Translocable Glucose Transporters in the Brain

Where Are We in 2006?

  1. Thierry Alquier12,
  2. Corinne Leloup1,
  3. Anne Lorsignol1 and
  4. Luc Pénicaud1
  1. 1Laboratoire de Neurobiologie, Plasticité Tissulaire et Métabolisme Energétique, Rangueil, France
  2. 2Montreal Diabetes Research Center, Centre de Recherche du Centre Hospitalier de l’Université de Montréal, Montréal, Québec, Canada
  1. Address correspondence and reprint requests to Luc Pénicaud, UMR5018 CNRS-UPS, IFR 31, BP 84225, 31432 Toulouse Cedex 4, France. E-mail: penicaud{at}toulouse.inserm.fr

Abstract

The discovery of the brain expression of the translocable glucose transporters, GLUT4 then GLUT8, led to the question of their putative role in the central nervous system, particularly in relation to insulin effect. The anatomical, cellular, and subcellular localization of these transporters has been described in detail. It has been shown that, as in peripheral tissues, these transporters are localized both in an intracellular pool and at the plasma membrane. This is coherent with a translocation phenomenon, but the data reporting the effect of insulin on that property of GLUT4 and GLUT8 remains very controversial. Glucose and insulin have been shown to modulate GLUT4 expression. Based on their anatomical features, different hypotheses have been proposed. Because of the colocalization with the insulin receptors, insulin might regulate glucose uptake in specific brain areas. A role in nutrient sensing has also been proposed, since both insulin and GLUT4 are expressed in gluco-excited neurons of the hypothalamus. Some studies suggest a role for GLUT8 in the endoplasmic reticulum stress. Whereas much has been learned about their cellular features in the central nervous system, many questions remain unanswered concerning their physiological functions. Gene knockout strategy specifically in the brain and even more in specific nuclei or type of cells should provide new clues to the physiological role of both transporters in the central nervous system.

Insulin effects in the central nervous system (CNS) are pleiotropic, including neurotrophic action, cognitive functions, and control of energy homeostasis. Thus insulin, considered as an adiposity signal, inhibits food intake, stimulates energy expenditure, and controls peripheral glucose homeostasis via its hypothalamic action (13). The mechanisms of action are partly known and the signaling pathways activated by insulin in the hypothalamus are similar to the one described in the peripheral insulin-sensitive tissues (muscle and adipose tissues), i.e., insulin receptor substrate 1 and 2, phosphatidylinositol 3-kinase, and protein kinase B (4,5). Central insulin control of energy homeostasis involves the regulation of orexigenic (neuropeptide Y) and anorexigenic neuropeptide (proopiomelanocortin) expression (6), modulation of the neuronal firing (7), and modulation of neurotransmission (8,9) in several hypothalamic nuclei, i.e., arcuate nucleus (ARC), ventromedial hypothalamus, and paraventricular nucleus. More recently, several studies have suggested that brain insulin may modulate the brain glucose-sensing mechanisms (7,1012). Indeed, insulin inhibits the activity of hypothalamic glucosensing neurons (7), brain insulin receptor is required for the counterregulatory responses to hypoglycemia (12), and data from our group showed that insulin increases the brain response to glucose (10). However, in these studies, whether insulin action involves modulation of glucose uptake in hypothalamic glucosensing neurons is yet to be demonstrated. In contrast to the insulin-responsive peripheral tissues (muscle and adipose tissues), brain glucose utilization and metabolism are considered predominantly to be insulin independent. However, using the hyperinsulinemic clamp, some studies showed that, depending on the glycemia achieved during the clamp or the length of the clamp, glucose utilization was either increased (13,14) or decreased (1517) in specific brain areas. Thus, the effect of insulin on brain cellular uptake and utilization has therefore remained controversial and elusive.

