Sleep disorders have become a global public health issue, affecting approximately 15–30% of adults and causing significant burdens on quality of life, as well as occupational, psychological, and economic well-being (Ohayon, 2002; Morin et al., 2006). In modern society, owing to the negative effects of modern work patterns and screen time, the prevalence of sleep disorders and circadian rhythm disorders is growing (Thomée et al., 2011; Jehan et al., 2017; Murdock et al., 2017; Harknett et al., 2021; Di et al., 2022). Moreover, recent studies have indicated that insufficient sleep and sleep disturbances are correlated with countless adverse outcomes (Itani et al., 2017; Jike et al., 2018; Sun et al., 2020; Winer et al., 2021; Sun et al., 2022; Han et al., 2023). However, the molecular mechanisms underlying sleep–wake cycles are unclear. Early research focused on the central nervous system’s role in sleep regulation and dysregulation (Raven et al., 2018; Van Someren, 2021; Sulaman et al., 2023). However, sleep is not only regulated by the central system but also affected by signals from peripheral tissues. Recently, researchers have focused on specific interactions between circadian rhythm processes and the gut microbiome.

The gut microbiome is a highly complex microbial community that may directly or indirectly participate in the regulation of the sleep–wake cycle through the microbiome–gut–brain axis (Neuman et al., 2015; Sen et al., 2021; Wang et al., 2022). Moreover, dietary composition, rhythms of feeding, and loss of the microbiome influence the composition of the gut cycling transcriptome and the expression of rhythm genes, independently and together (Leone et al., 2015; Zhang et al., 2023). For example, intraperitoneal injection of components of bacterial cell walls or bacteria-derived metabolites, such as lipopolysaccharides, lipoteichoic acid, and butyrate, were found to increase non-REM sleep in mice (Szentirmai and Krueger, 2014; Szentirmai et al., 2019, 2021). Some randomized controlled trials have suggested that Bifidobacteria and Lactobacillaceae may help enhance sleep quality, especially sleep induction, and relieve subclinical signs of anxiety and depression (Nishida et al., 2019; Ho et al., 2021; Lee et al., 2021). Additionally, sleep disturbances are associated with the disruption of gut bacteria, which results in a dysfunctional colonic barrier and the development of intestinal illnesses (Benedict et al., 2016; Gao et al., 2023). However, owing to the existence of confounding factors, such as lifestyle, diet, and age (Rinninella et al., 2019), and the limitations of experimental ethics, it is difficult to carry out randomized controlled trials (RCT) to uncover the causal association between gut microbiota and sleep. Moreover, previous observational studies were not robust because they contained small numbers of participants and the direction of the effects was difficult to judge. As a result, it is unclear whether gut microbiota and sleep disturbances are causally related.

In genetic epidemiology, Mendelian randomization is a method that involves using genetic variants to compose instrumental variables (IVs) of traits to investigate the causal relationship between traits and outcomes. Since genetic variants are generally randomly assigned at meiosis and are not affected by disease states, MR analysis can minimize common confounding factors, avoid confounding factors measurement error, and overcome reverse causation (Smith and Ebrahim, 2003; Davey Smith and Hemani, 2014).

We conducted a two-sample, bidirectional MR study using single nucleotide polymorphisms (SNPs) from the most recent genome-wide association studies (GWASs) to identify whether gut microbiota affect sleep disturbances and whether such associations are directional.

To the best of our knowledge, this is the first bidirectional MR analysis that comprehensively clarifies causal relationships between gut microbiota and sleep-related traits. As shown in Figures 2, 3, our findings indicate that a total of 68 bacterial taxa are causally associated with seven sleep-related traits. Furthermore, 13 bacterial taxa related to sleep-related features in forward MR analysis were regulated by sleep-related traits, including Alistipes, Alloprevotella, Anaerofilum, Bacteroides, Catenibacterium, Coprococcus 3, Oxalobacter, Parabacteroides, Ruminiclostridium 6, Ruminococcaceae NK4A214 group, Ruminococcaceae UCG-013, Ruminococcus gnavus group, and Slackia. Nevertheless, the potential causal effects of Coprococcus 3, Oxalobacter, and Slackia on daytime sleepiness, and the potential causal effect of Parabacteroides and Bacteroides on chronotype in forward MR analysis were not supported by the results of reverse MR analysis. However, these findings did not exclude the possibility that the effects are interactive. Moreover, owing to potential pleiotropy, some causal effects (Eisenbergiella on daytime napping, daytime napping on Fusicatenibacter, sleep duration on Lactobacillus, and binary chronotype on Alistipes) were not credible. After excluding the above uncertain causal effects, a total of 40 bacterial taxa had potential causal effects on seven sleep-related traits, which in the other direction may be related to 34 bacterial taxa, and most of the bacterial taxa were of Bacillota (Figure 4).

