 3. Bourguignon, J.-P.; Gerard, A.; Debougnoux, G.; Rose, J.; Franchimont, P. Pulsatile release of gonadotrophin-releasing hormone (GnRH) from the rat hypothalamus in vitro: calcium and glucose dependency and inhibition by superactive GnRH analogues. Endocrinology 121: 993-999; 1987. 4. Dick, A. P. K.; Harik, S. I. Distribution of the glucose transporter in the mammalian brain. J. Neurochem. 46:1406-1411; 1986. 5. Froehner, S. C.; Davies, A.; Baldwin, S. A.; Lienhard, G. E. The blood-nerve barrier is rich in glucose transporter. J. Neurocytol. 17:173-178; 1988. 6. Garris, D. R.; West, R. L.; Coleman, D. L. Morphometric analysis of medial basal hypothalamic neuronal degeneration in diabetes (db/db) mutant C57BLiKSJ mouse-relation to age and hyperglycemia. Dev. Brain Res. 20:161-168; 1985. 7. Gerhart, D. Z.; Levasseur, R. J.; Broderius, M. A.; Drewes, L. R. Glucose transporter localization in brain using light and electron microscopy. J. Neurosci. Res. 22:464-472; 1989. 8. Gross, P. M.; Blasberg, R. G.; Fenstennacher, J. D.; Patlak, C. S. The microcirculation of rat circumventricular organs and pituitary gland. Brain Res. Bull. 18:73-85; 1987. 9. Guth, L.; Watson, P. K. A correlated histochemical and quantitative study on cerebral glycogen after brain injury in the rat. Exp. Neurol. 22:590X102; 1968. 10. Harik, S. I.; Gravina, S. A.; Kalaria, R. N. Glucose transporter of the blood-brain barrier and brain in chronic hyperglycemia. J. Neurothem. 51:1930-1934; 1988. 11. Heidenreich, K. A.; Gilmore, P. R.; Garvey, W. T. Glucose transport in primary cultured neurons. J. Neurosci. Res. 22:397-7; 1989. 12. Jarvis, C. R.; Andrew, R. D. Correlated electrophysiology and morphology of the ependyma in rat hypothalamus. J. Neurosci. 8:3691-3702; 1988. 13. Kadekaro, M.; Savaki, H.; Sokoloff, L. Metabolic mapping of neural pathways involved in gastrosecretory response to insulin hypoglycemia in the rat. J. Physiol. (Lond.) 300:39347; 1980. 14. Kalaria, R. N.; Gravina, S. A.; Schmidle, J. W.; Perry, G.; Harik, S. I. The glucose transporter of the human brain and blood-brain barrier. Ann. Neurol. 24:757-764; 1988. 15. Kohler, C.; Okuno, E.; Flood, P. R.; Schwartz, R. Quinolinic acid
16.
17.
18.
19.
20.
21. 22.
23. 24.
25.
26.
27. 28.
