Brain glucose metabolism during hypoglycemia in type 1 diabetes: insights from functional and metabolic neuroimaging studies

Abstract

Hypoglycemia is the most frequent complication of insulin therapy in patients with type 1 diabetes. Since the brain is reliant on circulating glucose as its main source of energy, hypoglycemia poses a threat for normal brain function. Paradoxically, although hypoglycemia commonly induces immediate decline in cognitive function, long-lasting changes in brain structure and cognitive function are uncommon in patients with type 1 diabetes. In fact, recurrent hypoglycemia initiates a process of habituation that suppresses hormonal responses to and impairs awareness of subsequent hypoglycemia, which has been attributed to adaptations in the brain. These observations sparked great scientific interest into the brain's handling of glucose during (recurrent) hypoglycemia. Various neuroimaging techniques have been employed to study brain (glucose) metabolism, including PET, fMRI, MRS and ASL. This review discusses what is currently known about cerebral metabolism during hypoglycemia, and how findings obtained by functional and metabolic neuroimaging techniques contributed to this knowledge.

Keywords: Brain metabolism; Cerebral blood flow; Hypoglycemia; Impaired awareness of hypoglycemia; Neuroimaging; Type 1 diabetes mellitus.

Figures

Fig. 1

A simplified illustration of the multiple metabolic pathways of glucose in the brain and the metabolic signals used in different neuroimaging techniques. The initial step of glucose metabolism is phosphorylation of glucose to glucose-6-phosphate (Glc-6-P) by hexokinase. Glc-6-P can enter several metabolic pathways in the brain. It can be metabolized to produce energy via glycolysis or the TCA cycle. Glycolytic and TCA cycle intermediates are also used for the synthesis of amino acids and neurotransmitters. In addition, Glc-6-P is a precursor for glycogen. Lastly, metabolism of Glc-6-P via the pentose phosphate pathway (PPP) provides pentose for nucleotide synthesis and NADPH, required for reductive reactions, such as lipid synthesis and for protection against oxidative stress. Arteriovenous concentration differences (AV dif) can be used to estimate global cerebral metabolic rate from the disappearance of metabolites from the circulation. PET (depicted in orange) uses radiolabeled glucose analogues (such as FDG), which are trapped early in metabolism (for example fluorodeoxyglucose-6-phosphate/FDG-6-P), to estimate rates of glucose uptake and metabolism. 31P MRS (depicted in blue) provides information about ATP production and thus brain energy metabolism. 13C-MRS (depicted in green) is useful for estimating TCA cycle fluxes and CMRglc, derived from 13C label incorporation into specific metabolites (Glu, Gln). Both ASL and BOLD fMRI provide estimates of CBF

Fig. 2

Time series of 13C-MR spectra, acquired from a ~125 mL voxel, placed in the occipital cortex. Spectra are averaged over 20 min, after administration of [1-13C]glucose during a hypoglycemic clamp in one healthy subject. Once the infused [1-13C]glucose is taken up by the brain and incorporated into various glucose metabolites, an increase in signal over time is observed. Numbers indicate the position of the 13C label, as explained in more detail in Fig. 3. Asp aspartate, Gln glutamine, Glu glutamate, Lac lactate (from Ref. [143], with permission from Elsevier)

Fig. 3

One-compartment metabolic model describing the incorporation of 13C label from (infused) [1-13C]glucose into the TCA cycle and its metabolites. When taken up by the brain, the 13C-label is first incorporated into the C3 position of pyruvate and subsequently into the C3 position of lactate. Once the 13C-label continues through the TCA cycle, it is incorporated into the C4 position of αKG, glutamate and glutamine. In the second turn of the cycle, the label is equally distributed over the C2 and C3 positions of these metabolites. The TCA cycle flux can be estimated using a metabolic model where the time courses of the uptake of the 13C-label in glutamate and glutamine in the different carbon positions, measured with 13C-MRS, are used as input. Filled circles represent the carbon position that is labeled with 13C, white circles represent unlabeled carbons. αKG α-ketoglutarate, BBB blood–brain-barrier, Glc glucose, Gln glutamine, Glu glutamate, Lac lactate, LDH lactate dehydrogenase, Pyr pyruvate, Vgln exchange rate between glutamate and glutamine, VTCA TCA cycle rate, Vx exchange rate between α-ketoglutarate and glutamate

Fig. 4

Linear relationship between plasma and brain glucose levels under normo- and hypoglycemic conditions in healthy subjects (open squares) and patients with type 1 diabetes (closed circles). Brain glucose levels were measured with 13C-MRS. The plasma versus brain glucose relation was fitted with linear regression analysis to determine reversible Michaelis–Menten kinetics to show the best fit of the data with 95 % confidence intervals. R2 = 0.59, P < 0.001. Assuming continuation of this linear relationship between plasma and brain glucose levels, brain glucose approaches zero at a plasma glucose level of approximately 1.2 mmol/L (from Ref. [104], with permission from the American Diabetes Association)