Growing evidence from observational studies indicates that gut microbiota is correlated with sleep-related traits and disorders, and that the absence of gut microbes and their metabolites may alter sleep traits and architecture (Leone et al., 2015; Paulsen et al., 2017; Szentirmai et al., 2019). Szentirmai et al. found that the intestinal microbiome induces non-REM sleep through butyrate-sensitive mechanisms (Szentirmai et al., 2019). Cross-feeding is the central metabolic mechanism of the gut microbiota. Faecalibacterium can produce butyrate from acetate and lactate, which are produced by Bifidobacteria from fermented carbohydrates. Moreover, the coculture of Eubacterium hallii group with Bifidobacterium promotes the accumulation of butyrate through cross-feeding (Belenguer et al., 2006). Butyricimonas, Marvinbryantia, Holdemanella, Intestinibacter, Ruminococcaceae NK4A214 group, Clostridium sensu stricto 1, and Oscillibacter have also been associated with the production of butyrate, while Eggerthella is involved in the depletion of butyrate (Chen et al., 2021; Liu et al., 2022). Consistent with these previous studies, our study found that these butyrate-producing bacteria (Bifidobacterium, Clostridium sensu stricto 1, Eubacterium hallii group, Holdemanella, Intestinibacter, Marvinbryantia, Oscillibacter, and Ruminococcaceae NK4A214 group) all had a nominally significant causal association with sleep (Figure 5). This finding suggests that these intestinal microbiotas associated with butyrate metabolism are involved in sleep-related regulatory mechanisms (Figure 6). Lipopolysaccharides and peptidoglycans, which are components of bacterial cell walls, are released during the decomposition or division of bacteria and then produce an inflammatory response by activating the expression of proinflammatory factors (Motta et al., 2015; Krueger and Opp, 2016). Previous research has confirmed that these inflammatory responses are related to the sleep that occurs during bacterial infections. However, the butyrate produced by intestinal flora has a strong anti-inflammatory effect and can inhibit inflammation in the colon and liver as well as the expression of inflammatory factors induced by lipopolysaccharides and NF-κB activation (Perez et al., 1998; Thangaraju et al., 2009; D'Souza et al., 2017).

Metabolites of gut microbes such as short-chain fatty acids, butyrate and, acetate may also regulate circadian rhythms by influencing the expression of circadian clock genes (Figure 6). Wang et al. have demonstrated that the gut microbiome modulates the expression of the circadian transcription factor NFIL 3 (Wang et al., 2017). Furthermore, the absence of gut microbiota and gut microbial metabolites such as butyrate and acetic can led to significant differential expression of hepatic and central circadian clock genes regardless of dietary changes (Leone et al., 2015). An observational study determined that short-chain fatty acids, propionate, and butyrate in feces were associated with nighttime sleep duration in infants (Heath et al., 2020). Another observational study also found that changes in abundance of Lachnospira, Bacteroides, Faecalibacterium, and Blautia were significantly associated with sleep quality and disorders (Li et al., 2020), which correspond to our findings. Yu et al. found that after giving mice oral high doses of gamma-aminobutyric acid fermented milk, the relative abundance of Ruminococcus, Allobaculum, and Adlercreutzia, and the levels of short-chain fatty acids increased significantly and sleep time was significantly prolonged (Yu et al., 2020). This result suggests that diet may affect sleep by regulating the intestinal microbiota. These previous studies and our research show that the gut microbiota and its metabolites can participate in the regulation of sleep.

