Early-life sleep disruption in Shank3-deficient rats: A preclinical model for autism-related sleep mechanisms and interventions
Mei-Hong Qiu, Zhi-Gang Zhong, Pei-Wen Song, Gui-Jin Tao, Jin-Tao Zhang, Yong-Hua Chen, Tian-Jia Song, Wei-Min Qu, Rong Zhang, Zhi-Li Huang
TL;DR
This study explores sleep issues in rats with a genetic mutation linked to autism, showing how sleep problems may be an early sign of the disorder.
Contribution
The study introduces a preclinical model using Shank3-deficient rats to investigate autism-related sleep mechanisms and interventions.
Findings
Shank3-deficient rats show sex-specific sleep abnormalities, including fragmented sleep in males and prolonged wakefulness in females.
Both sexes exhibit reduced NREM sleep δ power and blunted homeostatic rebound after sleep deprivation.
Downregulation of Clock and Bmal1 mRNA suggests circadian dysregulation in corticostriatal circuits.
Abstract
Sleep disturbances are among the most prevalent and early-emerging features of autism spectrum disorder (ASD), often preceding core behavioral symptoms. Despite their clinical relevance, the neurobiological mechanisms driving early-life sleep disruption in ASD remain poorly understood. Shank3, encoding a synaptic scaffolding protein at excitatory synapses, is one of the most well-established monogenic risk factors for ASD. Here, we systematically investigated sleep architecture and homeostatic regulation in juvenile Shank3Δe11–21 rats, which lack Shank3 protein and display ASD-like behavioral and sensory phenotypes. EEG/EMG recordings revealed sex-specific abnormalities: males exhibited fragmented sleep with frequent brief arousals, whereas females showed prolonged wakefulness. Both sexes demonstrated reduced NREM sleep δ power, indicating diminished sleep depth. Following 6-h sleep…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —National Key Research and Development Program of China (2023YFC2306500) National Natural Science Foundation of China (81771430, 81971239)
- —Beijing Natural Science Foundation (J230013)
- —National Natural Science Foundation of China (82020108014) National Major Project of China Science and Technology Innovation 2030 for Brain Science and Brain-Inspired Technology (2021ZD0203400) Lingan
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsSleep and related disorders · Sleep and Wakefulness Research · Autism Spectrum Disorder Research
Introduction
Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by deficits in social communication and the presence of restricted, repetitive behaviors. Sensory abnormalities and varying degrees of intellectual disability are also commonly features of the disorder [1, 2]. The global prevalence of ASD has risen sharply across demographic groups [2]. In the U.S., 1 in 36 (2.76%) 8-year-old children were diagnosed with ASD in 2020, representing a 17.8% rise from 2018 (1 in 44) and a 33% increase from 2016 (1 in 54), with boys affected 3.8 times more frequently than girls [3–5]. Comparable increases have also been reported in China, where approximately 700,000 Chinese aged 6-12 are affected by ASD [6]. Despite this growing public health burden, the etiology of ASD remains incompletely understood, and effective interventions for its core symptoms are still lacking.
Genetic factors contribute substantially to ASD risk, with numerous ASD-associated genes involved in processes related to fetal brain development, chromatin modification, transcriptional regulation, and synaptic function [7]. Among these, Shank3, encoding a postsynaptic scaffolding protein critical for glutamatergic synapses, stands out as a high-confidence monogenic contributor, with truncating mutations account for approximately 0.69% of ASD cases [8]. However, the precise mechanisms linking Shank3 deficiency to ASD pathophysiology remain elusive.
To elucidate the pathophysiological mechanisms of Shank3-related ASD and explore potential therapeutic strategies, a variety of mouse models with Shank3 mutations have been developed, including exon deletions within the ankyrin, PDZ or Homer domains, as well as full knockouts [9–15]. These mouse models show varying degrees of reduced social behaviors, increased repetitive behavior, and other phenotypes related to comorbidities such as anxiety and motor impairments [9–14, 16–19]. Nonetheless, investigating certain complex behaviors, particularly during juvenile stages, is technically challenging in mice. Rats, with their larger brain size and richer behavioral repertoire, offer distinct advantages for practical and translational studies [20]. Therefore, we generated a novel rat model with a global deletion of Shank3 exons 11-21 (Δe11-21), featuring complete loss of Shank3 protein [21]. These Shank3^Δe11–21^ rats exhibit pronounced ASD-like phenotypes, including deficits in social memory and cognition, heightened anxiety-like behaviors and abnormal nociception, closely paralleling clinical observations in individuals with Shank3-related ASD. Thus, this rat model holds significant potential for advancing our understanding of Shank3 mutation contributions to ASD and evaluating therapeutic interventions.
Beyond the core behavioral abnormalities, it has been recognized that sleep disturbances are highly prevalent in ASD, affecting approximately 50–80% children, with rates as high as 90% among those with Shank3 mutations [22–27]. These sleep issues often emerge as early as 2 years of age, and manifest as reduced sleep duration, difficulties initiating and maintaining sleep, frequent nocturnal awakenings, and heightened bedtime anxiety [22–27]. Clinical evidence suggests that sleep disturbances can exacerbate ASD symptom severity, and the severity of the core symptoms of ASD can be predicted based on the presence and extent of sleep problems [28–30]. In line with these clinical observations, studies in Shank3-deficient mice and macaques have revealed marked alterations in sleep architecture and regulation [31–35]. Furthermore, experimental disruption of sleep or circadian rhythms in animal models aggravates ASD-like behaviors and synaptic abnormalities, whereas restoration of sleep partially ameliorates these phenotypes [36–40]. Yet, the reasons underlying increased susceptibility to sleep disturbances in ASD remain poorly understood. Most Shank3 mouse studies have focused on adult animals, limiting their relevance to early-onset ASD phenotypes [31–33, 41]. Moreover, stable EEG/EMG recordings during juvenile development are technically challenging in mice, and while non-human primate models provide valuable translational insights [35], they are resource-intensive and impractical for routine use.
