What this is
- This study examines the relationship between vitamin A status and sleep problems, core symptoms, and clock genes in children with autism spectrum disorder (ASD).
- It includes 361 children with ASD, assessing clinical symptoms through standardized questionnaires and measuring vitamin A levels and gene expression.
- Findings suggest that lower vitamin A levels correlate with more severe sleep issues and autistic symptoms, alongside altered expression of clock genes.
Essence
- Lower vitamin A levels are linked to more severe sleep problems and autistic symptoms in children with ASD, with potential involvement of clock genes.
Key takeaways
- Children with lower vitamin A levels exhibited more severe sleep problems, particularly in bedtime resistance, and greater autistic symptom severity.
- Vitamin A levels showed weak positive correlations with the expression of RARĪ² and BMAL1, suggesting a potential link between vitamin A and clock gene regulation.
- In a mouse model, downregulation of RARĪ² was associated with reduced expression of clock genes and impaired social behavior, indicating a potential mechanism linking vitamin A to sleep and social functions.
Caveats
- The study relies on caregiver-reported questionnaires for sleep assessment, which may introduce bias and limit accuracy compared to objective measures.
- Peripheral blood measurements may not fully reflect central nervous system clock gene expression, limiting the understanding of circadian rhythm changes.
- Single time-point sampling for blood and brain tissue may not capture the dynamic nature of circadian rhythms, necessitating further longitudinal studies.
Definitions
- Vitamin A deficiency (VAD): A condition where vitamin A levels are below 0.2 mg/L, potentially impacting health and development.
- Core clock genes: Genes such as CLOCK and BMAL1 that regulate circadian rhythms and influence sleep-wake cycles.
AI simplified
Introduction
Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder characterized by impaired social communication, restricted interests, and stereotypic behaviors (1). Beyond these core symptoms, children with ASD frequently experience sleep disturbances and nutritional deficiencies, which further exacerbate societal and familial burdens.
Sleep plays critical roles in brain development, cognition, and emotion (2), governed primarily by two mechanisms: sleep homeostasis and the circadian rhythm (3). CLOCK and BMAL1 represent core clock genes whose encoded proteins form heterodimers. These bind to E-box elements within gene promoter regions to initiate transcription of downstream clock-controlled genes. Previous studies reported a comorbidity rate exceeding 50% for sleep problems in children with ASD. These disturbances not only compromised sleep integrity but also contributed to daytime sleepiness, behavioral fluctuations, and exacerbation of core symptoms. Consequently, effectively managing sleep disturbances in ASD is critical for enhancing their quality of life and mitigating core symptoms.
Vitamin A influences neuronal development through the retinoic acid signaling pathway (4, 5) and has been linked to the pathogenesis of ASD (6, 7). Previous studies indicated that nutrients play significant roles in sleep regulation (8), but the relationship between vitamin A and sleep remains unclear. Only one study compared VA levels between those with and without sleep problems, finding no significant difference (9). Several animal studies have suggested an association between VA and sleep. For example, Vitamin A deficiency (VAD) altered sleep electroencephalography (EEG) in mice, with restoration to normal patterns upon VA supplementation (VAS) (10); VAD disrupted molecular oscillations of circadian rhythm molecules in the rat hippocampus (11, 12).
These findings imply potential VA-sleep interactions. Therefore, this study aimed to: (1) investigate the association between VA, sleep problems and core symptoms in children with ASD; (2) investigate the association between VA and the expression of RARĪ² and clock genes (CLOCK and BMAL1) measured in the morning; (3) explore whether downregulation of RARĪ² signaling is associated with altered brain clock gene expression and ASD-relevant social behavior.
Materials
Study design and ethical procedures
This observational study was conducted in Chongqing, China, from November 2019 to December 2024. The study received ethical approval from the Childrenās Hospital of Chongqing Medical University Ethics Committee [approval number: 121-1/2018]. Written informed consent was obtained voluntarily from all participantsā parents.
Study participants
Children with ASD were recruited from developmental-behavioral pediatric outpatient departments and the Specialized Learning Center. Inclusion criteria: 1) age 2ā7 years, 2) diagnosis confirmed by an experienced developmental-behavioral pediatrician using Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria, and 3) signed informed consent from primary caregivers. Exclusion criteria: 1) severe physical, neurological, or psychiatric comorbidities; 2) acute/chronic infection within 3 months; 3) incomplete questionnaire data.
Clinical measures of children with ASD
Standardized questionnaires were used to collect baseline information on children with ASD, including: 1) gender and birth date, 2)individual and familial history, and 3) nutritional supplementation in the last 3 months.
