Scheduled feeding improves behavioral outcomes and reduces inflammation in a mouse model of fragile X syndrome

Abstract

Fragile X syndrome (FXS), a leading inherited cause of intellectual disability and autism, is frequently accompanied by sleep and circadian rhythm disturbances. In this study, we comprehensively characterized these disruptions and evaluated the therapeutic potential of a circadian-based intervention in the fragile X mental retardation 1 (Fmr1) knockout (KO) mouse. The Fmr1 KO mice exhibited fragmented sleep, impaired locomotor rhythmicity, and attenuated behavioral responses to light, linked to an abnormal retinal innervation and reduction of light-evoked neuronal activation in the suprachiasmatic nucleus. Behavioral testing revealed significant deficits in social memory and increased repetitive behaviors in the mutants, which correlated with sleep fragmentation. Remarkably, a scheduled feeding paradigm (6 hr feeding/18 hr fasting) significantly enhanced circadian rhythmicity, consolidated sleep, and improved social deficits and repetitive behaviors in the Fmr1 KO mice. This intervention also normalized the elevated levels of some pro-inflammatory cytokines, including IL-12 and IFN-γ, in the mutants' blood, suggesting that its benefits extend to inflammatory pathways. These findings highlight the interplay between circadian disruption, behavior and an inflammatory response in FXS, and provide compelling evidence that time-restricted feeding may serve as a promising non-pharmacological approach for improving core symptoms in neurodevelopmental disorders.

Keywords: circadian rhythms; fragile X syndrome; inflammatory cytokines; mouse; neurodevelopmental disorders; neuroscience; scheduled feeding.

Figure 1.. The Fmr1 KO mice exhibit reduced and fragmented sleep during the light phase.

(A) Waveforms of daily rhythms in sleep behavior under standard 12:12 hr light–dark (LD) cycles in WT (blue circle) and Fmr1 KO (yellow triangle) mice. Zeitgeber Time (ZT) 0 corresponds to lights on, and ZT12 to lights off. Sleep was defined by immobility and binned in 1-hr intervals. Both genotypes exhibited clear diurnal rhythms, but the Fmr1 KO mice showed significantly less sleep during the light phase and at several times in the dark phase. Asterisks indicate significant differences between genotypes at individual time points (two-way ANOVA with genotype and time as factors, followed by the Holm–Sidak’s multiple comparisons test, *p < 0.05). The white/black bar on the top indicates the LD cycle; gray shading indicates the dark-phase time-period. (B–D) Immobility-defined sleep metrics during the light-phase. Compared to WT, Fmr1 KO mice showed a greater number of sleep bouts, indicating sleep fragmentation, and of shorter duration. Histograms show the means ± SEM with the values from individual animals overlaid. Genotypic differences were analyzed by t-test (*p < 0.05). See Table 1.

Figure 2.. The Fmr1 KO mice show unstable locomotor rhythms.

(A) Representative actograms showing daily rhythms in wheel-running activity under LD followed by constant darkness (DD) in WT (left) and Fmr1 KO (right) mice. Activity levels were normalized to 85% of the most active mouse. Each row represents two consecutive days, and the second day is repeated at the beginning of the next row. (B) Waveforms of daily rhythms in cage activity in WT (blue circle) and Fmr1 KO (yellow triangle) mice under the LD cycles. Activity under LD (1 hr bins) was analyzed by two-way ANOVA with genotype and time as factors followed by the Holm–Sidak’s multiple comparisons test (*p < 0.05). There were significant effects of both time (F = 8.84; p = 0.003) and genotype (F = 39.75; p < 0.001) on the temporal pattern of the locomotor activity rhythms. Note that significant genotypic differences were found before and after drawn. Activity metrics in LD (C) and DD (D). Rhythmic power under both LD and DD conditions was significantly reduced in the mutants, which also presented higher imprecision in the activity onset in DD. Histograms show the means ± SEM with the values from individual animals overlaid. Genotypic differences were analyzed by t-test (*p < 0.05). See Table 2.

