Myeloid deficiency of the intrinsic clock protein BMAL1 accelerates cognitive aging by disrupting microglial synaptic pruning

Abstract

Aging is associated with loss of circadian immune responses and circadian gene transcription in peripheral macrophages. Microglia, the resident macrophages of the brain, also show diurnal rhythmicity in regulating local immune responses and synaptic remodeling. To investigate the interaction between aging and microglial circadian rhythmicity, we examined mice deficient in the core clock transcription factor, BMAL1. Aging Cd11b^cre;Bmal^lox/lox mice demonstrated accelerated cognitive decline in association with suppressed hippocampal long-term potentiation and increases in immature dendritic spines. C1q deposition at synapses and synaptic engulfment were significantly decreased in aging Bmal1-deficient microglia, suggesting that BMAL1 plays a role in regulating synaptic pruning in aging. In addition to accelerated age-associated hippocampal deficits, Cd11b^cre;Bmal^lox/lox mice also showed deficits in the sleep-wake cycle with increased wakefulness across light and dark phases. These results highlight an essential role of microglial BMAL1 in maintenance of synapse homeostasis in the aging brain.

Keywords: BMAL1; Circadian clock; Microglia; Sleep–wake cycle; Synaptic plasticity.

Fig. 1

Microglia BMAL1 deficiency disrupts hippocampal-dependent behavior and increases anxiety in aged, but not young mice. A Preference in the novel object recognition task in young (3–6 months, CD11b^cre, n = 15; CD11b^cre;Bmal1^lox/lox, n = 12) and aged (18–20 months, CD11b^cre, n = 10; CD11b^cre;Bmal1^lox/lox, n = 16) mice. B Primary escape latency during the 5-day trial period for young and aged mice. C Escape latency in the testing phase of the Barnes maze task in young (3–6 months, CD11b^cre, n = 6; CD11b^cre;Bmal1^lox/lox, n = 6) and aged (18–20 months, CD11b^cre, n = 9; CD11b^cre;Bmal1^lox/lox, n = 10) mice. D Distance travelled in whole arena in the open field task in young (3–6 months, CD11b^cre, n = 21; CD11b^cre;Bmal1^lox/lox, n = 17) and aged (18–20 months, CD11b^cre, n = 12; CD11b^cre;Bmal1^lox/lox, n = 19) mice. E Time spent in the center area in the open field task in young (3–6 months, CD11b^cre, n = 21; CD11b^cre;Bmal1^lox/lox, n = 17) and aged (18–20 months, CD11b^cre, n = 12; CD11b^cre;Bmal1^lox/lox, n = 19) mice. Data are represented as the mean ± SEM. P-values were calculated using paired t-test, or two-tailed Student’s t-test.

Fig. 2

Deficits in long-term potentiation in the CA1 hippocampal region in aged Bmal1 cKO mice. A Input/output (I/O) curves as a measure of basal synaptic transmission in the CA1 region of the hippocampus (10 slices, 5 mice per aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice). B Paired-pulse ratio was recorded from CA1 pyramidal neurons with inter-stimulus intervals: 10, 20, 50, 100, 200, or 500 ms from aged CD11b^cre (18–20 months; n = 8) and CD11b^cre;Bmal1^lox/lox (n = 10) mice. C Long-term potentiation (LTP) in the CA1 hippocampal region over a 90-min recording interval (n = 10 slices, 5 mice per aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice). Arrow shows that induction of LTP is significantly reduced in CD11b^cre (228.669 ± 7.285) and CD11b^cre;Bmal1^lox/lox (176.359 ± 7.593; p < 0.0001). D Representative immunoblot of phospho-CAMKII and CAMKII in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months; n = 5/group). E Quantification of phospho-CAMKII and CAMKII immunoblot in D normalized to ß-actin. F Representative confocal images of GluA1 (magenta) expression in the hippocampal CA1 region of aged CD11b^cre (top, n = 6) and CD11b^cre;Bmal1^lox/lox (bottom, n = 5) mice. Scale bar, 50 µm. G Quantification of mean fluorescence intensity of GluA1 in the hippocampal CA1 region of aged CD11b^cre (n = 6) and CD11b^cre;Bmal1^lox/lox (n = 5) mice. H Diagram highlighting differences in synaptic release and numbers of receptors between genotypes in aged CA1 hippocampus. Data are represented as the mean ± SEM. P-values were calculated using two-way RM ANOVA or two-way ANOVA: effects of time and genotype and Sidak’s multiple comparisons test with Geisser–Greenhouse correction. Theta burst stimulation (TBS)

Fig. 3

Increased numbers of immature spines in aged Bmal cKO CA1 hippocampus. A Representative images of PSD95 (green) and SNAP25 (red) expression in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months). Scale bar, 50 µm. White dotted lines in A indicate region of interest quantified in B. B Mean fluorescence intensity of PSD95 and SNAP25 in aged CD11b^cre (n = 5) and CD11b^cre;Bmal1^lox/lox mice (n = 6). C Representative immunoblot of PSD95 and SNAP25 in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months; n = 3/group). D Quantification of PSD95 and SNAP25 normalized to ß-actin. E Representative images of Golgi stain of apical dendritic spines in the CA1 region of the hippocampus aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months; n = 3/group). Scale bar, 5 µm. F Dendritic spine density (10–12 neurons/CA1 region) in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox (n = 3/group) mice. G Dendritic spine density of filopodia-like, stubby and mushroom spines in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox (n = 3/group) mice. Data are represented as the mean ± SEM. P-values were calculated using two-tailed Student’s t-test

