Rhythmic TDP-43 affects RNA splicing of USP13, resulting in alteration of BMAL1 ubiquitination

As an internal rhythm in response to changes in the light–dark cycle from the external environment, the circadian clock is driven by interactions between internal biological clocks and external environmental signals such as light, temperature, and food intake (Patke et al., 2020). It affects most physiological processes including sleep, alertness, and cognitive performance. The core circadian clock in humans and mice contains a set of conserved clock proteins that form a transcriptional–translation feedback loop (TTFL) that mediates diurnal oscillations in gene expression. Brain and muscle Arnt-like protein-1 (BMAL1) forms a complex with circadian locomotor output cycle kaput (CLOCK) and binds to the E-box element in the promoters of period circadian protein (PER) and cryptochrome (CRY) genes. This leads to the rhythmic expression of PER, CRY, nuclear receptor subfamily 1 group D member 1 (NR1D1), and retinoic acid receptor–associated orphan receptor α (RORα) (Patke et al., 2020; Peng et al., 2022b; Richards and Gumz, 2013; Shrinivas et al., 2019).

Circadian rhythm disruption (CRD) is commonly observed in elderly individuals, particularly in those with neurodegenerative diseases, including Alzheimer’s disease (AD) and Parkinson’s disease (PD) (Musiek and Holtzman, 2016; Nassan and Videnovic, 2022; Ruan et al., 2021). Disruptions in circadian rhythms could be an initial sign of neurodegenerative disorders and increase the likelihood of developing neurodegeneration in individuals over the age of 60 who are otherwise healthy (Musiek and Holtzman, 2016). FUS, a heterogeneous nuclear ribonucleoprotein (hnRNP) closely linked to the onset of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), has the potential to regulate circadian rhythms (Jiang et al., 2018). It is transcriptionally regulated by NR1D1 and regulates PER and CRY expression (Jiang et al., 2018). The transactive response DNA-binding protein of 43 kDa (TDP-43) encoded by the TARDBP gene is another well-known ubiquitous conserved hnRNP associated with ALS and FTD. It plays important roles in transcription and RNA processing (Corbet et al., 2021; Flores et al., 2019; Polymenidou et al., 2011; Wang et al., 2018). The aggregation of hyperphosphorylated, ubiquitinated, and truncated TDP-43 in neurons and glial cells is a common pathological feature of various neurodegenerative diseases including ALS, FTD, and AD (Chen and Mitchell, 2021; Josephs et al., 2014; Neumann et al., 2006; Prater et al., 2022). Cytoplasmic aggregation of TDP-43 leads to its nuclear depletion (McGoldrick and Robertson, 2023). Although TDP-43 stabilized the blue light receptor CRY1 by regulating the ubiquitin–proteasome system (Hirano et al., 2016), there is still scarce evidence supporting that TDP-43 controlling circadian rhythm function is involved in regulating the development of neurodegenerative diseases.

The utilization of glucose and energy produced in organisms exhibits a diurnal rhythm (Peng et al., 2022a; Womac et al., 2009). The hypothalamic suprachiasmatic nucleus (SCN) regulates the circadian oscillations in endogenous glucose production and extracellular adenosine triphosphate (ATP) accumulation (Marpegan et al., 2011; Peng et al., 2022a; Schmitt et al., 2018; Womac et al., 2009). AMP-activated protein kinase (AMPK) is the central regulator of energy sensing. Its expression and cellular localization also exhibit a rhythmic pattern (Lamia et al., 2009). Platelet-type phosphofructokinase (PFKP) is an enzyme that regulates the rate-limiting step of glycolysis and promotes fructose-6-phosphate (F6P) and ATP production, forming fructose-1,6-bisphosphate (FBP). PFKP regulates glucose oxidation by interacting with AMPK, affecting cellular energy and redox homeostasis and promoting cell survival under glucose starvation (Chen et al., 2022).

In this study, we found that the TDP-43 protein was expressed rhythmically in vivo and in cultured cells. Knockdown of TDP-43 changed the expression of rhythm genes, including BMAL1, CLOCK, CRY1, and PER2, the circadian rhythm function, and the cognitive and balance abilities of the mice. We found that TDP-43 regulated the deubiquitination of BMAL1 by affecting the variable splicing of ubiquitin-specific peptidase 13 (USP13) and expression of functional USP13, which impairs BMAL1 stability. Meanwhile, we found that the loss of the Bmal1 gene in mice did not affect the rhythmic expression of TDP-43 in multiple tissues. This study has increased our understanding of the role of TDP-43 in circadian rhythm maintenance.

The intracellular loss and extracellular aggregation of TDP-43 are pathological characteristics of neurodegenerative diseases, including ALS, FTD, and AD (Josephs et al., 2014; Neumann et al., 2006). This may stabilize CRY1 in an F-Box and leucine-rich repeat protein (FBXL3)-dependent manner to maintain the circadian cycle (Hirano et al., 2016). However, whether this protein is regulated by circadian rhythms remains unclear.

We investigated the expression pattern of TDP-43 in mice brains over 24 h. C57BL/6 mice were photoperiodically synchronized for 4 wk before being sacrificed at four zeitgeist time (ZT) points, ZT0 (6 am), ZT6 (12 am), ZT12 (6 pm), and ZT18 (12 pm) (Fig. 1 A). Then levels of TDP-43 and BMAL1 in the cortex, hippocampus, and liver were analyzed by western blotting (Fig. 1 B). The results showed that the levels of TDP-43 were different at varies time points and displayed obvious circadian oscillation, which coincided with the level of BMAL1, a core circadian rhythmic protein, in the cortex, hippocampus, and liver of the mice (Fig. 1 B). Furthermore, immunofluorescence staining revealed a clear circadian rhythm in TDP-43 expression in the SCN, reaching its peak at ZT 20 (Fig. 1 F; and Fig. S1, A and B). These results strongly suggest that TDP-43 might be a novel rhythmic protein.

Sleep deprivation (SD) affects the expression of multiple core circadian rhythm genes. To investigate whether the rhythmic oscillation of TDP-43 was affected by sleep, similar to other rhythmic proteins, the mice were photoperiodically synchronized for 4 wk. They were then deprived of sleep for 6 h starting at ZT0, ZT6, ZT12, and ZT18, followed by sacrifice together with the non-sleep-deprived controls (Fig. S1 C). Immunoblots showed that the basal level of TDP-43 and the amplitude of oscillation were decreased in SD mice, with a trend similar to that of BMAL1 (Fig. S1 D). Using RT-qPCR, we found that the basal mRNA of TARDBP and the amplitude of oscillation were also decreased in the hypothalamus of SD mice (Fig. S1 E). These results suggest that the expression of TARDBP mRNA and the encoded TDP-43 are regulated by circadian rhythms.