phosphoribosyl-transferase: preferential glial localization in the rat brain visualized by immunocytochemistry. Proc. Natl. Acad. Sci. USA 84:3491-3495; 1987. Masters, B. A.; Judkins, J. H.; Clarke, D. W.; LeRoith, D.; Raizada, M. K. Insulin receptors and insulin action in dissociated brain-cells. Brain Res. 417:247-256; 1987. Meister, B.; Hokfelt, T.; Tsuruo, Y.; Hemmings, H.; Ouimet, C.; Greengard, P.; Goldstein, M. DARPP-32, a dopamine- and cyclic AMP-regulated phosphoprotein in tanycytes of the medial basal hypothalamus: distribution and relation to dopamine and luteinizing hormone-releasing hormone neurons and other glial elements. Neuroscience 27607622; 1988. Mueckler, M.; Caruso, C.; Baldwin, S. A.; Panico, M.; Blench, I.; Morris, H. R.; Allard, W. J.; Lienhard, G. E.; Lodish, H. F. Sequence and structure of a human glucose transporter. Science 229:941-945; 1985. Oka, Y.; Asano, T.; Shibasaki, Y.; Kasuga, M.; Kanazawa, Y.; Takaku, F. Studies with antipeptide antibody suggest the presence of at least two types of glucose transporter in rat brain and adipocyte. J. Biol. Chem. 263:13432-13439; 1988. Sandoz, P. A. Perfusion induced edema does not disrupt perivascular glial sheaths in the rat area postrema-evidence for an inconspicuous type of glial junction. Acta Anat. 124:217-226; 1985. Shimizu, N.; Kumamoto, T. Histochemical studies on the glycogen of the mammalian brain. Anat. Rec. 114:479498; 1952. Stemberger, N. H.; Stemberger, L. A. Blood-brain barrier protein recognized by monoclonal antibody. Proc. Natl. Acad. Sci. USA 84:8169-8173; 1987. Tucker, S. P.; Cunningham, V. J. Autoradiography of [3-H] cytochalasin B binding in rat brain. Brain Res. 450:131-136; 1988. Turner, B. H.; Wilson, J. S.; McKenzie, J. C.; Richtand, N. MPTP produces a pattern of nigrostriatal degeneration which coincides with the mosaic organization of the caudate nucleus. Brain Res. 47360-64; 1988. Wang, C. The D-glucose transporter is tissue-specific-skeletal muscle and adipose tissue have a unique form of glucose transporter. J. Biol. Chem. 262:15689-15695; 1987. Young, J. K. Current evidence for a role of glucose as a regulator of hypothalamic function and caloric homeostasis. Psychoneuroendocrinology 6:281-299; 1981. Young, J. K. A glial toxin reduces effects of goldthioglucose on the hypothalamus and area postrema. Brain Res. Bull. 20:97-104; 1988. Young, J. K. The glial drug methionine sulfoximine reduces goldthioglucose lesions in mice. Brain Res. Bull. 22:929-936; 1989.
0361-9230/90
$3.00 + .OO
Glucose Transporter Immunoreactivity in the Hypothalamus and Area Postrema JOHN K. YOUNG Department of Anatomy, Howard University, 520 W Street, NW, Washington, D.C. 20059 AND CHUNG WANG Department of Biochemistry and Molecular Biology, Harvard University, Cambridge, MA 02138 Received 6 November
1989
YOUNG, J. K. AND C. WANG. Glucose transporter immunoreactivity in the hypothalamus and area postrema. BRAIN RES BULL 24(3) 525-528, 1990.-To study the glucose transporter (GT) protein in two glucose-sensitive areas of the rat brain, frozen coronal sections at the level of the median eminence (ME) and area postrema (AP) were stained immunocytochemically with an antibody raised against human erythrocyte glucose transporter. Immunoreactivity was mainly confined to blood vessels in most brain areas but was lacking in those of the ME and AF’,which also lack a normal blood-brain barrier. This suggests that glucose entry into these brain areas, unlike others, is not limited or regulated by capillary glucose transport systems. Tanycyte processes stained strongly for GT and for glycogen and thus may have an unusual glucose metabolism. Hypothalamus
Area postrema
Glucose transporter
Tanycytes
investigation of patterns of GT immunoreactivity in these specialized brain areas. This was the goal of this study.
