Researchers routinely use FDG PET scans, which measure brain glucose consumption, as a marker of neural activity. Now, a new paper challenges the idea that this signal solely represents neuronal function. In the January 30 Nature Neuroscience, researchers led by Pedro Rosa-Neto at McGill University, Montreal, report that stimulating glucose uptake by astrocytes in the living rat brain enhanced the FDG PET signal. “This has implications for understanding what FDG PET means as an outcome measure in clinical trials. It may reflect intervention effects on astrocytes as well as neurons,” Rosa-Neto told Alzforum.

Others noted that the authors did not determine whether astrocytes contribute to the FDG PET signals in rats under normal physiological conditions. “The idea that a significant portion of the FDG PET signal can be driven by astrocytes needs confirmation from other labs and by other methods. I consider it an intriguing, provocative hypothesis worthy of further investigation,” William Klunk at the University of Pittsburgh wrote to Alzforum.

Active Astrocytes.

Stimulation of astrocytes intensified the FDG PET signal in many regions of rat brain, particularly the prefrontal cortex (green). [Courtesy of Zimmer et al., Nature Neuroscience.]

[18F]FDG, short for fluorodeoxyglucose, is a radiolabeled sugar analog that allows researchers to track glucose uptake by brain cells. While brain glucose consumption has been assumed to reflect primarily neuronal activity, other researchers have pointed out that activated neurons release glutamate, which stimulates astrocytes to take up glucose as well (see Sokoloff 1993; Pellerin and Magistretti, 1994; Pellerin and Magistretti, 2012). Several groups have noted that this process could in theory affect the FDG PET signal (see Magistretti and Pellerin, 1996; Nehlig and Coles, 2007; Figley and Stroman, 2011). 

However, no one had shown that astrocytes contribute to FDG PET in the living brain. To do this, first author Eduardo Zimmer used ceftriaxone. Besides fighting bacteria, this antibiotic drug happens to activate the astrocytic glutamate transporter GLT-1, which in turn triggers glucose uptake (see Lee et al., 2008). The authors injected ceftriaxone into the tail veins of adult rats, then injected [18F]FDG 30 minutes later and scanned the animals 40 minutes after that. Compared to rats injected with saline, FDG uptake values in rats on ceftriaxone ran about 15 percent higher in several brain regions, including hippocampus, striatum, thalamus, and temporoparietal cortex. The changes were statistically significant, with the prefrontal cortex notching the most notable boost, of about 20 percent (see image above). 

Several lines of evidence reinforced the idea that astrocyte activation drove the enhanced signal. The most affected regions are reported to express the highest levels of GLT-1, according to the Allen Brain Atlas. On the other hand, the FDG PET signal barely budged in the cerebellum, which is reported to express little GLT-1. Also, blood flowed at normal rates through the cerebrum in treated rats. Because active neurons release glutamate, which causes blood vessels to dilate, this to the authors indicated no difference in neuronal activity in treated rats.

The finding should stimulate more research, commenters suggested. “This is a great example of the interaction between the cell types, and demonstrates that using imaging can help elucidate those interactions,” Arthur Toga at the University of Southern California, Los Angeles, wrote to Alzforum.

What does this mean for AD studies? Brain glucose use goes down in people with prodromal AD, and is low in adult carriers of the ApoE4 allele regardless of their cognitive health (see Research Timeline 1980; Research Timeline 1996; Dec 2012 news). If astrocytes do play a role in FDG PET in the resting-state brain, that could tweak the interpretation of scans in past studies, but would not invalidate that research, commenters agreed.

“We know that glucose hypometabolism in AD at least partially reflects neuronal and synaptic loss and dysfunction, but there are other factors at work, such as disconnection of distant brain regions,” Gaël Chételat at INSERM-EPHE-University of Caen, France, told Alzforum. “This [paper] provides additional evidence that hypometabolism reflects a broader mechanism than just neuronal and synaptic loss.” For his part, Klunk noted that the new data creates a puzzle, because astrogliosis develops around plaques in AD brain. If activated astrocytes drive the FDG PET signal, why would it go down in AD and not up?