We further investigated whether TDP-43 affects the expression of circadian rhythm-related genes. Human neuroblastoma M17 cells were transfected with control vectors (NC) and gene-editing vectors targeting the TARDBP gene (TKD) followed by 2 h of serum-shock synchronization. Cells from each treatment group were collected at CT12 before being used for bulk RNA sequencing. Compared with the NC group, there were 1,005 upregulated and 945 downregulated genes in the TKD group (Fig. 2 A). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment pathway analysis showed that in addition to enrichment of ALS-, AD-, and PD-related functional genes in the TKD group (Fig. 2 A), the genes related to ubiquitin-mediated protein degradation, AMPK signaling, and circadian activity (Fig. 2 A) were abundantly enriched in the TKD group, suggesting that TDP-43 may be involved in the regulation of ubiquitin-proteosome degradation of proteins, energy homeostasis, and circadian activity.

Based on our RNA sequencing (RNA-seq) data uploaded to the gene expression omnibus (GEO) database (GSE260737), TDP-43 knockdown mainly affected the abundance of sequencing reads assigned to ARNTL/BMAL1, CLOCK, CRY, and PER in M17 cells (Fig. 2 B). We further verified that TDP-43 knockdown led to an increase in the mRNA levels of BMAL1, CLOCK, and NR1D1, meanwhile, those of PER2 and CRY1 decreased in M17 cells at CT12 (Fig. 2 C). Therefore, TDP-43 knockdown likely affected the expression of multiple core circadian genes. In primary mouse brain fibroblasts (MBFs), TDP-43 knockdown significantly decreased the protein levels of BMAL1 at CT8, CT12, and CT16, CLOCK at CT16, CRY1 at CT12, CT16, CT20, and CT24, and PER2 at CT16 (Fig. 2 D). TDP-43 knockdown also significantly altered the periodic expression of BMAL1, CLOCK, PER2, and CRY1 (Fig. 2 D). Similar results were observed in HEK-293T cells depleted of TDP-43 (Fig. S2 A). Immunofluorescence staining showed that at the lowest point of circadian amplitude (CT0) and the highest point (CT12), M17 cells transfected with TARDBP gene-editing vectors (TKD) had significantly lower expression levels of PER2, CLOCK, BMAL1, and CRY1 than the NC group (Fig. 2 E). The ectopic expression of human TDP-43 in synchronized TDP-43-knockdown HEK-293T cells restored the expression of PER2, CLOCK, and BMAL1 (Fig. S2 B). These results further support that TDP-43 regulates the expression of rhythm proteins.

BMAL1 is generally considered as the key gene of the molecular clock because the knockdown of BMAL1 downregulates the expression of other rhythm genes (Bolsius et al., 2021). Consistently, we found that the knockdown of BMAL1 significantly decreased the expression of PER2, CLOCK, and CRY1, but did not affect the expression of TDP-43 at both mRNA and protein levels (Fig. S2 C), suggesting that TDP-43 may act upstream of BMAL1.

SCN is the principal circadian clock in the brain. Therefore, we investigated whether TDP-43 knockdown affects the expression of core rhythmic genes in the SCN of mice. Control and TDP-43 knockdown adenoassociated virus (AAV) were intracerebroventricularly (icv) injected bilaterally into 8-wk-old C57BL/6 mice. 3 wk after injection, the mice were subjected to a 12/12 h light/dark (LD) cycle for 1 wk and sacrificed at ZT0, ZT5, ZT10, ZT15, and ZT20 (Fig. 3 A).

We first investigated whether TDP-43 knockdown affected circadian rhythm behavior and cognition in vivo. We evaluated the effect of TDP-43 knockdown on wheel-running activity, which correlated with circadian rhythm in mice. After familiarization with the running wheels for 3 wk in 12/12 h light–dark (LD), followed by 3 wk of complete dark–dark (DD) exposure, the mice in both control (CON) and TDP-43 knockdown (shTDP-43) groups retained free-running rhythms in DD (Fig. 3 A). The circadian periods, determined by analyzing the recorded data of animal wheel activity during the dark–dark cycle, in shTDP-43 infected (shTDP-43) mice were significantly shorter than those in the control virus-infected (CON) mice (Fig. 3, A and B). In the open field test, there was no difference between the control and TDP-43 knockdown mice in the distance and duration of movement (Fig. 3 C). This indicated that TDP-43 knockdown did not alter the spontaneous activity of animals and the observed differences in circadian behavior between the control and TDP-43 knockdown mice were not because of the changes in the locomotor activity from TDP-43 knockdown. These findings suggest that a reduction in TDP-43 alters the circadian rhythm in mice.

Motor learning and coordination of mice were assessed using an accelerating rotarod apparatus in which the latency and rate of fall were decreased in shTDP-43 virus-infected mice (Fig. 3 D). This indicated that TDP-43 knockdown might affect motor coordination ability. During the fourth training day in the Morris water maze task, shTDP-43 virus-infected mice took more time to locate the escape platform (Fig. 3 E). This indicated that the spatial learning and memory of the mice may be impaired. Western blotting showed that TDP-43 knockdown by shRNA downregulated the protein expression of BMAL1, PER2, CLOCK, and CRY1 (Fig. 3 F). BMAL1 protein expression was significantly decreased by TDP-43 knockdown at ZT0, ZT15, and ZT20, CLOCK at ZT15 and ZT20, CRY1 at ZT0 and ZT5, and PER2 at ZT0, ZT5, and ZT20 (Fig. 3 G). The effect of TDP-43 knockdown on the periodic expression patterns of circadian proteins (Fig. 3 F) was consistent with observations in the in vitro cellular model (Fig. 2 D). The changes in the mRNA expression patterns of circadian genes caused by TDP-43 knockdown in mouse brains (Fig. 3 H) were similar to those observed in an in vitro cellular model (Fig. 2 C).