SUBSTANTIAL advances in the understanding of glucose transport in various tissues has been made possible by the identification and sequencing of the glucose transport protein of human red blood cells, a protein first isolated by taking advantage of its ability to bind cytochalasin B in a manner specifically displaceable by glucose (18). Since then, numerous studies characterizing this protein in muscle, adipose, and brain tissue have appeared (2, 4, 5, 10, 19). Most studies of the transporter protein in brain have characterized it biochemically and reported a marked enrichment of glucose transporter (GT) protein in brain microvessel preparations (4,lO). Immunocytochemical studies of dog and human brains have also localized the GT protein to endothelial cells of blood vessels (7,14); thus far, the only analogous study of the rat brain has been a description of glucose-displaceable cytochalasin B binding in brain sections (23). A better understanding of glucose transport in two areas of the rat brain-neural tissue adjacent to the median eminence (ME) and area postrema (AP)-may be of particular value in understanding brain function. These two brain areas have a high blood-brain barrier permeability; also, elevations in blood glucose appear to influence the function of these areas and can cause localized cell death if blood glucose elevations are chronic (6, 13, 26). An antibody raised against the human erythrocyte glucose transporter was recently purified for biochemical and immunohistochemical study of the GT protein in muscle and in brain synaptosomes (2,25). This antibody was found to have a substantially greater affinity for rat brain microsomes than for microsomes from muscle and fat and thus would seem suitable for immunohistochemical
METHOD
Rabbit anibody against human erythrocyte GT protein was prepared and then affinity purified according to previously published methods (25). Afftnity purified antibody was then used, at a dilution of 1: 100, in a histochemical staining procedure described below. Brain tissue was obtained from five 3-month-old male SpragueDawley albino rats that were perfused through the heart with 4% paraformaldehyde in 0.2 M phosphate buffer (pH 7.3) under deep chloral hydrate anesthesia. Fixative was contained in a reservoir 166 cm above each rat, producing an estimated IV pressure of about 122 mm of mercury. Brains were removed, fixed overnight, and then transferred to 20% sucrose-formalin until sinking; then, 30 p frozen sections were cut and placed in phosphate buffered saline (PBS-8.5 g NaCl, 0.85 g NaaHPO,, and 0.54 g KH,PG, per 1 of distilled water). After washing for several hr in PBS, free-floating sections were processed through the following steps (24): 1) 30-min incubation in PBS + 0.3% Triton X-100 containing 3% normal goat serum, 2) 24-hr incubation (4 degrees C) in GT antibody, diluted 1:lOO in PBS + 2% normal goat serum + 0.3% Triton X-100, 3) three five-mm washes in PBS-O. 1% T&on, 4) 30-min incubation with biotinylated goat-anti-rabbit IgG (Vectastain ABC kit, Vector Laboratories, Burlingame, CA), 5) three
525
YOUNG AND WANG
FIG. 1, (A) Coronal section of hypothalamus, showing the distribution of immunoreactivity for glucose transporter protein. Immunoreactive blood vessels of various sizes are visible in all areas except the median eminence. Also, strongly staining, thin tanycyte processes can be seen (arrow). (B) Control section stained with nonimmune rabbit serum. (C) Section stained for glycogen using the periodic acid-Schiff method. Glycogen deposition in tanycyte processes is visible (arrow). Magnification bar= 100 IL.
five-mm washes in PBS-0.1% Triton, 6) 30-min incubation in PBS containing 0.3% hydrogen peroxide to nullify sources of endogenous pseudoperoxidase, and 7) three more washes in PBS, followed by a 60-min incubation in ABC reagent (avidin-biotinperoxidase complex). After completion of these steps, sections were again washed and incubated for 15 min in a solution of 0.05% diaminobenzidine in Tris buffer (pH 7.2) containing 0.03% hydrogen peroxide to produce a brown precipitate over rabbit IgG binding sites. Finally, sections were dried onto gelatinized slides and coverslipped. Control sections were incubated with nonimmune rabbit serum diluted in 2% normal goat serum-PBS to examine nonspecific binding. Special features of these brain areas were also illustrated by other methods. Glycogen in the arcuate area was demonstrated by rapidly perfusing 3 rats with fixative, fixing brains overnight, immersing tissue blocks in acidic alcohol + formalin for 2 hr, and then preparing 30 p frozen sections for staining (9). Sections were stained for glycogen via the periodic acid-Schiff method after blockade of nonspecific aldehyde groups by immersion in alcoholic dimedone (9). Specific staining of glycogen was checked by comparison with sections that had been exposed to saliva for 1 hr at 40 degrees C. In order to more clearly visualize the appearance of blood vessels in the area postrema, three additional rats were perfused with 0.5% pamformaldehyde-1.5% glutaraldehyde in 0.2 M phosphate buffer, small blocks of medulla were dehydrated, and then were embedded in EM Bed 812 resin (Electron Microscopy Sciences, Fort Washington, PA). One micron thick sections of this tissue were cut using a Sorvall JV-4 microtome and dry glass knives and were stained with 0.1% toluidine blue in borate buffer. RESULTS lrnrnunoreactivity
