Glucose, the major fuel in the brain, is transported across the cell membranes by facilitated diffusion mediated by glucose transporter proteins. Essentially two types of glucose transporters are localized in the membranes of brain endothelial cells, astrocytes, and neurons. Their densities are well adjusted to changes in local energy demand.
The importance of membrane transport proteins for the adequate supply of the cells with nutrients becomes more and more evident. In this review, we will focus on the main glucose transporter proteins in the brain, Glut1 and Glut3. Their expression and regulation during physiological and pathophysiological conditions have recently been extensively analyzed.
Importance of glucose transporters for glucose transport across cell membranes
Glucose is an essential metabolic substrate of all mammalian cells. It is taken up into cells by facilitated diffusion, which is mediated by members of the glucose transporter family. Therefore, glucose transporters can be found in the cell membranes of all mammalian cells. More than any other organ, the brain is entirely dependent on a continuous supply of glucose from the circulation since glucose is almost the sole substrate for energy metabolism. Ketone bodies can be used as an additional substrate when their plasma concentration is increased. The existence of the blood-brain barrier, which comprises tight junctions between the endothelial cells, necessitates the transcellular transport of glucose from the blood to the brain through the luminal and abluminal membranes of the brain endothelial cells (2). This transport is stereospecific for d-glucose. Pharmacological inhibition of glucose transporter function by substances like phloretin in vitro results in a reduction of glucose transfer across brain cell membranes to <5% of control values, indicating that glucose transporters mediate >95% of the glucose transfer to the brain. Besides pharmacological inhibition, a defective function of glucose transporters may also cause a reduced glucose transport to the brain. The defective function of glucose transporter Glut1 has been demonstrated in infants (14). Either of two genetic mutations, hemizygosity or nonsense mutations, results in severe deteriorations of brain metabolism and function. These infants suffer from seizures, delayed development, and acquired microcephaly. If the disease is properly diagnosed, the symptoms can be attenuated by a ketogenic diet that allows, to a certain extent, ketone bodies to replace glucose as a fuel. These ketone body nutrients are transported into the cells by specific transport proteins, the monocarboxylate transporters (MCT).
Glucose transporter isoforms in the different organs
Glucose transporter proteins can be divided into two large classes, the sodium-dependent and the facilitative glucose transporters. The sodium-dependent isoforms mediate the transport of glucose against a concentration gradient. The driving force is the flux of sodium along an electrochemical gradient that is directed opposite to the transport of glucose. Such sodium-dependent glucose transporters can be found in the kidney and other organs. Their existence in the brain has not been convincingly demonstrated. In contrast to the sodium-dependent isoforms, the facilitative glucose transporters can only support the flux of glucose along an existing concentration gradient for glucose. Among these sodium-independent transporters, six functional isoforms have been identified and characterized so far. Their more-or-less organ-specific distribution is outlined in Table 1⇓.
The amino acid sequence of each Glut isoform is nearly identical when different species are compared. For example, Glut1 is >95% identical between different mammalian species (10). In addition to the high conservation of each Glut isoform during evolution, comparing the amino acid sequence of the six glucose transporter isoforms seems to indicate that all glucose transporters originated from a joint predecessor. An identity of 39-65% can be found within a given species. Such a joint predecessor may even encompass additional transport proteins, such as MCTs; glucose transporters and MCTs share a common molecular arrangement in the cell membrane. The transport proteins contain 12-membrane-spanning domains, and their COOH terminal and NH2 terminal ends are located on the cytoplasmic side of the cell membrane.
Glucose transporter isoforms in the brain
Which types of glucose transporters exist in the brain? Figure 1⇓ shows the different types. The main glucose transporters in the brain are Glut1 and Glut3. Of these, Glut1 exists in two isoforms of different molecular weights. The 55-kDa isoform is located at the luminal and the abluminal membranes of the brain endothelial cells (7), whereas the 45-kDa isoform is expressed in the perivascular endfeet of the surrounding astrocytes (11). The higher molecular weight of the 55-kDa isoform is due to the existence of a glycosylation site at the first extracellular loop of the membrane-spanning molecule of Glut1. Whether this difference in glycosylation between the endothelial and astrocytic isoforms is of functional significance is not known. Another feature of Glut1 is its uneven distribution between the cytoplasm and the luminal and the abluminal membranes of the endothelial cells, as demonstrated by electron microscopy. The fractional distribution of Glut1 to the cytoplasm and to the cell membranes of the endothelial cells is shown schematically in Fig. 2⇓ (1, 7). Although the fractional distribution varies between different species, the lowest density of Glut1 molecules is preferentially found in the cytoplasm and the highest density in the abluminal membrane. Whether this fractional distribution of Glut1 in the brain is fixed over longer time periods or whether these glucose transporters can be redistributed between the cytoplasm and the abluminal or luminal membranes is not known. An acute redistribution of glucose transporters within the cell has been found for Glut4 in skeletal muscle and fat tissue under the influence of insulin. Besides the two Glut1 isoforms, Glut3 is of major importance for the transport of glucose within the brain. The low Michaelis-Menten constant value of Glut3 facilitates a continuous supply of glucose to the neurons even at low interstitial glucose concentrations. Traces of other isoforms of glucose transporters could be detected in specific brain cells. Of these isoforms, Glut5 fructose transporters have been detected in brain microglia and Glut4 glucose transporters in specific neuronal cell populations. The regional distribution pattern of Glut4 in the brain appears to correspond to that of insulin receptors and of insulin-containing cells, indicating the possibility of Glut4 regulation by insulin in the brain as well. Recently, a novel glucose transporter has been described and has been named GlutX1 (8). Glucose transporters of the type GlutX1 have only one-third homology with Glut1–5. GlutX1 mRNA has been detected in several brain structures and some other organs, like testis, adrenal gland, liver, and lung.
Heterogeneous distribution of glucose transporters in the brain
The brain is a heterogeneous organ with respect to cell density and functional activation. Does this heterogeneity also apply for the distribution of glucose transporters? A method was developed that allows us to detect the distribution of Glut1 and Glut3 densities in different brain regions. Using antibodies and autoradiography, local heterogeneities could be found for Glut1 and Glut3. The concentrations varied by a factor of two for both Glut1 and Glut3. Thus the autoradiographic techniques have shown a heterogeneous distribution of Glut1 and Glut3 in the brain.
Correlation between glucose transporter densities and capillary density
What could be the cause of the heterogeneous distribution of glucose transporters in the brain? Two possibilities should be considered. One possibility is that the density of glucose transporters reflects the distribution of capillaries. An uneven distribution of capillary densities was found in the brain during previous studies by our group. Capillary densities of different brain structures are directly related to their local blood flow. Brain structures in which resting blood flow is high have a high capillary density and vice versa. Thus the morphological basis of the different flow values in different brain structures is the amount of perfused capillaries per volume of brain tissue. This direct relationship between blood flow and capillary morphology is based on a continuous perfusion of capillaries under normal conditions. In contrast to previous assumptions of nonperfused capillaries and capillary recruitment, it is now generally accepted that all capillaries are perfused in the brain under normal, nonischemic conditions, although at different velocities. Therefore, glucose is taken up from all capillaries by the glucose transporters. Glucose transporters are associated with blood vessels since Glut1 is located in the capillary wall (55-kDa isoform) and close to capillaries (45-kDa isoform). Comparison of Glut1 density and capillary density shows a close correlation between local Glut1 density and local capillary density when plotted for the different brain structures. Thus capillary density is a major determinant of Glut1 density in the brain.
Correlation between local glucose transporter densities and local cerebral glucose utilization
As a second possibility, the heterogeneous distribution of glucose transporters in the brain could be related to the local glucose utilization. In this case, the different rates of glucose utilization in different brain structures would be reflected in differing densities of Glut1 and Glut3. This possibility was tested by using immunoautoradiography to detect Glut1 or Glut3 and the 2-[14C]deoxyglucose method to determine local cerebral glucose utilization. For different brain structures, the local values of glucose utilization were related to the local distribution of the glucose transporter proteins Glut1 and Glut3.
A positive correlation could be shown between the local densities of Glut1 and local cerebral glucose utilization as well as the local densities of Glut3 and local cerebral glucose utilization, as shown in Fig. 3⇓ (5). The most likely explanation for these results is that the local cerebral glucose utilization determines the local densities of both glucose transporters, Glut1 and Glut3. In addition, the local cerebral glucose utilization also determines the local capillary densities in the different brain structures. The stimulus for the varying expression of glucose transporters is not known. Besides the local glucose utilization, the local oxygen consumption in the different brain structures may be relevant. Because of the normally fixed stoichiometric relationship between oxygen consumption and glucose utilization (6 mol oxygen/1 mol glucose), it is not possible to differentiate between the impact that each of these two parameters has as a local stimulus of glucose transporter expression.
Regulation of glucose transporter expression in the brain: chronic activation
The correlations between the local densities of Glut1 or Glut3 and local cerebral glucose utilization had been determined in rats under normal experimental conditions. The correlations indicate that the densities of glucose transporters apparently have been adjusted to the local metabolic demand of the different brain structures. The time course of this long term adjustment was investigated by using several types of experiments in which either the local metabolic rate for glucose or the plasma glucose concentration was changed over several days. To stimulate various brain regions chronically, rats obtained a continuous infusion of a nicotine solution for 1 wk. Under these conditions, local glucose utilization is chronically increased in specific brain structures. Moreover, glucose transporter densities of Glut1 and Glut3 showed a distinct increase only in brain structures with an elevated local cerebral glucose utilization. Thus the density of glucose transporters can be chronically raised in specific brain regions as a response to an enhanced metabolic rate for glucose. The increase in the local densities of glucose transporters (mean +19%) is paralleled by an increase in local cerebral glucose utilization (mean +15%). The close correlation between Glut1 or Glut3 densities and local cerebral glucose utilization shown for control conditions was maintained during nicotine infusion (6).
Additional local activation studies confirmed these conclusions. In these studies, rats were water deprived for 3 days to stimulate the osmoregulatory system of the brain. Under these conditions, the densities of glucose transporters Glut1 and Glut3 increased in parallel to the increased local cerebral glucose utilization in some of the osmoregulatory structures. In this case, the increase in the transporter densities in these structures (mean Glut1 +26%, mean Glut3 +39%) was lower than the increase in local cerebral glucose utilization (mean +85%). Both types of studies, nicotine infusion and water deprivation (3), as well as studies by Vannucci et al. (17), indicate a moderate long term local upregulation of glucose transporters as a response to increased local cerebral glucose utilization.
Regulation of glucose transporter expression in the brain: chronic inactivation
Whereas these experiments had shown an upregulation of glucose transporters, it remained to be shown whether a chronic decrease in glucose utilization in specific brain structures is followed by a selective decrease of glucose transporter densities. The visual system was chosen as an example of a chronic deprivation of a functional system in the brain. Unilateral visual deprivation was induced by monocular enucleation in rats. After 1 wk, the contralateral structures of the visual system were analyzed for the densities of glucose transporters Glut1 and Glut3 and for local cerebral glucose utilization. Since >90% of the afferent fibers from the retina cross in the chiasma in the rat, the ipsilateral structures of the visual system served as controls. During chronic visual deprivation, Glut1 and Glut3 densities and local cerebral glucose utilization were moderately decreased in some structures of the visual system (mean Glut1 –1%, mean Glut3 –7%, mean local cerebral glucose utilization –14%). Primary visual structures receiving direct retinal projections (dorsal lateral geniculate nucleus, superior colliculus layer I-III) were more affected than secondary structures that receive no direct retinal input (e.g., visual cortex). Thus glucose transporters can be moderately downregulated as a consequence of a lowered metabolic rate for glucose (4).
Regulation of glucose transporter expression in the brain: chronic hypoglycemia
Another approach to the question of whether glucose transporters can be up- and downregulated is to measure the densities of glucose transporters during chronic changes in the plasma glucose concentration. Under these conditions, a more general increase or decrease of glucose transporter densities can be expected, if there is an effect at all. Therefore, chronic hypo- and hyperglycemia were tested. During chronic hypoglycemia, induced by a continuous infusion of insulin into rats for 1 wk, the mean density of Glut1 remained unchanged, whereas the mean density of Glut3 increased slightly, although significantly (+3%). A previous study of Uehara et al. (15), in which the quantitative immunoblot technique was used, had qualitatively but not quantitatively (+76%) the same result. To determine whether the increased density of Glut3 is related to a change in glucose metabolism, cerebral glucose utilization was quantified by the 2-deoxyglucose method. Mean glucose utilization was decreased by 15%. Local analysis of transporter densities (Glut1 and Glut3) and glucose utilization showed a significant correlation between local glucose transporter densities (Glut1 and Glut3) and local cerebral glucose utilization during hypoglycemia, as previously observed during normoglycemia. Therefore, 1 wk of hypoglycemia is a stimulus for the moderate induction of additional Glut3 in the brain. These additional neuronal glucose transporters may support the maintenance of glucose utilization even though glucose utilization remains decreased (5).
Regulation of glucose transporter expression in the brain: chronic hyperglycemia
Chronic hyperglycemia was induced for 3 wk by selective pharmacological destruction of pancreatic β-cells with streptozotocin. As a result, a moderate decrease of Glut1 (–8%) was measured, whereas Glut3 remained unchanged. In accordance with these findings, Pardridge et al. (12) found a downregulation of Glut1 after 1 wk of diabetes in a cerebral microvessel preparation that only contained the vascular Glut1. Thus chronic hyperglycemia is a stimulus for a moderate downregulation of Glut1.
Developmental regulation of glucose transporters
In the brain, glucose transporter densities are regulated during development. Whereas Glut1 already exists at the blood-brain barrier at birth in rats, its density increases rapidly between 10 and 20 days after birth, indicating the maturation of the blood-brain barrier during that period. Thirty days after birth, adult Glut1 levels are reached (18). Accordingly, the maturation of the blood-brain barrier is paralleled by the neuronal maturation, including a developmental increase of the neuronal glucose transporter Glut3. The most prominent increase of Glut3 density occurs between days 14 and 21. At day 30, ~90% of adult Glut3 levels are attained (16).
Mechanisms of glucose transporter regulation
The regulation of glucose transporter expression appears to occur at the transcriptional, posttranscriptional, and posttranslational levels. The genomic sequence of Glut1 consists of ~35,000 base pairs, of which the promotor region contains several elements for transcription factors. For an example of transcriptional control, Glut1 mRNA is increased during hypoxia. On the posttranscriptional level, glucose transporter density can be modified through regulation of the cellular content of transporter protein by alteration of transcript stability (13). Regulated Glut1 mRNA decay plays an important role in control of posttranscriptional gene expression. An essential element in this regulation is the 3′-untranslated region of the message. Alterations of the 3′-untranslated region by specific proteins can modify the transcript stability. For an example of translational regulation, insulin specifically upregulates Glut1 mRNA translation. An additional regulatory mechanism exists for Glut3; an increase in Glut3 can occur by a mechanism not affecting transcription or translation of new Glut3 protein but rather by stabilization of the protein, thus prolonging the half-life of Glut3 protein during an increased glucose demand (9). Due to an increased half-life of Glut3 at an unchanged synthesis rate, total Glut3 is then elevated.
The spatial heterogeneity of energy metabolism in the brain results in spatial heterogeneities of capillary density and glucose transporter density. The long term demand for glucose in each brain structure determines its glucose transporter density. Chronic changes in the local or global energy metabolism of the brain are followed by moderate changes in the glucose transporter densities, resulting in the support of energy metabolism.
Our work was supported by the Deutsche Forschungsgemeinschaft, Bonn, Germany, the Forschungsförderung of the University of Heidelberg, Heidelberg, Germany, and VERUM, Stiftung für Verhalten und Umwelt, Munich, Germany.
- © 2001 Int. Union Physiol. Sci./Am.Physiol. Soc.