Bulk RNA sequencing showed that genes related to ubiquitin-mediated proteolysis were enriched in the TKD group (Fig. 2 A). Moreover, TDP-43 knockdown suppressed BMAL1 protein expression and promoted its mRNA expression, leading us to first evaluate whether TDP-43 regulated the stability of BMAL1 and other circadian rhythm proteins. Using cycloheximide (CHX) to inhibit protein synthesis in M17 cells showed that the degradation of BMAL1, CLOCK and CRY1, but not that of PER2, was significantly lower in the TKD group than in the control group (Fig. 4, A and B). MG132, a ubiquitin–proteasome specific inhibitor, considerably blocked the degradation of BMAL1 in M17 cells (Fig. 4 C). The ubiquitination of BMAL1 was increased by TDP-43 knockdown (Fig. 4 C), directly supporting the hypothesis that the knockdown of TDP-43 affected the ubiquitination of BMAL1.

We also performed immunoprecipitation with cortex homogenates from shTDP-43 virus-infected mice collected at several zeitgeber times (ZT0, ZT5, ZT10, ZT15, and ZT20) using an anti-BMAL1 antibody and performed to immunoblot analysis (Fig. 4 D). Compared with the control group, we observed an increase in the ubiquitination level of BMAL1 in the TDP-43 knockdown group. Our findings suggest that TDP-43 regulates the ubiquitination of BMAL1 by modulating its protein stability (Fig. 4 D).

Then we wanted to know by which TDP-43 regulates the ubiquitination of BMAL1. The RNA-seq results (Roczniak-Ferguson and Ferguson, 2019) showed that the deubiquitinated ubiquitin-specific peptidase 13 (USP13) was altered by TDP-43 knockdown (Fig. 5 A). We confirmed that USP13 mRNA expression was decreased by TDP-43 knockdown and could be restored by the ectopic expression of wild-type TDP-43 in TDP-43 knockdown M17 cells (Fig. 5 B). USP13 was expressed periodically, and the periodicity disappeared after TDP-43 knockdown (Fig. 5 C).

We previously reported that TDP-43 regulates the alternative splicing of tau exon 10 and promotes tau instability (Gu et al., 2017a, 2017b). TDP-43 contains tandem RNA recognition motifs and acts as an alternative splicing regulator by binding to long UG-rich RNA sequences to modulate gene splicing (Lukavsky et al., 2013). Intron 1 of USP13 contains UG-rich RNA sequence (Fig. 5 D). Analysis of the differences in the sequence readers of USP13 splicing revealed a moderate difference between the vector control and TDP-43 knockdown groups. We found one UG-rich RNA sequence localized in intron 1 of the USP13 premature mRNA (Fig. 5 D). Therefore, TDP-43 may inhibit UG-rich RNA splicing. Using RT-PCR, we confirmed that TDP-43 knockdown led to more intron sequences being spliced into USP13 mRNA (Fig. 5 D). This could potentially elucidate the reason for the diminished normal mRNA formation of USP13 upon TDP-43 knockdown, leading to altered USP13 expression. Reducing the expression of USP13 in HEK-293T cells led to a significant decrease in BMAL1 level (Fig. 5 E). Meanwhile, introducing an increasing amount of USP13 into HEK-293T cells resulted in a dose-dependent increase in BMAL1 protein level (Fig. 5 F). Thus, we recognized that TDP-43 knockdown decreased the mRNA and protein expression of USP13, thereby affecting the protein level of BMAL1 in cells. Using MG132 to inhibit the ubiquitin–proteasome process showed that ectopic expression of USP13 almost completely removed the ubiquitin modification from BMAL1 and increased BMAL1 protein level in cells with TDP-43 knockdown (Fig. 5 G). These findings suggest that USP13 might affect BMAL1 protein level by modulating the ubiquitination of BMAL1.

The immunoprecipitation assay showed that anti-USP13 captured BMAL1 and also anti-USP13 captured BMAL1 from M17 cell lysates (Fig. 6 A). In vitro–purified GST-tagged USP13 protein interacted with His-tagged BMAL1 (Fig. 6 B). Overexpression of wild-type USP13 in HEK-293T (Fig. 5 F) and M17 cells (Fig. 6 C) increased the protein level of BMAL1 in a dose-dependent manner, but overexpression of mutant USP13 (USP13 CA) (Fang et al., 2017) did not increase BMAL1 expression (Fig. 6 C), and we verified that USP13 knockdown decreased the cellular protein levels of BMAL1 in M17 cells (Fig. 6 D). By using CHX to inhibit eukaryotic translation, we found that USP13 knockdown accelerated the decrease of BMAL1 protein level in M17 cells (Fig. 6 E). To determine whether USP13 deubiquitinates BMAL1, HEK-293T cells were transfected with BMAL1, USP13 (wild type), or the enzyme-deficient mutant USP13 C345A (USP13 CA) and 6×His-ubiquitin. The ubiquitination level of BMAL1 was determined by Ni-NTA purification of proteins conjugated to 6×His-ubiquitin and BMAL1 by western blotting. We found that wild-type USP13, but not USP13 CA, deubiquitinated BMAL1 (Fig. 6 F). In USP13-knockdown M17 cells, wild-type USP13 expression, not the mutant USP13, elevated the protein level of BMAL1 (Fig. 6 G). We also found that in TDP-43-knockdown M17 cells, wild-type USP13 elevated the protein level of BMAL1, but not the mutant USP13 (Fig. 6 H). In mice infected with TARDBP interference virus, restoring the expression of wild-type USP13 reduced the ubiquitination of BMAL1 in the hypothalamus but not the mutant USP13 (Fig. 6 I). Thus, we believe that USP13 directly interacts with BMAL1 and regulates the ubiquitination of BMAL1 (Fig. 5 G and Fig. 6 I).

BMAL1 is a core component of the mammalian circadian clock, and its absence leads to the loss of circadian behaviors (Partch et al., 2014; Ray et al., 2020). Since TDP-43 regulates the rhythmic expression of BMAL1 protein and its knockdown causes phenotypic changes in rhythms, we are curious whether BMAL1 affects the rhythmic expression of TDP-43.

The results revealed that in 4-wk-old mice with Bmal1 knockout, we did not observe significant changes in the rhythmic expression pattern of TDP-43 in the hypothalamus, cortex, hippocampus, and liver (Fig. 7, A–H). By western blotting, we found rhythmic oscillation of USP13 expression in M17 cells (Fig. S3). In the hypothalamus, cortex, hippocampus, and liver of 4-wk-old wild-type mice, we observed rhythmic oscillation of the protein level of USP13, although the amplitude of the oscillations was not massive; besides, the oscillations were not affected by BMAL1 knockout (Fig. 7, A–H).

