Immunization with a heat-killed bacterium, Mycobacterium vaccae NCTC 11659, prevents the development of cortical hyperarousal and a PTSD-like sleep phenotype after sleep disruption and acute stress in mice

Abstract

Study objectives: Sleep deprivation induces systemic inflammation that may contribute to stress vulnerability and other pathologies. We tested the hypothesis that immunization with heat-killed Mycobacterium vaccae NCTC 11659 (MV), an environmental bacterium with immunoregulatory and anti-inflammatory properties, prevents the negative impacts of 5 days of sleep disruption on stress-induced changes in sleep, behavior, and physiology in mice.

Methods: In a 2 × 2 × 2 experimental design, male C57BL/6N mice were given injections of either MV or vehicle on days -17, -10, and -3. On days 1-5, mice were exposed to intermittent sleep disruption, whereby sleep was disrupted for 20 h per day. Immediately following sleep disruption, mice were exposed to 1-h social defeat stress or novel cage (control) conditions. Object location memory (OLM) testing was conducted 24 h after social defeat, and tissues were collected 6 days later to measure inflammatory markers. Sleep was recorded using electroencephalography (EEG) and electromyography (EMG) throughout the experiment.

Results: In vehicle-treated mice, only the combination of sleep disruption followed by social defeat (double hit): (1) increased brief arousals and NREM beta (15-30 Hz) EEG power in sleep immediately post-social defeat compared to baseline; (2) induced an increase in the proportion of rapid-eye-movement (REM) sleep and number of state shifts for at least 5 days post-social defeat; and (3) induced hyperlocomotion and lack of habituation in the OLM task. Immunization with MV prevented most of these sleep and behavioral changes.

Conclusions: Immunization with MV ameliorates a stress-induced sleep and behavioral phenotype that shares features with human posttraumatic stress disorder.

Keywords: M. vaccae; NREM; PTSD; REM; cortical hyperarousal; multimodal stress; mycobacteria; repeated sleep disruption; social defeat; stress resilience.

Figure 1.

Experimental timeline and effect of sleep disruption on sleep in vehicle-treated and M. vaccae-treated mice. (A) Overview of the experimental protocol. Experimental protocol days are depicted along the axis of the timeline, with tick marks indicating ZT0 of that experimental day. (B) Detail of indicated time period that includes the sleep disruption protocol, social defeat stress (red arrowhead), object location memory (green arrowhead), and terminal sample collection (black arrowhead). Tick marks along the bottom indicate zeitgeber time on that day, white/black rectangles indicate 12 h light/dark phases, and yellow rectangles indicated time periods of sleep disruption for the sleep-disrupted groups (ZT6–ZT2). EEG/EMG recording intervals, as used in the forthcoming figures, are indicated above the timeline in quotation marks. (C) NREM, (D) REM, (E) REM:Sleep ratio, and (F) state shifts per 24 h (ZT6–ZT6) during the sleep disruption protocol. (G) NREM, (H) REM, and (I) REM:sleep ratio during the 4 h (ZT2–ZT6) ad libitum sleep opportunities. Data are mean ± SEM. Symbols: @ p < 1.0 × 10^–4 (overall effect of “sleep disruption”), linear mixed effects model; *p < 0.05, **p < 0.01, ***p < 0.001 (Tukey’s post hoc test). For post hoc tests, red asterisks indicate comparisons between the vehicle-injected control group and the vehicle-injected, sleep-disrupted group at that timepoint. Blue asterisks indicate comparisons between the MV-injected control group and the MV-injected, sleep-disrupted group at that timepoint. n = 29–30/group. Abbreviations: BL, baseline; Dis, sleep disruption; MV, Mycobacterium vaccae NCTC 11659 injection; NREM, nonrapid eye movement sleep; OLM, object location memory; R, recovery; REM, rapid eye movement sleep; VEH, vehicle injection; ZT, zeitgeber.

Figure 2.

The double hit causes a maladaptive sleep phenotype acutely after social defeat, which is ameliorated by M. vaccae immunization. EEG/EMG sleep recordings were started after social defeat or control manipulation (~ZT7) and continued until object location memory (OLM) testing began the next morning (ZT3), for a ~20 h window labeled “day 6.” (A) NREM, (B) REM, (C) REM:sleep ratio during day 6. (D) Brief arousals during the dark cycle of day 6. (E, F) NREM EEG spectral power and (G, H) REM EEG spectral power during day 6, expressed as a percent of the power during the identical ZT time window during baseline recording. Data are mean ± SEM. Symbols in panels (A–D): ⁺p < 0.05 (overall effect of “Treatment”); ^@p < 0.05 (overall effect of “sleep disruption”); ^%p < 0.05 (overall effect of “social defeat”); ^&p < 0.05 (“treatment” × “sleep disruption” × “social defeat” interaction), linear mixed effects model; *p < 0.05, **p < 0.01, ***p < 0.001 (Tukey’s post hoc test). Symbols in panels (E–H): ^p < 0.05 versus 100% (one-sample Wilcoxon Rank-Sum test), *p < 0.05 versus control (Tukey’s post hoc test). n = 10–14/group. Abbreviations: BL, baseline; EEG, electroencephalography; MV, Mycobacterium vaccae NCTC 11659 injection; NREM, nonrapid eye movement sleep; REM, rapid eye movement sleep; VEH, vehicle injection.

Figure 3.

