CircADAM9 activates autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors

Abstract

Background

Depression is a prevalent mental disorder, and its pathogenesis has not yet been fully elucidated. Circular RNAs (CircRNAs) have exhibit critical regulatory role in depression, while Unc-51-like kinase 1 (ULK1)-mediated autophagy participates in the pathophysiology of the disorder. Nevertheless, the potential involvement of circRNAs in modulating ULK1-mediated autophagy in depression remains unclear.

Objectives

This study aimed to identify depression-associated circRNAs regulating ULK1-medaited autophagy, and uncover the underlying mechanisms.

Methods

The depression model was established in mice using chronic unpredictable mild stress (CUMS) procedure. RNA sequencing and bioinformatics analysis were carried out to screen signaling pathways, circRNAs and microRNAs (miRNAs) related to depression. Overexpression and knockdown of circRNAs or miRNAs were conducted by lentivirus microinjection and cell transfection. The depression-like phenotypes were assessed through behavioral tests. Western blot, immunofluorescence and transmission electron microscopy (TEM) were employed to evaluated ULK1-mediated autophagy. RNA pull-down and dual-luciferase reporter assay was executed to explore the underlying mechanisms.

Results

CircADAM9 exhibited the most significant downregulation in CUMS mice among the differentially expressed circRNAs. Overexpression of circADAM9 relieved depressive-like behaviors, and promoted ULK1-mediated autophagy in vivo and in vitro. Importantly, circADAM9 sponged miR-302b-3p to regulate Ulk1 gene expression, and miR-302b-3p mimics diminished the behavioral improvements and pro-autophagic effect modulated by circADAM9.

Conclusion

We identified circADAM9 as a novel regulator in depression, which activated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors. Our findings would provide a new potential therapeutic target for the treatment of depression.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00738-y.

Introduction

Depression, clinically termed major depressive disorder, is a severe psychiatric syndrome characterized by a serious of emotional and physical disturbances including persistent despair, anxiety, anhedonia, suicidal ideation, sleep dysregulation, and chronic fatigue [1]. As a predominant contributor to global disease burden, depression affects approximately 350 million individuals worldwide, and over 1 million annual deaths are attributed to depression-induced suicide [2]. Current therapeutic interventions including pharmacotherapy and psychotherapy exhibit limited response and remission rates, resulting in treatment resistance in over 50% of patients [3]. Despite the considerable therapeutic hurdles in depression treatment, the precise pathogenesis underlying the disorder have yet to be fully uncovered. Elucidating the molecular underpinnings of depression is imperative for revealing its mechanistic basis and developing effective treatment strategies.

Recent studies increasingly reveal a critical association between Unc-51-like kinase 1 (ULK1)-mediated autophagy and depression [4–7]. Autophagy is an evolutionarily conserved catabolic process which maintains cell homeostasis and survival under stress conditions [8]. The process initiates with phagophore formation, which expands to encapsulate cytoplasmic cargo and forms autophagosome. The autophagosome then fuses with lysosome to form autolysosome, where misfolded proteins and damaged organelles are degraded by lysosomal hydrolase [9]. ULK1 is a master regulator of autophagy initiation. It promotes phagophore nucleation through phosphorylation of Beclin-1 and activation of the VPS34/PI3K-III complex, while facilitating phagophore expansion by phosphorylating ATG9 [10, 11]. Furthermore, ULK1 regulates autophagosome maturation by activating the ATG5-12-16L1 complex to mediate LC3 lipidation, a process that converts soluble LC3-I to membrane-associated LC3-II [12, 13]. Preclinical studies demonstrate significant downregulation of ULK1 in depression [4–7]. The suppression of autophagy is also observed in depressive patients and animals [14, 15]. These findings indicate ULK1-mediated autophagy is a pivotal node in the pathogenesis of depression. However, the upstream signaling cascades controlling ULK1 activity in depression has not been fully elucidated. Further investigation into upstream regulatory networks of ULK1 remains essential for depression research.

Emerging evidence suggests that circular RNAs (circRNAs) are critically involved in the pathophysiological process of depression [16–18]. CircRNAs are a unique class of endogenous non-coding RNAs generated through non-canonical back splicing of pre-mRNAs, resulting in their defining covalently closed circular structure [19]. The deficiency of 5’ caps and 3’ polyadenylated tails endows exceptional stability to circRNAs, rendering them resistant to cellular RNA degradation [20]. CircRNAs can regulate mRNA expression by microRNA (miRNA) sponging and RNA-Binding Protein interaction to modulate diverse biological processes [21, 22]. It has been reported that circRNAs regulate astrocyte dysfunction, microglial activation and neuron injury in depressive patients and animals [16–18]. Nevertheless, whether circRNAs influence ULK1-mediated autophagy in depression remains poorly understood. Deciphering the circRNA-ULK1 regulatory axis could provide novel insights into autophagy-associated pathogenesis of depression.