These findings support that USP13 expression also exhibits a rhythmic pattern, and USP13 interacts with BMAL1 to regulate ubiquitination of BMAL1, thereby affecting the protein level of BMAL1 in cells.

The circadian system controls glucose and energy metabolism in humans. TDP-43 truncation has been shown to induce metabolic reduction of metabolic intermediates that supply glycolysis and the tricarboxylic acid cycle (TCA), and TDP-43 proteinopathy also induced metabolic alterations, including the TCA cycle, in models of ALS (Loganathan et al., 2022; Maksimovic et al., 2023). However, there is insufficient data regarding the impact of TDP-43 knockdown on the TCA cycle and its role in regulating cellular energy production in the TCA cycle.

Here, we found that in metabolomics, TDP-43 knockdown led to a decrease in multiple glycolytic and TCA cycle intermediates including F6P, glucose-6-phosphate, FBP, phosphoenolpyruvate, 3-phospho-D-glyceric acid (3-PGA), pyruvate, and lactate (Fig. S4 A). TDP-43 knockdown also decreased versatile energy production intermediates (Fig. S4 B). RNA-seq analysis (GSE260737) showed that the sequence reads of multiple glycolysis- and TCA-related genes were also altered by TDP-43 knockdown (Fig. 2 A). This supports that TDP-43 may control cellular glycolysis and the TCA cycle. For glucose can provide energy to organisms through glycolysis and TCA cycle, we investigated whether TDP-43 knockdown affects cellular glucose uptake and the expression of glucose transporters (GLUTs). TDP-43 knockdown led to a decrease in the cellular glucose uptake capacity (Fig. S4 C). TDP-43 knockdown resulted in decreased expression of GLUT2 and GLUT3, which might explain the reason for TDP-43 knockdown attenuating cellular glucose uptake (Fig. S4 D).

Given that TDP-43 knockout decreased ATP production in metabolomics, and AMPK, which is also rhythmic (Lamia et al., 2009), is the most important molecule in organisms that senses energy levels and regulates metabolic homeostasis, we also examined whether the TDP-43–BMAL1 axis regulates AMPK signaling. TDP-43 and BMAL1 knockdown significantly induced AMPK activation at ZT4 and ZT8 and altered the pattern of AMPK activation over time (Fig. 8, A and B). In the TDP-43 knockdown cells, ectopic expression of BMAL1 or USP13 significantly restored ATP levels (Fig. S4 F). This supported that TDP-43 knockdown-induced ATP attenuation is dependent on cellular BMAL1 and USP13 levels. These findings let us deduce that TDP-43 could regulate glucose transporter expression and AMPK signaling in neurons, controlling glucose uptake, cellular glycolysis, and energy production in neuronal cells.

Based on the RNA-seq and metabolomic results, we further evaluated the effect of TDP-43 knockdown on the expression of glycolysis-related proteins. TDP-43 knockdown significantly decreased the expression of the glycolysis-related proteins PFKP, HXK1, and LDHA (Liu et al., 2017; Zheng et al., 2016) (Fig. S5 A). This result is consistent with the findings of a previous metabolomic study on the effects of TDP-43 on glycolysis and TCA (Fig. S4, A and B). We also found that PFKP was expressed rhythmically and TDP-43 knockdown led to the disappearance of periodic expression patterns (Fig. 8 C). In the promoter of PFKP, we found multiple conserved BMAL1 binding consensus sites and PFKP expression may be regulated by BMAL1. We also found that BMAL1 knockdown decreased PFKP protein expression (Fig. S5 B). According to the chromatin immunoprecipitation assay, BMAL1 could regulate the transcriptional activity of PFKP by directly binding to the promoter of the PFKP gene (Fig. S5 C). These results suggest that the TDP-43–BMAL1 signaling may regulate the rhythmic expression of activated AMPK and PFKP.

Based on our study, we concluded that TDP-43 is a new rhythmic protein that regulates the expression of multiple circadian genes and affects circadian behavior by inhibiting USP13 mRNA splicing. This affects the protein level of USP13 and ubiquitin-mediated degradation of BMAL1 and regulates the AMPK signaling pathway, glucose metabolism, and ATP production (Fig. 8 D).

A bidirectional regulatory relationship exists between neurodegenerative diseases and circadian rhythm disorders. Rhythm disorders and sleep disturbances have become the main symptoms of most neurodegenerative diseases. Meanwhile, circadian rhythms and sleep disturbances may contribute to drive pathology in the early stages of neurodegenerative diseases. TDP-43 dysfunction or its pathological accumulation is associated with neurodegenerative diseases, such as ALS. This mainly affects motor neuron function in the brain and spinal cord. Therefore, we evaluated whether TDP-43 affects the circadian rhythm and contributes to the development of neurodegenerative diseases.

We propose that TDP-43 is a new member that participates in maintaining rhythm homeostasis. TDP-43 was expressed in a 24-h periodic oscillation in the hypothalamus, hippocampus, cortex, and liver of the mice, with a pattern similar to that of BMAL1 being different from in the cortex (Fig. 1). SD affected the oscillatory pattern of TDP-43 expression in the brain (Fig. S1, C and D), supporting a potential link between TDP-43 and sleep functions (Perlegos et al., 2024). TDP-43 regulated the expression of circadian rhythm-related genes in vitro and in vivo (Fig. 2, Fig. 3, and Fig. S2 A). The expression of BMAL1, CLOCK, and CRY1 proteins was restored in TDP-43-depleted cells by the ectopic expression of TDP-43 (Fig. S2 B). These results led us to deduce that TDP-43 is a component of the rhythm maintenance system, and that cellular TDP-43, at least in part, controls the function of multiple rhythm proteins. There may be bidirectional regulation between circadian homeostasis and TDP-43 function. Knockdown of TDP-43 mainly affected mRNA expression of multiple genes related to neurodegenerative diseases, such as ALS and AD, in our bulk RNA-seq analysis (Fig. 2, A and B). This is consistent with previous reports (Chen and Mitchell, 2021; Prater et al., 2022; Riku et al., 2022). These results also shed light on TDP-43 dysfunction by affecting rhythmic homeostasis, contributing to the development of neurodegenerative diseases.