In muscle and adipose tissues, insulin stimulates glucose transport through the translocation of the insulin-sensitive glucose transporter GLUT4. The large stimulatory effect of insulin in these tissues results from the unique targeting of GLUT4 in the basal states in intracellular vesicles. In response to insulin, these vesicles translocate to the plasma membrane, resulting in an increased glucose transport. Gene knockout strategy demonstrated that reduction in GLUT4 levels in muscle or fat causes insulin resistance and diabetes (18). In the past 10 years, we and others have demonstrated the expression of the insulin-sensitive glucose transporter GLUT4 in several areas of the CNS (1926). More recently, the expression of a translocable glucose transporter, GLUT8, has been reported in the testis, the brain, and to a lesser degree in peripheral insulin-sensitive tissues, i.e., adipose tissue and muscle (2729). GLUT8 protein sequence shows 23% homology with GLUT4 sequence and a Km of 2.4 mmol/l for 2-deoxyglucose comparable to GLUT4 Km (27,29). Although the subcellular localization and translocation of GLUT8 has not been elucidated in muscle or adipose tissues, GLUT4 and GLUT8 harbor a similar dileucine internalization motif in the cytoplasmic tail responsible for their retention into the cytoplasm (27,30). Indeed, mutation of this motif into dialanine results in the redistribution of the GLUT8 transporter to the plasma membrane, strongly suggesting that GLUT8 could also be translocated (27). In agreement with this, an insulin-induced translocation of GLUT8 has been reported in mouse blastocysts (29), similarly to the GLUT4 translocation described previously in adipocytes or muscle cells.

Although basal glucose transport in the brain is ensured by the glial-endothelial GLUT1 transporter and the neuronal GLUT3 transporter, the presence of GLUT4 and GLUT8 in the brain is of particular interest. Considering the controversial effect of insulin on brain glucose uptake, the expression of both translocable glucose transporters in different brain areas raises several questions about their physiological relevance and roles in brain glucose metabolism and hypothalamic glucose-sensing mechanisms. The main goal of this review is to establish an update of the recent studies investigating GLUT4 and GLUT8 in the brain and to propose some clues and a hypothesis (based on their respective features) of their potential functional roles in the CNS.

CELLULAR FEATURES OF GLUT4 AND GLUT8 IN THE BRAIN

Anatomical, cellular, and subcellular localizations of GLUT4 and GLUT8 in the adult brain.

GLUT4 and GLUT8 are expressed in several brain areas including the cortex, the amygdala, the hippocampus, the hypothalamus, and the cerebellum (23,31). The highest levels of GLUT8 expression were found in the hippocampus, amygdala, hypothalamus, and pituitary gland. Anatomical localization demonstrates a more spread expression of GLUT4 in others areas such as the thalamus, the brainstem, and olfactory bulbs, with the highest levels of expression in areas associated with the control of motor activity such as the cortex and the cerebellum (23,25).

Some GLUT4 immunoreactivity has been observed in glial cells of the hippocampus and the cerebellum (23,25) and in the ependymal cells lining the ventricles (19) as well as endothelial cells of microvessels (22,32). The presence of GLUT8 in these cells has not been reported. However, the expression of GLUT8 and GLUT4 in the brain is essentially neuronal. Several colocalization studies provided new data on the chemical nature of neurons expressing GLUT4 or GLUT8. In the basal forebrain, GLUT4 staining was found selectively in cholinergic (using ChAT marker) and GABAergic (using parvalbumin marker) neurons (33). In the hippocampus, GLUT8 is present in principal neurons, pyramidal and granule neurons (which are glutamatergic), and nonprincipal GABAergic neurons (34). In the neurons of the hypothalamic pituitary system, GLUT8 was colocalized with vasopressin but not with oxytocin (31).