The double hit causes lasting changes in sleep that are prevented by M. vaccae immunization. Ad libitum recovery sleep was recorded from day 7 (after object location memory testing) until the end of the experiment (day 12). 24-h bins of (A) NREM sleep, (B) REM sleep, (C) REM:sleep ratio, (D) average number of epochs of NREM preceding a REM bout (i.e. REM transition latency), (E) state shifts, and (F) NREM EEG beta (15–30 Hz) power is reported for control, sleep disruption only, and sleep disruption plus social defeat (double hit) groups. Beta power is reported as a percent of 24-h baseline. Data are mean ± SEM. Symbols: ⁺p < 0.05 (overall effect of sleep disruption or double hit), ^#p < 0.05 (overall effect of time), ^&p < 0.05 (group × time interaction), linear mixed effects model; *p < 0.05, **p < 0.01, ***p < 0.001 (Tukey’s post hoc testing between groups at individual timepoints). n = 10–14/group. Abbreviations: BL, baseline; MV, Mycobacterium vaccae NCTC 11659 injection; NREM, nonrapid eye movement sleep; REM, rapid eye movement sleep; VEH, vehicle injection.

Figure 4.

The double hit causes a behavioral phenotype that is blunted by M. vaccae immunization. Approximately 24 h after acute social defeat (or control manipulation consisting of being moved to a clean cage in a quiet room for 1 h), mice were exposed to the object location memory (OLM) task, which consisted of a 5-min training session and a 5-min testing session 90 min later. (A) Total locomotion during the training session and (B) total locomotion during the testing session. (C) Habituation to the testing arena over the two sessions as measured by the percent change in locomotion from the training to the testing session. (D) Percent of the 5-min training session spent in the center of the arena. (E) Object location memory, measured as percent of exploration time devoted to exploring the displaced object during the first 3 min of the testing session (Location Index). Significant deviations from 50% indicate learning. Data are mean ± SEM. Symbols: ⁺p < 0.05 (overall effect of “treatment”); ^%p < 0.05 (overall effect of “social defeat”); ^ap < 0.05 (“treatment” × “sleep disruption” interaction); ^bp < 0.05 (“treatment” × “social defeat” interaction); ^cp < 0.05 (“social defeat” × “sleep disruption” interaction), linear mixed effects model; *p < 0.05, ** p < 0.01 (Dunnett’s post hoc test). n = 13–15/group. Abbreviations: BL, baseline; MV, Mycobacterium vaccae NCTC 11659 injection; NREM, nonrapid eye movement sleep; REM, rapid eye movement sleep; VEH, vehicle injection.

Figure 5.

Impact of the double hit and M. vaccae on serum cytokines, mesenteric lymph node cytokines, and spleen weight. At the end of the experiment, serum and other tissues were collected. Serum levels of (A) IL-6, (B) MCP-1 (i.e. CCL2), and (C) CINC-1 (i.e. CXCL-1) were assessed using a magnetic bead multiplex. (D) Spleens were weighed and normalized to body weight. Mesenteric lymph nodes were homogenized and concentrations of (E) IL-17A, (F) IL-1beta (G) IL-6, (H) IFNγ, (I) IL-4, and (J) IL-10 were assessed using a magnetic bead multiplex and normalized to total protein concentration. Data are mean ± SEM. Symbols: ^@p < 0.05 (overall effect of “sleep disruption”); ^%p < 0.05 (overall effect of “social defeat”); ^ap < 0.05 (“treatment” × “sleep disruption” interaction), linear mixed effects model; *p < 0.05, ** p < 0.01 (Dunnett’s post hoc test); ^#p < 0.01 (Wilcoxon Rank-Sum test). n = 11–15/group. Abbreviations: IL, interleukin; CINC-1, cytokine-induced neutrophil chemoattractant-1; MCP-1, monocyte chemoattractant protein-1; MLN, mesenteric lymph node; MV, Mycobacterium vaccae NCTC 11659 injection; VEH, vehicle injection.

Figure 6.

Associations among immediate sleep changes, lasting sleep changes, behavior, and physiological measures in vehicle-treated and M. vaccae-treated mice. Pairwise Spearman’s rank order correlations of 33 measures of interest from throughout the experiment were computed for (A) vehicle-treated and (E) M. vaccae (MV)-treated mice that received either control manipulations or sleep disruption plus social defeat (double hit). The value of the Spearman’s rho for each comparison is represented by the size and shade of circle in each square of the correlation plot, with darker blue meaning more positive rho and darker red meaning more negative rho. Asterisks indicate uncorrected p values that were less than 0.05, while black boxes indicate q values that were less than 0.05. REM:sleep ratio on day 11 versus NREM EEG beta (15–30 Hz) power on day 6, spleen weight per gram of body weight at the end of the experiment versus brief arousals during the dark phase of day 6, and time spent in the center of the arena during the OLM training session versus REM:sleep on day 6 are presented in (B–D) vehicle-treated and (F–H) MV-treated mice. n = 9–12/group. Abbreviations: EEG, electroencephalography; IFNg, interferon gamma; IL, interleukin; CINC-1, cytokine-induced neutrophil chemoattractant-1; MCP-1, monocyte chemoattractant protein-1; MLN, mesenteric lymph node; MV, Mycobacterium vaccae NCTC 11659 injection; NREM, nonrapid eye movement sleep; REM, rapid eye movement sleep; VEH, vehicle injection.