We obtained data supporting a regulatory relationship between BMAL1 and TDP-43. Although BMAL1 was not found to have a role in controlling the circadian rhythm of the body (Koronowski et al., 2019), it is still believed to be a key molecular clock. BMAL1 knockdown leads to the downregulation of other rhythm genes (Partch et al., 2014; Ray et al., 2020). BMAL1 knockdown did not change the protein expression of TDP-43 (Fig. S2 C) and its knockout also did not change the rhythmic expression of TDP-43 (Fig. 7). However, overexpression of TDP-43 in BMAL1 knockdown cells restored CLOCK, BMAL1, PER2, and CRY1 expression (Fig. S2 C). Because Bmal1 knockout mice still exhibit a diurnal pattern of locomotion and feeding activity (Ray et al., 2020), and the rhythmic expression of TDP-43 was not influenced by the absence of Bmal1, we would deduce that the cyclic nature of TDP-43 expression is driven by rest–activity rhythms (RAR) (Innominato et al., 2023), either directly or indirectly, rather than by the circadian clock.

We identified the mechanism through which TDP-43 regulates BMAL1 signaling. The ubiquitin pathway plays a central role in regulating the circadian clock (Stojkovic et al., 2014). The ubiquitin–proteasome pathway was regulated by TDP-43 in transcriptomics (Fig. 2 A), and TDP-43 knockout promoted the ubiquitination of BMAL1 (Fig. 4, C and D). RNA-seq showed that TDP-43 regulated the mRNA and protein expression of USP13 (Fig. 2 A; and Fig. 5, B and C), and ectopic expression of TDP-43 restored the expression of USP13 (Fig. 5 B). USP13 regulated ubiquitin modification and the protein level of BMAL1 (Fig. 5 G and Fig. 6). Its knockdown increased USP13 intron 1 being spliced into USP13 mRNA (Fig. 5 D), which decreased the amount of normal USP13 protein and attenuated the deubiquitination of BMAL1. Therefore, TDP-43 may inhibit USP13 intron splicing by binding to the UG-rich RNA of USP13 (Lukavsky et al., 2013). These findings have improved our understanding of the posttranslational regulation of BMAL1 protein (Fig. 8 D) and facilitated the investigation of the role of TDP-43 in modulating physiological and pathological functions through its impact on RNA processing (Tollervey et al., 2011).

We showed that the TDP-43–BMAL1 axis regulated cellular glucose use and energy production (Fig. S4). TDP-43 regulates the expression of glucose transporter genes, including GLUT2 and GLUT3, resulting in an altered glucose uptake (Fig. S4 D); this phenomenon was also endorsed in neurons (Manzo et al., 2019). PFKP, one of the key enzymes involved in glycolysis, was also identified as being regulated by TDP-43 and BMAL1 (Fig. 8 C; and Fig S5, A and B). TDP-43 knockdown led to a lower protein level of BMAL1 in cells under high glucose conditions (Fig. S4 E). This suggests that cellular rhythms may be more unstable in a hyperglycemic environment. Our findings indicate that circadian rhythms regulate glucose metabolism within organisms. We deduced that this regulation may be universally present in various organisms (Dyar et al., 2013; Marcheva et al., 2010). TDP-43 and BMAL1 also regulated the signal transduction of the energy sensor AMPK (Fig. 8, A and B). BMAL1 knockdown could induce energy stress by activating AMPK in cells (Lin and Hardie, 2018). This elucidates the mechanism by which the TDP-43–BMAL1 axis regulates cellular glucose usage to generate energy.

We found that TDP-43 regulated mammalian memory and balance in mice (Fig. 3, D and E), which is consistent with previous reports (Ma et al., 2021). Because icv-injected AAV carrying TDP-43 shRNA could circulate to the entire brain, TDP-43 knockdown was not limited to the hypothalamus. Therefore, we could not exclude the possibility that TDP-43 knockdown affected circadian behavior and that memory and balance abilities were induced by TDP-43 knockdown in other brain regions.

Based on these findings, we conclude that TDP-43, which is identified as a new rhythmic protein, regulates the expression of multiple circadian genes and affects circadian behavior, and cognitive and balance abilities in mice. TDP-43 may be an upstream regulator of BMAL1 because BMAL1 knockout did not change the rhythmic expression of TDP-43. We also propose that TDP-43 regulates the expression and alternative splicing of USP13, thereby affecting the protein level of USP13 and the ubiquitin-mediated degradation of BMAL1 and regulating the AMPK signaling pathway. This may result in changes in cellular glucose usage and ATP production. Our findings have enriched our understanding of the role of TDP-43 dysfunction in CRD in neurodegenerative diseases and provided new evidence supporting the interaction between CRD and neurodegeneration. Further, in-depth research will help to better understand more specific functions and finer regulatory signals in the rhythm function regulation of TDP-43.

C57BL/6 mice were purchased from Wuxi AppTec. The Bmal1-floxed strain (JAX007668) was from Jackson Laboratory and was bred with EIIa-cre (JAX003724) to achieve whole body heterozygous Bmal1-deletion. The heterozygous Bmal1-deleted mouse line was then back-crossed to C57BL/6 for five generations prior to the standard heterozygote × heterozygote breeding for whole body Bmal1-knockout mice. All the mice were maintained at Nantong University Animal Colony. Mice were raised at temperatures of 21–22°C and subjected to a fixed light–day (12 h light/12 h dark) cycle with access to food and water ad libitum. Neonatal (P0) FVB mice used for fibroblast culture were bred in the Nantong University Animal Colony. All the animal experiments followed the ethical guidelines for animal research at the Nantong University (S20210302-032).