In the present work, we identified circADAM9 (mmu_circ_0001665) in chronic unpredicted mild stress (CUMS) model, and established its first reported association with the pathogenesis of depression. The functional role of circADAM9 in regulating depressive-like behaviors and ULK1-mediated autophagy were systematically investigated in CUMS mice and astrocytes. Utilizing bioinformatics analysis and dual-luciferase reporter assays, we explored the potential mechanism by which circADAM9 modulated Ulk1 through sponging miR-302b-3p. Furthermore, we investigated whether circADAM9 regulated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors.

Materials and methods

Animals

Adult male C57BL/6J mice (18–22 g, 4–5 weeks old) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China, License No. SCXK2021-0006) and randomly assigned to experimental groups. Prior to experiments, the mice were acclimated for 7 days under a 12-hour light/dark cycle at room temperature with free access to water and food.

Stereotaxic microinjection

LV-circADAM9 and SH-circADAM9 lentivirus were obtained from LandM Biotechnology Co., Ltd. (Guangzhou, China) for overexpression and knockdown of circADAM9. MiR-302b-3p mimics was purchased from GenePharma (Shanghai, China). One week prior to the CUMS procedure, the lentivirus and mimics were bilaterally microinjected into the prefrontal cortex of the mice at the following coordinates relative to bregma: anterior-posterior + 1.70, medio-lateral ± 0.35, dorso-ventral − 2.20 [23].

CUMS procedure

The CUMS procedure was performed based on our previous study with minor adjustments [24]. The mice were subjected to unpredictable and mild stressors twice daily over a 4-week period. The applied stressors included (1) food deprivation for 24 h, (2) water deprivation for 24 h, (3) empty bottle exposure for 2 h, (4) cage tilt (45°) for 7 h, (5) overnight illumination, (6) bedding removal for 24 h, (7) damp cage (200 mL water/100 g bedding) for 24 h, and (8) forced swim in ice water for 6 min.

Behavioral tests

The behavioral tests included sucrose preference test (SPT), tail suspension test (TST), forced swimming test (FST), and open field test (OPT). All the behavioral tests were conducted according to our previous study [25, 26].

For the SPT, mice were habituated to two bottles of 1% sucrose solution for 24 h Following 24 h of food and water deprivation, the mice were individually housed with free access to two bottles containing different liquid (1% sucrose solution and distilled water) for 24 h. The consumption of the liquid in the two bottles were recorded, and the sucrose preference was calculated as sucrose solution intake / (sucrose water intake + distilled water intake) × 100%.

In the TST, the tail of each mouse was secured to a suspension bracket using adhesive tape, which was placed about 1 cm from the tail tip. The test duration was 6 min, with the initial 2 min excluded from analysis. Immobility time was recorded during the final 4-min testing period.

The FST was conducted in a transparent glass cylinder (25 cm diameter × 20 cm height) containing 15 cm of water maintained at 23–25 °C. Each mouse was placed in the cylinder for 6 min, with the first 2 min serving as an acclimation period. Immobility time was measured during the subsequent 4-min testing phase.

The OPT was performed in a 40 cm × 40 cm × 35 cm plastic chamber with a floor divided into 25 squares. Following preconditioning to the test environment, mice were placed in the chamber for 5-min locomotion, and their time spent in the central area, rearing frequency, and total movement distance were recorded. The chamber was cleaned with 75% ethanol between trials to eliminate odor cues.

Tissue collection and preparation

Mice were deeply anesthetized using isoflurane and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde. Brains were immediately extracted, and the prefrontal cortex was carefully dissected on ice according to an established procedure [27]. All tissue samples were rapidly frozen in liquid nitrogen and maintained at − 80 °C for subsequent analysis.

Cell culture and transfection

The C8-D1A mouse astrocyte cell line was obtained from iCell Bioscience Inc (Shanghai, China). Cells were cultured in high glucose DMEM (GIBICO, USA) containing with 10% fetal bovine (GIBICO, USA), 100 U/mL penicillin, and 100 mg/ml streptomycin sulfate at 37 °C in 5% CO₂. Lipopolysaccharide (LPS) was obtained from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). The circADAM9 overexpression vector (LV-circADAM9), short hairpin RNA of circADAM9 (SH-circADAM9), and their corresponding negative controls (LV-NC and SH-NC) were constructed by LandM Biotechnology Co., Ltd. (Guangzhou, China). The miR-302b-3p mimics, miR-302b-3p inhibitor, and their respective negative controls were sourced from GenePharma (Shanghai, China). The transfection was performed using Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer’s instructions. The sequences were listed in Table S1 and Table S2.

