Adult male and female C57Bl/6 mice were purchased from the Jackson Laboratory (JAX stock #0664) and maintained in the AAALAC-accredited vivarium at the Texas A&M University Health Science Center. All animals were maintained in vivarium rooms under controlled temperature (22–25 °C) and lighting (LD 12:12) conditions with food (standard mouse chow) and water available ad libitum. All animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Animal procedures used in this study were conducted in compliance with Animal Use Protocol 2022 − 0211 as reviewed and approved by the Institutional Animal Care and Use Committee at Texas A&M University.

To analyze the effects of circadian dysregulation, experiments used a chronic light-dark (LD) cycle shift paradigm that has been shown to be effective in desynchronizing circadian rhythms and in inducing pro-inflammatory responses of immune cells, leading to a persistent inflammatory condition [17, 18, 23]. After baseline acclimation under standard LD 12:12 conditions (lights-on at 0800 h; light intensity = 110–170 lx at 500–580 nm) for about 2 weeks, young C57Bl/6 mice (≈ 3mo) were randomly divided into 2 groups and exposed for 80 days to either the same “fixed” LD 12:12 cycle (n = 20) or to a “shifted” LD 12:12 cycle (n = 16). During exposure to the shifted LD paradigm (Fig. 1A), lights-on was advanced by 12 h (at 2000 h) every 5 days and these shifts in the LD cycle were repeated for 8 full cycles. At the conclusion of experimental LD cycle manipulations (“treatment period”), animals (~ 7mo) in both groups were exposed to the same standard LD 12:12 schedule (lights-on at 0800 h) for ~ 7 additional months (“post-treatment period”). Then following completion of behavioral assays, all animals were anesthetized early in the 12-hour photoperiod (0900–1200 h) with isoflurane and tissues were collected for flow cytometry and immunohistochemistry analyses.

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)

To confirm the long-term effects of circadian dysregulation on the rhythm of wheel-running behavior, a separate cohort of mice were housed individually in cages equipped with running wheels and divided at ~ 3mo of age into two treatment groups exposed to fixed (n = 7) or shifted (n = 7) LD cycles as described previously. Wheel-running activity was continuously recorded, stored in 10-minute bins, graphically depicted in actograms, and analyzed using ClockLab data collection and analysis software (ActiMetrics, Evanston, IL). Entrainment and qualitative parameters of the activity rhythm in mice exposed to fixed or shifted LD cycles were measured at middle age over the same interval (40 days) following exposure to experimental LD cycles when both groups were exposed to the same standard LD 12:12 cycle. During LD entrainment, the onset of activity for a given cycle was identified as the first bin during which an animal attained 10% of peak running-wheel revolutions (i.e., intensity). To measure phase angle of entrainment (Ψ), least squares analyses was used to establish a regression line through the daily onsets of activity during the period of entrainment (40 days), and then the number of minutes before (positive) or after (negative) the time of lights-off in the LD cycle (2000 h) was determined for each animal. Total daily activity was calculated by averaging the number of wheel revolutions per 24 h over the interval of analysis. Group differences were established by separately analyzing these entrainment and qualitative parameters of the activity rhythm during the LD “treatment” (fixed, shifted) phase, and the subsequent “post-treatment” phase of the study when both groups were exposed to the same fixed LD cycle.

To assess the effects of environment-induced circadian dysregulation on cognition in relation to normal aging, Barnes maze data from middle-aged (13mo) fixed (n = 20, 9 males and 11 females) and shifted (n = 16, 6 males and 10 females) LD mice in this study was compared with our published observations from aged (18-22mo; n = 29, 14 males and 15 females) C57Bl/6 mice that were constantly maintained on a standard LD 12:12 cycle (lights-on at 0800 h) [14]. The Barnes circular platform maze is a 91.44 cm diameter circular platform on a 1.4 m stand with 20 evenly spaced 5.08 cm diameter holes around the circumference, where a black box (escape tunnel) was placed underneath one of the holes (San Diego Instruments, CA). Four bright lights were positioned above the maze as an aversive stimulus to cause the mice to seek out the escape box using spatial cues. Between each trial, the table and escape box were cleaned with 70% ethanol, and video was acquired using a Color GigE camera (model: acA1300-30gc). Data were quantified using Ethovision XT 16 video tracking software (Noldus, Leesburg, VA).

