The current study is the first to investigate both the levels of neurotransmitters and the mRNA levels of clock genes, and substantiate associations between clock genes, neurotransmitters, and sleep disturbance in preschool children with ASD. There were four crucial findings: (1) Children with ASD had a higher prevalence of sleep disturbances; (2) lower mRNA levels of clock genes and abnormal levels of orexin and leptin were detected in children with ASD; (3) CLOCK mRNA was a strong predictor of sleep disturbance score; and (4) dysregulation of clock genes may lead to abnormal levels of neurotransmitters and cause sleep disturbance.
4.1. Sleep disturbance
The present study used subjective and objective measures to evaluate sleep disturbance. The CSHQ results indicated that the prevalence of sleep disturbance in ASD children was 88.64%, which was considerably higher than that in TD children (63.64%). There were marked differences in scores for bedtime resistance, sleep duration, sleep anxiety, parasomnias, and sleep disordered breathing, which is concordant with results reported by Chen et al. (2021). In that study the prevalence of sleep disturbance in ASD participants was 67.4%, and in TD participants it was 51.0%. The common sleep problems in ASD children were bedtime resistance, sleep anxiety, sleep onset delay, and daytime sleepiness. However, the prevalence of sleep disturbance in ASD differed slightly from that reported in other countries (Inthikoot & Chonchaiya, 2021; Malhi et al., 2019; Sultana et al., 2021), and the difference may be related to racial variation, sleeping habits, ASD heterogeneity, and ages (Takahashi et al., 2018; Wang et al., 2016).
The current study utilized Apple Watches to accurately assess sleep. Compared with TD children, ASD children had shorter total sleep time, lower sleep efficiency, longer awake time at night, and longer total awake time, which is consistent with results reported by Mughal et al. (2020). There were inconsistencies in sleep onset delay when comparing CSHQ results with Apple Watch results. ASD children had higher scores for sleep onset delay in the CSHQ than TD children. There was no significant difference in sleep-onset latency however, which may be related to differences in subjective and objective measures, and different emphases. We should combine the results of the two methods to ensure a comprehensive evaluation that reflects the real situation in ASD children, given that the CSHQ can reflect sleep disordered breathing problems and the Apple Watch can quantify them.
4.2. mRNA levels of clock genes
Previous studies have mostly focused on clock gene mutation when investigating mechanisms of sleep disturbance in ASD (Hoang et al., 2021; Yang et al., 2016), without investigating clock gene transcription and translation levels. We investigated whether mRNA levels of clock genes were changed, and relationships between these mRNA levels and sleep disturbance. mRNA levels of the clock genes CLOCK, BMAL1, CRY1, CRY2, PER2, TIMELESS, and RORA were significantly lower in ASD children. CLOCK mRNA (β = -0.514) was a predictor of CSHQ score in a regression model. Previous studies have indicated that the mutation of clock genes in ASD patients is associated with sleep disturbance. Yang et al. (2016) reported that mutations in PER2, PER3, TIMELESS, and CLOCK were only detected in ASD patients with sleep disturbances, and suggested there was a probable link between clock genes and sleep disturbances in ASD. In the current study lower mRNA levels of clock genes were detected in ASD children who had significant sleep disturbances, indicating that clock genes play a critical role in sleep disturbance. The circadian system is regulated by the rhythmic expression of clock genes, thus abnormal expression of clock genes will lead to circadian system dysfunction and sleep disturbance (Charrier et al., 2017). In the present study lower mRNA levels of these clock genes suggested abnormal rhythmic expression of clock genes, which may lead to sleep disturbance in ASD.
In correlational analysis mRNA levels of clock genes were significantly associated with multiple sleep problems including night waking, low sleep efficiency, and poor sleep quality. mRNA levels of CLOCK, CRY2, and PER2 were negatively correlated with total CSHQ score. There was also a significant positive correlation between PER2 mRNA and sleep efficiency. Although no children with ASD exhibited symptoms of either disorder in the present study, the children did exhibit many sleep problems and significant sleep disturbances, which were related to abnormal expression of clock genes. Specific clock genes may be targeted in future interventions designed to treat specific sleep problems.
4.3. Levels of neurotransmitters
Inconsistent results have been reported with respect associations between sleep disturbances and levels of melatonin and cortisol in ASD (Goldman et al., 2017; Tomarken et al., 2015). Associations between sleep disturbances and levels of other neurotransmitters including acetylcholine, dopamine, S100B, orexin, and leptin are yet to be investigated. Therefore, the current study investigated the levels of these neurotransmitters in ASD children.
In the present study orexin levels in ASD children were significantly lower than those in TD children. Orexin is a wake-promoting neuropeptide that plays a vital role in maintaining the sleep-wake cycle (Anaclet et al., 2009; Zhu et al., 2020). Messina et al. (2018) speculated that dysregulation of orexinergic neurotransmission was associated with sleep disturbance in ASD, and that orexin may be a useful biomarker in ASD. However, sleep disturbances in children with ASD and its correlation with orexin were not investigated in that study. In the current study ASD children had lower orexin levels as well as low sleep efficiency as determined via the Apple Watch, indicating that orexin is potentially involved in a mechanism of sleep disturbance. Leptin levels were also higher in ASD children. Leptin is a peptide encoded by an obesity gene that participates in the sleep-wake cycle. Serum leptin levels reportedly increased in sleep-deprived mice, indicating that leptin may play a role in sleep (Wu et al., 2014). Leptin may lead to sleep disturbance by binding with its receptor and acting on orexin, such that concentrations of leptin and orexin are inversely regulated (López et al., 2000). The low levels of orexin and high levels of leptin observed in ASD children in the present study were consistent with this type of relationship. ASD children also had higher disordered breathing during sleep scores in the CSHQ, which manifests as snoring, choking, apnea, and other problems during sleep. Li and He (2021) reported that as a neuroregulatory factor leptin could affect respiration in the respiratory center. High leptin levels were significantly positively correlated with the Apnea-Hypopnea Index, suggesting that leptin may be involved in a mechanism that results in sleep breathing problems in ASD children.
Levels of other sleep-related neurotransmitters, including melatonin, cortisol, acetylcholine, dopamine, and S100B did not differ significantly between ASD and TD children. In previous studies levels of these neurotransmitters were heterogeneous in ASD. Goldman et al. (2017) reported that there were no significant differences in salivary cortisol levels (morning and evening) or melatonin levels between ASD and TD children, but Tomarken et al. (2015) reported that in ASD children salivary cortisol levels at night differed from those in TD children. There are inconsistencies in reported results pertaining to melatonin, cortisol, acetylcholine, dopamine, and S100B in ASD, which may be related to sample processing methods, detection methods, collection times, and sample heterogeneity (Zheng et al., 2021). The concept of co-neurotransmitters was proposed by Oh et al. (2019), who noted that neurotransmitters are not simply responsible for unidirectional adjustment, but adjust each other according to circadian rhythms and environmental factors. We hypothesized that sleep disturbance may involve a combination of these neurotransmitters, rather than a single substance. More effective detection methods that can accurately reflect neurotransmitter levels should be considered in the future. The regulatory mechanisms involving neurotransmitters in sleep are still being investigated.
The current study had some limitations. Due to substantial heterogeneity within the study population the results are not directly generalizable to other countries. Large population studies and multi-center studies need to be conducted. Serum neurotransmitter levels and clock gene mRNA levels were only detected at a single timepoint in children with ASD and TD, and evaluation of dynamic changes in these parameters needs to be conducted in future studies. Lastly, it was a cross-sectional study and could not clarify causal relationships between sleep disturbance, neurotransmitters, and clock genes.