Acute exercise alters brain glucose metabolism in aging and Alzheimer's disease

Abstract

There is evidence that aerobic exercise improves brain health. Benefits may be modulated by acute physiological responses to exercise, but this has not been well characterized in older or cognitively impaired adults. The randomized controlled trial 'AEROBIC' (NCT04299308) enrolled 60 older adults who were cognitively healthy (n = 30) or cognitively impaired (n = 30) to characterize the acute brain responses to moderate [45-55% heart rate reserve (HRR)] and higher (65-75% HRR) intensity acute exercise. Each participant received two fluorodeoxyglucose positron emission tomography (FDG-PET) scans, one at rest and one following acute exercise. Change in cerebral glucose metabolism from rest to exercise was the primary outcome. Blood biomarker responses were also characterized as secondary outcomes. Whole grey matter FDG-PET standardized uptake value ratio (SUVR) differed between exercise (1.045 ± 0.082) and rest (0.985 ± 0.077) across subjects [Diff = -0.060, t(58) = 13.8, P < 0.001] regardless of diagnosis. Exercise increased lactate area under the curve (AUC) [F(1,56) = 161.99, P < 0.001] more in the higher intensity group [mean difference (MD) = 97.0 ± 50.8] than the moderate intensity group (MD = 40.3 ± 27.5; t = -5.252, P < 0.001). Change in lactate AUC and FDG-PET SUVR correlated significantly (R² = 0.179, P < 0.001). Acute exercise decreased whole grey matter cerebral glucose metabolism. This effect tracked with the systemic lactate response, suggesting that lactate may serve as a key brain fuel during exercise. Direct measurements of brain lactate metabolism in response to exercise are warranted. KEY POINTS: Acute exercise is associated with a drop in global brain glucose metabolism in both cognitively healthy older adults and those with Alzheimer's disease. Blood lactate levels increase following acute exercise. Change in brain metabolism tracks with blood lactate, suggesting it may be an important brain fuel. Acute exercise stimulates changes in brain-derived neurotrophic factor and other blood biomarkers.

Keywords: Alzheimer's disease; PET; exercise; glucose metabolism; lactate; metabolism.

Figure 1.

Consort Diagram. Study flow from initial assessment for eligibility through study completion is given in this diagram. Of 201 participants initially screened, 60 participants were recruited and 59 participants completed the study.

Figure 2.

Study visit flow. For the AEROBIC study, there were 4 in-person study visits beginning with a graded exercise test to screen for safety and determine heart rate reserve, which was used to accurately administer the correct intensity of exercise on an individualized basis. Visits 2 and 3 were the resting and exercise FDG-PET visits, which were randomized in terms of order (whether resting or exercise was done first) as well as intensity (whether exercise was at moderate (45–55% HRR) or higher (65–75% HRR) intensity). The final visit was an MRI scan which was used for spatial normalization of the FDG-PET scan.

Figure 3.

Change in brain glucose metabolism with exercise. A) Diagnosis group means and standard deviations of raw FDG-PET SUVR are shown in each condition (exercise vs. rest). B) Threshold-free cluster-enhanced voxels tested in a paired t-test design as having lower SUVR (cool colors) or higher SUVR (warm colors) in the exercise condition. Only voxels that tested significant at an FWE-controlled p-value of 0.05 are shown. The lower panels show the mean differences between conditions for each diagnosis group individually. No clusters were found that suggested the mean differences between conditions changed based on diagnosis group. Abbreviations: HC, Cognitively Healthy; AD, Alzheimer’s Disease; SUVR, Standardized Uptake Value Ratio; A.U., arbitrary units.

Figure 4.

Effect of intensity of change in brain glucose metabolism. Here, individual change values for whole-GM SUVR are shown for each participant by diagnosis by intensity group. Individual values are shown with the group mean. Error bars represent SD. (*p<.05 pairwise comparison).

Figure 5.

Blood lactate responses during resting and exercise conditions. A) Lactate curve showing Intensity group (Blue, Moderate; Red, Higher) means and standard errors for each time point under each Condition (Solid lines, Exercise; Dotted lines, Rest). The dashed grey line corresponds to the time of radiotracer uptake (injection taking place immediately after the 5min blood draw). The dotted grey line corresponds the time of scan. B) Comparison of Lactate AUCs with pairwise comparisons for both between-group (Intensity, ****p<.0001) and within-group (Condition, ####p<.0001).

Figure 6.

Relationship between lactate response and change in brain glucose metabolism. A) Scatterplot showing relationship between the change in Lactate AUC Change between conditions (exercise – rest) and change in whole-GM SUVR. The slopes for each group did not test as significantly different, so one pooled slope was plotted. Panel B) shows the voxel-wise results from FSL’s randomise for the parameter estimate corresponding to the change in Lactate AUC predictor. The color bar legend includes both the ranges of each voxel’s t-statistic (above color bar) and equivalent standardized beta weight (below color bar).