Shift work schedules alter immune cell regulation and accelerate cognitive impairment during aging

Abstract

Background: Disturbances of the sleep-wake cycle and other circadian rhythms typically precede the age-related deficits in learning and memory, suggesting that these alterations in circadian timekeeping may contribute to the progressive cognitive decline during aging. The present study examined the role of immune cell activation and inflammation in the link between circadian rhythm dysregulation and cognitive impairment in aging.

Methods: C57Bl/6J mice were exposed to shifted light-dark (LD) cycles (12 h advance/5d) during early adulthood (from ≈ 4-6mo) or continuously to a "fixed" LD12:12 schedule. At middle age (13-14mo), the long-term effects of circadian rhythm dysregulation on cognitive performance, immune cell regulation and hippocampal microglia were analyzed using behavioral, flow cytometry and immunohistochemical assays.

Results: Entrainment of the activity rhythm was stable in all mice on a fixed LD 12:12 cycle but was fully compromised during exposure to shifted LD cycles. Even during "post-treatment" exposure to standard LD 12:12 conditions, re-entrainment in shifted LD mice was marked by altered patterns of entrainment and increased day-to-day variability in activity onset times that persisted into middle-age. These alterations in light-dark entrainment were closely associated with dramatic impairment in the Barnes maze test for the entire group of shifted LD mice at middle age, well before cognitive decline was first observed in aged (18-22mo) animals maintained on fixed LD cycles. In conjunction with the effects of circadian dysregulation on cognition, shifted LD mice at middle age were distinguished by significant expansion of splenic B cells and B cell subtypes expressing the activation marker CD69 or inflammatory marker MHC Class II Invariant peptide (CLIP), differential increases in CLIP+, 41BB-Ligand+, and CD74 + B cells in the meningeal lymphatics, alterations in splenic T cell subtypes, and increased number and altered functional state of microglia in the dentate gyrus. In shifted LD mice, the expansion in splenic B cells was negatively correlated with cognitive performance; when B cell numbers were higher, performance was worse in the Barnes maze. These results indicate that disordered circadian timekeeping associated with early exposure to shift work-like schedules alone accelerates cognitive decline during aging in conjunction with altered regulation of immune cells and microglia in the brain.

Keywords: Activity rhythm; Adaptive immune cell; Aging; B cells; Barnes maze; Circadian rhythm dysregulation; Cognition; Mice; Microglia; T cells.

Fig. 1

LD treatment groups and experimental design. Mice were separated into two cohorts: (1) control group (left) was maintained throughout on a fixed LD 12:12 schedule and (2) experimental treatment group (right) was exposed to shifted LD cycles (12 h advance/5d) for 80 days and then was placed back on the same standard LD 12:12 cycle. At middle age (13mo), both groups were assessed in the Barnes maze and euthanized at 14 months for flow cytometry analysis. (Created with BioRender.com)

Fig. 2

Effects of experimental LD cycles and aging on light-dark entrainment and other properties of the circadian rhythm in wheel-running activity. (A) Representative records of wheel-running activity in adult mice (≈ 3mo) that were maintained in a fixed LD 12:12 cycle (left) or exposed to a shifted (12 h/5d) LD 12:12 cycle (right). Actograms are plotted over a 24-hour period. The open and closed bars at the top respectively signify the timing of the light and dark phase in the fixed and shifted LD 12:12 cycles. Red arrows on the right denote the interval when exposure to the shifted LD cycles was initiated (“treatment” phase) and when shifted LD animals were returned to the same regular LD 12:12 schedule as the fixed LD group (post-treatment phase). (B) The phase angle (Ψ) between daily activity onsets and lights-off, (C) absolute day-to-day variability, and (D) total daily wheel-running activity (wheel revolutions/24hr) were later analyzed during the post-treatment phase in fixed (n = 7) and shifted (n = 7) LD mice at middle age (13-14mo). Then these entrainment and qualitative parameters of the activity rhythm in middle-aged mice from both treatment groups were compared to similar published data obtained from aged mice on fixed LD cycles (Souza et al., [14]). In panel B, negative phase angle values (in minutes) indicate that daily onsets of activity occur after lights-off. Bars (in B-D) depict mean values (+ SEM). Circles indicate individual data values for each mouse. (*p < 0.01; Tukey’s multiple comparisons)

Fig. 3

Effects of experimental LD cycles on cognitive performance in the Barnes maze. (A) Distance traveled (cm) to reach the escape is depicted across days of learning trials, (B) cognitive index for training trials, and (C) percent path in target quadrant (i.e., quadrant in which the escape was localized during training trials, during the first 30 s of the trial) were analyzed in fixed (n = 20) and shifted (n = 16) LD mice at middle age (13-14mo) and compared to similar data from aged mice on fixed LD cycles (Souza et al., [14]). Plotted values (in A-C) represent mean ± SEM. Circles (in B-C) indicate individual data values for each mouse. (*p < 0.05, **p < 0.01; ***p < 0.0001; Fisher’s PLSD post hoc analysis)

