Pioglitazone does not increase cerebral glucose utilisation in a murine model of Alzheimer’s disease and decreases it in wild-type mice
- 581 Downloads
Clinical trials are in progress to test thiazolidinediones in neurodegenerative diseases such as Alzheimer’s disease that involve deficiencies in brain glucose metabolism. While thiazolidinediones enhance glucose uptake in non-cerebral tissues, their impact on brain energy metabolism has not been investigated in vivo. We thus determined whether the thiazolidinedione pioglitazone reverses the decrease in cerebral glucose utilisation (CGU) in a model of brain metabolic deficiency related to Alzheimer’s disease. Results are relevant to diabetes because millions of diabetic patients take pioglitazone as an insulin-sensitising drug, and diabetes increases the risk of developing Alzheimer’s disease.
Materials and methods
The regional pattern of CGU was measured with the 2-deoxy [14C] glucose autoradiographic technique in adult awake mice overexpressing transforming growth factor β1 (TGFβ1), and in wild-type littermates. Mice were treated with pioglitazone for 2 months.
Measurement of CGU in 27 brain regions confirmed that TGFβ1 overexpression induced hypometabolism across the brain. Pioglitazone did not reverse the effect of TGFβ1 overexpression and decreased regional CGU in control animals by up to 23%. The extent of the regional CGU decrease induced by pioglitazone, but not that induced by TGFβ1, correlated strongly with basal CGU, suggesting that the higher the local metabolic rate the greater the reduction of CGU effected by pioglitazone.
In contrast to its stimulatory effect in non-cerebral tissues, chronic treatment with pioglitazone decreases CGU in vivo. This evidence does not support the hypothesis that pioglitazone could act as a metabolic enhancer in Alzheimer’s disease, and raises the question of how thiazolidinediones could be beneficial in neurodegenerative diseases.
KeywordsAlzheimer’s disease Cerebral glucose utilisation Murine model Neurodegenerative diseases Peroxisome proliferator-activated receptor gamma Transforming growth factor beta-1
cerebral glucose utilisation
peroxisome-proliferator activated receptor
transforming growth factor β1
Thiazolidinediones such as pioglitazone are commonly used to treat type 2 diabetes because they exert a number of effects on glucose and lipid metabolism, resulting in reduced insulin resistance and increased glucose uptake and utilisation in non-cerebral tissues [1, 2]. The finding that thiazolidinediones are potent inhibitors of inflammation [3, 4, 5] has prompted several clinical trials testing the effects of thiazolidinediones in neurological diseases such as Alzheimer’s disease and multiple sclerosis, in which inflammation contributes to pathogenesis . The anti-inflammatory properties of pioglitazone in the brain have also been shown in murine models of acute  and chronic  inflammation, when this thiazolidinedione was administered to the animals orally. Both the metabolic and inflammatory actions of thiazolidinediones have been attributed to changes in gene expression mediated by peroxisome proliferator-activated nuclear receptor (PPAR)-γ, of which thiazolidinediones are ligands . However, several recent studies point to the existence of PPAR-independent effects of thiazolidinediones, such as the regulation of mitochondrial respiration [9, 10, 11, 12]. Impaired mitochondrial activity has been related to insulin resistance  and correlated with the clinical state in Alzheimer’s disease . These data raise the question of whether thiazolidinediones can determine the metabolic rate of the brain.
So far, the evaluation of cerebral glucose utilisation (CGU), a quantitative measure of neuronal activity that principally reflects integrated synaptic function , has not been used to demonstrate whether thiazolidinediones alter brain energy metabolism in vivo. In vitro studies showing that pioglitazone increases glucose uptake and consumption in muscle cells , adipocytes  and astrocytes  have suggested that the capacity to stimulate glucose metabolism could be a general property of thiazolidinediones that is also applicable to the brain. Considering the strict dependence that the brain has on glucose as a fuel , a stimulatory action of thiazolidinediones on glucose metabolism would provide an additional therapeutic benefit in pathologies such as Alzheimer’s disease. A growing body of evidence supports the idea that alterations in brain glucose metabolism and insulin sensitivity or signalling contribute to the pathogenesis of Alzheimer’s disease. Regional decreases in the activity of glycolytic enzymes and in glucose utilisation have been detected in sporadic Alzheimer’s disease [18, 19], and the administration of glucose  or insulin [21, 22] can facilitate memory in patients. This evidence has led to the notion of metabolic insufficiency or glucoregulatory impairment in Alzheimer’s disease [23, 24, 25] and has provided a strong rationale for the therapeutic use of drugs, such as thiazolidinediones, that increase insulin sensitivity and glucose consumption. Preliminary results consistently suggest that restoring adequate levels of insulin and glucose by using a thiazolidinedione facilitates memory in patients with Alzheimer’s disease . Since type 2 diabetes increases the risk of Alzheimer’s disease and causes cognitive decline [27, 28], facilitated glucose utilisation in the brain would provide additional therapeutic benefits to patients taking thiazolidinediones regularly to treat diabetes.