HEK-293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Invitrogen). Human neuroblastoma M17 cells were cultured in DMEM/F12, also supplemented with 10% FBS. All the transfections were performed with Lipofectamine 2000 (Thermo Fisher Scientific) or jetPRIME (Polyplus Transfection) according to the manufacturer’s instructions. For primary mouse brain fibroblast (MBF) culture, the cerebral meninges of fetal FVB mice (P0) were peeled off and placed in a Petri dish with precooled phosphate-buffered saline (PBS). Cerebral meninges were then cut into pieces followed by digestion with 0.25% trypsin at 37°C for 20 min and preparation of a cell suspension. Following filtration, the cells were collected by centrifugation before being resuspended in DMEM containing 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin. The cells were adjusted to ∼8 × 10⁴/ml and maintained in a 3.5 cm Petri dish at 37°C with 5% CO₂.

pCI/TDP-43 tagged with hemagglutinin (HA) was constructed by our group (Gu et al., 2017a). pcDNA3.1/Flag-BMAL1 and pcDNA3.1/His-USP13 were provided by Dr. Xu of Soochow University and Dr. Sun of Wuhan University, respectively (Sun et al., 2017; Ullah et al., 2020). pAdeno-associated virus (AAV) serotype2/9-CBG-enhanced green fluorescent protein (EGFP)-3×FLAG-WPRE-H1-shRNA targeting TARDBP, pSLenti-U6-shRNA (TARDBP)-CMV-EGFP-F2A-Puro-WPRE (mouse), and their corresponding control viruses were constructed and packaged by OBiO Technology Co., Ltd. Using pcDNA3.1/His-USP13 as a template, PCR amplification was used to obtain the USP13 coding sequence carrying XhoI and NotI clone site; the resulting PCR fragment was cut with two enzymes and was then inserted into the pGEX-4T-1 vector between the sites, obtaining pGEX-4T-USP13. To generate His-BMAL1, BMAL1 cDNA was produced by PCR with full-length cDNA BMAL1 as a template. The resulting BMAL1 PCR fragment was inserted into the pET-30a vector between BglII and NotI sites, obtaining pGEX-4T-BMAL1. pGEX-4T-USP13 and pGEX-4T-BMAL1 were used for producing recombinant GST fusion USP13 and BMAL1. For pcDNA3.1/USP13-C345A construction, the mutant USP13 coding sequence was amplified by PCR using the following primer pairs: 5′-GAC​AGA​GCT​GAG​ATA​GGC​GCT​GTT​GCC​CAG​GTT​C-3′ and 5′-GAA​CCT​GGG​CAA​CAG​CGC​CTA​TCT​CAG​CTC​TGT​C-3′, and be cloned into pcDNA3.1 vector. LentiCRISPR v2/TDP-43_KO and pSpCas9(BB)-2A-GFP (PX458)/TDP-43_KO (TKD) were constructed by clustered regularly interspaced short palindromic repeat–associated endonuclease Cas9 (CRISPR-CAS9), and the gDNA sequence was 5′-GGT​TAC​AGC​CCA​GTT​TCC​AG-3′. The lentiviruses shTDP-43, shBMAL1, and shUSP13 were constructed by inserting interfering sequences into pLL3.7 (Addgene). The target sequences for shRNAs were TDP-43₁ (human): 5′-GCA​GGT​CAA​GAA​AGA​TCT​TAA-3′, TDP-43₂ (human): 5′-AGA​TCT​TAA​GAC​TGG​TCA​TTC-3′, TDP-43 (mouse): 5′- GCA​AAC​TTC​CTA​ATT​CTA​AGC-3′, BMAL1 (human): 5′-CTT​CTA​GGC​ACA​TCG​TGT​TAT-3′, and USP13 (human): 5′-CGC​CTG​ATG​AAC​CAA​TTG​ATA-3′.

Primary antibodies used in this study are listed in Table S1. 2-(n-7-nitro-2,1, 3-benzoxadiazole-4-amino)-2-deoxy-glucose (NBDG) glucose uptake assay kit was purchased from Invitrogen. Protein G beads, ECL western blotting substrates, Alexa 555-conjugated goat anti-mouse IgG, Alexa 488-conjugated goat anti-rabbit IgG, and Cy3-conjugated goat anti-rabbit IgG were purchased from Thermo Fisher Scientific. Horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit IgG antibodies were from Jackson ImmunoResearch Laboratories. MG-132 (Z-Leu-Leu-Leu-al) and CHX (CHX) was from MedChemExpress LLC.

8-wk-old C57BL/6 mice were intraperitoneally anesthetized with a 1.25% solution of 2,2,2-tribromoethanol (Sigma-Aldrich). After hair removal, a sagittal scalp incision was made and a burr hole was drilled in the skull. Stereotaxic injection of 1 μl of AAV viruses was performed bilaterally into the intracerebroventricular (icv) region using coordinates of 0.5 mm A/P and ±1.0 mm M/liter relative to the bregma and 2.5 mm beneath the skull surface. The viruses were injected at a rate of 0.4 μl/min using a Hamilton syringe equipped with a 31-gauge cemented needle, which was automated using an injection pump. The needle was left in place for 5 min before slow withdrawal. The skin incision was closed using a surgical staple. During and after surgery, the animals were kept warm on a heating pad until they fully recovered from the anesthesia. The survival of the animals was monitored daily. 1 mo after the injection, the mice underwent synchronization treatment for 1 wk and were then subjected to behavioral assays. Subsequently, the mice were sacrificed and their brains were collected for further experiments.

8-wk-old C57BL/6 mice were synchronized for 4 wk and randomly assigned to two groups without blinding. SD was performed at four zeitgeist times of the day, that is, ZT0, ZT6, ZT12, and ZT18, This was conducted using an SD instrument (Sansbio) (Curie et al., 2015), in which the animals were kept in a cage with a motor system controlling a sweeper bar at torque, speed, and operating frequency. After 5.5 h of SD, the treated or non-sleep-deprived control mice were sacrificed in 30 min, and the brain and liver were collected and stored at −80°C.

1 mo after the AAV virus injection, the mice were raised in a cage with a running wheel and an independent ventilation system for 3 wk in a 12-h light/12-h dark (LD) sequence. This was followed by a continuous dark (DD) status for another 3 wk. To examine the phase shift in locomotor activity, a wireless running wheel system (80820s; Campden) carrying a sensor hub was used to measure the wheel-running activity of the mice and circadian phase differences, according to the manufacturer’s instructions. Actograms and chi-square periodograms were analyzed and generated using ClockLab Analysis 6 software (Actimetric).

Open field and Morris water maze (MWM) experiments were performed using a video monitoring system (ANYmaze; Stoelting Co.) for recording and analysis. In the open field test, the mice were set in a black plastic arena (“L × W × H” is 48 cm × 48 cm × 40 cm) and recorded every 150 s over four sequential sessions. A square area of 225 cm³ in the central region of the open field was designated as the central zone. The cumulative distance traveled, duration of immobility, average speed, time spent within the central zone, frequency of entry into the central zone, and average distance from the central zone were recorded for each animal.

During the rotarod test, mice were subjected to three trials per day for five consecutive days using a rotarod apparatus (Ugo Basile). The rotarod gradually increased its speed from four revolutions per minute (rpm) to 40 rpm within a span of 2 min. Each trial had a maximum duration of 500 s and the mice were given a 30-min rest period between each trial.