Fig. 4

Effects of experimental LD cycles and aging on splenic B cell populations. (A) Representative dot plots (top) using CD19 (x-axis) as a B cell marker versus CD90.2 as a T cell marker to compare populations of B cells in isolated splenocyte samples from middle-aged (MA) mice on fixed (left) and shifted (12 h/5d, center) LD 12:12 cycles, and from an aged (18-22mo) animal (right) maintained on a fixed LD cycle. Bar graphs (bottom) depict group comparisons (MA fixed, n = 12; MA shifted, n = 13; Aged fixed, n = 13) of the percentage (mean ± SEM) of B cells identified from Quadrant 3 (left), and of CD19+ B cell populations expressing the activation marker CD69 (center) or cell surface CLIP (right). (B) Representative dot plots (top) of live B cells gated from quadrant 3 (see panel A) displaying MHCII (y-axis) and CD74 (x-axis) expression. Bar graph (bottom) depicts group comparisons (MA fixed, n = 9; MA shifted, n = 10; Aged fixed, n = 10) of the percentage (mean ± SEM) of CD74+ B cells identified from quadrant 2 and 3 (left). (*p < 0.05, **p < 0.01; Tukey’s multiple comparisons)

Fig. 5

Effects of experimental LD cycles and aging on splenic T cell populations (A) Flow cytometry analysis using CD90.2 (y-axis from Fig. 4A, quadrant 1) as a T cell marker to compare populations of T cells in isolated splenocyte samples from middle-aged (MA) mice on fixed (left) and shifted (12 h/5d, center) LD 12:12 cycles, and from an aged (18-22mo) cohort (right) of fixed LD mice. Representative dot plots (top) depict gated CD4+ T cells exhibiting expression of CD44 (y-axis) and CD62L (x-axis). Bar graphs (bottom) depict group comparisons (MA fixed, n = 10; MA shifted, n = 11–13; Aged fixed, n = 13–14) of the percentage (mean ± SEM) of T cells (left) identified in quadrant 1 of Fig. 4A and of CD4+ naïve T cells (right) that were gated from quadrant 3 as CD62L+/CD44-. (B) Representative dot plots (top) of CD4+ T cells displaying expression of FoxP3 (y-axis) and CD25 (x-axis). FoxP3 + and CD25+ (high) expression was used to gate regulatory T cells, as shown in the outlined box. Bar graph (bottom) depicts group comparisons (MA fixed, n = 8; MA shifted, n = 10; Aged fixed, n = 3) of FoxP3+ and CD25+ regulatory T cells. (*p < 0.05, **p < 0.01; Tukey’s multiple comparisons)

Fig. 6

Effects of experimental LD cycles on meningeal B cell populations. (A) Representative dot plots using CD19 (x-axis) as a B cell marker versus CD90.2 as a T cell marker to compare populations of B cells in isolated meningeal samples from mice on fixed (left) and shifted (12 h/5d, right) LD 12:12 cycles. Bar graph depicts group comparisons of the percentage (mean ± SEM) of CD19+ B cells (right). (B) Bar graphs depict group comparisons (MA fixed, n = 9–10; MA shifted, n = 16) of the percentage (mean ± SEM) of CD19+ B cell subsets expressing 41BBL (left), CLIP (center), or CD74 (right) as indicated. (*p < 0.05, **p < 0.01; Mann-Whitney test)

Fig. 7

Relationship between splenic B cell populations and cognitive performance in Barnes maze for middle-aged (MA) mice that were exposed to fixed or shifted LD cycles. Pearson correlation coefficients comparing cognitive index with the percentage of: (A) B cells, (B) CLIP + B cells, (C) CD74 + cells and (D) CD69 + cells in the middle-aged cohorts of fixed (n = 5) and shifted (n = 6–8) LD mice. Circles represent individual data values for each mouse. Lines in each graph denote simple linear regression for the data set with corresponding p values

Fig. 8

Effects of experimental LD cycles on the morphology of hippocampal microglia. (A) High-magnification (60X) confocal images of IBA1+ hippocampal microglia located within the dentate gyrus (DG) of representative fixed (left) and shifted (right) LD mice at middle age. Bar graphs depict quantitative analysis of the morphological profiles of IBA-1+ microglia in the DG (mean ± SEM) from regions of interest (ROI) in fixed (n = 5) and shifted (n = 5) LD mice with regard to: (B) integrated fluorescent intensity, (C) number of immunopositive cells, (D) soma area, and (E) the number of primary immunopositive processes with Sholl intersections (at 5 μm intervals of the distance from the microglial soma). For all morphological analyses, three representative hemibrain sections through the hippocampus were analyzed from each animal and three randomized FOV within the dentate gyrus (DG) of the hippocampus were captured in each section. In turn, the data points represent individual IBA-1+ microglia (approximately 10–20/FOV) in the DG that were selected as ROI. (*p < 0.05, **p < 0.01; Mann-Whitney test)