Dogma says that brain metabolism slows down with age: Blood flow, and the total demand for fuel in the form of glucose, decline with passing years. A new study shows that waning glucose uptake is just part of the story—how that glucose gets used also changes dramatically with age.
Writing in the August 1 issue of Cell Metabolism, Manu Goyal, Andrei Vlassenko, and Marcus Raichle of Washington University School of Medicine in St. Louis report that the young brain favors a quick burn, in the form of the partial breakdown of glucose known as aerobic glycolysis. But whole-brain glycolysis declines steadily with normal aging, from its peak at age 20 to nearly zero by age 60. At that age, nearly all of the glucose entering the brain gets metabolized through mitochondria in a longer process that oxidizes glucose completely to water and carbon dioxide. Regional measures indicate that brain areas that start out with the highest rates of glycolysis decline the most, and by age 60, topological differences all but disappear.
“We see that with aging, one of the principal changes in brain metabolism is loss of aerobic glycolysis, and the loss of a youthful pattern of aerobic glycolysis,” said Goyal.
Aerobic glycolysis declines fastest in brain regions that finish developing last. This is consistent with the idea that aerobic glycolysis helps meet the special needs of the maturing brain by supporting synaptic remodeling and myelination, and providing neuroprotection, then lessens once maturation is complete.
The highest activity and fastest drops are seen in areas of the default mode network, a set of interconnected brain hubs that are voracious consumers of glucose through life and that experience amyloid deposition and network dysfunction early in Alzheimer’s disease. “This suggests that the cells in these particular parts of the brain are different from others. Maybe hidden within that difference is something that increases their vulnerability over the life span,” said Raichle.
One source of vulnerability may be a diminished buffer against oxidative stress. “While this may be a normal physiological development of old age, an associated aspect of decreasing aerobic glycolysis is a loss of neuroprotection via the pentose phosphate pathway that protects against oxidative stress, thus increasing the risk for oxidative damage. This, in turn, may be a key predisposing factor to neurodegenerative diseases such as Alzheimer’s and Parkinson’s. This may be one reason why these diseases are age-related,” said Richard Caselli, Mayo Clinic Arizona, Scottsdale, in an email to Alzforum. Amyotrophic lateral sclerosis (ALS), too, is marked by oxidative stress in neurons.
Previous work from Raichle’s lab and others demonstrated high levels of aerobic glycolysis in children’s brains, especially in regions undergoing synapse formation and growth. In young adults, glycolysis persists in the resting brain in default mode network regions, which are also the regions most prone to amyloid deposition (Goyal et al., 2014; Sep 2010 news; Research Timeline).
In the new study, Goyal and Vlassenko asked what happens to glycolysis as the brain ages and developmental activity ramps down. The investigators first meta-analyzed previously published quantifications of glucose uptake, oxygen use, and cerebral blood flow in healthy people across age. Some of the 15 studies they drew from used PET scanning; others were classical studies dating back to the 1950s involving the placement of venous and arterial catheters for direct measurement of tracers in blood. The technique, too invasive for general use today, still provides the most quantitative measure of glucose metabolism. Putting the data together, the researchers found an age-related decrease in total glucose use, as expected. However, while the fraction of total glucose burned through the slower route—mitochondrial oxidative phosphorylation—barely changed over the 40-year age span in the study, the fraction going specifically through glycolysis declined steadily from a high of about 20 percent in 20-year-olds, to zero in 60-year-olds.
“The decrease in glucose utilization with age is not new, but the finding that the decrease was largely restricted to loss of aerobic glycolysis was a surprise,” Raichle said.
To explore regional changes, Goyal and colleagues moved to modern PET, using 18F-fluoro-deoxyglucose (FDG) and 15O to measure glucose and oxygen use, respectively, in resting brain scans of 205 healthy men and women between the ages of 20 and 82. The subjects, all cognitively normal, were part of ongoing studies in St. Louis, including some from the Dominantly Inherited Alzheimer Network (DIAN) observational trial, and the ongoing Adult Children Study. Only amyloid-negative subjects were included in the final data analysis of 184 PET sessions in 165 participants.
For each subject, the scientists calculated local rates of glycolysis in 78 brain regions, which revealed that the topology of glycolysis across the brain varied dramatically with age. “Once people got to that 60-and-above age group, you saw some resemblance to the pattern of young adults, but it was just a resemblance,” Goyal said.
The regions that showed highest glycolysis in young adults, such as the precuneus and precentral cortex, fell the most in older people. Areas that started out lower, including the cerebellum, hippocampus, and corpus callosum, remained relatively stable. Because the areas with the highest rates dropped the most, aging effectively “flattened” the topographical map of glycolysis across the brain (see image).
The rapidly declining regions had something in common—they were among the last parts of the brain to mature. That led Goyal to think that the reason the human brain loses aerobic glycolysis across adulthood is because it finally grows up. “It might be that regions with high aerobic glycolysis in young adults are still developing. When they reach maturity, they don’t need aerobic glycolysis anymore, and that’s why we see this flattening,” Goyal said.
In support of this idea, the group had previously demonstrated that, in 16 regions of the brain where they had gene-expression data, the persistence of genes involved in childhood brain development (a characteristic known as neoteny) correlated with glycolysis. In the current work, the age-related decreases also correlated with the degree of neoteny.