The relative abundance of gut microbiota is unique between individuals, and under healthy conditions, the gut microbiota displays resilience and stability. However, the “healthy” microbiome can be disrupted by changes in age, disease and environmental factors (Hou et al., 2022). Sleep disturbances or circadian rhythm disturbances have also been reported to disrupt the balance of the gut microbiota. Circadian rhythms are essential for maintaining normal physiological functions of the gastrointestinal tract, and circadian rhythm disorders are closely related to certain diseases of the digestive system (Schernhammer et al., 2003; Hoogerwerf et al., 2007; Nojkov et al., 2010). In vitro experiments by Summa et al. proved that circadian rhythm disturbance and sleep fragmentation lead to the destruction of the integrity of intestinal barrier function, which in turn increased intestinal permeability (Summa et al., 2013). This increased intestinal permeability may lead to translocations of gut microbiota and its metabolites, which alters the variety and abundance of gut microbiota. Poroyko et al. also demonstrated through in vitro experiments that sleep fragmentation can induce selective changes in intestinal flora, such as reducing the abundance of Bacteroidetes, Actinobacteria, and Bifidobacteriaceae (Poroyko et al., 2016). However, the disruption of the integrity of the intestinal barrier caused by sleep disorder may be related to its suppression of melatonin levels (Gao et al., 2019). Gao et al. found that the decrease in the abundance of Faecalibacterium, Bacteroides, and Akkermansia caused by sleep deprivation was associated with decreased levels of melatonin (Gao et al., 2019). Consistent with these conclusions, we also found nominally causal effects of sleep disorder or sleep fragmentation on these bacterial taxa (Figure 6).

The previous studies mentioned above demonstrate that gut microbiota are involved in circadian rhythm regulation and show that circadian rhythm disturbances may cause changes in gut microbiota abundance by damaging the intestinal barrier or inducing changes in melatonin. This is consistent with the results from 13 bacterial taxa found in our study that were both involved in and subject to sleep regulation, which suggests that regulation between gut microbiota and sleep may be bidirectional.

A major advantage of this study was that MR analysis effectively excluded the interference of reverse causation and possible confounding factors in inferring causal effects between gut microbiota and sleep-related traits. The SNPs of the intestinal microbiota came from the largest GWAS meta-analysis available, and the sample sizes were large enough to ensure the strength of the IVs and the robustness of the results. The design of the two-sample MR further avoided bias resulting from overlapping data on exposure and outcome pools. Utilization of various statistical models (such as IVW, weighted median, and maximum likelihood method) as well as sensitivity analyses (such as MR-PRESSO and MR-Egger regression intercept term tests) ensured the confidence of causal effect estimates.

However, our study also had some limitations. First, genus level was the lowest classification level in the data for gut microbiota, which limited the ability to uncover causal relationships between gut microbes and sleep at the species level. Second, the SNPs we utilized in the forward MR analysis did not meet the conventional GWAS threshold (p < 5 × 10⁻⁸), but were rescued by Bonferroni correction, which ruled out false positive results to the greatest extent. Third, the participants of gut microbiota and sleep in the GWAS meta-analysis were primarily of European ancestry. The same genetic variant may have different pleiotropic effects in different ethnic populations; therefore, the inference of causal effects derived in our study may not be applicable in non-European populations. Fourth, in the reverse MR analysis, estimates of effects may have been biased by weak IVs because of the small sample size of the GWAS meta-analysis for sleep-related traits. Finally, we applied a number of exclusion criteria to select IVs; however, many internal and external factors affect gut microbes and sleep. Thus, bias resulting from SNPs being associated with potential risk factors cannot be completely ruled out.

In conclusion, this study represents the first bidirectional MR analysis to systematically reveal the causal association between gut microbiota and sleep. Our findings suggest the possible causal effect of 42 bacterial genera on sleep-related traits. Conversely, sleep-related traits may also be involved in the regulation of the abundance of 39 bacterial genera. In addition, 13 of these bacterial genera overlapped, which provides suggestive evidence for a reciprocal role between gut microbiota and sleep. The demonstration of a causal relationship between sleep and gut microbiota provides support for techniques to modify sleep by manipulating the gut microbiome. However, the basic mechanism of gut microbiota on sleep is still unknown, and more research is needed to provide theoretical support for targeted intervention in sleep by regulating specific gut microbiota.

Publicly available datasets were analyzed in this study. This data can be found at: Summary statistics for the sleep traits are available at: (http://sleepdisordergenetics.org/). Summary statistics for the gut microbiota are available at: (https://mibiogen.gcc.rug.nl/).

DO and JW designed the research. JW, BZ, and SZ collected and analyzed the data and drafted the manuscript. DO, XZ, and XL supervised the study. JW, ZH, and YX were involved in writing the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the National Natural Science Foundation of China (NO. 82271336) and Taizhou “500 Elite Plan” high-level talent project.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

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.