for the GT protein
in most of the coronal
brain sections was restricted to a thin rim on the inner surfaces of blood vessels of all sizes (Figs. 1 and 2). This thin layer of reaction product was not wider in larger blood vessels than in capillaries, suggesting that the smooth muscle layers of these larger blood vessels did not contribute to this immunoreactivity. In the hypothalamus, intense staining of ependymal tanycyte processes traversing the arcuate nucleus was also observed (Fig. 1A). Examination of similar sections after periodic acid-Schiff staining revealed dense accumulations of PAS positive material in tanycyte processes (Fig. 1C) that presumably was glycogen, since prior exposure of sections to saliva prevented this staining reaction (not shown). Immunoreactive blood vessels were absent from two brain areas. Inspection of 10 sections in which the median eminence had not been tom during sectioning failed to find a single blood vessel with clear immunoreactivity, although numerous immunoreactive blood vessels were present in the nearby arcuate nucleus and irregular patches of immunoreactivity were present in the median eminence (Fig 1). Also, 21 sections through the area postrema showed a striking absence of immunoreactive vessels in this region (Fig. 2). Semi-thin sections of this area showed a marked perivascular edema around vessels within the AP itself, but a normal appearance of the neuropil around vessels just a short distance away from the AP (Fig. 2B and C). Inspection of other circumventricular organs (organum vasculosum of the lamina terminalis in the anterior hypothalamus and the subfornical organ) that are also known to have unusually permeable capillaries failed to reveal any distinctive patterns of immunoreactivity: immunoreactive blood vessels close to the ventricular margin were seen in both areas (not shown). In general, sections stained with nonimmune rabbit serum showed only a low level of light brown, nonspecific staining (Fig. 1B).
GLUCOSE TRANSPORTERS
527
IN THE HYPOTHALAMUS
FIG. 2. (A) Coronal section of the area postrema in the medulla, showing the distribution of immunoreactivity for glucose transporter protein. Blood vessels within the triangular-shaped area postrema proper are unstained. (B) Semi-thin (1 u) section of area uostrema embedded in EM Bed 812 resin. Blood vessels within the area-postrema proper show pericapillary edema (arrow). (C) Portion of B at a higher magnification, showing capillary with pericapillary edema (arrow). Note the more normal appearing capillary at the right, outside of the area postrema. Magnification bars= 100 p.
DISCUSSION
The basic finding of this study-that the GT protein bound by this antibody is mainly detected in blood vessels of rat brain tissue and is absent from blood vessels of the ME and AP-is in accord with biochemical and histochemical studies (4,5, 10) and with the absence of GT immunoreactivity in AP vessels in the human brain reported by Kalaria et al. (14). The absence of GT immunoreactivity in blood vessels of the median eminence and area postrema
is likely related to the high permeability of these vessels: a specialized glucose transport system across these vessels would seem unnecessary (8). Other authors have also reported that rat AP vessels fail to stain for another protein associated with blood-brain barrier function in brain capillaries (22). In contrast to these two circumventricular organs, capillaries near the organum vasculosum lamina terminalis and subfomical organ did appear immunoreactive; this may reflect a functional difference from the AP and MB, or may partly be due to the absence of a clear anatomical border around the most permeable areas of these latter organs that made it difficult to pinpoint a specific area that would contain specialized capillaries. Some apparent variation in capillary characteristics between different circumventricular organs is supported by staining of pineal, but not AP capillaries, for GT immunoreactivity in the human brain (14). Another indication of a specialized function of capillaries in the area postrema was the marked pericapillaty edema found around AP vessels. This is probably not a simple artifact of excessive perfusion pressure, since nearby capillaries had a normal appearance. The ultrastructure of this phenomenon was studied by Sandoz, who concluded that the pericapillary edema was the result of both highly permeable capillaries and a relatively impermeable perivascular glial sheath that surrounded