Sleep disturbances were assessed using the Childrenās Sleep Habits Questionnaire (CSHQ) (13). It was completed by primary caregivers based on the childās sleep patterns during a typical week in the last month. The eight dimensions of CSHQ were bedtime resistance, sleep disordered breathing, daytime sleepiness, sleep anxiety, sleep duration, sleep onset delay, night wakings, and parasomnias. Based on a previous study (14), these subscales were categorized as medically based or behaviorally based sleep problems. According to recent studies, a total scoreā„ 48 suggests a sleep problem (15ā17).
Core symptoms of ASD were assessed by the Childhood Autism Rating Scale (CARS) (18) and the Social Responsiveness Scale (SRS) (19). The CARS scale evaluates ASD symptom severity through 15 items, each scored on a 4-point scale (1-4) (20). The SRS scale reflects the social behavior of ASD and contains 65 items, with each item scored between 0 and 3. It was divided into social awareness, social cognition, social communication, social motivation, and autistic mannerisms.
Lab measurements
Children in the study underwent fasting venous blood collection between 8:00 and 10:00 AM. Vitamin A levels were quantified using High Performance Liquid Chromatography-Tandem Mass Spectrometry (HPLC-MS/MS). Vitamin A normal (VAN): ā„ 0.3 mg/L; Marginal Vitamin A deficiency (MVAD): 0.2-0.3 mg/L; Vitamin A deficiency (VAD): < 0.2 mg/L. The blood cells were used to extract total RNA, using the Trizol method. The PrimeScript RT reagent Kit (TaKaRa) was used for reverse transcription. Real-time quantitative polymerase chain reaction was employed to assess the relative mRNA expression of CLOCK, BMAL1, and RARĪ². Primer sequences were designed using Primer 6.0 software (Premier Biosoft International) and synthesized by the Beijing Genomics Institute. Primer sequences are listed in. 1
Animals
We obtained C57BL/6 mice from Chongqing Enswell Biologicals and housed them in Specific pathogen free (SPF) grade environment. Adeno-associated virus (AAV) was injected stereotaxically into the lateral ventricle of 3-week-old mice. They were divided into a negative control group (short hairpin Negative Control RNA, sh-NC) and a down-regulated RARĪ² group (short hairpin RNA targeting RARĪ², sh-RARĪ²). All animal experiments complied with the Guidelines for the Care and Use of Laboratory Animals developed by the National Research Council and approved by the Ethics Committee for Animal Experiments of Chongqing Medical University (CHCMU-IACUC20250217007).
Behavioral tests and laboratory analyses in animals
Behavioral tests were performed when mice were reared until 7 weeks of age, including three-chamber and open-field tests. We utilized the three-chamber test (a commonly used indicator of socialization in mice) to assess perceived social novelty and social interactions in mice (13). The open-field test was used to assess exploratory behavior and repetitive tendencies of mice in a novel environment (14). Recordings were made using the ANY-Maze Animal Behavioral Video Analysis System (ANY-Maze, USA).
Mice were reared until 8 weeks of age for tissue collection (prefrontal cortex, PFC), with all sampling performed between 8:00 and 9:00 AM. Total RNA from the mouse PFC was extracted using the Bioer Simply P Total RNA Extraction Kit according to the manufacturerās protocol. RT-PCR and qPCR procedures followed the methodology described in the human cohort section, with primer sequences provided in. The total protein in PFC was extracted using a radio immunoprecipitation assay (RIPA) lysis buffer (KeyGEN Biotech) containing 0.1% protease inhibitor cocktail (KeyGEN Biotech). The protein concentrations were determined using the BCA protein assay kit (KeyGEN Biotech). Western blotting was performed to detect the protein expression levels of RARĪ² (HuaBio), CLOCK (HuaBio), and BMAL1 (HuaBio). The JASPAR database was used to predict binding sites of the transcription factor RARĪ² within the mouse Clock gene promoter region, with the two highest relative score sites selected for subsequent experiments (predicted sites detailed in). Chromatin immunoprecipitation (ChIP) assays were conducted using the ChIP kit (Abclonal), with qPCR quantification of ChIP results. Primers targeting the predicted binding sites were designed as specified in. 1 1 1
Statistical analysis
Statistical analysis and graphing were performed using R-4.4.1 and GraphPad Prism 8.0. Normality was assessed using the Kolmogorov-Smirnov (KS) and Shapiro-Wilk (SW) tests combined with graphical methods (histograms, PP plots, and QQ plots). Continuous variables were described as mean Ā± standard deviation (M Ā± SD) or median (25th to 75th percentile) [M (P25-P75)], depending on whether they conformed to normal distribution. Categorical variables were described as frequency (percentage) [n, (%)]. For between-group comparisons, two independent samples t-test was used for comparison of two groups conforming to normal distribution; Mann-Whitney U test was used for non-normally distributed data; The chi-square test was used for categorical variables. McNemarās test was used for categorical variables. Linear regression analysis was used for the association with VA, sleep scores and autistic symptoms. Spearman correlation analyse was used to assess the relationship among vitamin A, RARĪ², and clock gene expression. All statistical tests were two-sided, with P < 0.05 considered statistically significant.