Barnes maze testing was performed during the early portion of the 12-hour photoperiod (0900–1300 h) as described previously [14]. Our protocol consists of habituation, acquisition training (learning) and probe testing (memory).

Habituation. During habituation, mouse was placed on the table for 5 min and allowed to explore the maze without an escape box, under dim lights. Next, mice were placed in a 2 L transparent glass beaker, under aversive lighting. After 1 min, the mouse was gently guided to the escape box and lights are turned off.

Learning training trials. On all subsequent trials, the animals were placed into the center of the table under a dark container for 30 s before the container was lifted and the mouse was allowed to navigate the maze using spatial cues under aversive lighting. Each mouse was allowed 180 s to locate and enter the escape box per trial. If the mouse was unable to locate the escape hole after 180 s, it was gently guided to the correct hole location and allowed to enter the escape box. Once the mouse entered the escape box (either guided or on their own), it remained in the box for one minute before returning to its home cage. Each day, the animal was subjected to 4 trials spaced 15 min apart for a total of sixteen learning training trials, for 4 days.

Probe Trial. Seventy-two hours after learning trials, the mouse is given a single probe trial, in which escape box is not available. The animal’s search behavior is analyzed for 180 s, after which the mouse is removed and placed back into its holding cage.

Search Strategies and Cognitive Index. Detailed description of search strategies is provided in Souza et al. [14]. In brief, the hippocampal-dependent strategies are: direct (no error; score = 1), corrected (searched + or – 1 immediate hole, score = 0.75), focused (searched + or – 3 immediate holes, score = 0.5) and long correction (mouse searches across the target and immediately corrects toward correct hole, score = 0.5). Non-hippocampal strategies include the serial search (animal methodically searches holes one by one, score = 0.25), random (search without a clear strategy and target hole identified by chance, score = 0), and failure (animal searches but does not find the target, score = 0). Scores were summed to produce the Cognitive Index.

To compare the effects of circadian dysregulation and normal aging on immune cell activation, immune profiling using flow cytometry was performed on single cell suspensions [24] of spleen-derived lymphocytes from shifted LD mice at middle age (~ 14mo; n = 14, 6 males and 8 females) following exposure to a standard LD 12:12 cycle, and from groups of middle-aged (~ 14mo; n = 12, 7 males and 5 females) and aged (18-22mo; n = 13, 8 males and 5 females) mice that were continuously maintained on fixed LD cycles. Parallel analyses were performed on lymphocytes extracted from the cranial meninges of fixed (n = 12, 7 males and 5 females) and shifted (n = 16, 8 males and 8 females) LD mice at middle age (~ 14mo). Isolated lymphocytes from the spleen and cranial meninges were treated with Ammonium-Chloride-Potassium (ACK) Lysis Buffer to lyse red blood cells, resuspended in phosphate-buffered saline (PBS) with 3% heat-inactivated Fetal Bovine Serum (FBS), and then incubated with FC block to prevent non-specific cell staining. Using our previously established procedures for flow cytometry [25], cells were stained with the following fluorochrome-conjugated antibodies: Ghost Dye Red 780 Viability stain, FITC CLIP (15G4; Santa Cruz Biotechnology, Inc., Dallas, TX), BV421 CD19, BV510 CD90.2, Alexa Fluor 700 CD4, BV 785 CD69, PE 41BBL, PerCP MHCII(IA-IE), Alexa Fluor 647 CD74, PE CD44, BV 421 CD62L, APC CD25 and PE FoxP3 (Biolegend, San Diego, CA). Staining for subsets of T cells, B cells, MHCII + or CLIP + lymphocyte subsets and monocytes/macrophages was performed using the fluorochrome-conjugated antibodies. Cellular analysis data was collected using the Beckton Dickson LSR Fortessa flow cytometer (Franklin Lakes, NJ) and analyzed using FlowJo software (Tree Star Inc., Ashland, OR).