them and promoted perivascular fluid accumulation (20). According to this data, glucose entry into the area postrema would thus be regulated by glia and not by capillary endothelia. This interpretation would be consistent with data indicating an important influence of glia upon the reactivity of tissue near the area postrema to a toxic form of glucose, goldthioglucose (27,28). Staining of tanycyte processes for the GT protein, in addition to glycogen deposition in tanycytes, a phenomenon known for some time (21), suggests that tanycytes may have a specialized glucose metabolism. The functional significance of this is uncertain. Tanycytes have a number of unusual features, including positive staining for iron and for proteins associated with GABA, dopamine, and glutamate uptake and metabolism (1, 15, 17, 26). Tanycyte processes show a close anatomical relationship with neural processes containing dopamine or LHRH and may influence the excitability of these processes (12,17). Uptake of glucose by tanycytes, causing a change in tanycyte function, may thus be one way in which glucose influences LHRH secretion (3). A more precise localization of GT immunoreactivity in semi-thin and thin sections would provide more information about brain glucose transport; however, we and others have found that exposure of tissues to glutaraldehyde or osmium abolishes GT antigenicity, making adequate preservation of ultrastructure problematic. The only study successfully overcoming this problem has reported that 89% of GT immunoreactivity is associated with luminal and abluminal plasma membrane of endothelial cells (7). The reasons for the absence of GT immunoreactivity in the neuropil in this study, as well as in other studies, are not clear. Neurons are known to have a lowered content of GT protein compared to most cells, and unlike glia or other cell types, fail to increase glucose uptake in response to insulin (11,16). Altered amounts, molecular configuration (glycosylation), and antigenicity of neural GT protein may thus underlie the lack of neuropil immunoreactivity (11,19). Other cytochemical/biochemical approaches may be required to investigate the basis of neuronal glucose uptake in the hypothalamus and elsewhere in the brain.
ACKNOWLEDGEMENT Supported ton, D.C.
by a grant from the Sugar
Association,
Inc.,
Washing-
528
YOUNG
AND WANG
REFERENCES 1. Alho, H.; Fremeau, R. T., Jr.; Tiedge, H.; Wilcox, J.; Bovolin, P.; Brosius, J.; Roberts, J. L.; Costa, E. Diazepam binding inhibitor gene expression: location in brain and peripheral tissue of rat. Proc. Natl. Acad. Sci. USA 85:7018-7022; 1988. 2. Bhattacharyya, M. V.; Brodsy, J. L. Characterization of the glucose transporter from rat brain synaptosomes. Biochem. Biophys. Res. Commun. 155:685-691; 1988. 3. Bourguignon, J.-P.; Gerard, A.; Debougnoux, G.; Rose, J.; Franchimont, P. Pulsatile release of gonadotrophin-releasing hormone (GnRH) from the rat hypothalamus in vitro: calcium and glucose dependency and inhibition by superactive GnRH analogues. Endocrinology 121: 993-999; 1987. 4. Dick, A. P. K.; Harik, S. I. Distribution of the glucose transporter in the mammalian brain. J. Neurochem. 46:1406-1411; 1986. 5. Froehner, S. C.; Davies, A.; Baldwin, S. A.; Lienhard, G. E. The blood-nerve barrier is rich in glucose transporter. J. Neurocytol. 17:173-178; 1988. 6. Garris, D. R.; West, R. L.; Coleman, D. L. Morphometric analysis of medial basal hypothalamic neuronal degeneration in diabetes (db/db) mutant C57BLiKSJ mouse-relation to age and hyperglycemia. Dev. Brain Res. 20:161-168; 1985. 7. Gerhart, D. Z.; Levasseur, R. J.; Broderius, M. A.; Drewes, L. R. Glucose transporter localization in brain using light and electron microscopy. J. Neurosci. Res. 22:464-472; 1989. 8. Gross, P. M.; Blasberg, R. G.; Fenstennacher, J. D.; Patlak, C. S. The microcirculation of rat circumventricular organs and pituitary gland. Brain Res. Bull. 18:73-85; 1987. 9. Guth, L.; Watson, P. K. A correlated histochemical and quantitative study on cerebral glycogen after brain injury in the rat. Exp. Neurol. 22:590X102; 1968. 10. Harik, S. I.; Gravina, S. A.; Kalaria, R. N. Glucose transporter of the blood-brain barrier and brain in chronic hyperglycemia. J. Neurothem. 51:1930-1934; 1988. 11. Heidenreich, K. A.; Gilmore, P. R.; Garvey, W. T. Glucose transport in primary cultured neurons. J. Neurosci. Res. 22:397-7; 1989. 12. Jarvis, C. R.; Andrew, R. D. Correlated electrophysiology and morphology of the ependyma in rat hypothalamus. J. Neurosci. 8:3691-3702; 1988. 13. Kadekaro, M.; Savaki, H.; Sokoloff, L. Metabolic mapping of neural pathways involved in gastrosecretory response to insulin hypoglycemia in the rat. J. Physiol. (Lond.) 300:39347; 1980. 14. Kalaria, R. N.; Gravina, S. A.; Schmidle, J. W.; Perry, G.; Harik, S. I. The glucose transporter of the human brain and blood-brain barrier. Ann. Neurol. 24:757-764; 1988. 15. Kohler, C.; Okuno, E.; Flood, P. R.; Schwartz, R. Quinolinic acid
16.