Results
Study population
The study cohort included 361 children diagnosed with ASD and aged 2ā7 years. They were enrolled from 2019 to 2024 in Chongqing (Table 1).
| Variables | ASD (n=361) |
|---|---|
| Age (years), M (P25-P75) | 3.88 (3.23, 4.57) |
| Gender, n (%) | |
| Male | 295 (81.7) |
| Female | 66 (18.3) |
| Ethnicity, n(%) | |
| Han | 329 (91.1) |
| Others | 32 (8.9) |
| picky about food, n(%) | |
| Yes | 125 (35.2) |
| No | 230 (64.8) |
| Residence, n(%) | |
| City | 320 (88.6) |
| Countryside | 41 (11.4) |
| Motherās education level, n(%) | |
| Middle school or below | 62 (17.2) |
| High school | 89 (24.7) |
| College | 192 (53.1) |
| Master degree or above | 13 (3.6) |
| Missing | 5 (1.4) |
| Fatherās education level, n(%) | |
| Middle school or below | 64 (17.8) |
| High school | 82 (22.7) |
| College | 193 (53.5) |
| Master degree or above | 17 (4.7) |
| Missing | 5 (1.3) |
| Annual family income, n(%) | |
| < 60,000 | 90 (24.9) |
| 60,000 to 150,000 | 200 (55.4) |
| ā„ 150,000 | 63 (17.5) |
| Missing | 8 (2.2) |
Association between vitamin A levels, CSHQ dimension scores and ASD symptoms
After adjusting for age and gender and applying FDR correction, vitamin A levels in children with ASD were negatively correlated with bedtime resistance (Ī²=-4.188, 95%CI: -7.072, -1.305, FDR q=0.041) and social awareness (Ī²=-5.355, 95%CI: -9.248, -1.462, FDR q=0.043). Although other sleep dimensions and ASD symptoms did not reach statistical significance, they all exhibited a trend toward negative correlation (Table 2).
| Variables | Ī²(95%CI) | RawP | FDR q |
|---|---|---|---|
| Sleep scores | |||
| Bedtime resistance | -4.188 (-7.072, -1.305) | 0.005 | 0.041 |
| Sleep onset delay | -0.578 (-1.531, 0.375) | 0.234 | 0.526 |
| Sleep duration | -0.809 (-2.642, 1.023) | 0.386 | 0.579 |
| Sleep anxiety | -0.806 (-2.996, 1.384) | 0.47 | 0.604 |
| Night waking | -0.138 (-1.274, 0.997) | 0.811 | 0.811 |
| Parasomnia | -0.277 (-2.219, 1.666) | 0.78 | 0.811 |
| Sleep-disordered breathing | -0.409 (-1.303, 0.486) | 0.37 | 0.579 |
| Daytime sleepiness | -2.848 (-6.473, 0.777) | 0.123 | 0.37 |
| Total score | -9.346 (-17.115, -1.577) | 0.019 | 0.083 |
| SRS | |||
| Social awareness | -5.355 (-9.248, -1.462) | 0.007 | 0.043 |
| Social cognition | -2.936 (-8.613, 2.741) | 0.31 | 0.465 |
| Social communication | -8.527 (-19.825, 2.771) | 0.139 | 0.317 |
| Social motivation | -2.02 (-9.196, 5.156) | 0.58 | 0.58 |
| Autism behavior mannerisms | -3.131 (-11.123, 4.861) | 0.442 | 0.53 |
| SRS total score | -21.968 (-52.555, 8.619) | 0.159 | 0.317 |
| CARS total score | -1.349 (-9.334, 6.636) | 0.74 | 0.74 |
Association between vitamin A, RARĪ², and clock genes
As shown in Table 3, VA levels in children with ASD exhibited a weak positive correlation with RARĪ²(Ļ=0.112, P = 0.034) and BMAL1 (Ļ=0.165, P = 0.002). RARĪ² was also weakly correlated with CLOCK (Ļ=0.211, P<0.001) and BMAL1 (Ļ=0.299, P<0.001). Although these correlations are weak, they suggest a potential link between VA, RARĪ², and clock genes. Therefore, we further explored their association in a RARĪ² knockdown animal model.