Microglial activation in the hippocampus was examined using immunohistochemical localization of the microglial marker ionized calcium-binding adapter molecule 1 (IBA-1). Anesthetized fixed and shifted LD mice (n = 10; fixed LD: 2 females, 3 males; shifted LD: 2 females, 3 males) were immediately perfused transcardially with 50 ml of 0.1 M phosphate buffer (pH = 7.2). After perfusion, brains were removed and divided into hemibrains that were post-fixed in 4% paraformaldehyde for 48 h at 4oC and stored in 20% sucrose solution at 40C until sectioning. Hemibrains were then frozen and sectioned serially through the entire hippocampus on a freezing microtome (coronal plane, 30 μm). Sections were stored in cryoprotectant solution (25% glycerin, 25% ethylene glycol, 50% 0.1 M phosphate buffer, pH 7.4) until subsequent immunohistochemical processing. With interceding rinses in Tris-buffered saline (TBS; 100 mM Tris-HCl, 150 mM NaCl, pH 7.5), free-floating sections were sequentially incubated in: blocking solution containing 10% bovine serum albumin (BSA) and 5% NGS normal goat serum (NGS) in TBS for 1 h, rabbit anti-IBA-1 pAb (1:2000; FUJIFILM Wako Pure Chemical Corp.) in TBS with 3% triton at 4oC for 48 h, and then Alexa Fluor 555-conjugated goat anti-rabbit IgG cross-absorbed secondary antibody (1:150; Invitrogen) in TBS for 2 h at room temperature in the dark. Sections were rinsed in TBS and mounted on Superfrost Plus slides (Fisher Sci.), air-dried at room temperature and coverslipped using Vectashield hardset antifade mounting medium with DAPI (Vector Labs). Slides were sealed with clear fingernail polish, and stored in the dark at 4°C.

High-quality images of microglial somas and cellular processes were acquired by confocal fluorescence microscopy (Olympus Fluoview FV3000) at maximum intensity projection (MIP). Using the 10X objective, images were captured from three representative sections for each animal (n = 10), and then three randomized fields of view (FOV) within the dentate gyrus (DG) of the hippocampus were analyzed in each section (9 DG per hemibrain). Within each FOV, individual microglia (approximately 10–20) in the DG were selected as regions of interest (ROI) according to the following criteria: (i) complete cell soma and processes and (ii) no overlap with other cells. The DG region was selected due to its involvement in memory consolidation [26, 27] and well-characterized localization of microglia [28, 29]. Next, 60X images were acquired by multi-place virtual Z-mode. On average the z-stack of each image was composed of 45–75 layers, depending on the Z-step size, which was optimized for each scan. Once the z-stack for randomized regions of interest was collected it was converted to a MIP format.

Manual morphological analyses were performed using FIJI software (Version1.54), specifically with the neuroanatomy and Sholl analysis plug-ins [30]. Before analysis, images were converted to 16-bit grayscale and binary processing was used to generate black and white images (Pixel radius: 3, mask weight: 0.6, radius: 1.0 pixel). Thresholds were manually adjusted between 15 and 20% depending on staining intensity. For each ROI, concentric circles with an increasing step size of 5 μm from the soma center were generated and then used to count the number of microglial processes at each radius length. In addition, ROIs were used to measure the soma area of each microglia and the integrated intensity of IBA-1 staining in the DG region.

Image acquisition and Sholl analysis were independently conducted by two investigators blinded to treatment groups. For each of the three FOVs in the DG of a given section, Sholl analysis was performed on ROIs encompassing approximately 10–20 cells, resulting in the morphological assessment of 90–180 IBA-1+ microglia per animal.

The significance of LD treatment differences in circadian entrainment and quantitative parameters of the activity rhythm was determined by one-way ANOVA adjusted for multiple comparisons, followed by Tukey’s post-hoc pairwise analysis. For behavioral measures, statistical analyses were performed on the raw data using repeated measures ANOVA across days and one-way ANOVAs were performed on all other comparisons. For probe trials, the percent of the path (distance) in the target quadrant is used to quantify the animal’s insistence about the escape hole’s location. Quadrants contain five possible locations for escape, and the escape hole is in the middle of the target quadrant. Only data from the first 30 s are analyzed because mice typically give up searching after approximately 30 s. Group means from the probe trial were analyzed with one-way ANOVAs. Fisher’s PLSD post hoc analysis was used for more comprehensive Barnes maze analysis. Statistical analysis was performed on all flow cytometric data to determine the significance of LD treatment- and age-related differences using a one-way ANOVA adjusted for multiple comparisons in conjunction with Tukey’s post-hoc pairwise analysis. Pearson’s correlation coefficients were determined to analyze the relationship between the proportions of different subtypes of adaptive immune cells (B and T cells) in the spleen and cognitive index scores. In each case, LD treatment- and age-related differences in circadian entrainment parameters, cognitive behavior, adaptive immune cell profiling and IBA-1 immunostaining parameters were considered significant at p < 0.05 (GraphPad, San Diego, CA).