We thus sought to determine whether thiazolidinediones affect brain glucose metabolism in vivo. To this end we tested the effect of pioglitazone treatment on regional CGU using the 2-deoxyglucose autoradiographic technique . We compared the cerebral pattern of CGU in normal mice with that of mice overexpressing transforming growth factor β1 (TGFβ1). When produced in excess, TGFβ1 causes robust astrocytosis and microglia activation, vascular deposition of fibrillar material, decreased CGU and impaired cerebral blood flow [29, 30, 31]. This recently corroborated murine model  presents a condition of chronic inflammation  and oxidative stress  associated with metabolic deficiency which makes them suitable for testing in vivo the combined anti-inflammatory and metabolic actions of thiazolidinediones. Our findings reveal unexpected actions of thiazolidinediones on brain glucose metabolism, which raise the question of whether thiazolidinediones are appropriate for the treatment of neurodegenerative diseases in which deficiencies in glucose metabolism occur.
Materials and methods
Animal groups and treatment
Physiological parameters in the four experimental groups
Wild-type, control food
Wild-type, pioglitazone treatment
TGFβ1 overexpression, control food
TGFβ1 overexpression, PIO treatment
Weight before treatment (g)
Weight after treatment (g)
Plasma glycaemia (g/l)a
Packed cell volume (%)a
Measurement of glucose utilisation
Glucose utilisation was measured by the 2-deoxy-d-[14C]glucose ([14C]DG) technique , adapted to the awake mouse . A dose of 5.5 MBq/kg of [14C]DG (2,035 MBq/mmol; NEN/Perkin Elmer Life Sciences: Courtaboeuf-Les Ulis, Essonne, France) dissolved in 45 μl of saline was infused through the femoral vein. The concentration of [14C]DG in the brain was assessed by densitometric analysis of autoradiograms of 20 μm-thick brain sections . For methodological reasons, animals were fasted overnight before CGU measurement. Data are expressed as regional CGU with respect to the value found in the white matter (genu of the corpus callosum) of the same animal, and calculated as a ratio of tracer concentration in Bq/g. Absolute CGU values (mean±SEM) for the genu of the corpus callosum (μmol 100 g−1 min−1) were 34.7±5.0 and 38.7±4.4 for wild-type mice on the control and pioglitazone diets, respectively, and 39.7±5.1 and 33.2±3.3 for TGFβ1-overexpressing mice on the control and pioglitazone diets, respectively. Groups did not differ significantly in this CGU value.
Data were analysed by one- or two-way ANOVA, as appropriate, using StatView software (SAS Institute, Cary, NC, USA); p values are given for the ANOVA. Linear regression analysis was used to evaluate the relationship between the changes in CGU induced by either TGFβ1 overexpression or pioglitazone and regional CGU values. The p value of the correlation coefficient is given.
The main finding of this study is that pioglitazone decreased CGU in multiple brain regions, and that an essential feature of this action is its proportionality to the basal CGU. This is the only case in which pioglitazone has been found to reduce glucose metabolism, since all previous reports showed that this and other thiazolidinediones increase glucose uptake in cultured cells [9, 16, 17] and in vivo, as recently shown in the myocardium of diabetic patients . In addition, pioglitazone did not reverse the decreases in CGU associated with overexpression of TGFβ1, despite the fact that this thiazolidinedione efficiently inhibits the inflammatory reaction in this model . Together, these data support a dual but distinctive action of pioglitazone on glucose metabolism and inflammatory processes in the brain. As glucose utilisation reflects neuronal activity in normal, resting physiological conditions [15, 34], this finding prompts the question of the effects of thiazolidinediones on brain function. This issue is of current interest since (1) several million diabetic patients take pioglitazone, (2) these patients suffer from a greater rate of cognitive decline [27, 28], and (3) pioglitazone is being tested for the treatment of neurological diseases (for information about clinical trials in progress see ). Below we discuss the possible mechanisms underlying the changes in brain glucose utilisation induced by pioglitazone in vivo, compared with the effects in cultured cells, and the therapeutic implications.
An increase in glucose uptake is suggestive of a neural activation . The increases in glucose uptake in cultured astrocytes or in brain slices were not restricted to pioglitazone, as they were mimicked by other thiazolidinediones, although with different potencies of action . A clue to the underlying mechanism is provided by the findings that pioglitazone reduces mitochondrial respiration and that the lactate content increases in parallel . The triple event of increased glucose uptake and lactate release concomitant with decreased respiration has also been described in skeletal muscle . Such a combination suggests a cause–effect relationship in which the increased glucose uptake may represent an adaptive response, producing ATP via a highly glucose-consuming pathway, which generates lactate. This response would compensate for the decreased oxygen-dependent production of ATP resulting from impaired respiration. If so, the increased CGU could be considered an acute reaction that protects against alterations in mitochondrial function. Interestingly, the effect of thiazolidinediones on mitochondrial respiration appeared to be independent of PPAR [9, 11].