| Variables | VA | RARĪ² | ||
|---|---|---|---|---|
| Ļ | P | Ļ | P | |
| RARĪ² | 0.112 | 0.034 | ||
| CLOCK | 0.044 | 0.408 | 0.211 | <0.001 |
| BMAL1 | 0.165 | 0.002 | 0.299 | <0.001 |
Results of behavioral tests, mRNA and protein expression levels, and CHIP-qPCR in sh-NC and sh-RARĪ² groups of mice
We conducted preliminary experiments to investigate the regulatory mechanism of the retinoic acid receptor RARĪ² on clock genes. We down-regulated RARĪ² signaling in the mouse PFC by brain stereotaxic injection of AAV. In the open-field test, no statistically significant differences were observed between the sh-RARĪ² and sh-NC groups in total distance moved, time spent in the center zone, or self-grooming duration (Figure 1A, all P > 0.05). In the three-chamber test, the sh-RARĪ² group spent significantly more time in the object zone than in the stranger mouse zone on day one. Furthermore, their time in the object zone was significantly longer than that of the control group (Figures 1B, C, all P < 0.05), suggesting impaired social interaction in the sh-RARĪ² group. These results indicate that down-regulating RARĪ² signaling in the mouse PFC may impair social interaction ability. Additionally, cortical mRNA levels of RARĪ², Clock, Bmal1, and Cry2 were significantly decreased in the sh-RARĪ² group compared to controls (all P < 0.05). Clock protein levels were also significantly reduced (P < 0.05), while Bmal1 protein expression showed a non-significant decreasing trend (P > 0.05) (Figures 1DāF).
ChIP-qPCR was performed to assess the binding of RARĪ² to the predicted binding sites within the Clock gene promoter region in the mouse PFC. The results showed that in the mouse PFC, there was a significant difference in the binding of transcription factor RARĪ² at predicted site 1 of the Clock promoter region of the target gene, with a significant decrease in the binding of the sh-RARĪ² group (P < 0.05, Figure 1G). Binding enrichment at predicted site 2 also showed a decrease in the sh-RARĪ² group, although this difference was not statistically significant (Figure 1G). These preliminary findings suggest that the transcription factor RARĪ² may regulate Clock gene expression by specifically binding to Site 1 in its promoter region. However, further functional validation is required to confirm this regulatory relationship.
Results of behavioral tests, mRNA and protein expression levels, and CHIP-qPCR in sh-NC and sh-RARĪ² groups of mice.Total distance traveled, duration in the center zone, and self-grooming duration in the open field test;Duration in the object and stranger mouse zones on the first and second days in the three-chamber test;mRNA expression levels of RARĪ²and circadian rhythm molecules in the prefrontal cortex of the two groups of mice;Protein expression levels of RARĪ² and circadian rhythm molecules in the prefrontal cortex of the two groups of mice;Chromatin immunoprecipitation quantitative PCR (ChIP-qPCR) analysis of RARĪ² binding to the Clock gene promoter *<0.05, **<0.01. (A) (B, C) (D) (E, F) (G) P P
Discussion
In this study, we found that lower vitamin A levels were associated with more severe sleep problems and autistic symptoms, as well as with altered expression of RARĪ² and clock genes in children with ASD. In the animal model, we further observed that downregulation of RARĪ² was associated with changes in brain clock gene expression and autism-like social behaviors. These findings suggest a potential link among vitamin A status, retinoic acid signaling, clock gene expression, and clinical manifestations in ASD, but the causal relationships among these factors remain to be established.
We found that children with lower VA levels exhibited more severe sleep problems, particularly in the dimensions of bedtime resistance. This finding differs from the only previous cross-sectional study examining the VA-sleep relationship in ASD children (9). That study used a CSHQ total score cutoff of 41 (sleep problem group: n=163 vs. normal sleep group: n=18), found no difference in VA levels. The inconsistency may be attributable to differences in grouping strategies and the specific sleep dimensions investigated. The prior study potentially suffered from limitations, including insufficient subscale analysis and statistical bias due to imbalanced group sample sizes. Therefore, our study incorporated a more detailed examination of CSHQ subscale dimensions. Furthermore, a large-scale Japanese study involving more than 3000 participants suggested an association between VA intake and sleep (15), and several population-based studies have reported an association between vitamin A or carotenoid (provitamin A, metabolized to retinol in vivo) levels and sleep parameters (16ā18), lending support to our findings.
Previous studies have extensively reported the regulatory role of core clock genes in sleep (19, 20). Therefore, we examined the expression levels of BMAL1 and CLOCK in both human and animal models. In human cohorts, we observed a modest time-specific association among retinoic acid signaling, social behavior, and morning clock genes expression. In a mouse model, we similarly found that downregulation of RARĪ² signaling in the PFC was associated with reduced Clock expression and impaired social ability. The relatively consistent findings from both humans and animals provide preliminary evidence for an interaction among nutrition, clock genes, and neurodevelopment.
Some animal studies have suggested the presence of retinoic acid response elements (RAREs) within the promoter regions of several circadian rhythm molecules, but direct experimental confirmation was lacking (11, 12). Our ChIP-qPCR results indicate RARĪ² occupancy at a predicted Clock regulatory region and reduced enrichment following RARĪ² knockdown, which is consistent with a potential regulatory relationship. RARĪ± and RARĪ², as the two primary subtypes of nuclear retinoic acid receptors, have similarities in physiological functions. An in vitro experiment has demonstrated that RARĪ± can directly interact with the CLOCK protein, potentially modulating circadian rhythms by influencing its heterodimerization with BMAL1 (21). BMAL1, a core clock gene, enhances the stabilization of rhythmic output from the suprachiasmatic nucleus (SCN) when its expression is increased, thereby potentially improving sleep initiation and maintenance (22).
Our study has several limitations. First, the CSHQ questionnaire introduces inherent limitations. It was difficult for us to use objective tests (sleep EEG or somatic sleep recorders) because of time, expense, and the unique nature of children with ASD. This prevented us from accurately assessing the relationship between vitamin A nutritional status and sleep. Second, peripheral blood clock gene mRNA levels may not fully reflect changes in circadian rhythm molecules within the central nervous system. Third, blood samples from patients and brain tissue from mice were both collected at a single morning time point, which precluded assessment of circadian rhythmicity. Finally, technical limitations prevented us from monitoring sleep-wake cycles or sleep phenotypes in animal models through behavioral experiments, making it impossible to verify the causal relationship between retinoic acid signaling and circadian rhythms or sleep in animals. Future studies should further explore these associations through longitudinal follow-up, multi-time point sampling, and objective sleep assessments in both humans and animal models.
Conclusion
This study is the first to explore the associations among retinoic acid signaling, sleep, and ASD, and to examine the expression of RARĪ² and specific clock genes in morning peripheral blood samples. The results showed that lower vitamin A levels were associated with more severe sleep problems, greater symptom burden, and altered clock gene expression. RARĪ² knockdown led to reduced Clock expression and impaired social ability. Although these findings cannot establish causality, they provide a preliminary framework for future longitudinal studies and further mechanistic exploration.
Acknowledgments
We sincerely thank the participating families and children. We also acknowledge the Childrenās Nutrition Research Center team for their support throughout this investigation.
Funding Statement
The author(s) declared that financial support was received for this work and/or its publication. This survey was supported by the National Natural Science Foundation of China (No.81771223, 82372559, 82304119), and Chief Medical Expert Studio of Chongqing (No.YWBF2018263).
Footnotes
Data availability statement
The original contributions presented in the study are included in the article/. Further inquiries can be directed to the corresponding author. 1
Ethics statement
The studies involving humans were approved by the Childrenās Hospital of Chongqing Medical University Ethics Committee [approval number: 121-1/2018]. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participantsā legal guardians/next of kin. The animal study was approved by the National Research Council and approved by the Ethics Committee for Animal Experiments of Chongqing Medical University (CHCMU-IACUC20250217007). The study was conducted in accordance with the local legislation and institutional requirements.
Author contributions
XX: Writing ā review & editing, Writing ā original draft, Data curation. HC: Investigation, Writing ā original draft. BY: Investigation, Writing ā original draft. QW: Writing ā original draft, Investigation. BH: Writing ā original draft, Investigation. DZ: Investigation, Writing ā original draft. DA: Investigation, Writing ā original draft. TY: Writing ā original draft, Supervision. JC: Writing ā original draft, Supervision. TL: Supervision, Writing ā original draft. YD: Resources, Writing ā review & editing, Conceptualization.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisherās note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyt.2026.1805599/full#supplementary-materialā
References
Associated Data
Supplementary Materials
Data Availability Statement
The original contributions presented in the study are included in the article/. Further inquiries can be directed to the corresponding author. 1