Our Shank3^Δe11-21^ rat model provides a technically feasible and developmentally relevant platform for the precise examination of sleep-wake behavior during the juvenile stage. In this study, we systematically characterized sleep-wake architecture and homeostatic regulation in juvenile Shank^Δe11-21^ rats, aiming to gain a deeper insight of the function of Shank3 in early-life sleep physiology and reinforce the preclinical relevance of this model for exploring ASD-related pathophysiology and potential sleep-based interventions.
Materials and methods
Animals and experimental procedures
Animals were housed under controlled environment conditions (temperature 22-23 °C, relative humidity 60 ± 2%) and maintained on a 12:12 h light–dark cycle, with light on at 7:00 a.m. (ZT0, ~ 100 lux) and lights off at 7:00 p.m. (ZT12). Food and water were available ad libitum. Shank3 wild-type (+/+), heterozygous (+/−) and homozygous knockout (−/−) littermates, generated from Shank3^Δe11-21^ heterozygous breeder pairs, were used in all experiments. Genotyping was performed on tail tissue collected at postnatal days (PND) 7–10. The generation of the Shank3 knockout rats and the genotyping protocols were previously described [21] and are further detailed in the supplemental information. All experiment procedures were approved by the Committee on the Ethics of Animal Experiments of the School of Basic Medical Sciences, Shanghai Medical College, Fudan University (Permit number: 20170223-043), and were conducted in accordance with relevant guidelines and regulations. Every effort was made to minimize any pain and discomfort experienced by the subjects.
Experiments were conducted in three cohorts of juvenile rats (< PND 32). Cohort 1 underwent spontaneous locomotor activity (LMA) monitoring at PND23 and baseline electroencephalogram (EEG)/electromyography (EMG) acquisition at PND30-31. Cohort 2 was employed for sleep-deprivation testing at PND30–31. Cohort 3 was used for the assessment of circadian transcription factor expression in brain regions with high Shank3 expression (prefrontal cortex, hippocampus, and striatum) at PND 30 (Fig. 1). Animals were assigned to experimental groups based on their genotype; no further randomization or blinding procedures were applied. No animals or samples were excluded from the analyses.Fig. 1. Timeline of experimental procedures.A The 24-h continuous spontaneous LMA of juvenile rats was monitored at PND23. The EEG/EMGs collected at PND30 or PND31 were used to assess the baseline sleep-wake architecture of the juvenile rats. B 6-h total sleep deprivation was performed on PND31 from 13:00 to 19:00 (ZT6–ZT12). EEG/EMGs recorded at PND30 served as each animal’s baseline control. C Brain dissection for qRT-PCR analysis. The Prefrontal cortex, striatum, and hippocampus of the juvenile rats were dissected at 9:00 (ZT2) on PND30 for quantification of Clock, Bmal1, Per1, Per2, Cry1, and Cry2 mRNA expression. White and black bars represent the light and dark phases, respectively. LMA: locomotor activity; PND: postnatal day; SD: sleep deprivation.
Locomotor activity
Spontaneous locomotor activity (LMA) was monitored using an infrared beam- break detection apparatus (Biotex, Kyoto, Japan) as previously described [42]. Briefly, at PND21, juvenile rats were individually housed in transparent cages (28 × 23 × 21 cm) placed inside soundproof cabinets. Following two days of habituation, LMA was continuously recorded for 24 h beginning at 7:00 a.m. on PND23 using an infrared sensor mounted above each cage.
EEG/EMG recordings and vigilance state analysis
On PND24, rats were anesthetized with sodium pentobarbital (50 mg/kg, i.p.) and implanted with electrodes for EEG/EMG recordings as described previously [43]. Two stainless-steel screw electrodes for EEG were placed over the right frontal and parietal cortex, with a reference electrode over the left frontal cortex. Two insulated wire electrodes were inserted into the trapezius muscles for EMG measurements recording. All electrodes were connected to a micro-connector, which was secured to the skull with dental cement.
After recovering from anesthesia, rats were individually housed in transparent cylinders within the recording chamber. Beginning at PND28, they were connected to the recording cable and habituated for two days. Continuous EEG/EMG recordings were then performed for 48 h starting at 7:00 a.m. on PND30 using a slip-ring commutator to allow unrestricted movement.
EEG/EMG signals were amplified and filtered (EEG: 0.5–30 Hz, EMG: 20–200 Hz), digitized at a sampling rate of 128 Hz, and recorded using VitalRecorder software (Kissei Comtec, Nagano, Japan). Vigilance states were automatically scored offline in 10-s epochs as wakefulness, non-rapid eye movement (NREM) sleep, or rapid eye movement (REM) sleep using SleepSign software (Kissei Comtec, Japan), followed by manual verification of all epochs to ensure scoring accuracy. EEG power spectra for each epoch were computed within SleepSign using fast Fourier transformation (FFT) with a 256-point Hanning window (0–24.5 Hz, 0.5-Hz resolution). Power density at each frequency bin was expressed as a percentage of the total EEG power in the 0–24.5 Hz range. Frequency bands were defined as δ (0–4 Hz), low θ (5–7 Hz), high θ (7.5-10 Hz), α (12-13.5 Hz), and β (14-24.5 Hz). Stage transitions were defined as changes in vigilance state between two consecutive 10-s epochs and quantified as the total number per 12-h period. The behavior of the animals was simultaneously monitored with infrared cameras.
Sleep deprivation
Rats were implanted with EEG/EMG electrodes on PND 24. Baseline recordings were collected for 24 h beginning at 7:00 a.m. on PND30. On PND31, animals were subjected to 6 h of total sleep deprivation from 13:00–19:00 (ZT6-12) by gentle handling with a soft brush, as described previously [43].