Figure 2—figure supplement 1.. Graphic presentation of the experimental design.

(A) Mice were entrained to a 12:12 hr light–dark (LD) cycle followed by 14 days of cage activity and 3 days of immobility sleep recordings, then the animals were released in constant darkness (DD) to measure their activity rhythms driven by the endogenous clock (free-running activity) and not by external cues. Behavioral tests were conducted after the sleep/wake cycles recordings in a different group of mice. (B) A different cohort of WT and Fmr1 KO mice after entrainment to the LD cycle for 2 weeks was divided into two groups, one was held on ad libitum feeding (ALF) and one on a scheduled feeding regimen (time-restricted feeding, TRF). Mice were allowed to freely eat for 6 hr between ZT15 and ZT21. After cage activity and immobility sleep were recorded, the mice were tested for social and repetitive behaviors. Blood was collected in the early light phase.

Figure 3.. The Fmr1 KO mice display deficits in light-regulated circadian behaviors.

(A, B) Photic-suppression (masking) of activity in response to a 1 hr light (300 lx, 4500 K) pulse at ZT 14 (lights off; n = 10/genotype). The activity level during the light exposure was compared to the activity level during the equivalent hour (ZT 14–15) on the day before the treatment (baseline activity). (A) The genotypic difference in the fold change was determined by t-test, with the mutants showing a significantly reduced suppression of activity as compared to the WT (*p = 0.05). (B) Changes in the activity levels of each individual mouse during the baseline window and the light masking were analyzed using a paired t-test. Within-genotype comparisons showed significant suppression in WT (p < 0.001), but not in KO mice (p = 0.12). (C, D) Entrainment induced by a 6-hr-phase advanced LD cycle. Representative actograms of light-induced phase shifts of wheel-running activity rhythms (C). The white/black bars on the top of actograms indicate the LD cycle before (upper) and after (lower) the 6-hr phase advance. The gray shading in the waveforms indicates the dark phase time--period. The arrows next to the actograms indicate the day when the 6-hr phase advance was applied. (D) Quantification of the days to re-entrain shows that the KO mice required more time to adjust (two-way ANOVA: genotype effect F_(1,285) = 130.157, p < 0.001; followed by the Holm–Sidak’s multiple comparisons test *p < 0.001). The entrainment shifting in the WT (blue circle) and the Fmr1 KO (yellow triangle) was quantified by the difference between the activity onset and the new ZT12 on each day. The yellow and blue arrowheads in the graph indicate the day when the activity rhythms are considered well entrained. See Table 3.

Figure 4.. The Fmr1 KO mice exhibit difficulty in adapting to the skeleton photic period (SPP).

(A) Representative actograms of daily rhythms in cage activity under standard LD cycles (2 weeks) followed by the SPP (1 hr light:11 hr dark:1 hr light:11 hr dark) in WT (left) and Fmr1 KO (right) mice. The white/black bars on the top of actograms indicate the baseline LD cycle (upper) and the SPP LD cycles (lower). The gray shading in the waveforms indicates the time of the dark phases. (B) Measures of locomotor activity rhythms under SPP. Many of the parameters measured were significantly different between the genotypes, with the mutants being more impacted and showing lower rhythmic power and increased variability of the activity onset. In addition, the mutants displayed higher activity during their day. Histograms show the means ± SEM with the values from each individual animal overlaid. Significant differences (p < 0.05), determined by t-test or Mann–Whitney test, are indicated with an asterisk. (C, D) Light-induced phase delay of free-running activity rhythms in mice exposed to light (300 lx, 4500 K) for 15 min at circadian time (CT) 16. Mice were held in constant darkness. By definition, CT 12 is the beginning of the activity cycle in DD for a nocturnal organism. Examples of light-induced phase shifts of wheel-running activity rhythms (C) of WT (left) and Fmr1 KO (right) and quantified phase delay (D). In the representative actograms, the yellow lines indicate the best-fit line of the activity onset across the 10 days before and after the light pulse. The amount of phase delay is determined by the difference between the two lines on the day after the light pulse. The sunny-shape symbols indicate when the mice were exposed to light (CT16). Compared to WT, the Fmr1 KO showed reduced phase shift of their activity rhythms (Mann–Whitney U, *p = 0.011). See Table 3.