17.
18.
19.
20.
21. 22.
23. 24.
25.
26.
27. 28.
phosphoribosyl-transferase: preferential glial localization in the rat brain visualized by immunocytochemistry. Proc. Natl. Acad. Sci. USA 84:3491-3495; 1987. Masters, B. A.; Judkins, J. H.; Clarke, D. W.; LeRoith, D.; Raizada, M. K. Insulin receptors and insulin action in dissociated brain-cells. Brain Res. 417:247-256; 1987. Meister, B.; Hokfelt, T.; Tsuruo, Y.; Hemmings, H.; Ouimet, C.; Greengard, P.; Goldstein, M. DARPP-32, a dopamine- and cyclic AMP-regulated phosphoprotein in tanycytes of the medial basal hypothalamus: distribution and relation to dopamine and luteinizing hormone-releasing hormone neurons and other glial elements. Neuroscience 27607622; 1988. Mueckler, M.; Caruso, C.; Baldwin, S. A.; Panico, M.; Blench, I.; Morris, H. R.; Allard, W. J.; Lienhard, G. E.; Lodish, H. F. Sequence and structure of a human glucose transporter. Science 229:941-945; 1985. Oka, Y.; Asano, T.; Shibasaki, Y.; Kasuga, M.; Kanazawa, Y.; Takaku, F. Studies with antipeptide antibody suggest the presence of at least two types of glucose transporter in rat brain and adipocyte. J. Biol. Chem. 263:13432-13439; 1988. Sandoz, P. A. Perfusion induced edema does not disrupt perivascular glial sheaths in the rat area postrema-evidence for an inconspicuous type of glial junction. Acta Anat. 124:217-226; 1985. Shimizu, N.; Kumamoto, T. Histochemical studies on the glycogen of the mammalian brain. Anat. Rec. 114:479498; 1952. Stemberger, N. H.; Stemberger, L. A. Blood-brain barrier protein recognized by monoclonal antibody. Proc. Natl. Acad. Sci. USA 84:8169-8173; 1987. Tucker, S. P.; Cunningham, V. J. Autoradiography of [3-H] cytochalasin B binding in rat brain. Brain Res. 450:131-136; 1988. Turner, B. H.; Wilson, J. S.; McKenzie, J. C.; Richtand, N. MPTP produces a pattern of nigrostriatal degeneration which coincides with the mosaic organization of the caudate nucleus. Brain Res. 47360-64; 1988. Wang, C. The D-glucose transporter is tissue-specific-skeletal muscle and adipose tissue have a unique form of glucose transporter. J. Biol. Chem. 262:15689-15695; 1987. Young, J. K. Current evidence for a role of glucose as a regulator of hypothalamic function and caloric homeostasis. Psychoneuroendocrinology 6:281-299; 1981. Young, J. K. A glial toxin reduces effects of goldthioglucose on the hypothalamus and area postrema. Brain Res. Bull. 20:97-104; 1988. Young, J. K. The glial drug methionine sulfoximine reduces goldthioglucose lesions in mice. Brain Res. Bull. 22:929-936; 1989.