Quantitative real-time PCR
At 9:00 a.m. on PND30, Shank3^−/−^ and wild-type rats were sacrificed. Brains were rapidly removed and placed on an ice-chilled dissection surface. The prefrontal cortex, hippocampus, and striatum were dissected for total RNA extraction using Total RNA extraction Reagent (Novizan, China). cDNA was synthesized using the HiScript® III RT SuperMix for qPCR with gDNA Wiper (Novizan, China). Quantitative real-time PCR (qRT-PCR) was then carried out using SYBR Green master mix (Novizan, China) on an Eppendorf RealPlex PCR system (Germany). Primers specific to circadian rhythm genes (Clock, Bmal1, Per1, Per2, Cry1, Cry2) and the reference gene Gapdh were used (as detailed in supplementary Table S1). The reaction program included initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. All reactions were run in triplicate. Relative expression level of each target gene was calculated using the 2^−ΔΔCt^ method with Gapdh as the internal control. The primers used in this study were referenced from Sun et al. [44].
Statistical analysis
Data were analyzed using GraphPad Prism 10 (GraphPad Software, San Diego, CA, USA). Results are expressed as mean ± standard error of the mean (SEM). Normality of data distributions was assessed using the D’Agostino–Pearson omnibus test. Time-course data (e.g., locomotor activity and sleep–wake profiles) and frequency-domain data (EEG power spectra) were analyzed using repeated-measures analysis of variance (RM ANOVA), with time (1-h bin) or frequency (0.5-Hz bin) as the within-subject factor, and sex and genotype as between-subject factors. Bar-graph datasets involving two or three independent variables were analyzed using two-way or three-way ANOVA, as appropriate, with factors including genotype, sex, treatment or phase (light/dark). For comparisons of genotype effects within each sex, one-way ANOVA was additionally performed. Mixed-effects models with Geisser–Greenhouse correction were applied when sphericity was violated or missing values occurred. When ANOVA revealed a significant main effect or interaction, Fisher’s LSD or Tukey’s post hoc tests were applied. In all analyses, differences were considered statistically significant at p < 0.05. The sample size and p value for each graph was stated in figure legends. No statistical methods were used for sample size estimate.
Results
Male juvenile shank3−/− rats exhibit hypoactivity
Locomotor activity (LMA) abnormalities have been reported in several Shank3 mutant mouse lines [18, 45, 46]. To assess whether juvenile Shank3 deficiency rats show similar deficits, we continuously monitored 24-h LMA at PND23. A three-way ANOVA (time × sex × genotype) revealed a significant sex × genotype interaction (F(1,45) = 4.26, p = 0.045; supplementary Table S2), indicating that Shank3 loss differentially impacted activity levels between males and females (Fig. 2A).Fig. 2. Male juvenile shank3^−/−^ rats exhibit hypoactivity.A Hourly time courses of spontaneous LMA across 24-h of male (upper panel) and female (lower panel) shank3^+/+^, shank3^+/–^ and shank3^–/–^ rats at PND23. Each circle represents the hourly mean of LMA. Open bars along the x-axis denote the light phase, and dark gray bars denote the dark phase. B Total LMA counts during the light phase, dark phase, and over the 24-h period in male (blue) and female (purple) Shank3^+/+^, Shank3^+/–^ and Shank3^–/–^ rats at PND23. Data are presented as mean ± SEM., n = 13, 25 and 8 for male +/+, +/– and –/–, respectively; and n = 15, 18 and 13 for female +/+, +/– and –/–, respectively. * p < 0.05, ** p < 0.01, compared between Shank3^+/+^ and Shank3^–/–^ rats. ^##^ p < 0.01, compared between Shank3^+/+^ and Shank3^+/–^ rats, assessed by two-way RM ANOVA followed by Sidak’s post hoc test (A), and by two-way ANOVA with sex and genotype (WT vs. KO) as factors followed by Fisher’s LSD post hoc test for light, dark, and 24-h total LMA counts (B). Three-way ANOVA [time/phase × sex × genotype (WT: KO)] and Phase-wise one-way ANOVAs across WT, HET, KO within each sex are provided in the supplementary materials. m: male. f: female. LMA: locomotor activity. Detailed raw data and statistical outputs are available in Fig. 2–source data 1 and Fig. 3–source data 2. Figure 2 – source data 1: Raw data and summary statistics used for generating plots in Fig. 2. Figure 2**– source data 2**. Statistical results of Fig. 2.
Phase-specific two-way ANOVAs (sex × genotype) showed consistent interaction patterns across the 24-h cycle, with trend-level effects observed in both the light (p = 0.055) and dark (p = 0.082) phases (Supplementary Table S4). Post hoc tests revealed that male Shank3^−/−^ rats were less active than wild-type males during the light phase (p = 0.029) and across 24 h (p = 0.037), with a trend toward hypoactivity in the dark phase (p = 0.085; Fig. 2B, blue bars). In females, activity did not differ by genotype. Across sexes, Shank3^+/−^ juveniles did not differ from either wild-type or Shank3^−/−^ littermates. Together, these findings indicate that Shank3 deficiency preferentially reduces spontaneous LMA in male juveniles.
Juvenile shank3−/− rats exhibit higher arousal
Previous studies have reported difficulties in initiating and maintaining sleep in both individuals with Phelan-McDermid syndrome (PMS) carrying Shank3 mutations and in adult male Shank3-deficient mice [27, 31]. To assess whether spontaneous sleep is altered in juvenile Shank3 mutant rats, undisturbed baseline EEG/EMG were obtained at PND30–31in male and female rats of all genotypes. We found that one-month-old rats of both sexes already exhibited a clear circadian sleep-wake pattern, albeit with a relatively high REM fraction typical of juveniles (Figure S2A, Fig. 3A). No sex differences were observed in total sleep amount or EEG power spectra among wild-type rats (Figure S2).Fig. 3. Juvenile Shank3^–/–^ rats sleep less, especially during the dark period.A Hourly time courses of wakefulness, NREM sleep, and REM sleep across 24 h of the male (blue) and female (purple) shank3^+/+^, shank3^+/–^ and shank3^–/–^ rats at PND30. Each circle represents the hourly mean of the respective vigilance state. B Total time spent in wakefulness, NREM sleep, and REM sleep during the light phase, dark phase, and over the 24-h period. Data are presented as mean ± SEM. n = 8/genotype for males and 9, 9 and 8 for female +/+, +/– and –/–, respectively. * p < 0.05, ** p < 0.01, *** p < 0.001, and # p < 0.05, difference between genotypes or sexes, assessed by two-way RM ANOVA (time × genotype) with Sidak’s post hoc test (A), or by two-way ANOVA (sex × genotype) followed by Fisher’s LSD post hoc test for phase-based comparisons (light, dark, 24 h; B). Three-way ANOVA [time/phase × sex × genotype (WT: KO)] and within-sex one-way ANOVAs across WT, HET, KO groups are provided in the Supplementary Materials. m: male. f: female. Detailed raw data and statistical outputs are available in Fig. 3–source data 1 and Fig. 3–source data 2. Figure 3–source data 1. Raw data and summary statistics used for generating plots in Fig. 3. Figure 3–source data 2. Statistical results of Fig. 3.
In male juveniles, Shank3^−/−^ rats spent more time awake than wild-type littermates (Fig. 3A, upper-left). Repeated-measures ANOVA across 24 h revealed a genotype main effect on wake (F(2,21) = 4.40, p = 0.025). Specifically, wake time increased by 14.4% during the dark phase (449.3 ± 19.9 vs. 392.8 ± 10.2 min; p = 0.021) and by 10.0% across the full 24 h (720.4 ± 16.6 vs. 655.1 ± 8.5 min; p = 0.008; Fig. 3B, upper panel), with corresponding NREM sleep reductions during both the dark phase (210.1 ± 15.9 vs. 250.5 ± 10.2 min; p = 0.041) and the 24-h period (597.3 ± 13.6 vs. 642.0 ± 7.8 min; p = 0.035) (Fig. 3B, middle panel). Although REM sleep tended to decrease in Shank3^+/−^ and Shank3^−/−^ rats during the dark phase, one-way ANOVA within males did not detect significant genotype effects (Fig. 3B, lower panel).
In female juveniles, two-way RM ANOVA across the 24-h period revealed robust genotype effects on wakefulness (F(2,23) = 22.59, p < 0.0001) and NREM sleep (F(2,23) = 15.77, p < 0.0001, Supplementary Table S5). Shank3^−/−^ females spent significantly more time awake across both phases, with a particularly strong effect during the dark period, with a particularly strong effect during the dark phase (light: F(2,23) = 4.09, p = 0.030; dark: F(2,23) = 12.72, p = 0.0002; Supplementary Table S7) (Fig. 3A and B, upper panels). Compared with wild-type littermates, wake time increased by 15.3% in the light phase (275.9 ± 11.0 vs. 239.3 ± 7.8 min; p = 0.024), 21.8% in the dark phase (486.2 ± 12.7 vs. 399.3 ± 11.5 min; p < 0.0001), and 19.3% across the full 24 h (762.0 ± 6.8 vs. 638.6 ± 9.1 min, p < 0.0001). NREM sleep was correspondingly reduced (light: F(2,23) = 4.02, p = 0.032; dark: F(2,23) = 13.71, p = 0.0001; Fig. 3A and B, middle panels). Notably, even Shank3^+/−^ females displayed altered sleep–wake profiles, with increased wake (light: 276.1 ± 12.5 vs. 239.3 ± 7.8 min, p = 0.020; dark: 447.2 ± 12.2 vs. 399.3 ± 11.5 min, p = 0.009) and reduced NREM sleep (light: 375.0 ± 10.9 vs. 406.0 ± 6.0 min, p = 0.032; dark: 210.7 ± 9.3 vs. 242.6 ± 7.7 min, p = 0.014). As in males, REM sleep showed only a trend-level reduction that did not reach significance within females.
To further evaluate sex effects, we performed three-way repeated-measures ANOVA with sex, genotype, and either time (1-h bins; Fig. 3A) or phase (light/dark; Fig. 3B) as factors. Across the 24-h time course, there were robust main effects of time (all vigilance states, p < 0.0001) and genotype (all vigilance states, p < 0.05), together with significant sex × genotype interactions for wakefulness (p = 0.012) and NREM sleep (p = 0.010). Phase-based analyses yielded convergent results, showing strong phase effects (wake and NREM, p < 0.0001) and consistent genotype influences (Supplementary Tables S8–S9).
Overall, these results suggest that the sleep–wake disturbances associated with Shank3 deficiency were more pronounced in females, especially during the dark (active) phase.