For the MWM test, mice were placed in a black circular pool and separated into four equal quadrants. In each quadrant, a submerged escape platform, measuring 13 cm in diameter, was positioned 1 cm below the water surface. During each training trial, mice were allowed a maximum of 90 s to explore the maze freely. If the mouse failed to reach the platform within the allocated time, it was gently guided to the platform. After each trial, the mouse remained on the platform for 20 s. This training routine was repeated four times per day for four consecutive days, with a break of >30 min between each trial. 24 h after the final training session, a probe trial was conducted for 90 s without an escape platform. Spatial memory was assessed by measuring the time taken to reach the platform zone, swimming speed, and distance traveled until the mouse first reached the target zone.

For the CHX (MCE) treatment, M17 cells were transfected with PX458/TDP-43_KO (TKD). 48 h later, the cells were synchronized by being cultured in DMEM/50% horse serum for 2 h and then DMEM/1% FBS for 22 h. Thereafter, the cells were treated with 200 μg/ml CHX at CT8, and the cell lysates were extracted every 4 h. For the proteasome inhibition experiments, M17 cells were infected with Lenti/shTDP-43. After a 24 h period, the cells were synchronized and exposed to 10 μM MG132 (MCE) for a duration of 12 h. Following this treatment, the cells were harvested at CT12 and their lysates were used for western blot analysis.

Cells were seeded onto glass coverslips in 24-well plates one day prior to transfection, reaching ∼30% confluence. Cells were fixed with 4% paraformaldehyde in PBS for 30 min at 25°C. After PBS washing, the cells were blocked for 2 h at 37°C in 10% goat serum diluted in PBS containing 0.2% Triton X-100. The cells were incubated overnight at 4°C with monoclonal anti-BMAL1 (1:500), anti-CLOCK (1:500), polyclonal anti-PER2 (1:500), or anti-CRY1 (1:500) antibodies. After washing, the cells were incubated at 25°C for 2 h with Alexa 488-conjugated goat anti-rabbit IgG or Alexa 555-conjugated goat anti-mouse IgG, both at a dilution of 1:500, along with Hoechst dye at 1:5,000 dilution for nuclear staining. Following a PBS wash, the cells were mounted using ProLong Gold antifade reagent (Thermo Fisher Scientific) and observed using a Leica TCS-SP2 laser scanning confocal microscope.

For the staining of a mouse brain, adult C57BL/6 mice were anesthetized and fixed by cardiac infusion with 4% PFA at ZT0, ZT4, ZT8, ZT12, ZT16, and ZT20, and then frozen sections were performed after gradient dehydration with sucrose. Free-floating sections, 40 μm in thickness, of the cortex were washed with three changes of PBS, subsequently subjected to blocking with 10% goat serum in 0.5% Triton X-100/PBS for 2 h at 37°C after permeabilization, and then incubated with polyclonal anti-TDP-43 (Proteintech, 1:1,000) overnight at 4°C. Then, they were washed with PBS and incubated for 2 h with Cy3-conjugated goat anti-rabbit IgG (Thermo Fisher Scientific, 1:400) at room temperature. After washing with PBS, the sections were mounted with DAPI Fluoromount-G (SouthernBiotech; Amersco), and observed with OLYMPUS FV300 laser scanning confocal microscope.

GST and His-fusion proteins were purified as described (Zhang et al., 2014). Briefly, BL21 cells harboring the GST or various GST and His recombinant expression plasmids were grown to log phase and induced with 0.1 mM IPTG for 2 h at 37°C. Bacteria were collected and resuspended in lysis buffer and sonicated. Solubilized proteins were recovered by centrifugation and incubated with glutathione-agarose beads or Ni-NTA beads for 4 h at 4°C and washed several times with ice-cold PBS. The resulting bead-bound proteins were eluted with an elution buffer. Then, the eluted proteins were subjected to dialysis against PBS overnight at 4°C.

For in vitro binding assays, GST-USP13 or GST was mixed with His-BMAL1 in a GST pull-down buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, 1 mM DTT, 10% glycerol) at 4°C overnight. Then glutathione sepharose was added and samples were rotated at 4°C for 1 h. After washing three times, the bound proteins were separated on SDS-PAGE followed by western blotting analysis.

For IP, the cells or mice brains were lysed in a lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.2% NaDox, 0.1% NP-40, 0.1% Triton X-100, 1 mM Na₃VO₄, 50 mM NaF, 2 mM EDTA, and protease inhibitor cocktail (1:1,000). For co-IP, the cells were lysed using sonication in lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM Na₃VO₄, 50 mM NaF, 2 mM EDTA, and cocktail (1:1,000). The extract was then centrifuged at 15,000g, for 5 min at 4°C. The resulting lysate was incubated overnight at 4°C with protein G beads precoupled with monoclonal antibodies. The beads were washed twice with lysis buffer and Tris-buffered saline (TBS). The bound proteins were eluted by boiling in 2× Laemmli sample buffer and subjected to western blotting analysis.

M17 cells were infected with viruses engineered to deliver guide RNA (gRNA) targeting the TARDBP gene (see LentiCRISPR v2/TDP-43_KO in “Plasmids and viruses” section) or control viruses, synchronized for 36 h, and then collected with TRIzol reagent at CT12. The samples were sent to Huada Gene Company for digital gene expression analysis. For sequencing, Illumina technology was used with minor adjustments, resulting in millions of raw reads generated in each sequencing tunnel with a sequencing length of 49 bp. Quality evaluation of the RNA-seq data was conducted by analyzing the tag distribution of each group. To functionally annotate the differentially expressed genes, the Huada Gene Company performed gene ontology (GO) annotation and KEGG pathway analyses. The raw data of the RNA-seq was uploaded to gene expression omnibus (GEO) database (GSE260737).

10% homogenate of mouse tissue was prepared in a buffer containing 20 mM Tris-HCl (pH 8.0), 0.32 M sucrose, 10 mM β-ME, 5 mM MgSO₄, 1 mM EDTA, 1 mM Na₃VO₄, and 50 mM NaF, along with a cocktail of protease inhibitors (diluted 1:1,000). The homogenate was then centrifuged at 15,000g for 30 min at 4°C. The resulting extract and cells were lysed in 1× Laemmli buffer and boiled for 5 min in a thermostatic metal bath (JOANLAB). The samples were resolved using 10% SDS-PAGE and transferred onto an Immobilon-P transfer membrane (Merck Millipore). The blots were incubated overnight with primary antibodies, followed by incubation with the corresponding horseradish peroxidase-conjugated secondary antibodies at room temperature for 2 h. The blots were developed using Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific) and visualized using a Tanon-5200 Chemiluminescence imager (Biotanon Ltd.). The specific immunosignal was quantified using Multi Gauge software (V3.0; Fuji Film).