At the subcellular level, the transporters are mainly present in the neuronal cell bodies and the proximal apical dendrites. Contrary to GLUT8, GLUT4 is also abundantly present in the dendrites, in the dendrite spines, and at the synaptic level (23). In unstimulated conditions, GLUT4 immunoreactivity is localized on the membranes of transport vesicles, Golgi apparatus, and rough endoplasmic reticulum (ER) of all the brain areas analyzed (23) (Fig. 1A). Immunoblot analysis of hippocampal neurons revealed that GLUT8 is expressed in high-density microsomes and low-density microsomes, albeit at lower levels. Using electron microscopy, Piroli et al. (35) confirmed that GLUT8 was mainly present near and adjacent to the rough ER and absent in the Golgi apparatus (Fig. 1A). However, Ibberson et al. (31) demonstrated recently that GLUT8 immunoreactivity was differentially localized depending on the brain region analyzed. Indeed, GLUT8 was restricted to the nerve terminals and synaptic vesicles of the supraoptic nucleus, whereas it appeared associated with the cell surface in the cortex (and dentate gyrus of the hippocampus) (31).

Altogether, these studies demonstrated that 1) despite a specificity of expression in some brain regions, there is an important overlap between the cerebral areas where GLUT4 and GLUT8 are present, and 2) both transporters exhibit an intracellular localization suggesting a translocation phenomenon to the plasma membrane. However, more investigations will be needed to determine whether GLUT4 and GLUT8 are colocalized in the same neurons and to determine more precisely the chemical nature of the neurons expressing the transporters.

Regulation of GLUT4 and GLUT8 expression.

Insulin is one of the main regulators of GLUT4 expression at the mRNA and protein levels in the peripheral insulin-sensitive tissues, muscle, and adipose tissues (36). Indeed, a large body of literature has described changes of GLUT4 expression in insulin-resistant states in humans and rodents. For instance, in rodents, the expression of GLUT4 is decreased in the adipose tissues of several hyperinsulinemic models such as diet-induced obese rats or obese Zucker fa/fa rats, whereas the protein levels are unchanged in muscle (37).

Several studies have shown that the cerebral expression of GLUT4 was also modified in insulin-resistant models and demonstrated that the changes were specific to the area considered. For instance, Vannucci et al. (24) showed that GLUT4 protein levels were increased in the cerebellum of hyperinsulinemic-hyperglycemic db/db mice, whereas the expression was unchanged in the cortex or the olfactory bulbs. These data are in agreement with previous work of Campbell et al. (38), who demonstrated an increased GLUT4 cerebellar content in hyperinsulinemic Milan rats. In support of these findings, our studies showed increased GLUT4 mRNA levels in the same area of hyperinsulinemic fa/fa Zucker rats without changes in the cortex (T.A., L.P., unpublished data; Fig. 2A). Finally, recent work by Komori et al. (39) showed increased levels of GLUT4 protein in the hypothalamus of ob/ob mice. However, at the mRNA level, GLUT4 expression was unchanged in the ARC and lateral hypothalamus of obese Zucker rats compared with lean Zucker rats (T.A., L.P., unpublished data; Fig. 2A). The reason for the discrepancy between the data in ob/ob mice and fa/fa rats is unknown but could be related to the species difference as well as the different metabolic phenotype. However, altogether, these studies demonstrate that chronic insulin-resistant states are associated with increased expression of GLUT4 in the cerebellum.

In our experiments, when insulin levels were acutely increased by a 48-h glucose infusion in rats, GLUT4 mRNA content was unchanged in the cerebellum, whereas GLUT4 expression was specifically decreased in the lateral hypothalamus and ARC (Fig. 2B) (40). When insulin levels were lowered acutely by streptozotocin treatment or exercise in rats, the cerebellar GLUT4 protein levels were diminished, suggesting a clear correlation between circulating insulin levels and GLUT4 protein content in the cerebellum. However, at the mRNA level, hypothalamic (lateral hypothalamus) and cerebellar GLUT4 levels were unchanged in hypoinsulinemic-hypoglycemic fasted rats (unpublished data; Fig. 2C). All the data mentioned above are summarized in Table 1.