In future work, Rosa-Neto will try to determine the precise role of astrocytes in the FDG PET signal by combining PET scans with fMRI and electrophysiology. He will also study how therapeutic interventions affect aerobic metabolism and astrocyte-neuron interactions.—Madolyn Bowman Rogers


  1. We thank Dr. Klunk for sparking an important discussion related to the role of astrocytes in AD by asking the question “If activated astrocytes drive the FDG PET signal, why would it go down in AD and not up?” As molecular imaging is based on pharmacokinetic compartments, inferences regarding cellular compartmentalization of PET signals should be carefully considered. Specifically regarding our results, one can link [18F]FDG uptake to the astrocytic compartment since GLT-1 is highly expressed in astrocytes (Dehnes et al., 2008). In pathological conditions, such as AD, one might predict regional coexistence between hypometabolism and astrocytosis, if the astrocyte is injured (expressing reduced or dysfunctional GLT-1). In fact, data from the human postmortem literature suggest abnormal GLT-1 expression in the brain of AD patients (Scott et al., 2011; Masliah et al., 2006). Animal model data shows that GLT-1 is a target of Aβ toxic effects (Scimemi et al., 2013). Indeed, an elegant study by Hefendehl and colleagues (recently covered in Alzforum) shows reduced GLT-1 levels in the vicinity of amyloid plaques (Hefendehl et al., 2016). These data suggest a link between astrocytic GLT-1 abnormalities and AD pathophysiology that might contribute with the reduced [18F]FDG uptake observed in AD brain. We believe that the molecular mechanism underling hypometabolism in AD is complex and involves the local deleterious actions of Aβ aggregates on astrocytes and neurons as well as depletions of remote cortical inputs. We hope that our results would contribute with the interpretation of FGD PET in normal and abnormal conditions. 


    . A quantitative assessment of glutamate uptake into hippocampal synaptic terminals and astrocytes: new insights into a neuronal role for excitatory amino acid transporter 2 (EAAT2). Neuroscience. 2008 Nov 11;157(1):80-94. Epub 2008 Aug 27 PubMed.

    . Glutamate transporter variants reduce glutamate uptake in Alzheimer's disease. Neurobiol Aging. 2011 Mar;32(3):553.e1-11. PubMed.

    . Deficient glutamate transport is associated with neurodegeneration in Alzheimer's disease. Ann Neurol. 1996 Nov;40(5):759-66. PubMed.

    . Amyloid-β1-42 slows clearance of synaptically released glutamate by mislocalizing astrocytic GLT-1. J Neurosci. 2013 Mar 20;33(12):5312-8. PubMed.

    . Mapping synaptic glutamate transporter dysfunction in vivo to regions surrounding Aβ plaques by iGluSnFR two-photon imaging. Nat Commun. 2016 Nov 11;7:13441. PubMed.

  2. The authors demonstrate nicely that astrocytic glutamate transport affects glucose uptake and that this can be measured using [18F]FDG PET. This is a great example of the interaction between cell types and demonstrates that using imaging can help elucidate those interactions. The implications for PET studies in AD are also important given the fact that neuron/astrocyte communication is an important regulatory mechanism in a variety of brain functions.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Does ApoE4 Lower Brain Metabolism Independently of Aβ?

Paper Citations

  1. . Function-related changes in energy metabolism in the nervous system: localization and mechanisms. Keio J Med. 1993 Sep;42(3):95-103. PubMed.
  2. . Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc Natl Acad Sci U S A. 1994 Oct 25;91(22):10625-9. PubMed.
  3. . Sweet sixteen for ANLS. J Cereb Blood Flow Metab. 2012 Jul;32(7):1152-66. Epub 2011 Oct 26 PubMed.
  4. . The contribution of astrocytes to the 18F-2-deoxyglucose signal in PET activation studies. Mol Psychiatry. 1996 Dec;1(6):445-52. PubMed.
  5. . Cellular pathways of energy metabolism in the brain: is glucose used by neurons or astrocytes?. Glia. 2007 Sep;55(12):1238-50. PubMed.
  6. . The role(s) of astrocytes and astrocyte activity in neurometabolism, neurovascular coupling, and the production of functional neuroimaging signals. Eur J Neurosci. 2011 Feb;33(4):577-88. PubMed.
  7. . Mechanism of ceftriaxone induction of excitatory amino acid transporter-2 expression and glutamate uptake in primary human astrocytes. J Biol Chem. 2008 May 9;283(19):13116-23. Epub 2008 Mar 7 PubMed.

Other Citations

  1. Research Timeline 1980

External Citations

  1. Allen Brain Atlas

Further Reading

Primary Papers

  1. . [(18)F]FDG PET signal is driven by astroglial glutamate transport. Nat Neurosci. 2017 Jan 30; PubMed.