A major difference between in vitro and our in vivo studies is that the former provide information about effects occurring as early as 2 h after the addition of pioglitazone, whereas the latter reflect the effects of 2 months of treatment aimed at modelling the chronic actions of the thiazolidinedione. In brain slices the stimulatory effect of pioglitazone on glucose uptake dissipates several days after initial exposure to the thiazolidinedione . This reveals the transient nature of the rapid increase in glucose uptake. Another difference is that the metabolically available glucose, which is in large excess in vitro, may be limited in vivo. Hence, increased glucose utilisation in vivo, if any, may be short-lasting and limited. Conceivably, reduction in glucose utilisation in pioglitazone-treated mice could represent chronic down regulation of aerobic glucose metabolism in the brain resulting from a reduction in mitochondrial respiration, as shown in cultures. This process would be initiated to maintain the balance between aerobic and anaerobic utilisation of glucose so as to avoid excessive glucose consumption. It presumably consists of multiple adaptive changes, which would account for the regionally differentiated sensitivity to pioglitazone, including local re-equilibration of the glucose metabolic pathways and improvement of the insulin signalling pathway [35, 36]. As a result, the highest, potentially ‘excessive’ metabolic rate in vivo would be preferentially depressed. Why this action of pioglitazone is specific to the brain remains unclear. The differing effects of TGFβ1 overexpression on CGU throughout the brain indicate the involvement of other mechanisms of action. The decrease in CGU induced by TGFβ1 is probably due to glia inflammation and fibrosis-derived damage. Indeed, the hippocampal formation, in which several regions showed large decreases in CGU, also displayed the strongest inflammation . In regions where a significant interaction was found between pioglitazone and TGFβ1, no further hypometabolism occurred with pioglitazone treatment, as if a minimal metabolic rate had already been attained owing to TGFβ1.
Although the rationale for the therapeutic use of thiazolidinediones to compensate for hypometabolism in Alzheimer’s disease has received strong support [20, 21, 22], the data presented here suggest that long-term administration of pioglitazone may not enhance the capacity of the brain to metabolise glucose in Alzheimer’s disease. However, this may not necessarily mean that pioglitazone is deleterious to the brain. First, the possibility cannot be ruled out that the brain utilises more of a fuel other than glucose, such as lactate of peripheral origin . Second, pioglitazone treatment may be protective by limiting the highest rates of regional glucose oxidative metabolism and the concomitant oxidative stress [14, 32, 38, 39], in association with the benefits derived from the attenuation of inflammation . Finally, pioglitazone can induce the expression of proteins related to the stress response  and mitochondrial membrane hyperpolarisation that can safeguard the cells against an excessive metabolic rate . This scenario would be akin to a ‘preconditioning’ effect in which a drug is beneficial by mildly impairing mitochondrial respiration. In support of these ideas, pioglitazone has been shown to be protective in situations of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine administration in the mouse model of Parkinson’s disease  and in acute metabolic stress caused by ischaemia . Extrapolating these findings to diabetic patients should nevertheless be done with caution. Indeed, no cognitive impairment has been reported in patients treated with pioglitazone, and first results from a clinical trial testing the thiazolidinedione rosiglitazone in Alzheimer’s disease point to preserved cognition . A very recent study in mice modelling Alzheimer’s disease suggests that rosiglitazone attenuates learning and memory deficits by lowering glucocorticoid actions . However, unlike pioglitazone, rosiglitazone does not cross the blood–brain barrier. Therefore, any effect on cognition by rosiglitazone ought to be initiated outside the brain, while pioglitazone would act both centrally and peripherally. Hence, it is premature to conclude that pioglitazone would mimic the beneficial effects of rosiglitazone. Forthcoming studies should address more adequately the repercussions on brain function of the use of different thiazolidinediones, and define the mechanisms by which thiazolidinediones could provide therapeutic benefits in neurological disorders.
The authors thank G. Landreth and M. Heneka for providing the animal food containing pioglitazone and R. Sercombe for critical reading of the manuscript. E. Galea was the recipient of an award from Association France Alzheimer.
Duality of interest
The authors declare that they have no duality of interest.
- 36.Steen E, Terry BM, Rivera EJ et al (2005) Impaired insulin and insulin-like growth factor expression and signaling mechanisms in Alzheimer’s disease – is this type 3 diabetes? J Alzheimers Dis 7:63–80Google Scholar
- 42.Pedersen WA, McMillan PJ, Kulstad JJ, Leverenz JB, Craft S, Haynatzki GR (2006) Rosiglitazone attenuates learning and memory deficits in Tg2576 Alzheimer mice. Exp Neurol (Epub ahead of print) DOI 10.1016/j.expneurol.2006.01.018