Altogether, these data point out that insulin may not be the only factor controlling GLUT4 expression in the brain, especially in the cerebellum, since depending on glycemia, the insulinemic state does not lead to identical results. Also, these data strongly suggest a region-specific control of the expression of GLUT4 in the CNS.

The regulation of the expression of GLUT8 in the brain has been poorly studied so far. The group of Reagan et al. (28) showed that GLUT8 mRNA levels were increased in the hippocampus of hypoinsulinemic streptozotocin-treated rats. This increase was partially normalized by short-term restraint-induced stress in streptozotocin rats, but the GLUT8 protein content was unchanged in the same area (28,41). These data suggest that insulin and glucocorticoids may be involved in the regulation of GLUT8 expression in the hippocampus.

Recent data have shown that GLUT4 and GLUT8 are expressed in the developing mammalian brain and that their expression is modulated throughout the fetal to adult stages (4245). GLUT4 is present in the embryo brain as early as embryonic day 9 (E9) and E14 in rats, and the expression level peaks at E14 and is lower at E19 and remains stable thereafter up to adulthood (42,43). After birth, GLUT4 levels are increased during the suckling phase between postnatal day 14 (P14) and P21 in rats and mice and lower thereafter, similarly to GLUT8 (43,44). Gestational hypoxia or hyperglycemia during the last stages of gestation increased GLUT4 mRNA levels in the brain, suggesting that intervention modulating glucose or oxygen availability during pregnancy affects the expression of the transporter in the fetal brain (42,45).

Whereas these data demonstrate that metabolic and hormonal parameters are involved in the control of GLUT4 expression, further investigations are needed to determine the specific factors involved in the control of GLUT4 and GLUT8 expression in the CNS.

Translocation of GLUT4 and GLUT8.

To date, insulin has been the best characterized stimulus involved in the translocation of GLUT4 in peripheral tissues (adipose tissues and muscles) and recently in the translocation of GLUT8 in the mouse blastocyst (29). However, other stimuli such as membrane depolarization or hyperosmotic shock have been shown to stimulate GLUT4 translocation in muscle or adipose cells (46). It is important to note that the translocation of GLUT4 in peripheral tissues requires the SNARE (soluble N-ethylmaleimide–sensitive factor attachment protein receptor) protein complex, in which proteins similar to those involved in exocytosis of synaptic vesicles in neurons are used (47,48).

GLUT4.

The subcellular localization of GLUT4 in vivo in intracellular vesicles close to the plasma membrane suggests that the transporter could be translocated to the cell surface (23) (Fig. 1A and B). In support of this, a recent report showed, using Western blot and immunocytochemistry, that insulin induces the translocation of GLUT4 to the plasma membrane in human neuroblastoma cells (SH-SY5Y) (49). Also, the insulin-induced translocation of GLUT4 in SH-SY5Y cells enhances glucose uptake, and this augmentation is abolished in the presence of a phosphatidylinositol 3-kinase inhibitor. These data suggest that GLUT4 redistribution in response to insulin is a phosphatidylinositol 3-kinase–dependent mechanism similar to the mechanism described in muscle or adipose tissues. Interestingly, leptin treatment also stimulates the redistribution of the transporter to the membrane (49). To our knowledge, this the first report describing the translocation of GLUT4 in neural cells in vitro. In collaboration with the group of Y. Lemarchand, a DsRed-tagged GLUT4 was used to investigate the localization and translocation of GLUT4 in primary cultures of fetal cortical neurons transfected by the construct. Our preliminary data showed that the localization of DsRed-GLUT4 was mainly somato-dendritic as the one described in vivo (23). The protein was present in the cytoplasm showing a perinuclear localization. However, no change of this localization was observed upon insulin stimulation (4 μmol/l during 20 min) (T.A., Y. Lemarchand, L.P., unpublished data). Cortical neurons in culture are known to express the insulin receptor, suggesting that, in contrast with the report of Benomar et al. (49), insulin does not stimulate the translocation of GLUT4 in vitro. However, our preliminary studies need further investigations to be confirmed, and the effect of stimuli other than insulin needs to be investigated.

Although some GLUT4-enriched brain areas exhibit a high level of the insulin receptor, a direct demonstration of the insulin-induced translocation has not been reported in vivo yet. Several observations suggest that an insulin-induced translocation might occur also in the brain. A recent study showed that the insulin-regulated aminopeptidase shows a brain distribution resembling that of GLUT4 and that the transporter is colocalized with insulin-regulated aminopeptidase in hippocampal neurons (50). In insulin-responsive tissues, insulin-regulated aminopeptidase is a marker of GLUT4 vesicles and is involved in the trafficking of these specialized vesicles in adipocytes (51). Recently, the group of B. Levin demonstrated that 60–75% of the glucose-sensitive neurons coexpressed GLUT4 and the insulin receptor mRNA in dissociated neurons of the ventromedial nucleus, strongly pointing out a role of insulin in GLUT4 translocation (11). Using electron microscopy, Komori et al. (39) reported that in the ARC, the amount of GLUT4 associated with the plasma membrane was increased in hyperinsulinemic ob/ob mice (unstimulated conditions), suggesting a potential increase of GLUT4 translocation. However, the total amount of cellular GLUT4 protein was also increased in the same nucleus, which could explain the increased GLUT4 content in the cell membrane.

GLUT8.

In contrast to GLUT4, different groups have investigated the translocation of GLUT8 in neural cells under the action of stimuli known to induce GLUT4 redistribution to the cell membrane in adipocyte or muscle cells (35,52,53). The only demonstration of the insulin-sensitive redistribution of GLUT8 from intracytoplasmic storage to the cell membrane has been in blastocysts. In mouse blastocysts, insulin exposure induced the translocation of GLUT8 to the plasma membrane, resulting in an increased glucose transport that ensured viability (29,54). However, in primary rat adipocytes, 3T3-L1 adipocytes, and CHO cells, insulin did not stimulate GLUT8 translocation to the cell surface (55,56). The reason for this discrepancy is unknown, but the latter is in agreement with two studies using mouse neural cells in vitro. Recent work from the group of B. Thorens, using a myc-tagged GLUT8 construct transfected into PC12 cells and hippocampal neurons, demonstrated no change of GLUT8myc intracellular localization in response to insulin (53). Brain-derived neurotrophic factor, hyperosmotic shock, potassium- or glutamate-induced depolarization, glucoprivic stress induced by 2-deoxyglucose, or AMP-activated protein kinase activation by 5-amino-imidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) failed to stimulate the translocation of GLUT8myc (53). In agreement with these data, Shin et al. (52) showed an absence of translocation of a green fluorescent protein–tagged GLUT8 in N2A neuroblastoma cells in response to insulin, IGF-I, depolarization, or hypoxia treatment.

However, in vivo studies from Piroli et al. (35,41) showed that glucose administration, inducing increased serum insulin levels, stimulated the translocation of GLUT8 from light-density microsomes to high-density microsomes in hippocampal neurons in vivo. Using electron microscopy, Piroli et al. (35) confirmed that glucose administration induces the redistribution of GLUT8 into the membranes of the rough ER (Figs. 1A and B). In contrast, streptozotocin-treated rats, characterized by hyperglycemia and hypoinsulinemia, have a decreased amount of GLUT8 associated with the high-density microsome fraction. These studies suggest that circulating insulin might stimulate the redistribution of GLUT8 from a cytoplasmic localization into the membranes of the rough ER in the neurons of the hippocampus. Because the translocation of GLUT8 has been only studied in the hippocampal formation in vivo, it is difficult to extend Piroli’s data to other brain areas. Importantly, GLUT8 was present in the nerve terminals and synaptic vesicles of the supraoptic nucleus and vasopressin neurons, suggesting that in these specific neurons, a redistribution of GLUT8 to the plasma membrane could occur upon a stimulus that remains to be identified (31).

