Effects of Exercise on Progranulin Levels and Gliosis in Progranulin-Insufficient Mice

Abstract

Loss-of-function mutations in progranulin (GRN) are one of the most common genetic causes of frontotemporal dementia (FTD), a progressive, fatal neurodegenerative disorder with no available disease-modifying treatments. Through haploinsufficiency, these mutations reduce levels of progranulin, a protein that has neurotrophic and anti-inflammatory effects. Increasing progranulin expression from the intact allele is therefore a potential approach for treating individuals with GRN mutations. Based on the well-known effects of physical exercise on other neurotrophic factors, we hypothesized that exercise might increase brain progranulin levels. We tested this hypothesis in progranulin heterozygous (Grn^+/-) mice, which model progranulin haploinsufficiency. We housed wild-type and progranulin-insufficient mice in standard cages or cages with exercise wheels for 4 or 7.5 weeks, and then measured brain and plasma progranulin levels. Although exercise modestly increased progranulin in very young (2-month-old) wild-type mice, this effect was limited to the hippocampus. Exercise did not increase brain progranulin mRNA or protein in multiple regions, nor did it increase plasma progranulin, in 4- to 8-month-old wild-type or Grn^+/- mice, across multiple experiments and under conditions that increased hippocampal BDNF and neurogenesis. Grn^-/- mice were included in the study to test for progranulin-independent benefits of exercise on gliosis. Exercise attenuated cortical microgliosis in 8-month-old Grn^-/- mice, consistent with a progranulin-independent, anti-inflammatory effect of exercise. These results suggest that exercise may have some modest, nonspecific benefits for FTD patients with progranulin mutations, but do not support exercise as a strategy to raise progranulin levels.

Keywords: BDNF; exercise; frontotemporal dementia; progranulin.

Figure 1

Exercise (3-4 weeks) produces a small increase in hippocampal progranulin in young adult wild-type mice. Two- to 3-month-old wild-type mice were randomized to sedentary or exercise groups for three to four weeks (A). B, Exercise produced a robust increase in BDNF (****p < 0.0001). C, Representative BDNF and α-tubulin blots. D, Exercise did not significantly increase frontal cortex progranulin protein. E, Exercise produced a small but statistically significant increase in hippocampal progranulin (*p = 0.038). F, Exercise did not affect plasma progranulin. n = 8–16 mice per group. Values in B–F are expressed relative to the sedentary group.

Figure 2

Progranulin deficiency does not affect wheel-running behavior. Mice aged 6 months were randomized to sedentary or exercise groups for 7.5 weeks. A, No genotype differences were detected in wheel running distance between Grn^+/+, Grn^+/−, and Grn^−/− mice. B, Exercise reduced body weight similarly in all three genotypes (***ANOVA exercise effect, p = 0.003). n = 10–15 mice per group.

Figure 3

Exercise (7.5 weeks) does not increase progranulin in wild-type or Grn^+/− mice. Exercise for 7.5 weeks did not increase progranulin mRNA in frontal cortex (A) or thalamus (B), or progranulin protein levels in the hippocampus (C) or plasma (D). E–G, In contrast, exercise increased hippocampal BDNF levels across all three genotypes (E, ANOVA effect of exercise, p = 0.035), with no change in α-tubulin as a loading control (F). G, Representative BDNF and α-tubulin blots. H, Exercise also increased the number of doublecortin (DCX)-positive neurons in the dentate gyrus (ANOVA effect of exercise p < 0.0001, Sidak’s post hoc test exercise > sedentary Grn^+/+, p = 0.0089, Grn^+/−, p = 0.0198). Representative images (20×) are shown of doublecortin immunostaining in the dentate gyrus. Scale bar, 25 µm. Mice were 6-months-old on average when beginning the study, and 8-months-old on average when samples were collected. n = 10–15 mice per group. Values in A–F are expressed relative to the wild-type sedentary group.

Figure 4

Exercise (4 weeks) does not increase progranulin protein levels in multiple brain regions of Grn^+/− mice. Four- to 8-month-old Grn^+/− mice were randomized to sedentary or exercise groups for 4 weeks (A). Exercise did not increase progranulin protein levels in frontal cortex (B), thalamus (C), or hippocampus (D), despite producing the expected increase in the number of doublecortin-positive neurons in the dentate gyrus (E, ** p < 0.01). Values in B–D are expressed relative to sedentary Grn^+/+ mice, with sedentary Grn^+/− mice set at 0.5 to maintain a consistent scale with Figure 3. n = 11–12 per group.

Figure 5

Group housing does not potentiate the effects of exercise on progranulin. Six weeks of wheel running (A) failed to increase frontal cortex progranulin mRNA (B) or hippocampal progranulin protein (C) in either solo- or group-housed wild-type mice aged 3–6 months. D-F, Hippocampal BDNF was increased by exercise (D; ANOVA effect of exercise, p = 0.041). F, Representative BDNF and α-tubulin blots for each group. n = 8–14 mice per group. Values in B–E are expressed relative to the solo-housed sedentary group.

Figure 6

Exercise (7.5 weeks) reduces cortical microgliosis in Grn^−/− mice. A, Grn^−/− mice had elevated CD68 immunoreactivity in the frontal cortex, hippocampus, and thalamus (ANOVA genotype effect, p < 0.0001). Global analysis of CD68 immunostaining revealed a region x genotype interaction (p < 0.0001), so each region was analyzed with a separate ANOVA. Analysis of CD68 immunostaining in the frontal cortex revealed an effect of genotype (ANOVA, p < 0.0001) and a strong trend for an exercise effect (ANOVA exercise effect, p = 0.051). Subsequent post hoc analysis revealed significantly lower CD68 immunoreactivity in exercised relative to sedentary Grn^−/− mice (*p < 0.05). B, Representative images (20×) of CD68 immunostaining from the frontal cortex. Scale bar, 25 µm.