Male juvenile shank3−/− rats exhibit sleep fragmentation, whereas female juvenile shank3−/− rats show prolonged wakefulness
To further characterize sleep-wake alterations induced by Shank3 deficiency in juvenile stage, we further analyzed sleep-wake architecture of the rats. In Male juveniles, shank3^−/−^ rats showed fewer stage transitions from NREM to REM sleep during the dark phase (Fig. 4A, right panel), and fewer REM sleep bouts (Fig. 4B, right panel). Mean NREM episode duration was shorter (F(2,21) = 4.51, p = 0.023), reflecting a 19% reduction relative to wild-type (65.1 ± 3.3 s vs. 80.4 ± 4.0 s; p = 0.009; Fig. 4C, right panel). Bout-length distribution analysis further revealed shorter NREM sleep bouts (0–30 sec; 58.88 ± 4.01 vs. 45.13 ± 4.89, p = 0.031) during the dark phase (Fig. 4D, middle-right panel), indicating impaired NREM sleep consolidation in male juvenile shank3^−/−^ rats.Fig. 4. Juvenile Shank3^–/–^ rats exhibit increased short sleep episodes in males and longer wake episodes in females.A Number of state transitions during the light and dark phases. N, W, and R denote NREM sleep, wakefulness, and REM sleep, respectively. B Number of wake, NREM sleep and REM sleep bouts during the light and dark phases. C Mean duration of each vigilance state during the light and dark phases. D Distribution of wake, NREM sleep and REM sleep episode counts across different duration bins during the light and dark phases. Data were from Shank3^+/+^, Shank3^+/–^ and Shank3^–/–^ rats at PND30, represented as mean ± SEM. n = 8/genotype for male rats and 9, 9 and 8 for female^+/+^, female^+/–^ and female^–/–^ rats. * p < 0.05, ** p < 0.01, *** p < 0.001, difference between genotypes, assessed by two-way ANOVA with sex and genotype (WT: KO) as factors followed by Tukey’s post hoc test. Three-way ANOVA [phase × sex × genotype (WT: KO)] and within-sex one-way ANOVAs across WT, HET, KO groups are provided in the Supplementary Materials. m: male. f: female. Detailed raw data and statistical outputs are available in Fig. 4–source data 1 and Fig. 4–source data 2. Figure 4–source data 1. Raw data and summary statistics used for generating plots for Fig. 4. Figure 4 – source data 2. Statistical results for Fig. 4.
In female juveniles, shank3^−/−^ rats showed significantly fewer transitions from wake to NREM sleep (dark: 141.0 ± 4.86 vs. 175.0 ± 8.60, p = 0.006), together with reduced wake and NREM sleep bout numbers (Wake: dark 142.75 ± 4.85 vs. 177.22 ± 8.84, p = 0.017; NREM: Light 257.88 ± 10.95 vs. 285.00 ± 12.15, p = 0.046; dark: 143.88 ± 5.23 vs. 180.44 ± 9.01, p = 0.010) (Fig. 4A, B). While NREM sleep mean duration remains unchanged, wake episode duration was markedly prolonged in both the light and dark phase (light period: F (2, 23) = 8.010, p = 0.0023; dark period: F (2, 23) = 12.77, p = 0.0002). Compared with wild-type females, the mean duration of wake episodes increased by 35.6% during light (66.13 ± 4.29 sec vs. 48.78 ± 2.50 sec, p = 0.003) and 48.1% during dark (205.75 ± 9.6 sec vs. 138.89 ± 10.64 sec, p < 0.001) (Fig. 4C). Short wake bouts in the range of 0–30 sec were fewer in shank3^−/−^ females, and NREM sleep bouts in the range of 60–120 sec were also significantly decreased (all p < 0.05, Fig. 4D), further supports a phenotype of extended wake and impaired sleep initiation. The female juvenile Shank3^+/−^ showed no significant differences from wild-type counterparts.
To further assess whether these alterations were sex-dependent, a two-way ANOVA with sex × genotype as factors across state transition, bout number and mean duration were performed (Supplementary Table S16–S19). The results confirmed distinct sex-specific effects of Shank3 deficiency: significant main effect of sex were observed for NREM→wake transitions, Wake and NREM bout numbers, and wake mean duration (all p < 0.01). Moreover, Significant sex × genotype interactions emerged for NREM sleep bouts and (p < 0.05) and wake mean duration (p < 0.01), indicating that Shank3 deficiency differentially affects sleep continuity across sexes.
Together, these findings indicate that Shank3 deficiency differentially disrupts sleep stability, leading to fragmented NREM sleep in males while prolonging wakefulness in females.
Juvenile shank3−/− rats exhibit altered EEG power spectra
Fourier analysis of the EEG revealed marked alterations in power spectra in both male and female juvenile shank3^−/−^ rats (Fig. 5). During REM sleep, EEG power spectra were right-shifted in both light and dark phases (Fig. 5A, C, lower panels), characterized by a significant reduction in low-θ (5–7 Hz) and an increase in high-θ (7.5–10 Hz) activity (Fig. 5B, D, lower panels). During NREM sleep, both sexes exhibited substantial changes in spectral composition (Fig. 5A–D, middle panels). The δ (0–4 Hz) power—a hallmark of slow-wave activity—was significantly reduced in shank3^−/−^ rats compared with wild-type controls [males: -20.85% (light, p = 0.008) and -24.75% (dark, p = 0.0005); females: -22.09% (dark, p = 0.013)], accompanied by enhanced θ and α power, suggesting reduced slow-wave activity and a relative shift toward high-frequency oscillations in shank3^−/−^ rats. In Wake, low-θ power was also attenuated, particularly in females (Fig. 5A–D, upper panel). The temporal dynamics of NREM δ activity remained consistently lower across the 24-h cycle in male shank3^−/−^ rats (Fig. 5E) and was moderately reduced in females (Fig. 5F). In juvenile shank3^+/−^ rats, EEG power spectrum changes followed similar trends but were generally non-significant.Fig. 5. Juvenile shank3^−/−^ rats exhibit altered EEG power spectra.A–D Averaged EEG power density (A, C) and relative power ratio of each frequency band (B, D) of wakefulness, NREM sleep, and REM sleep during the light and dark phase in male (blue) and female (pink) Shank3^+/+^, Shank3^+/−^, and Shank3^−/−^ rats at PND30. Shaded color regions in (A, C) denote the respective frequency bands (δ, θ, α, β). E–H Hourly dynamics of normalized NREM δ power ratio across 24 h in male (E) and female (F) rats at PND 30. Data are represented as mean ± SEM. n = 8/genotype for males and 9, 9 and 8 for female +/+, + /–, –/– rats, respectively. Thin horizontal lines indicate significant differences between Shank3^+/−^ and Shank3^+/+^ rats, thick horizontal lines indicate significant differences between Shank3^−/−^ and Shank3^+/+^ rats, and the dashed lines indicate significant differences between Shank3^+/−^ and Shank3^+/+^ rats, assessed by two-way RM ANOVA (frequency × genotype) with Sidak’s test (A, C). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, difference between genotypes, assessed by two-way ANOVA (sex × genotype) followed by Tukey’s multiple comparisons test (B, D), or two-way RM ANOVA (time × genotype) with Sidak’s test (E, F). Three-way ANOVA [frequency/time × sex × genotype (WT: KO)] and within-sex one-way ANOVAs across WT, HET, KO groups are provided in the Supplementary Materials. m: male. f: female. Detailed raw data and statistical outputs are available in Fig. 5–source data 1 and Fig. 5–source data 2. Figure 5–source data 1. Raw data and summary statistics used for generating plots for Fig. 5. Figure 5 – source data 2. Statistical results for Fig. 5.
Although spectral changes appeared more pronounced in males, multi-factor ANOVA assessing the combined effects of genotype, sex, and time or frequency revealed that genotype was the primary determinant of EEG power alterations. Sex effects were relatively weak and state-dependent, and no consistent genotype × sex interaction was detected (Supplementary Tables S23–24). Overall, these findings demonstrate that Shank3 deficiency leads to state- and frequency-specific alterations in cortical oscillatory activity, characterized by weakened slow-wave activity and relatively enhanced faster rhythms.
Juvenile Shank3−/− rats display abnormal homeostatic responses to sleep deprivation
To assess whether Shank3 deficiency impact sleep homeostatic regulation during the juvenile stage, we performed 6-h sleep deprivation (SD) from 13:00 to 19:00 in juvenile wild-type and Shank^−/−^ rats. The SD procedure effectively reducing wakefulness by 96.21 ± 0.79%, 96.04 ± 0.87%, 97.52 ± 0.46% and 95.88 ± 1.07% in male juvenile wild-type and Shank^−/−^ rats, and female wild-type and Shank^−/−^ rats, respectively. As expected, juvenile wild-type rats of both sexes showed a pronounced NREM sleep rebound during the first 2 h of recovery (Fig. 6A, B, left panel; Fig. 6D, E). In contrast, male juvenile Shank3^−/−^ rats showed a blunted NREM sleep rebound (Fig. 6A, right panel), with total NREM sleep during the first 2 h of recovery significantly reduced (43.00 ± 4.46 min vs. 66.50 ± 2.92 min, p < 0.0001; Fig. 6D), as well as a smaller NREM increase relative to baseline (16.36 ± 10.32 min vs. 31.91 ± 9.14 min, p = 0.007; Fig. 6E). Juvenile female Shank3^−/−^ rats also displayed reduced NREM sleep rebound than wild-type females (46.41 ± 3.73 min vs. 64.79 ± 2.52 min, p = 0.035, Fig. 6D), but the relative NREM increase did not differ significantly between genotypes (Fig. 6E).Fig. 6. Juvenile Shank3^−/−^ rats display abnormal homeostatic responses to sleep deprivation.A, B Time-course of NREM sleep under baseline (PND 30) and SD (PND 31) in male (A) and female (B) juvenile Shank3^+/+^ and Shank3^−/−^ rats. The SD period (13:00–19:00) is indicated by gray shading. C Latency to enter NREM and REM sleep after SD in juvenile male and female Shank3^+/+^and Shank3^–/–^ rats. D–F Total NREM sleep time (D), absolute NREM sleep increase (E), and the percentage increase in NREM sleep relative to baseline (F) during the first 2 h of recovery following SD in each genotype. G–J Time-course of NREM δ (0–4 Hz) power ratio (G, I) and normalized NREM δ power ratio (%baseline; H, J) during the 12-h recovery period after 6-h SD in juvenile male and female Shank3^+/+^and Shank3^–/–^ rats. The dark gray bars along the x-axis denote the dark phase. K, L Normalized NREM spectra during the first 2 h of recovery after SD in male (K) and female (L) rats of each genotype. M Normalized δ (0–4 Hz) power ratio in male and female Shank3^+/+^ and Shank3^−/−^ rats. Data represented as mean ± SEM. n = 9 male^+/+^, 7 male^−/−^, 9 female^+/+^, 6 female^−/−^. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.001, difference from own baseline control or from Shank3^+/+^ controls; # # p < 0.01, difference between sex, assessed by two-way RM ANOVA (time/frequency × treatment/genotype) followed by Sidak’s (A, B, G, I), or Fisher’s LSD post hoc test (H, J, K, L) or two-way ANOVA (sex × genotype) followed by Tukey (E, F), or Fisher’s LSD (M) post hoc test. m: male, f: female. Detailed raw data and statistical outputs are available in Fig. 6–source data 1 and Fig. 6–source data 2. Figure 6–source data 1. Raw data and summary statistics used for generating plots for Fig. 6. Figure 6 – source data 2. Statistical results for Fig. 6 and Figure S3.
In wild-type rats, 6-h SD triggered a robust δ-power rebound, lasting approximately 4 h in male and 3 h in females (Fig. 6G–I). This compensatory δ enhancement was absent in Shank3^−/−^ rats, whose NREM sleep δ power activity remained near baseline levels throughout recovery (Fig. 6G–I). Consistently, normalized NREM sleep spectral during the first 2 h of recovery (19:00–21:00) revealed that male Shank3^−/−^ rats exhibited lower δ power (120.18 ± 13.11% vs. 132.17 ± 12.02%, p = 0.045) accompanied by enhanced higher-frequency (6–24.5 Hz) activity (Fig. 6K, M). Whereas female mutants primarily showed high-frequency enhancement without δ suppression (Fig. 6L). A mild REM rebound occurred in juvenile Shank3^+/+^ rats but was absent in Shank3^−/−^ rats (Figure S3).
Although post-SD sleep rebound changes appeared more evident in males, three-way ANOVA (genotype, sex, and time/treatment factors) confirmed that genotype was the primary determinant of sleep homeostatic alterations, with no significant sex effects or interactions across NREM sleep time, δ-power dynamics, or normalized spectra (Supplementary Table S26–S29).
Together, these findings demonstrate Shank3 deficiency markedly impairs sleep homeostasis in juvenile rats, characterized by attenuated NREM sleep rebound and blunted δ-power recovery following sleep loss.
Juvenile shank3−/− rats show reduced expression of circadian rhythm genes
To further investigate the possible molecular basis of sleep disturbances in juvenile rats with Shank3 loss, we detected the mRNA expression of six core circadian genes (Clock, Bmal1, Per1, Per2, Cry1, and Cry2) in the prefrontal cortex (PFC), striatum, and hippocampus—regions known to exhibit high Shank3 expression. Two-way ANOVA (sex × genotype) revealed significant genotype effects on Clock in both the PFC and striatum, and on Bmal1 in the striatum, but not in the hippocampus (Fig. 7, Supplementary Table S37). Specifically, Clock mRNA was significantly reduced in the PFC (95% CI = 0.11–0.56, p = 0.009 for males; 0.07–0.52, p = 0.018 for females) and striatum (95% CI = 0.04–0.88, p = 0.035 for males; 0.29–1.13, p = 0.004 for females) of Shank3^−/−^ rats compared with wild-type littermates. In contrast, Bmal1 expression was selectively decreased in the striatum (95% CI = 0.05–0.84, p = 0.033 for males; 0.34–1.14, p = 0.003 for females), while remaining unchanged in the PFC and hippocampus. No main effect of sex or sex × genotype interaction was detected.Fig. 7. Juvenile shank3^−/−^ rats show reduced expression of circadian rhythm genes.Quantitation of Clock, Bmal1, Cry1, Cry2, Per1, and Per2 mRNA levels in the prefrontal cortex, striatum and hippocampus of male (blue) and female (purple) rats, measured by real-time RT-PCR. Each gene’s expression level was normalized to the endogenous reference gene Gapdh. Data are presented as mean ± SEM. n = 3/genotype. * p < 0.05, difference from Shank3^+/+^ rats, assessed by two-way ANOVA (sex × genotype) followed by Fisher’s LSD post hoc test. m: male, f: female. Detailed raw data and statistical outputs are available in Fig. 7–source data 1 and Fig. 7–source data 2. Figure 7–source data 1. Raw data and summary statistics used for generating plots for Fig. 7. Figure 7 – source data 2. Statistical results for Fig. 7.
Together, these findings indicate that Shank3 deficiency leads to consistent downregulation of Clock across cortical and striatal regions, and selectively reduces Bmal1 in the striatum, implicating disrupted molecular circadian regulation within corticostriatal circuits as a potential contributor to sleep–wake abnormalities in juvenile mutants.
Discussion
The present study demonstrates that Shank3 plays an important regulatory role in sleep during early neurodevelopment. Global deletion of Shank3 resulted in pronounced yet sex-dependent alterations in sleep architecture, EEG power spectra, and homeostatic regulation in juvenile rat. Specifically, male Shank3^–/–^ rats exhibited fragmented sleep characterized by frequent brief arousals, whereas female Shank3^–/–^ rats displayed sustained insomnia-like wakefulness. Although the gross sleep–wake organization remained preserved, these consistent abnormalities indicate that Shank3 is crucial for maintaining sleep stability and depth during a critical developmental window. Collectively, our results extend the phenotypic spectrum of Shank3 deficiency beyond core ASD-like behaviors to include early-onset disturbances in sleep regulation, a domain increasingly recognized as clinically relevant in ASD.
A particularly striking observation is the sex divergence in sleep phenotypes. Both sexes display hyperarousal, yet their manifestations differed: male mutants exhibited more fragmented sleep, characterized by shorter NREM bouts and a higher frequency of brief sleep episodes, indicating impaired sleep maintenance. In contrast, female mutants remained awake for abnormally long durations, suggesting difficulty in initiating sleep once aroused. These sex-specific patterns mirror clinical subtypes of insomnia observed in individuals with ASD, where some experience frequent awakenings and others difficulty falling asleep [22, 23, 25]. Although clinical data on sex differences in ASD sleep remains limited, emerging evidence suggests that sleep problems may be as severe or even more pronounced in autistic girls [47–49], despite ASD being more frequently diagnosed in boys. For instance, girls with ASD have been reported to exhibit lower sleep efficiency and longer wake time after sleep onset compared with boys [48]. Our findings in juvenile Shank3^–/–^ rats therefore mirror these human patterns and imply that biological sex modulates the neurobehavioral impact of Shank3 deficiency on sleep regulation. The observed differences may stem from sex-dependent variations in neuroendocrine environment, synaptic plasticity, or circadian gene expression, all of which could differentially shape arousal and sleep-maintenance circuits. Together, these findings highlight the importance of incorporating both sexes in preclinical ASD research to capture the full spectrum of sleep phenotypes and their translational relevance.
Both male and female Shank3^–/–^ rats exhibited lighter, less restorative sleep (Fig. 5). EEG Power spectral analysis revealed reduced δ power during NREM sleep, reflecting diminished slow-wave activity (SWA)—a hallmark of deep sleep and a key indicator of homeostatic sleep drive [50]. This mirrors polysomnographic findings in children with ASD, who often show reduced deep sleep and lower sleep efficiency [22, 23]. A recent study of Shank3 mutant mice also reported reduced NREM sleep intensity in both sexes [32], supporting a cross-species effect of Shank3 deficiency on sleep depth. Moreover, following 6-h SD, Shank3^–/–^ juveniles exhibited attenuated rebound in both NREM sleep time and δ power, suggesting a blunted homeostatic response. These findings parallel those in adult Shank3 knockout mice, which also fail to show normal recovery sleep [31]. Given the role of rebound sleep in restoring cognitive and physiological functions [51], impaired sleep homeostasis in Shank3 mutants may exacerbate neurodevelopmental deficits. These abnormalities may stem from disruptions in synaptic or molecular mechanisms that govern SWA and sleep pressure regulation, such as adenosine signaling or cytokine-mediated mechanisms [31, 52]. These hypotheses warrant further investigation in future studies.
Our findings both confirm and extend previous work in Shank3 mouse models. The heightened arousal observed in our juvenile Shank3^–/–^ rats during the active phase aligns with findings in adult Shank3 exon-21 knockout mice (Shank3^ΔC^) [31]. However, contrasting results have been reported in other Shank3 models, such as increased sleep during the active phase in exon 4–9 mutants [41], likely reflecting differences in isoform-specific targeting. These discrepancies highlight the need for careful interpretation of model-specific results. Regarding younger animals, a study of pre-adolescent Shank3^ΔC^ mice reported increased REM sleep [33], which we did not observe. In fact, increased wakefulness in our juvenile rats would inherently reduce REM time. Notably, invasive electrode implantation in that mouse study commenced on PND 18, an age when the skull is thin and fragile in mice, potentially causing significant disruption to normal brain function and confounding sleep data. Thus, the current juvenile rat model provides more dependable insights into the developmental trajectory of sleep disturbances associated with Shank3 mutations, making it potentially more translatable to clinical manifestations seen during childhood in ASD patients. Furthermore, unlike previous mouse studies, two of which focused solely on adult males [31, 33], and one briefly addressing females but lacking detailed analysis of sex differences [41]—our study systematically examined both sexes during the juvenile stage in a Shank3 deficient model, and reveal distinct patterns of sleep disruption in male and female Shank3^–/–^ rats, underscoring the role of sex in shaping neurodevelopmental outcomes. These findings emphasize the need to include both sexes in preclinical ASD research to more comprehensively reflect the heterogeneity of clinical presentations and improve translational relevance.
In exploring potential mechanisms underlying the observed sleep disturbances, we found a marked downregulation of circadian genes Clock and Bmal1 in PFC and striatum of Shank3^–/–^ rats. Both regions are recognized as key regions involved in sleep-wake regulation [53, 54]. Interestingly, the expression levels of Per and Cry—key repressors in the transcription-translation feedback loop [55]—remained unchanged, which may partly explain why gross circadian rhythmicity appeared preserved despite pronounced sleep disruption. These findings are consistent with previous reports implicating Shank3 in circadian regulation [31, 56]. As a synaptic scaffold, Shank3 may influence neural circuits that modulate clock gene expression, such as corticostriatal and suprachiasmatic pathways. The region-specific pattern observed here, including differences between the PFC and hippocampus, suggest circuit-level complexity in Shank3’s interaction with circadian machinery. These transcriptional results should, however, be interpreted with caution given the limited sample size (n = 3 per group). Follow-up studies with larger cohorts and cell-type–specific validation approaches (e.g., RNAscope or double immunofluorescence) are needed to confirm these preliminary findings and clarify whether the observed reductions reflect region- or neuron-specific modulation by Shank3. Nonetheless, the consistent direction of change suggests that impaired circadian signaling may contribute to the altered sleep homeostasis in Shank3 mutants and could represent a potential therapeutic target in ASD.
Overall, our findings carry several implications for ASD research and intervention, particularly in the context of early neurodevelopment. First, they establish Shank3 deficiency as a causal factor for early-life sleep disturbances, highlighting the synaptic function in neurodevelopmental regulation of sleep. Second, these disturbances occurred during a critical developmental window, highlighting their primary neurobiological origin rather than being secondary to chronic ASD symptoms or medication exposure. This reinforces the notion that early-life sleep problems are an intrinsic and early component of ASD pathology, especially in cases involving Shank3 deficiency. Finally, given the strong link between poor sleep and core ASD symptom severity [28–30], and sleep interventions in ASD children have shown benefits not only for sleep itself but also for daytime behavior and family well-being [57], our Shank3^Δe11–21^ rat model offers a valuable preclinical platform for evaluating sleep-targeted therapeutic strategies and dissecting how sleep disturbances influence broader neurodevelopment outcomes.
In conclusion, our study introduces a novel juvenile rat model that links a high-confidence autism gene, Shank3, to early-onset sleep disturbances. Juvenile Shank3^Δe11–21^ rats faithfully recapitulate sleep phenotypes observed in ASD children and provide a valuable bridge between clinical manifestations and underlying biological mechanisms. By shedding light on sex-specific alterations, this work opens new avenues for personalized interventions targeting sleep problems in ASD. Improving sleep in neurodevelopmental disorders is not only vital for enhancing quality of life, but may also yield cascading benefits for cognition, mood regulation, and the core behavior symptoms of autism. Collectively, our findings underscore the importance of recognizing and addressing sleep issues as an integral component of autism research and therapy, especially in the context of genetic subtypes like Shank3 deficiency.
Supplementary information
Supplementary Materials Supplementary Tables Dataset 1A Dataset 1B Dataset 2A Dataset 2B Dataset 3A Dataset 3B Dataset 4A Dataset 4B Dataset 5A Dataset 5B Dataset 6A Dataset 6B