Altogether, these findings demonstrate that any of the known signals inducing GLUT4 translocation in the peripheral tissues are able to stimulate the redistribution of GLUT8 into the neuronal cell membrane, in vivo or in vitro. Consistently with the studies in neural cells, it was recently shown in 3T3-L1 adipocytes and CHO cells that GLUT4 and GLUT8 do not colocalize and that GLUT8 does not recycle to the plasma membrane, suggesting that GLUT8 is unlikely to respond to a stimulus that leads to GLUT4 translocation (56). This strongly points out that the role of GLUT8 may not be to increase neuronal uptake of extracellular glucose. These unexpected observations raise several questions concerning the physiological roles of GLUT4 and GLUT8 in the CNS that we will attempt to address in the following section.

FUNCTIONAL ROLE(S) OF GLUT4 AND GLUT8 IN THE BRAIN

GLUT4.

The physiological role of GLUT4 is still unknown, but different hypotheses based on its neuroanatomical distribution can be proposed. First, GLUT4 is preferentially and highly expressed in brain areas associated with the control of motor activity such as the cortex and the cerebellum. In these areas, translocation of GLUT4 may provide additional glucose to motor neurons under conditions of high energy demand such as a high rate of neuronal firing. This hypothesis is consistent with data showing that physical exercise stimulates local cerebral glucose utilization in the motor system (57). Hence, the electrophysiological activity of the motor neurons may directly stimulate GLUT4 translocation to provide energy substrate required for neuronal firing. Also, in the basal forebrain, GLUT4 is mainly localized in cholinergic and GABAergic neurons (33). Because glucose serves as a substrate for the synthesis of acetylcholine, it could be hypothesized that GLUT4 in these neurons may provide glucose under high firing activity to replenish acetylcholine contents. Also, GLUT4 is highly present in neuronal processes and at the synaptic level, pointing to a role in neurotransmission. Once again, the exact mechanisms of how GLUT4 is translocated in those neurons and may affect neuronal glucose homeostasis are not yet elucidated.

Second, the anatomical localizations of GLUT4 often relate with those of the insulin receptors, suggesting that the hormone could induce GLUT4 translocation and increase neuronal glucose transport. As we mentioned previously, whole brain glucose uptake is considered insulin independent. However, depending on the glycemia achieved during the hyperinsulinemic clamp, glucose uptake was either increased (13,14) or decreased (1517) in different brain areas. The reason for such discrepancies is unknown, but when hyperinsulinemia was accompanied by hyperglycemia, glucose uptake was increased in several brain areas including the hypothalamus (14). Thus, we cannot rule out that insulin may modulate directly or indirectly (through modulation of neuronal firing) glucose uptake in discrete brain areas.

Finally, the localization of GLUT4 in the hypothalamus indicates that the transporter may be involved in glucose-sensing mechanisms. Those mechanisms are critical for the control of whole-body glucose homeostasis. Increases of glucose levels activate the hypothalamic glucose-excited neurons, whereas they inhibit the glucose-inhibited neurons in the ARC or the ventromedial nucleus (58). Several studies, using electrophysiology, have demonstrated that central insulin modulates the activity of hypothalamic glucosensing neurons. Indeed, insulin inhibits the electrical activity of glucose-excited neurons in the ARC and the ventromedial hypothalamus in a glucose-dependent way (7,59). In the presence of 10 or 2.5 mmol/l glucose, insulin has no effect on glucose-excited neuron activity, whereas at 0.1 mmol/l glucose, insulin activates glucose-excited neurons of the ARC (7,59). The effect of insulin may be related to increased glucose transport into glucose-excited neurons. In support of this, the same group showed that 75% of the glucose-excited neurons in the ventromedial nucleus express mRNA coding for both the insulin receptor and GLUT4 (11). Interestingly, Fisher et al. (12) showed that the counterregulatory response to hypoglycemia was impaired in mice with a neuron-specific insulin receptor knockout, suggesting that central insulin action may be critical in the setting of low glucose concentrations. On the other hand, several rodent models characterized by altered glucose-sensing mechanisms have an alteration of GLUT4 expression in the hypothalamus (39,40).

Taken together, these data suggest that central insulin action might be critical for glucose-sensing mechanisms. However, the precise role of GLUT4 in insulin effect and in glucose-sensing mechanisms is yet to be elucidated.

GLUT8.

The results of several studies in neural and non-neural cells have demonstrated specific features of GLUT8, such as a subcellular localization of GLUT8 in a cytoplasmic compartment that differs from the one described for GLUT4 (56), the absence of recycling (56), and translocation to the plasmic membrane upon different stimulations (52,53). Interestingly, studies from Piroli et al. (35,41) showed that the transporter is translocated from the cytoplasm into the rough ER upon glucose stimulation in vivo and that streptozotocin-induced diabetes disrupts GLUT8 trafficking in the ER. This observation is of particular interest in the context of recent studies demonstrating that obesity and diabetes causes ER stress (6062). An essential function of the ER is the synthesis and processing of secretory and membrane proteins. Disruption of ER homeostasis, collectively termed ER stress, leads to the accumulation of unfolded protein and protein aggregates in the ER lumen, which is detrimental for cell survival (63). Based on their findings, Piroli et al. (35,41) proposed that in the hippocampus, GLUT8 transports the glucose molecules removed from glycoproteins during protein processing out of the rough ER lumen into the cytoplasm. Thus, it is tempting to speculate that somehow GLUT8 in the rough ER of hippocampal neurons may control ER glucose homeostasis, which is critical for glycosylation processes. Disruption of GLUT8 trafficking and the subsequent glucose equilibrium could lead to an ER dysfunction and stress affecting neuron viability (64). Studying the involvement of GLUT8 in ER glucose homeostasis is of particular importance, since neuronal damage observed during cerebral ischemia/reperfusion or during neurodegenerative diseases is associated with ER stress (64,65).

GLUT8 is also expressed in the hypothalamus, and a potential role in hypothalamic glucose-sensing cannot be ruled out (31). However, similarly to GLUT4, this hypothesis has not yet been demonstrated.

As mentioned previously, GLUT4 and GLUT8 are expressed in the developing mammalian brain. Dynamic changes of expression have been described throughout the fetal to adult stages (4245). For instance, GLUT4 levels peak at E14 in rat brain embryos, are further decreased at E19, and remain stable until adulthood (42). Brain GLUT8 expression in fetal stages has not been investigated. After birth, GLUT4 levels are increased during the suckling phase between P14 and P21 in rats and mice and further decreased until adulthood, similarly to the GLUT8 pattern of expression (43,44). The brain requires important supplies of fuel to support neuroglial growth and process formation during early postnatal development (66); thus, the increase in GLUT4 and GLUT8 expression during this period might be critical for glucose supply. In support of this, Cheng et al. (67) demonstrated that knocking out the neurotrophic factor IGF-I (whose expression peaks during postnatal brain development) in mice induces a dramatic decrease of glucose uptake in several brain regions of young mice (P10). Decreased glucose uptake was associated with a specific decrease of GLUT4 expression, whereas GLUT1 and GLUT3 expression was unchanged in the brain of IGF-I knockout mice. Taken together, these observations strongly suggest that GLUT4 and probably GLUT8-mediated glucose transport may play critical roles during fetal and postnatal brain development.

While energy needs are highly increased during brain development, aging processes and neurodegenerative diseases (Alzheimer’s, Parkinson’s, and Huntington’s diseases) are characterized by decreased glucose metabolism in the CNS (6870). Insulin and glucose are known to contribute to and stimulate cognitive functions such as learning and memory (71,72). Also, the insulin receptor, GLUT4, and GLUT8 are densely expressed in areas supporting cognition (such as the hippocampus). More interestingly, numerous lines of evidence suggest that an increased prevalence of insulin resistance similar to that seen in type 2 diabetes may contribute to the pathophysiology of Alzheimer’s disease (69). Based on these observations, it is tempting to speculate that insulin action on cognitive functions is directly and/or indirectly dependent of GLUT4 and/or GLUT8 in those areas and that impairment of insulin effects during insulin-resistant states might be responsible for the decreased brain glucose metabolism observed in Alzheimer’s disease.

Finally, an alternative hypothesis that cannot be ruled out is that glucose is not the primary substrate for GLUT4 and GLUT8 in the brain and that it transports other substrates.

SUMMARY

It is now clear that the translocable glucose transporters GLUT4 and GLUT8 are present in neurons of several brain areas. Numerous studies have addressed the specific anatomical and subcellular localizations of both transporters, suggesting that GLUT4 and GLUT8 may be involved in different mechanisms within the neurons. Whereas much has been learned about their cellular features in the central nervous system, many questions remain unanswered concerning their physiological functions. Based on their anatomical features, are the physiological roles of different brain areas specific? Do transporters participate in brain glucose uptake during basal states and/or during specific conditions such as brain development or high neuron firing? Are transporters involved in hypothalamic glucose sensing and control of energy balance? These and many other questions still need to be addressed. Gene knockout of GLUT4 and/or GLUT8 specifically in the brain and even more in specific nuclei or types of cells should provide new clues on the physiological role of both transporters in the CNS.

FIG. 1.

Basal and stimulated subcellular localizations of GLUT4 and GLUT8 in neurons. Although GLUT4 and GLUT8 are present in similar brain areas, no colocalization of the transporters in the same neuron has been reported yet. A: In the basal state, GLUT4 and GLUT8 are mainly present in the neuronal cell bodies. Whereas GLUT8 is mainly localized in the neuronal soma and proximal dendrites, GLUT4 is also abundantly present in vesicles in the dendrites, in the dendrite spines, and at the synaptic level. At the subcellular level, GLUT4 is associated with the membranes of the rough ER, Golgi apparatus, and transport vesicles in the vicinity of the plasma membrane. GLUT8, however, shows a more specific localization near or adjacent to the rough ER. B: In stimulated states, such as a glucose challenge, GLUT8 is translocated to the membranes of the rough ER. It has been proposed that this redistribution may ensure the transport of glucose from the ER lumen into the cytosol during protein glycosylation. The translocation of GLUT4 to the cell membrane has been recently demonstrated in vitro upon insulin stimulation, but the redistribution of the transporter in neurons in vivo has not yet been reported. The potential translocation of GLUT4 to the plasma membrane at the dendrite level is not represented here. 1, Rough ER; 2, nucleus; 3, Golgi apparatus; 4, transport vesicles; 5, dendrites.

FIG. 2.

GLUT4 mRNA quantification by quantitative competitive RT-PCR in different rat brain regions. A: fa/fa obese Zucker rats and lean control fa/fa. B: 48-h glucose-infused Wistar rats and saline infused control rats. C: 4-day fasted Wistar rats and fed control rats. CxP, cortex parietal; LH, lateral hypothalamus. Data are means ± SE. **P < 0.01 vs. control (unpaired t test).

TABLE 1

Regulation of GLUT4 mRNA and protein levels in different brain areas in animal models characterized by chronic or acute changes in circulating insulin levels

Footnotes

  • This article is based on a presentation at a symposium. The symposium and the publication of this article were made possible by an unrestricted educational grant from Servier.

    The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

    • Accepted May 2, 2006.
    • Received March 23, 2006.

REFERENCES

| Table of Contents