The authors suggest that loss of glycolysis could serve as a biomarker for brain aging. But as Marc Dhenain of the CNRS in Fontenay aux Roses, France, pointed out, much of the absolute decrease occurs early, between 20 and 40 years of age. “I think this is more a marker of brain development, and the end of development, rather than aging,” Dhenain said.
Goyal said he thinks of glycolysis as a marker of the brain’s youthfulness. In older people, he sees a lot of variability in the pattern of glycolysis. “In some people in their 70s, the pattern of aerobic glycolysis looks more or less like it does in the young adult, whereas in another older person it looks completely different,” he said. Are those differences important? “My suspicion is that people with a less youthful metabolic brain profile might be more vulnerable to the neurodegeneration seen in Alzheimer’s disease, and might develop it faster,” he said. He and Vlassenko would like to do longitudinal studies to try to answer that question.
In the meantime, they are measuring glycolysis in people with amyloid pathology, both presymptomatic and mildly symptomatic. Vlassenko presented preliminary data at the recent AAIC in London, showing that, in cognitively normal people with amyloid, aerobic glycolysis and tau went in opposite directions. In other words, people with higher glycolysis had less tau pathology.
The current study looked at resting-state metabolism; however, the brain ramps up glycolysis when at work, to fuel synaptic function. Vlassenko said they are starting to look at this dynamic activity as well. A previous study in mice linked neuronal activity, glycolysis, and Aβ deposition (May 2011 news). Goyal is also interested in how changes in glucose use in the default mode network relate to changes in functional connectivity seen on fMRI.
Gwenaëlle Douaud, University of Oxford, England, U.K., said the work provides a strong metabolic hypothesis for what she and others have previously observed: that regions of the brain that develop later are more vulnerable to oxidative stress and age-related neurodegeneration (Douaud et al., 2014). “It will be very interesting to see, as the authors point out, whether this might also be related to loss of myelination. This is something for which Big Data imaging projects, such as the Lifespan Human Connectome Project or the U.K. Biobank Brain Imaging Study, might provide direct, quantifiable answers,” she wrote in an email to Alzforum.
“The findings raise interesting and important questions about the relationship of these physiological changes to other biological changes associated with normal brain aging, such as declines in the density, connectivity, function, or turnover of terminal neuronal fields that innervate these regions, or the density, connectivity, or function of peri-synaptic astroglial cells,” said Eric Reiman, Banner Alzheimer’s Institute, Phoenix. “For instance, prior studies of normal aging have noted a preferential reduction of terminal neuronal fields innervating frontal cortex—not to mention how some of these changes may conspire with other factors in the predisposition to Alzheimer’s disease,” Reiman added.
The study suggests that asessing total glucose uptake with FDG PET offers but a partial view of brain metabolsim. Ai-Ling Lin, University of Kentucky College of Medicine in Lexington, told Alzforum that she believes measuring glycolysis is a better biomarker for aging because it gives more information and changes more dramatically than FDG PET. However, measuring glycolysis with 15O PET is technically complex, and done at only a few centers in the world. Several labs are now developing MRI techniques for measuring oxygen metabolism that will make it easier and more accessible to additional centers, Lin said.
Peter Nelson, a pathologist at the University of Kentucky, Lexington, told Alzforum the study shows why descriptive studies of the human brain remain so important and scientifically valid. “There’s a vogue that is against anything that is not based on clinical hypotheses, and a several-decades-long trend of valuing mouse models. This study shows that you get this incredible benefit from simply studying what goes on in the human brain. Doing it carefully in a large series like this is what it’s all about.”—Pat McCaffrey
- Brain Aβ Patterns Linked to Brain Energy Metabolism
- Do Overactive Brain Networks Broadcast Alzheimer’s Pathology?
- Goyal MS, Hawrylycz M, Miller JA, Snyder AZ, Raichle ME. Aerobic glycolysis in the human brain is associated with development and neotenous gene expression. Cell Metab. 2014 Jan 7;19(1):49-57. PubMed.
- Douaud G, Groves AR, Tamnes CK, Westlye LT, Duff EP, Engvig A, Walhovd KB, James A, Gass A, Monsch AU, Matthews PM, Fjell AM, Smith SM, Johansen-Berg H. A common brain network links development, aging, and vulnerability to disease. Proc Natl Acad Sci U S A. 2014 Dec 9;111(49):17648-53. Epub 2014 Nov 24 PubMed.
- Vlassenko AG, Raichle ME. Brain aerobic glycolysis functions and Alzheimer's disease. Clin Transl Imaging. 2015 Feb 1;3(1):27-37. Epub 2014 Dec 10 PubMed.
- Goyal MS, Vlassenko AG, Blazey TM, Su Y, Couture LE, Durbin TJ, Bateman RJ, Benzinger TL, Morris JC, Raichle ME. Loss of Brain Aerobic Glycolysis in Normal Human Aging. Cell Metab. 2017 Aug 1;26(2):353-360.e3. PubMed.
- Dagher A, Misic B. Holding Onto Youth. Cell Metab. 2017 Aug 1;26(2):284-285. PubMed.