Figure 5.. Abnormal retinal-suprachiasmatic nucleus (SCN) connectivity in the Fmr1 KO mice.

To trace the projections from the retina to the SCN via the retino-hypothalamic tract (RHT), WT and Fmr1 KO mice received a bilateral intravitreal injection of Cholera Toxin (β-subunit) conjugated to Alexa Fluor555 and were perfused 72 hr later. (A) Lower intensity of the fluorescently labeled RHT projections can be observed both laterally and medially to the ventral part of the SCN in the Fmr1 KO mice (white arrows) as compared to WT, suggesting a loss of afferent projections to the SCN. (B, C) Densitometric analysis of the distribution of the Cholera Toxin fluorescence intensity in the ventral SCN (see also Figure 5—figure supplement 1) of WT and Fmr1 KO mice. The intensity peaks of the profile plot of four to five consecutive coronal sections containing the middle SCN were aligned and then averaged to obtain a single curve per animal. Results are shown as the mean ± standard deviation (SD) for the left (B) and the right (C) SCN of each genotype. (D, E) Light-induction of cFos was greatly reduced in the SCN of the Fmr1 KO mice compared to WT. Mice held in DD were exposed to light (300 lx, 4500 K) for 15 min at CT 16 and perfused 45 min later (CT 17). (D) Representative serial images of light-evoked cFos expression in the SCN. The inset in the lower left panel shows the lack of cFos immunopositive cells in the SCN of mice held in DD but not exposed to light. Dotted magenta lines delineate the SCN. OC = optic chiasm, V = third ventricle. (E) The number of immune-positive cells in the left and right SCN from 3 to 5 consecutive coronal sections per animal were averaged to obtain one number per animal and are presented as the mean ± SD per genotype. One-way ANOVA followed by Bonferroni’s multiple comparisons test, *p = 0.0201. See Table 4.

Figure 5—figure supplement 1.. Histomorphological Analysis and SCN Landmarks.

Top Panel: The distribution of the Cholera Toxin fluorescent signal was obtained for each left and right suprachiasmatic nucleus (SCN) using the Profile Plot Analysis feature of ImageJ. A rectangular box of fixed size (415.38 μm × 110.94 μm, width × height) was created to include the entire ventral part of the SCN. A column plot profile was generated whereby the x-axis represents the horizontal distance through the SCN (lateral to medial for the left SCN and medial to lateral for the right SCN, as indicated by the arrows) and the y-axis represents the average pixel intensity per vertical line within the rectangular box. Bottom Panels: Images with outlined the shell (green) and core (white) of the SCN, the master circadian clock, located in the anterior hypothalamus: OC = optic chiasm, V = third ventricle.

Figure 6.. The deficits in social recognition and repetitive behaviors of the Fmr1 KO mice correlate with altered sleep behavior.

Tests of social behavior: (A) In the first stage of the three-chamber social test, when the testing mouse is given a choice between a stranger mouse and an inanimate object, the Fmr1 KO mice spent less time with the stranger mouse and had a lower social preference index (SPI) than the WT. (B) In the second stage, the testing mouse is given the choice between a chamber with a novel mouse and the one with the familiar mouse. The mutants spent less time with the novel mouse compared to the familiar one, and also in this phase exhibited lower social novelty preference (SNPI) as compared to the WT. (C) The possibility of reduced social recognition was further tested with the five-trial social test. In this test, the stranger mouse becomes a familiar mouse after four exposures to the testing mouse, then a novel mouse is introduced in the fifth trial. The WT mice showed a higher interest in the novel mouse compared to the Fmr1 KO mice. Tests of repetitive behaviors: The amount of digging in the bedding (D) and the percentage of marbles buried (E) were measured with the marble bury test. The Fmr1 KO mice spent longer time digging and buried more marbles compared to WT. (F) Grooming behavior, assessed in a novel arena, was significantly higher in the Fmr1 KO mice as compared to WT. Histograms show the means ± SEM with the values from the individual animals overlaid. Significant differences (*p < 0.05) were determined by t-test or Mann–Whitney test. See also Table 5. Sleep duration (min, G, H) correlated with impaired social recognition and abnormal grooming behaviors. Sleep fragmentation, measured by number of sleep bouts, correlated only with grooming (I, J) (Pearson correlation test). See Table 6.

Figure 6—figure supplement 1.. To assess social recognition memory, Fmr1 KO and WT mice underwent a five-trial social interaction paradigm in a neutral open-field arena.

Each trial lasted 2 min and was separated by a 5-min inter-trial interval. During trials 1–4, the test mouse was exposed to the same unfamiliar conspecific (mouse A) enclosed within a wire cup to permit olfactory but limited tactile interaction. In trial 5, a novel conspecific (mouse B, stranger) was introduced. Time spent investigating the stimulus mouse (defined as sniffing or directing the nose toward the enclosure in close proximity) was manually scored. A progressive decrease in investigation time across trials 1–4 reflects habituation, while a significant increase in trial 5 indicates dishabituation and suggests intact social recognition memory. (A) Box plots: time spent by the testing mice interacting in each of the five trials. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median and the boundary of the box farthest from zero indicates the 75th percentile. The whiskers above and below the box indicate the 90th and 10th percentiles. (B) Plots showing the interaction time for each mouse during fourth (familiar mouse) and fifth (novel mouse) trials. The WT mice exhibited significant increases in the time spent interacting with the stimulus mouse in the fifth trial as compared to the between the fourth (paired t-test, T₍₁₄₎ = 2.604, p = 0.021), while no differences were observed for the KO mice between the fourth and the fifth trials (T₍₁₄₎ = 0.624, p = 0.542). *P<0.05 forth vs fifth trial.

Figure 7.. Amelioration of sleep/wake rhythms in the Fmr1 KO mutants by time-restricted feeding (TRF).

(A, B) Waveforms of daily rhythms in cage activity using infrared (IR) detection in the WT (circle) and Fmr1 KO (triangle) mice under ad libitum feeding (ALF) or TRF. The activity waveforms (1 hr bins) were analyzed using a three-way ANOVA with genotype, feeding regimen, and time as factors followed by Holm–Sidak’s multiple comparisons test. There were significant effects of genotype (F_{(1, 767)} = 13.301; p < 0.001) and time (F_{(23, 767)} = 94.188; p < 0.001), as well as significant interactions between genotype and time (p < 0.001) and feeding regimen and time (p < 0.001) on the locomotor activity rhythms of both WT and Fmr1 KO mice. The green area indicates the time period when the food hoppers were opened for 6 hr between ZT 15 and ZT 21. (C–E) Measures of locomotor activity rhythms: Both genotypes exhibited an increase in the power of the rhythms under TRF compared to ALF controls. The increase in early-day and late-night activity as well as the onset variability seen in the Fmr1 KO mice was corrected by the TRF. Data are shown as the means ± SEM; two-way ANOVA followed by Holm–Sidak’s multiple comparisons test with genotype and feeding regimen as factors, *p < 0.05 significant differences within genotypes (different feeding regimen); ^#p < 0.05 significant differences between genotypes (same feeding regimen). (F, G) Waveforms of daily rhythms of immobility-defined sleep. The sleep waveforms (1 hr bins) were analyzed by two-way ANOVA with time and feeding regimen as factors followed by the Holm–Sidak’s multiple comparisons test. There were significant effects of time for both WT (F_{(23, 351)} = 9.828, p < 0.001) and Fmr1 KO (F_{(23, 351)} = 1.806, p = 0.014) mice, but not of feeding regimens. Missing data points precluded the use of three-way ANOVA for these measures. (H–J) Measures of immobility-defined sleep in the light phase. Both genotypes held on TRF exhibited an increase in sleep duration and in sleep bout length as well as a reduction in sleep fragmentation, measured by the number of sleep bouts, compared to their ALF counterparts. Data are shown as the means ± SEM; two-way ANOVA followed by Holm–Sidak’s multiple comparisons test with genotype and diet as factors, *p < 0.05 significant differences within genotype – between diet regimens; ^#p < 0.05 significant differences between genotypes – same feeding regimen. See Table 7.

Figure 7—figure supplement 1.. Both genotypes well adapt to the feeding regimen.

(A) Food intake was similar across genotypes and conditions. Over the 2 weeks of scheduled feeding, no significant differences were found in the total amount of food consumed between genotypes (F_{(1, 31)} = 3.086, p = 0.090) or feeding schedules (F_{(1, 31)} = 0.307, p = 0.584). If we just analyzed the food consumed over the first 3 days, there were no effects of genotype (F_{(1, 31)} = 2.737, p = 0.109), albeit the time-restricted feeding (TRF) groups consumed less (F_{(1, 31)} = 85.912, p < 0.001). (B) Prior to TRF, the mice had similar weights (WT: 22.1 ± 0.4 g; KO: 23.1 ± 0.4 g; t₍₃₀₎ = –1.748, p = 0.090). After 2 weeks on TRF, both WT and Fmr1 KO mice exhibited lower weights than their counterparts on ALF (F = 30.551; p < 0.001). Two-way ANOVA followed by the Holm–Sidak’s multiple comparisons test (*p < 0.01).

Figure 8.. Improved social memory and stereotypic grooming behavior in the Fmr1 KO mice after 2 weeks of time-restricted feeding (TRF).

(A) Social memory was evaluated with the five-trial social interaction test as described above. The social memory recognition was significantly augmented in the Fmr1 KO by the intervention, suggesting that the treated mutants were, perhaps, able to distinguish the novel mouse from the familiar mouse. The time spent in social interactions with the novel mouse in the fifth trial was increased to WT-like levels in the mutants on TRF. Paired t-tests were used to evaluate significant differences in the time spent interacting with the test mouse in the fourth (familiar mouse) and fifth (novel mouse) trials. *p < 0.05 indicates the significant time spent with the novel mouse compared to the familiar mouse. (B) Grooming was assessed in a novel arena in mice of each genotype (WT, Fmr1 KO) under each feeding condition and the resulting data analyzed by two-way ANOVA followed by the Holm–Sidak’s multiple comparisons test with feeding regimen and genotype as factors. *p < 0.05 indicates the significant difference within genotype – between diet regimens, and ^#p < 0.05 those between genotypes – same feeding regimen. (C) TRF did not alter the overall locomotion in the treated mice. See Table 8.

Figure 9.. Time-restricted feeding (TRF) repristinates the levels IL-12 and IFN_Ƴ in the plasma of Fmr1 KO mice to WT levels.

(A) The levels of selected plasma pro-inflammatory markers are shown. The increase in IL-12 and IFN_Ƴ in the mutants is lowered by TRF to WT levels. The full list of the assayed makers is reported in Table 9. Data were analyzed with two-way ANOVA followed by the Holm–Sidak’s multiple comparisons test with feeding regimen and genotype as factors. *p < 0.05 indicates the significant difference within genotype – between feeding regimen, and ^#p < 0.05 between genotypes – same feeding regimen. (B, C) Correlations between IL-12 or IFN-γ levels and social recognition or grooming behavior. IL-12 levels correlated with both behaviors. Data were analyzed using the Pearson Correlation, and the coefficients are reported in Table 10.