Fig. 4

Microglial BMAL1 deficiency decreases C1q and increases C3 in aged CA1 hippocampus. A Representative images of C1q (green), PSD95 (magenta) and CD68 (red) expression in the hippocampal CA1 region of aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months). Scale bar, 10 µm. White arrow shows colocalization of C1q, PSD95 and CD68 present in aged CD11b^cre that is absent in CD11b^cre;Bmal1^lox/lox mice. B Mean fluorescence intensity (MFI) of C1q, PSD95 and CD68 (n = 6/group). C Schematic illustrating C1q activation of C3. D Representative confocal images of C1q (green), PSD95 (magenta) and C3 (red) expression in the CA1 hippocampal area in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice. Scale bar, 20 µm. E MFI of C3 in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (n = 6/group). Data are represented as the mean ± SEM. P-values were calculated using two-tailed Student’s t-test

Fig. 5

Microglial BMAL1 deficiency decreases synaptic engulfment in aged mice. A Representative images of IBA1 + /CD68 + cell engulfment of PSD95 + particles in microglia in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice. Scale bar, 2 µm. B Orthogonal views and rotated 3D projection of PSD95 + particles in IBA + /CD68 + microglia from aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice. C Pearson’s coefficient of colocalization of PSD95 and CD68 in IBA1 + cells (n = 6/group) mice. Data are represented as the mean ± SEM. P-values were calculated using two-tailed Student’s t-test

Fig. 6

Myeloid Bmal1 deletion increases microglial activation but decreases lysosomal function in aged mice. A Representative confocal images of IBA1 (green) and CD68 (red) expression in the hippocampal CA1 region of aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months, n = 6/group). Scale bar, 50 µm. White dotted lines in A indicate region of interest quantified in C. B Higher magnification of microglia in CA1 hippocampal region from aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice.Scale bar, 5 µm. C Mean fluorescence intensity of IBA1, CD68, proportion of IBA1-positive ( +) cells that are CD68 + , and the number of DAPI + IBA1 + microglia (n = 6/group). D Representative immunoblot of IBA1 in aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice (18–20 months; n = 6/group). E Quantification of IBA1 immunoblot normalized to ß-actin. F Representative images of skeletonized microglia overlaid on original image from aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice. G Quantification of microglial complexity. Scale bar, 5 µm. Every measurement that contained ≤ 2 endpoints with a maximum branch length of less than the cutoff value of 0.5 µm was removed from the analysis. Data are represented as the mean ± SEM. P-values were calculated using two-tailed Student’s t-test

Fig. 7

Microglial gene expression changes in young and aged WT and Bmal1 cKO mice. A Principal component analysis (PCA) shows distinct clustering of young CD11b^cre and CD11b^cre;Bmal1^lox/lox (n = 6/group) mice versus aged CD11b^cre (n = 6) and CD11b^cre;Bmal1^lox/lox (n = 7) mice. B Venn diagram of number of differentially expressed genes in young and aged CD11b^cre;Bmal1^lox/lox mice. C Volcano plot of -log10 p-adjusted value versus −log2 fold change of normalized counts between aged CD11b^cre and CD11b^cre;Bmal1^lox/lox mice. Each dot represents a single transcript. Cyan dots denote significantly differentially expressed genes. Orange dots denote genes that are unchanged between the genotypes (adjusted p-value < 0.05)

Fig. 8

Sleep–wake behavior is disrupted in aged Bmal1 cKO mice. Baseline sleep–wake behavior was assessed over continuous 24-h recordings of aged CD11b^cre (18–20 months; n = 6) and CD11b^cre;Bmal1^lox/lox (n = 7) mice under normal light–dark (12-h light:12-h dark) conditions. A–C Percentage of time spent in wake (A), NREM sleep (B), and REM sleep (C) across the entire 12-h light/dark period. D–F The average duration of individual wake (D), NREM (E), and REM (F) bouts across the entire 12-h light/dark phase. G–J The total number of wake (G), NREM (H), and REM (I) bouts across the 12-h light/dark phases. J Transitions between sleep–wake states across the 12-h light/dark phases. K The average REM bout latency during the 12-h light/dark periods. Data are represented as the mean ± SEM. P-values were calculated using two-tailed Student’s t-test or two-way ANOVA. ZT zeitgeber time

Fig. 9

Microglial BMAL1 deficiency alters recovery from sleep deprivation. A–C Percent of time per hour spent in wake (A) NREM sleep (B) and REM sleep (C) during a 4-h sleep deprivation (pink) and subsequent recovery during the remaining light phase and full 12-h dark phase in aged CD11b^cre (18–20 months; n = 6) and CD11b^cre;Bmal1^lox/lox (n = 7) mice. D–F Comparisons of the percentage of time spent in wake (D) NREM sleep (E) and REM sleep (F) across the entire 4-h light phase recovery period (left) and 12-h dark phase recovery period (right) during baseline (BL) and sleep deprivation (SD) recordings. G–I Cumulative amount (total hours) of wake (G), NREM (H) and REM (I) across the entire recovery period during both BL and SD recordings. Data are represented as the mean ± SEM. Significant between group differences in A, B are expressed as P-values calculated using two-way ANOVA. All significant within group differences between baseline (BL) and sleep deprivation (SD) are expressed as adjusted P-values generated from Sidak’s post hoc analysis following two-way ANOVA. ZT zeitgeber time