Total RNA was isolated from cells or tissues using the RNA isolator Total RNA Extraction Reagent (Vazyme) and then reverse-transcribed using BL699A reverse transcription kit (Biosharp). The RNA splicing of USP13 was measured by PCR (Vazyme) with forward primer 5′-ACG​AGA​AGG​GAA​GCA​GAA​GC-3′ and reverse primer 5′-TTG​GTT​CAT​CAG​GCG​ATC​TT-3′. The PCR products were resolved on 1.5% agarose gels, visualized by gel red staining, and quantitated using the Molecular Imager system (Bio-Rad).

The qPCR reactions were performed using AceQ qPCR SYBR Green Master Mix (Vazyme) with LightCycler 96 Real-Time PCR System (Roche Diagnostics) according to the manufacturer’s instructions. The primers used for qPCR are listed in Table S2. All the reactions were performed in triplicate. The qPCR reactions were conducted as follows: 3 min at 95°C for the initial denaturation, followed by 45 cycles of amplification at 95°C for 10 s and 60°C for 15 s, and then 95°C for 10 s, 65°C for 1 min, and 97°C for 1 s. The 2^−∆∆CT method, with GAPDH as the internal control, was used to determine the relative expression.

M17 cells cultured on 10-cm plates at 90% confluence were treated with formaldehyde added to the culture medium at a final concentration of 1% (vol/vol). The reaction was maintained at 25°C with shaking for 10 min, after which glycine (0.125 M) was added. After 10 min, the cells were rinsed twice with cold PBS supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF). The cells were harvested and lysed in SDS lysis buffer containing 1 mM PMSF, 2 mg/ml pepstatin A, and 2 mg/ml aprotinin. DNA was fragmented into 400–1,100 bp fragments using sonication. The chromatin was precleared and incubated overnight at 4°C with agarose beads coupled with either control antiserum or anti-BMAL1 (Abcam).

The agarose beads were collected by centrifugation and washed sequentially with low-salt, high-salt, LiCl, and TE buffers (twice each). After reversing the protein/DNA cross-links, DNA was isolated using phenol/chloroform extraction and ethanol precipitation and used as a template for standard PCR amplification. The PCR products were separated on a 2.0% agarose gel and stained with ethidium bromide. The resulting samples were visualized under UV light and the PCR products were purified and sequenced.

Cellular TCA metabolism intermediates and energy metabolites were detected using mass spectrometry with multiple reaction monitoring (MRM) techniques. Briefly, M17 cells infected with viruses engineered to deliver gRNA targeting the TARDBP gene (see LentiCRISPR v2/TDP-43_KO in “Plasmid and virus” section) or control viruses for 72 h. After synchronization for 36 h, the cells in each group were washed with PBS three times, scraped, and centrifuged at 1,000g for 5 min. More than 10 million cells were rapidly frozen in liquid nitrogen and metabolic analyses were performed (Shanghai Applied Protein Technology).

M17 cells in the 24-well plates were transfected with the TDP-43 gene editor vector, and after 48 h, the medium was changed to sugar-free DMEM with 200 μM 2-NBDG (Invitrogen) for 30 min. The cells were washed three times with PBS for 3 min, the green fluorescence of the cells was observed, and images were captured under a microscope. The fluorescence intensity of 2-NBDG was measured using a luminometer. Liver homogenates from AAV tail vein-injected mice were subjected to a glucose assay kit with O-toluidine (Beyotime Biotechnology) to measure the concentration of the glucose (μg/mg protein) according to the manufacturer’s instructions.

The data are presented as the mean ± S.D. (Standard Deviation), and GraphPad Prism 8 (GraphPad Software Inc.) was used for statistical analysis. The results were analyzed using one-way ANOVA, unpaired t test, two-way RM ANOVA, or JTK_Cycle analysis. A value of P < 0.05 was considered statistically significant.

Fig. S1 shows the representative area of the suprachiasmatic nucleus (SCN) where used to quantify and the quantification results for Fig. 1 C, along with the observation that sleep deprivation alters the rhythmic expression of TDP-43 in the hypothalamus of mice. Fig. S2 shows that TDP-43 regulates the expression of BMAL1, CLOCK, PER2, and CRY1 in cultured cells. Fig. S3 shows that USP13 exhibits rhythmic oscillations in expression in M17 cells. Fig. S4 shows that TDP43 regulates cellular glucose uptake, glycolysis, TCA cycle, and energy production. Fig. S5 shows TDP-43 controlled neuronal glycolysis possibly by directly regulating the transcription of PFKP through BMAL1. Table S1 affords the information of primary antibodies used in this study. Table S2 shows the primers used for qPCR and ChIP in this study. Data S1 shows all source files for immunoblot experiments in excel file. Data S2 shows all source files for immunoblot experiments in PDF file.

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, H. Cui ([email protected]).

We are grateful to Drs. Hung-Chun Chang, Shujia Zhu, and Tingyu Shen (Shanghai Center for Brain Science and Brain-inspired Technology, Shanghai, China) for providing Bmal1 knockout mice and assistance during the wheel-running experiment. We would like to thank Editage (https://www.editage.cn) for English language editing.

This study was supported by the National Natural Science Foundation of China (32171258, 31870772), the Natural Science Foundation of Jiangsu Province (BK20211329), and funds from the Co-Innovation Center of Neuroregeneration and the NMPA Key Laboratory for Research and Evaluation of Tissue Engineering Technology Products of Nantong University.

Author contributions: J. Gu: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing, M. Yang: Investigation, L. Zhang: Investigation, Validation, Visualization, Y. Liu: Data curation, Software, R. Yan: Investigation, D. Pan: Validation, X. Qian: Project administration, Validation, H. Hu: Methodology, D. Chu: Resources, C. Hu: Conceptualization, Investigation, Methodology, Project administration, Resources, Validation, Writing - review & editing, F. Liu: Conceptualization, Supervision, Visualization, Writing - review & editing, H. Cui: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing.