Glucose transporter gene expression in rat brain: Pretranslational changes associated with chronic insulin-induced hypoglycemia, fasting, and diabetes

MOLECULAR

AND CELLULAR

NEUROSCIENCES

2, 244-252

(1991)

Glucose Transporter Gene Expression in Rat Brain: Pretranslational Changes Associated with Chronic Insulin-Induced Hypoglycemia, Fasting, and Diabetes LASZLO KORANYI, RAYMOND E. BOUREY, DAVID JAMES,* MIKE MUECKLER,* FREDERICK T. FIEDOREK, JR., AND M. ALAN PERMUTT Department

of Internal Washington

Medicine, University

Division School

of Metabolism, and *Department of Medicine, St. Louis, Missouri

Received for publication

Steady-state levels of the major glucose transporter gene (GLUT1) of the brain were evaluated under three conditions that induced chronic changes in plasma glucose and insulin in adult rats: (i) repeated injection of insulin for 6 days, resulting in plasma glucose levels of 60-70 mg/dl for at least 3 days; (ii) fasting for 3 days; and (iii) moderate streptozotocin-induced diabetes of 1 week duration. Brain GLUT- 1 mRNA was measured by dot blot hybridization with a HepGB/erythrocyte (GLUT1) [92P]cRNA probe, and GLUT-l protein by immunoblot analysis with a polyclonal antibody (R493). Insulin injection resulted in hypoglycemia, increased GLUT-l mRNA (143 f 15%, P -C O.OS), and increased GLUT-l protein (141 f 6%, P < 0.06). The increase in GLUT-l mRNA was specific for brain, as no change was observed in liver or kidney. Fasting resulted in mild hypoglycemia, lower plasma insulin, increased GLUT-l mRNA (131 f 17%, P < 0.05 vs control), and no change in GLUT-l protein (125 + 9%, N.S.). Mild streptozotocin diabetes resulted in hyperglycemia, undetectable plasma insulin, decreased GLUT- 1 mRNA (65 + 6%, P < 0.05 vs control), and no change in GLUT-l protein (84 f 9%, N.S.). A negative correlation (r = -0.61, P < .OOOl) between GLUT- 1 mRNA levels in brain and plasma glucose concentrations was observed among the three experimental groups and control animals, suggesting that the plasma glucose concentration may be at least one determinant of GLUT-l levels in rat brain. The importance of these results is the finding that GLUT-l gene expression in rat brain is regulated in uiuo by the nutritional and endocrine status of the animal. o 1991 Academic PEU.S, IIIC.

INTRODUCI-ION Glucose is the major fuel for energy metabolism in the brain, accounting for at least 90% of oxidative metabolism

1044-7431/91$3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

244

of Cell Biology, 63110

April 22, 1991

(1,Z). Because glucose is not stored by the brain, normal metabolism is dependent on glucose supplied to neuronal and glial cells through brain microvessels that are characterized by tight endothelial junctions (2-4). The brain microcirculation is poorly permeable to most molecules, thus forming the so-called blood-brain barrier. In viva studies of brain glucose uptake assess predominantly transport across the blood-brain barrier that is mediated by a facilitated diffusion process (2-5). Subcellular distribution studies using cytochalasin B binding indicate that the density of glucose transporter molecules is at least lo-fold higher in rat microvessels than in particulate fractions of cerebrum, cerebellum, or cerebral synaptosomes (6,7). The maximum transport capacity of this carrier is insulin-independent and exceeds by twoto threefold the normal rate of glucose utilization by brain (3,4,8,9). The glucose transport system appears to play little role in normal brain oxidative metabolism, yet during periods of acute hypoglycemia, brain glucose and oxidative metabolism can be reduced to levels that result in permanent brain damage (10). There is now substantial evidence, however, that under conditions of chronic hypoglycemia, glucose transport across the blood-brain barrier is enhanced (ll), while in chronic hyperglycemia this process is decreased (12-15). Recent cloning of glucose transporter genes has facilitated the study of adaptive changes of the brain glucose transport system at the molecular biochemical level. The first glucose transporter cDNA clones isolated were for the human erythrocyte glucose transporter (16) of hepatoma cells (HepG2/erythrocyte or GLUT-l) and the rat brain glucose transporter (17). Both cDNAs predict 492 amino acid proteins that are 97% identical. The GLUT1 gene was shown to be expressed most abundantly in brain and to a lesser extent in most other rat tissues as a 2.8-kb mRNA. At least four other glucose transporters

GLUCOSE

TRANSPORTER

GENE

have been described. These include the liver-type glucose transporter (GLUT-2) that is present in splanchnic tissues, kidney, and pancreatic islets (18-20) and the insulinregulatable muscle/adipose tissue (GLUT-4) glucose transporter that is expressed in tissues that exhibit insulin-dependent glucose transport (21-24). A brain glucose transporter was isolated from human brain (GLUT3) and shown to have a tissue distribution similar to that of GLUT-l in humans (25). In rats, however, GLUT-3 appears to be expressed only in brain as a single 4.1-kb mRNA, and to a considerably lower level than GLUT-l (25). While GLUT- 1 mRNA and protein are abundant in rodent microvessels and in cells with occluding junctions at the blood-eye barriers, it is unknown whether GLUT3 is also expressed in these tissues. In contrast to GLUT1, the immunochemical localization of GLUT-3 in brain has not been determined. Little is known about regulation of expression of the GLUT-l gene in u&o, except that the GLUT-l gene appeared to be constitutively expressed in adipose tissue of streptozotocin diabetic rats and fasted rats, while a striking decrease in GLUT-4 levels was observed (26-28). Now that the major glucose transporter of brain has been defined, and because significant changes in brain glucose transport occur with chronic hyperglycemia and hypoglycemia, the present in viva experiments were designed to determine whether these conditions would be associated with changes in GLUT-l mRNA and/or protein. Consequently we evaluated these parameters (i) in rats made diabetic by a single injection of streptozotocin, (ii) in rats made hypoglycemic by repeated injections of exogenous insulin, and (iii) in 3-day fasted rats. We observed significant increases in GLUT-l mRNA and protein in insulin-injected chronically hypoglycemic rats. These increases were specific for brain tissue. In addition, a striking correlation between GLUT-l mRNA and plasma glucose concentration was observed in all the experimental groups. METHODS

Animals Male Sprague-Dawley rats (150-250 g) were purchased from SASCO (Omaha, NE), caged in groups, and fed a standard diet of Purina Rat Chow (Ralston Purina, St. Louis, MO) and tap water ad lib&urn except as described. Animals were exposed to light between 6~00 AM and 8~00 PM daily. Diabetes was induced by a single intraperitoneal injection of streptozotocin (75 mg/kg body wt, Sigma, St. Louis, MO). The animals were observed for 7 days and the existence of diabetes was confirmed by daily measurements of blood glucose using Dextrostix strips and a glucometer (Ames, Miles Laboratory, Elkhart, IN).

EXPRESSION

IN

RAT

245

BRAIN

Hypoglycemia was produced by daily insulin injection (Ultra-lente, Squibb-Novo, Princeton, NJ, 40 u/kg body wt/day). Blood glucose levels in rats were determined three times a day and the insulin dose was adjusted (f5 u/kg body wt/day) to maintain the blood glucose levels at 60-70 mg/dl for 3 days prior to analysis. Animals in one group were individually housed and fasted for 3 days. Blood and Tissue Sampling The control (n = ll), diabetic (n = ll), and hypoglycemic (n = 11) animals were killed in the fed state, and the fasted (n = 10) animals in the fasted state, using pentobarbital sodium (100 mg/kg of body wt, ip). Before injection whole blood was collected from tail veins in tubes containing Paroxan (0.04%, Sigma, St. Louis, MO), EDTA (5.0 mA4), and Trasylol(9000 U/ml of blood, FBA, New York, NY), and before cardiorespiratory arrest, tissues were quickly removed, washed, and frozen in liquid nitrogen and stored at -70°C until extraction. The tissues were homogenized in guanidine thiocyanate and total RNA was prepared by multiple ethanol precipitation of guanidine thiocyanate, guanidine hydrochloride, and water extracts of homogenized tissues as described (29). For GLUT-l protein determinations (see below), control and experimental samples were analyzed by SDSPAGE in adjacent lanes in every case. We were concerned that repeated freezing and thawing of the protein samples from the control animals may have caused some protein degradation, so we therefore performed an additional experiment using three more groups of rats: (i) controls (n = 6), (ii) insulin-injected hypoglycemic (n = 6), and (iii) 3-day fasted animals (n = 6). The animals did not differ in initial and final weights or in glycemic levels compared to animals from the same experimental condition in the first experiment. cDNA Probes and Synthetic

mRNA

The probes used for study were as follows: (i) human erythrocyte/brain (GLUT-l) glucose transporter cDNA (2300 bp), isolated from a HepG2 cDNA library as described (16), subcloned into an RNA transcription vector (Bluescript SK+, Stratagene, La Jolla, CA), and (ii) chicken /3-actin cDNA (1200 bp) subcloned into another RNA transcription vector, pGEM-1 (Promega, Madison, WI). Transcription of uniformly labeled [32P]cRNA and synthetic mRNA with T7, SP6, or T3 RNA polymerase was performed according to protocols provided by suppliers. Quuntitation

of mRNAs

The quality of extracted total tissue RNA was checked by electrophoresis and ethidium bromide staining on aga-

246

KORANYI

rose-formaldehyde gels and by hybridization with [32P]actin cRNA. All samples had 28 S/l8 S RNA ratios over 2.0 and discrete actin mRNA bands on Northern analysis. Initial studies to establish appropriate stringency for hybridization and to validate specificity were performed using Northern analysis. Quantitation of mRNA levels was performed using dot blot analysis. Aliquots of total tissue RNA (l-10 pg) and dilutions of corresponding synthetic mRNA (0.5-1000 pg) and cDNA (l-1000 pg) as standards were dissolved in 15% formaldehyde, 10X SSC; blotted onto Nytran membranes (Schleicher & Schuell, Keene, NH); baked at 85°C for 2 h; prehybridized for 1 h at 55°C; hybridized with the corresponding cRNA probe (l-10 X lo6 cpm/ml) in 50% formamide, 5X SSPE, 5~ Denhardt’s, 1% SDS, 200 pg/ ml denatured salmon sperm DNA overnight at 55°C; and then washed in 0.1X SSC, 0.1% SDS at 60°C according to the protocol recommended by the vendor. Blots were exposed to Kodak XAR5 film at -70°C with an intensifying screen (Cronex Lightening Plus, E. I. Dupont de Nemours Co., Wilmington, DE). Samples were analyzed in duplicate and the amount of mRNA was measured by densitometric analysis (Bio-Rad EIA Reader) of autoradiographs comparing the intensity of the sample to standards. Autoradiographs were developed for various periods of time so that the intensity of the samples fell within the linear range of the standards. Measurement

of Plasma Parameters

Plasma insulin concentrations were determined by a double antibody radioimmunoassay using rat insulin standards ( NOVO, Copenhagen, Denmark) and plasma glucose by a glucose oxidase method using a Beckman Glucose Analyzer-2 (Beckman Instruments, Fullerton, CA). The plasma free fatty acid levels were measured by microfluorometric assay (30) and P-hydroxybutyrate and acetoacetate (31) in the laboratory of Dr. Irene Karl. Data Analysis Statistical analysis among experimental groups was performed by analysis of variance and between groups by unpaired Student’s t test and regression analysis. Data were managed and analyzed using the CLINFO Data Analysis System of Washington University (G.C.R.C. RR00036) supported by the Division of Research Resources of the NIH. Determination

of Brain GLUT-l

Protein Content

Brain was homogenized (Brinkman, Westbury, NY) on ice at maximum speed for 45 seconds in phosphate-buffered saline (1:20, wt/vol). Cerebral microvessels were prepared as previously described (32). Protein concentration was determined by a modification of the method of

ET

AL.

Lowry (33). Protein (50-100 pg) was subjected to SDSPAGE and electrophoretically transferred (Polyblot, ABN, Hayward, CA) to nitrocellulose paper (Schleicher & Schuell, Keene, NH). After blocking with 5% dry milk (Carnation Co., Los Angeles, CA) in PBS for 90 min, the blots were incubated in 1% dry milk in PBS at 37°C for 1 h with R493 polyclonal anti-rat GLUT-l (21). The blots were then washed with PBS and incubated for 1 h at 37°C with 2.5 &i ‘251-labeled protein A (Amersham, Arlington Heights, IL), rewashed, dried, and subjected to autoradiography. The labeled bands were traced on the nitrocellulose, excised, and analyzed by liquid scintillation counting. Equal areas away from the band of interest were excised and counted to establish background. The mean cpm above background/100 pg protein was 560 f 40 (n = 109). The value obtained for each sample was divided by the mean of the six control values for the blot to allow direct comparison of values between blots. The intra- and interblot coefficients of variation were 13% (the same sample run in 12 adjacent lanes) and 18% (mean of six samples, 10 blots) respectively. The GLUT-l protein content of each brain was determined as the mean of values from at least three blots. RESULTS Metabolic Alterations

in the Experimental

Models

Insulin-treated rats (hypoglycemic) gained more weight than did controls (P < 0.05) during the week of treatment (Table 1). There was a marked lowering of the plasma glucose concentration to 61 -t 14 mg/dl at the time of tissue sampling. No differences were noted in free fatty acid or keto acid levels between the insulin-treated and control animals. Rats fasted for 3 days lost weight relative to their initial body weights and weighed less than the control animals (P < 0.05) as a result. Plasma glucose and insulin concentrations were decreased and P-hydroxybutyrate and acetoacetate were increased relative to levels in control animals and increased relative to those in insulin-injected rats (P < 0.05). One week after the injection of streptozotocin, animals gained significantly less weight than control animals (P < 0.05). There was a marked increase in plasma glucose (P < 0.05), and plasma insulin was below the limits of detection (co.1 rig/ml) by the radioimmunoassay. While plasma free fatty acids were unchanged, P-hydroxybutyrate and acetoacetate were increased in the diabetic animals relative to levels in controls. Brain Glucose Transporter Concentrations

Messenger RNA

Total RNA and actin mRNA concentrations in brain (Table 2, Fig. 1A) did not differ among the four groups

GLUCOSE

TRANSPORTER

GENE

TABLE Metabolic

Alterations

28.5 142 1.2 1.08 128 28

of body weight (g) glucose (mg/dl) insulin (rig/ml) FFA (mmol/liter) B-OHBA (pmol/liter) AcOAc (pmol/liter)

f + f f + f

Hypoglycemic (n = 11) 44.1 k 11.4* 61 k 14* 0.66 k 0.08 109 * 5 28 + 2.5

3.5 5.2 0.1 0.4 21 3.0

IN

RAT

241

BRAIN

1

in the Experimental

Control (n = 11) Change Plasma Plasma Plasma Plasma Plasma

EXPRESSION

Models Fasted (n = 10) -52.9 100 0.18 1.68 397 40

f f + + f +

cko* 12* 0.04* 0.6 84** 9.4**

Diabetic (n = 11) 7.1 f 4.2* 523 3~ 32*
Note. Values are X f SEM. * P i 0.05 vs control. ** P < 0.05 vs control and hypoglycemic.

(ANOVA, N.S.). In contrast, GLUT-l mRNA in brain was increased in insulin-treated (147 +- lo%, P < 0.05) and in fasted rats (131 + 17%, P 0.05) and decreased in diabetic animals (65 f 5%, P < 0.05) relative to concentrations in controls. In liver, there were no changes in total RNA, actin mRNA, or GLUT-l mRNA concentrations among the four groups (Table 2, Fig. 1B). In kidney, there were no differences in total tissue RNA or in actin mRNA concentrations in any of the experimental conditions (Table 2, Fig. 1C). In kidney GLUT-l mRNA concentration was significantly increased (146 + 9%, P c 0.05) only in fasted animals relative to the GLUT-l mRNA concentration of control animals. Brain Glucose Transporter Immunoblotting

Protein Determinations

by

The rat brain glucose transporter protein has been defined by cytochalasin B photo-labeling and immunoblot-

TABLE Total

RNA, and GLUT-l, Control (n = 11)

Brain pg pg pg Liver pg/ pg pg Kidney pg pg pg

and Actin

ting to be a glycosylated protein with an apparent molecular weight on SDS-PAGE gel electrophoresis of 4355 kDa (6-8, 17, 34-43). The concentration of GLUT-l protein in total brain protein in control and experimental animals was measured by electrophoresis, followed by immunoblotting with a polyclonal antibody specific for the rat GLUT-l protein (21). Quantitation performed in at least three separate analyses for each experimental condition (Table 3) showed no difference in GLUT-l protein between diabetic and fasted animals, while there was an increase (141+- 6%, P = 0.002) in insulin-injected hypoglycemic rats relative to the concentration in control animals. An illustrative immunoblot is shown in Fig. 2. While a number of immunocytochemical studies of brain have suggested that microvessels are enriched in GLUT-l protein (34, 44-47), the relative amount of microvessel GLUT-l protein in whole brain is unknown. To

2 mRNA

in Brain,

Hypoglycemic (n = 11)

Liver, and Kidney Fasted (n = 10)

Diabetic (n = 11)

RNA/g tissue GLUT-l mRNA/rg RNA actin mRNA/pg RNA

923 2 21 49 5 0.8 218 f 11

881 * 94 70 f 5.0* 217 + 33

793 + 110 64 + 9.2* 249 rf- 31

830 5~ 61 32 + 2.1* 220 * 39

RNA/g tissue GLUT-l mRNA/ag RNA actin mRNA/pg RNA

779 f 23 2.26 f 0.1 373 * 31

943 If: 96 2.2 f 0.4 254 rt_ 84

574 + 65 5.9 f 2.8 423 2 59

779 f 93 3.1 f 0.7 416 + 104

RNA/g tissue GLUT-l mRNA/pg RNA actin mRNA/pg RNA

875 + 50 2.41 + 0.1 415 + 58

784 5~ 36 2.35 + 0.3 298 f 36

966 + 42 3.52 + 0.2* 327 + 30

872 f 56 2.36 f 0.2 338 2 35

Note. Values are X + SEM. * P < 0.05 vs control.

248

KORANYI

ET

AL.

too 80 60 40 20 0

HFCHFCHFCHFCHFC

FIG. 2. Illustrative immunoblot. Brain tissue from insulin-injected hypoglycemic (H), 3-day fasted (F), and control (C) rats was homogenized; protein was subjected to SDS-PAGE, electroblotted to membranes, and treated with anti-GLUT-l antibody and ‘251-staph A protein; and autoradiographs were obtained as described under Materials and Methods. Molecular weights were estimated with protein standards.

whole brain migrated as a wide band with an average size of 43 kDa, while GLUT-l protein in cerebral microvessels migrated as a larger and sharper band of approximately 50 kDa. Thus it appears that microvessel GLUT-l protein does not account for the majority of GLUT-l protein detected by immunoblotting of whole brain. Correlation Anulysis

FIG. 1. Changes in total tissue RNA, brain glucose transporter (GLUT-l) mRNA, and actin mRNA in (A) brain, (B) liver, and (C) kidney under the three experimental conditions as percentage of change in control animals (X f SEM, n = lo-11 for each group, *P c 0.05 vs control), from the data in Table 2.

estimate the fraction of whole brain GLUT-l protein that could be accounted for by microvessel GLUT-l protein, whole brain and microvessel-enriched fractions were subjected to immunoblot analysis with the polyclonal GLUT1 antibody (Fig. 3). Immunoreactive GLUT-l protein in

TABLE Rat

Brain

Plasma glucose (mddl) GLUT-l protein (% control)

Glucose

Diabetic (n = 6)

118 -c 6 100 + 6

Note. Values are X + SEM. * P < 0.05 vs control. ** P = 0.02 vs control.

3

Transporter

Control (n = 12)

A correlation matrix was constructed between plasma glucose, ,8-hydroxybutyrate, and acetoacetate concentrations and GLUT-l mRNA concentrations in brain, liver, and kidney for all determinations in all groups of animals (Table 4). The strongest correlation for GLUT-l mRNA in brain was with plasma glucose concentration (r = -0.61, P < O.OOOl), when considering all four groups including the markedly hyperglycemic diabetic rats (Fig. 4A). The correlation was considerably less when the diabetic group was excluded from analysis (Fig. 4B, r = -0.37, P < 0.02). A less strong correlation for GLUT1 mRNA in brain with plasma acetoacetate was also noted (r = -0.51, P < 0.01).

(GLUT-l)

Protein

Hypoglycemic (n = 6)

Fasted (n = 6)

523 2 35*

53 + 5*

109 + 4

84 k 9

141 f 6**

125 + 9

I

2

3

FIG. 3. Immunoblot analysis of (1) whole brain (2) whole brain (50 pg) + brain microvessel protein brain microvessel protein (100 pg) with anti-GLUT-l scribed in Fig. 2.

protein (100 gg), (50 pg), and (3) antibody as de-

GLUCOSE

TRANSPORTER

GENE

TABLE

EXPRESSION

IN

RAT

249

BRAIN

4

Correlation Matrix Plasma glucose Plasma Plasma

glucose (3-OHBA

Plasma @-OHBA

Plasma AcOAc

Brain

GLUT-l mRNA

Liver

GLUT-l mRNA

Kidney

GLUT-l mRNA

1.0 1.0

0.36

(ns) Plasma

AcOAc

Brain

GLUT-l

mRNA

Liver

GLUT-l

mRNA

Kidney

GLUT-l

0.69 (<0.ooo5) -0.18 bs) -0.13 (ns)

0.70

(<0.6001) -0.61 (<0.0001) -0.09 (ns)

mRNA

-0.26

1.0 -0.51 (
1.0

-0.22

-0.06

(ns)

0.03

-0.21

(ns)

(ns)

(ns)

DISCUSSION In this study repeated insulin injections into rats resulted in marked hypoglycemia (Table l), an increase (43%, P < 0.05) in the concentration of GLUT-l mRNA, and a comparable increase (41%, P < 0.05) in the concentration of the GLUT-l protein in brain. The effects of chronic hypoglycemia on GLUT-l mRNA and protein in whole brain may reflect changes in microvessel glucose transporters, as these have been shown to be lo-20 times more abundant than those in neuronal or glial cells (69, 48). Recent RNA hybridization (49) and immunocytochemical studies (44-47,50) suggested that the GLUT1 gene is expressed predominantly in microvessels. Northern blot analysis of RNA from microvessel-enriched

A 000 8

600

^

400

9 F

200

0

0

20

30

40

xl

60

70

80

90

100

IO

20

30

40

50 GLUT-i PQ +g

60

70 mRNA RNA

80

90

loo

FIG. 4. Correlation analysis of plasma glucose and GLUT-l mRNA in (A) control, diabetic, hypoglycemic, and fasted rate and in (B) control, hypoglycemic, and fasted rats, with diabetic rats excluded. For A, r = -0.61, P < 0.0001, and for B, r = -0.37, P < 0.02.

1.0

(ns) -0.42 (<0.02)

0.26

1.0

(ns)

brain fractions revealed that GLUT-l was expressed almost exclusively in microvascular endothelium (49). These studies were confirmed by in situ hybridization analysis further suggesting minimal expression of GLUT1 in neuronal or glial cells (51). There is reason to suspect, however, that whole brain GLUT-l mRNA and protein as assessed in the current study do not necessarily reflect changes in brain microvessels exclusively. Abundant GLUT-l immunofluorescence has been observed in brain outside endothelial cells (44,45,47). To assess whether whole brain GLUT-l protein represented that predominantly in microvessels, whole brain and microvessel-enriched protein were immunoblotted with GLUT-l antibody (Fig. 3). GLUT-l protein in whole brain migrated as a broad band of 43 kDa, while GLUT-l protein in microvessels was larger (50 kDa). These data suggested that most of the immunoreactive GLUT-l protein in Western blots of whole brain protein was not in microvessels. The current findings, that most of the GLUT-l protein in whole brain is not microvessel GLUT-l, conflict with the observations of the abundance of GLUT-l mRNA in microvessels published by Boado and Pardridge (49) and Pardridge et al. (51). With regard to protein analysis, the two studies differ in methodology in that we measured whole brain GLUT-l protein without fractionation, and Pardridge et al. (51) measured GLUT-l protein in membranes purified from microvessels and in membranes from microvessel-depleted brain cells. Interestingly, both studies showed a larger size for GLUT-l protein in microvessels (52 kDa, Pardridge et al. (51), and 50 kDa, this study) compared to the size in brain cells (47 kDa (51)) or in whole brain (43 kDa, this study). In this respect the two studies are similar, as both suggest that GLUT-l protein differs between the two tissues. The discrepancy remains, however, that GLUT-l mRNA abundance appeared greater in microvessels (49,51), and GLUT-l pro-

250

KORANYI

tein appeared to be more abundant in extra-microvessel tissue (Fig. 3). One possible explanation for this discrepancy is that GLUT-l gene expression is greater in microvessels, but that the stability of the protein may be greater in neuronal and glial cells. That post-translational changes in GLUT-l protein can occur in rat brain is evidenced by the observation that microvessel GLUT-l mRNA was increased, and GLUT-l protein decreased, in diabetic vs control rats (53). Thus while the changes noted in whole brain GLUT1 in the current experiments may partially reflect changes in microvessels, this does not account for the majority of GLUT-l protein. Changes in microvessel GLUT-l mRNA and protein were not assessed, because microvessel isolation is a fairly lengthy multistep procedure, which may not be readily quantitated. In contrast, whole brain measurements were established on rapidly frozen tissue. The precise cellular location of the changes in brain GLUT1 gene expression noted under conditions of chronic hypoglycemia thus remain to be defined. In contrast to the results of insulin-induced hypoglycemia, the differences in changes in GLUT-l expression in fasting and in streptozotocin-induced diabetic rats are of questionable significance. While significant changes in GLUT-l mRNA were noted under all experimental conditions relative to control levels (Table 2), GLUT-l protein did not significantly change in fasted or streptozotocin diabetic rats (Table 3). Fasting and insulin-induced hypoglycemia differ in several respects that might account for concomitant increases in mRNA, but increased GLUT-l protein only in insulin-injected hypoglycemic rats. First in insulin-injected vs fasted rats the hypoglycemia was more severe (61 mg/dl vs 100 mg/dl), and the increase in GLUT-l mRNA greater (43% vs 31%). Second, while insulin-induced hypoglycemia is associated with suppression of lipolysis, fasting is associated with accelerated lipolysis, and the brain adapts to enhanced utilization of keto acids, thereby being less dependent on glucose for metabolism (54). Finally, the levels of mRNA and protein might reflect differences in sensitivities of the two assays. RNA was measured by direct dot blot analysis, while GLUT-l protein was measured by immunoblotting that can be associated with variable recovery of protein (55). The coefficients of variation of GLUT1 protein determinations (Table 3) were such that only differences greater than 27% would be detected between groups at the P < 0.05 level. In contrast, the coefficients of variation for GLUT-l mRNA levels were less (Table 2), and differences between the means of two groups as little as 6% would be detected at the P < 0.05 level. For streptozotocin diabetic rats the decrease in GLUT1 mRNA was small (-35%, P < 0.05 vs controls), and the change in GLUT-l protein even smaller (-16%, NS vs controls). Both alloxan and streptozotocin diabetic rats have been reported to have a 33-45% reduction in brain

ET

AL.

glucose transport (11-13). Both cytochalasin B binding (15) and immunoreactive GLUT-l protein (51) have been shown to be significantly reduced in isolated brain microvessels of streptozotocin diabetic rats. Failure to note a decrease in GLUT-l protein in the current study may reflect methodological differences (i.e., microvessel membranes vs whole brain). The current results thus indicate that the previously noted decrease in brain glucose transport in chronically diabetic rats is not accompanied by changes in whole brain GLUT-l protein. The difference in apparent molecular weight of whole brain (43 kDa) and microvessel-enriched (50 kDa) GLUT1 protein on SDS-gel electrophoresis has not been previously reported. As noted above, Pardridge et al. (51) showed mobilities for GLUT-l protein from microvesselenriched and microvessel-depleted tissues similar to those in the current study. Some previous studies have noted two immunoreactive forms of brain GLUT-l protein (17, 35, 36, 43), with reduction to a single size following endoglycosidase treatment (17, 36). The decrease in GLUT-l mRNA in brain tissue seen with streptozotocin-induced diabetes and the increase of GLUT-l mRNA in fasted and insulin-injected hypoglycemic animals together suggest a regulatory effect of plasma glucose concentrations on brain glucose transporter GLUT-l gene expression. Rats made severely hyperglycemic by streptozotocin treatment had a 35% reduction in GLUT-l mRNA concentration relative to controls (Table 2, Fig. 1). Suppression of GLUT-l mRNA in diabetic rats by hyperglycemia rather than by low plasma insulin was suggested by the observation that plasma insulin levels were also low in fasted rats, and GLUT-l mRNA levels increased (Table 2, Fig. 1). For all four groups of rats, a negative correlation (r = -0.61, P < 0.05) between plasma glucose concentrations and GLUT-l mRNA levels was observed (Fig. 4). While many factors may regulate GLUT-l expression in brain, these data suggest that chronic adaptation to changes in plasma glucose concentrations occurs at least partially through an adaptive response in GLUT-l expression. Chronic adaptation to changes in plasma glucose have also been shown to occur at the post-transcriptional level (52, 53), however, as noted above, and thus many levels of control may be playing a role in brain glucose transporter function. ACKNOWLEDGMENTS This work was supported by NIH (M.M.1, andDK07140 (F.T.F. and Grant from the Juvenile Diabetes Career Development Award from L.K. was a recipient of a Research Diabetes Foundation. The authors preparing the manuscript, Joanne Dr. Irene Karl for fatty acid and

grants DK16746 (M.A.P.), DK36495 R.B.). D.J. is a recipient of a Research Foundation. M.M. is a recipient of a the Juvenile Diabetes Foundation. Fellowship 367269 from the Juvenile thank Jeannie Wokurka for help in Ochoa for technical assistance, and keto acid determinations.

GLUCOSE

TRANSPORTER

GENE

EXPRESSION

2. 3. 4. 5.

6.

Owen, 0. E., A. P. Morgan, H. G. Kemp, J. M. Sullivan, M. G. Herrera, and G. F. Cahill (1967). Brain metabolism during fasting. J. Clin. Invest. 46: 1589-1959. Lund-Anderson, H. (1979). Transport of glucose from blood to brain. Physiol. Reu. 59: 305-352. Gjedde, A. (1983). Modulation of substrate transport to the brain. Acta Neural hand 67: 3-25. Pardridge, W. M. (1983). Brain metabolism: A perspective from the blood-brain barrier. Physiol. Rev. 63: 1481-1535. Buschiazzo, P. M., E. B. Terrell, and D. M. Regen (1970). Sugar transport across the blood-brain barrier. Am. J. Physiol. 219: 1505-1513. Dick, A. P. K., S. I. Harik, A. Klip, and D. M. Walker (1984). Identification and characterization of the glucose transporter of the blood-brain barrier by cytocholasin B binding and immunological reactivity. Proc Nat1 Acad Sci USA 81: 7233-7237.

21.

22.

23.

24.

25.

I.

Dick, A. P. K., and S. I. Harik transporter in the mammalian 1411.

8.

Goldstein, G. W., J. Caejtey, and I. Diamond (1977). Carrier mediated glucose transport in capillaries isolated from rat brain. J. Neurochem. 28: 725-728. Matthaei, S., J. M. Olefsky, W. T. Garvey, and R. Horuk (1987). Biochemical characterization and subcellular distribution of the glucose transporter from rat brain microvessels. Biochim. Biophys. Acta 905: 417-725. Auer, R. N., Y. Olsson, and B. K. Siesjo (1984). Hypoglycemic brain injury in the rat-Correlation of density of brain damage with the EEG isoelectric time: A quantitative study. Diabetes 33: 1090-1098. McCall, A. L., L. B. Fixman, N. Fleming, K. Tornheim, W. Chick, and N. B. Ruderman (1986). Chronic hypoglycemia increases brain glucose transport. Am. J. Physiol. 251: E442-E447. Gjedde, A., and C. Crone (1981). Blood-brain glucose transfer: Repression in chronic hyperglycemia. Science 214: 456-457. McCall, A. L., W. R. Millington, and R. J. Wurtman (1982). Metabolic fuel and amino acid transport into the brain in experimental diabetes mellitus. Proc. Natl. Acad. Sci. USA 79: 5406-5410. McCall, A. L., J. B. Gould, and N. B. Ruderman (1984). Diabetesinduced alterations of glucose metabolism in rat cerebral microvessels. Am. J. Physiol. 247: E462-E467. Matthaei, S., R. Horuk, and J. M. Olefsky (1986). Blood-brain glucose transfer in diabetes mellitus. Diabetes 35: 1181-1184. Mueckler, M., C. Caruso, S. A. Baldwin, M. Panico, I. Blench, H. R. Morris, W. J. Allard, G. E. Lienhard, and H. F. Lodish (1985). Sequence and structure of human glucose transporter. Science 229: 941-945.

26.

Birnbaum, M. J., H. C. Haspel, and 0. M. Rosen (1986). Cloning and characterization of a cDNA encoding the rat brain glucose transporter protein. Proc. Natl. Acad. Sci. USA 83: 5784-5788. Fukumoto, H., S. Seino, H. Imura, Y. Seino, R. L. Eddy, Y. Fukushima, M. G. Byers, T. B. Shows, and G. I. Bell (1988). Sequence, tissue distribution and chromosomal localization of mRNA encoding a human glucose transporter-like protein. Proc. Natl. Acad. Sci. USA 85: 5434-5438. Thorens, B., H. K. Sarkar, H. R. Kaback, and H. F. Lodish (1988). Cloning and functional expression in bacteria of a novel glucose transporter present in liver, kidney and beta pancreatic islet cells. Cell 55: 281-290. Permutt, M. A., L. Koranyi, K. Keller, P. E. Lacy, D. W. Scharp, and M. Mueckler (1989). Cloning and functional expression of a

35.

9.

10.

11.

12. 13.

14.

15. 16.

17.

18.

19.

20.

(1986). brain.

RAT

Distribution of the glucose J. Neurochem. 46: 1406-

251

BRAIN

human pancreatic islet glucose Sci. USA 86: 8688-8692.

REFERENCES 1.

IN

transporter

cDNA.

Proc.

Natl.

Acad.

James, D. E., M. Strube, and M. Mueckler (1989). Molecular cloning and characterization of an insulin regulatable glucose transporter. Nature 338: 83-87. Birnbaum, M. J. (1989). Identification of a novel gene encoding an insulin-responsive glucose transporter protein. Cell 57: 305315. Charron, M. J., F. C. Brosius, S. L. Alper, and H. F. Lodish (1989). A glucose transport protein expressed predominately in insulinresponsive tissues. Proc. Natl. Acad. Sci. USA 86: 2535-2539. Fukumoto, H., T. Kayano, J. B. Buse, Y. Edwards, P. F. Pilch, G. I. Bell, and S. Seino (1989). Cloning and characterization of the major insulin-responsive glucose transporter expressed in human skeletal muscle and other insulin-responsive tissues. J. Biol. Chem. 264: 716-779. Kayano, T., H. Fukumoto, R. L. Eddy, Y.-S. Fan, M. G. Byers, T. B. Shows, and G. I. Bell (1988). Evidence for a family of human glucose transporter-like proteins: Sequence and gene localization of a protein expressed in fetal skeletal muscle and other tissues. J. Biol. Chem. 263: 15,245-15,248. Berger, J., C. Biswas, P. P. Vicario, H. V. Strout, R. Saperstein, and P. F. Pilch (1989). Decreased expression of the insulin-responsive glucose transporter in diabetes and fasting. Nature 340: 70-72.

27.

Sivitz, W. I., S. L. DeSautel, T. Kayano, G. I. Bell, and J. E. Pessin (1989). Regulation of glucose transporter messenger RNA in insulin-deficient states. Nature 340: 72-74.

28.

Garvey, W. T., T. P. Huecksteadt, and M. J. Birnbaum Pretranslational suppression of an insulin-responsive transporter in rats with diabetes mellitus. Science 245:

29.

Chirgwin, J. M., A. E. Przybyla, R. J. MacDonald, and W. J. Rutter (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18: 1594-1599.

30.

Miles, J., R. Glesscock, J. Aikens, J. Gerich, and M. Haymond (1983). Microfluoro-metric assay for the determination of free fatty acids in plasma. J. Lipid Res. 24: 96-99. Williamson, D. H., J. Mellanby, and H. A. Krebs (1962). Enzymic determination of D(-)-@-hydroxybutyric acid and acetoacetic acid in blood. Biochem. J. 82: 90-96. Goldstein, G. W., J. S. Wolinsky, J. Czejtek, et al. (1975). Isolation of metabolically active capillary from rat brain. J. Neurochem. 25: 715-717. Peterson, G. L. (1977). A simplification on the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83: 346-356.

31.

32.

33.

34.

36.

37.

38.

(1989). glucose 60-63.

Froehner, S. C., A. Davies, S. A. Baldwin, and G. E. Lienhard (1988). The blood-nerve barrier is rich in glucose transporter. J. Neurocytol. 17: 173-178. Walker, P. S., J. A. Donovan, B. G. Van Ness, R. E. Fellows, and J. E. Pessin (1988). Glucose-dependent regulation of glucose transport activity, protein, and mRNA in primary cultures of rat brain glial cells. J. Biol. Chem. 263: 15,594-15,601. Sivitz, W., S. DeSautel, P. S. Walker, and J. E. Pessin (1989). Regulation of the glucose transporter in developing rat brain. Endocrinology 124: 1875-1880. Matthaei, S., J. M. Olefsky, and R. Horuk (1987). Biochemical characterization and subcellular distribution of the glucose transporter from rat brain microvessels. Biochim. Biophys. Acta 905: 417-425. Kalaria, R. N., S. A. Gravina, J. W. Schmidley, G. Perry, and S. I. Harik (1988). The glucose transporter of the human brain and blood-brain barrier. Ann. Neurol. 24: 757-764.

252

KORANYI

39.

Kasanicki, M. A., M. T. Cairns, A. Davies, R. M. Gardiner, and S. A. Baldwin (1987). Identification and characterization of the glucose-transport protein of the bovine blood/brain barrier. Biochem. J. 247: 101-108.

40.

Chundu, the brain

41.

Ika, Y., T. Asano, Y. Shibasaki, M. Kasuga, Y. Kanazawa, and F. Takaku (1988). Studies with antipeptide antibody suggest the presence of at least two types of glucose transporter in rat brain and adipocyte. J. Biol. Chem. 263: 13,432-13,439.

42.

43.

44.

K., and I. Simpson and heart glucose

(1990). Developmental transporter. Pediatr.

Bhattacharyya, M. V., and J. L. Brodsky of the glucose transporter from rat brain Biophys. Res. Commun. 166: 685-691.

regulation Res. 27: 41.

ET

AL.

47.

Gerhart, D. Z., R. J. LeVasseur, M. A. Broderuis, and L. R. Drewes (1989). Glucose transporter localization in brain using light and electron immunocytochemistry. J. Neurosci. Res. 22: 464-472. Flier, J. S., M. Mueckler, A. L. McCall, and H. F. Lodish (1987). Distribution of glucose transporter messenger RNA transcripts in tissues of rat and man. J. Clin. Znoest. 79: 657-661. Boado, R. J., and W. M. Pardridge (1990). The brain-type glucose transporter mRNA is specifically expressed at the blood-brain barrier. Biochem. Biophys. Res. Commun. 166: 174-179. Bagley, P. R., S. P. Tucker, C. Nolan, et al. (1989). Anatomical mapping of glucose transporter protein and pyruvate hydrogenase in rat brain: An immunogold study. Brain Res. 499: 214-224. Pardridge, W. M., Boado, R. J., and Farrell, C. R. (1990). Braintype glucose transporter (GLUT-l) is selectively localized to the blood-brain barrier. J. Biol. Chem. 265: 18,035-18,040. Choi, T. B., Boado, R. J., and Pardridge, W. M. (1989). Bloodbrain barrier glucose transporter mRNA is increased in experimental diabetes mellitus. Biochem. Biophys. Res. Commun. 164: 375-380. Pa&ridge, W. M., Triguero, D., and Farrell, C. R. (1990). Downregulation of blood-brain barrier glucose transporter in experimental diabetes. Diabetes 39: 1040-1044. Gjedde, A., and C. Crone (1975). Induction processes in bloodbrain barrier transfer of ketone bodies during starvation. Am. J. Physiol. 229: 1165-1169. Kakita, K., K. O’Connell, and M. A. Permutt (1984). Immunoelectrophoretic detection of insulin. In Methods in Diabetes Research (J. Larner and S. L. Pohl, Eds.), pp. 419-429. Wiley, New York.

48.

of

(1988). Characterization synaptosomes. Biochem.

49.

50.

51.

Wang, C., and W. A. Brennan, Jr. (1988). Rat skeletal muscle, liver and brain have different fetal and adult forms of the glucose transporter. Biochim. Biophys. Acta 946: 11-18.

52.

Kasanicki, M. A., K. R. Jessen, S. A. Baldwin, J. M. Boyle, A. Davies, and R. M. Gardiner (1989). Immunocytochemical localization of the glucose-transport protein in mammalian brain capillaries. Histochem. J. 21: 47-51.

53.

45.

Davies, A., T. L. Ciardelli, G. E. Lienhard, J. M. Boyle, A. D. Whetton, and S. A. Baldwin (1990). Site-specific antibodies as probes of the topology and function of the human erythrocyte glucose transporter. Biochem. J. 266: 799-808.

46.

Vilaro, S., M. Palacin, P. F. Pilch, X. Testar, and A. Zorzano (1989). Expression of an insulin-regulatable glucose carrier in muscle and fat endothelial cells. Nature 342: 798803.

54.

55.