RNA sequencing and bioinformatics analysis

The RNA sequencing was conducted by Fenghui Biotechnology Co., Ltd. (Hunan, China). Total RNA was isolated from the tissues using TRIeasy Reagent (YEASEN, Shanghai, China). Differentially expressed mRNAs were analyzed using the DEGseq tool, with an adjusted p value < 0.05 and fold change ≥ 2 as the thresholds. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were conducted using DAVID database (https://david.ncifcrf.gov/) with p < 0.05 as the significance cutoff. The miRNA target prediction was carried out using starBase (https://rnasysu.com/encori/) and TargetScan (https://www.targetscan.org/vert_72/).

Fluorescence in situ hybridization (FISH)

FISH assays were performed using a FISH Kit (Bersinbio, Guangzhou, China) according to the manufacturer’s protocol. Briefly, cells were fixed in 4% paraformaldehyde (25 °C, 30 min) and washed twice with PBS. After treatment with Proteinase K solution (37 °C, 15 min), cells were incubated with 1% fixation solution (25 °C, 10 min). Samples were then pre-hybridized (37 °C, 30 min), followed by hybridization with Cy3-labeled circADAM9 probe and FAM-labeled miR-302b-3p probe at 37 °C for 18 h. DAPI (Beyotime, Shanghai, China) was employed for nuclei staining after washing with SSC buffer. Fluorescence images were captured by a laser scanning confocal microscope (Olympus, Tokyo, Japan). The sequences of the probes were shown in Table S3.

RNA stability analysis

RNA stability was assessed using RNase R digestion and actinomycin D treatment. For RNase R digestion, total RNA was incubated with or without 3 U/µg RNase R (Aladdin, Shanghai, China) for 20 min. For Act D treatment, cells were exposed to 2 µg/mL actinomycin D (Aladdin, Shanghai, China) for 0, 4, 8, 12, or 24 h prior to RNA extraction. Adam9 and circADAM9 expression were measured by RT-qPCR.

RNA pull-down assay

Biotin-labeled circADAM9 probe and miRNAs probes (RiboBio, Guangzhou, China) were incubated with streptavidin magnetic beads (Invitrogen, USA) for 2 h at room temperature. The circADAM9-binding RNAs were extracted using Trizol reagent (Invitrogen, USA). The miRNA expression levels were measured by RT-qPCR analysis. The sequences of probes for RNA pull-down assay were listed in Table S4.

Dual-luciferase reporter assay

The sequences of circADAM9 and 3’UTR of Ulk1 containing wild-type or mutant binding sites (Promega, USA) were cloned into pmirGLO luciferase reporter vector. The above vectors were co-transfected with miR-302b-3p mimics, inhibitor, and their corresponding negative controls into cells using Lipofectamine 3000 (Invitrogen, USA). After 48 h, the luciferase activity was detected by Dual-Luciferase Reporter Assay Kit (Promega, USA).

Real-time quantitative PCR (RT-qPCR)

Total RNA was extracted from the samples using TRizol reagent (Invitrogen, USA). Subsequently, 1 µg extracted total RNA was reverse-transcribed into cDNA using High Capacity cDNA Archive Kit (Invitrogen, USA). RT-qPCR was performed with PowerUp SYBR Green Mix (Invitrogen, USA) on a qTOWER³ real-time PCR system (Analytik Jena, Germany). The expression levels of circRNAs and mRNAs were normalized to Gapdh, while miRNAs expression levels were normalized to U6. The primer sequences for RT-qPCR assay were shown in Table S5.

Western blot

Total protein was extracted from the samples using RIPA Lysis Buffer (Beyotime, Shanghai, China). The protein concentration was determined by a BCA Protein Assay Kit (Beyotime, Shanghai, China). Equal amount of protein from each sample were separated on 10% or 12% SDS-PAGE and transferred to PVDF membranes. After blocking with 5% non-fat milk in TBST for 1 h at room temperature, the membranes were incubated with primary antibodies overnight at 4 °C. The membranes were subsequently washed and incubated with HRP-labeled Goat Anti-Rabbit IgG (Beyotime, Shanghai, China) for 1 h at room temperature. Protein bands were visualized with ECL Kit (Beyotime, Shanghai, China) and photographed using a chemiluminescence imaging system (Bio-Rad, USA). The gray values were analyzed by ImageJ. The employed primary antibodies were listed as follows: ULK1 (Affinity, DF7588), ATG5 (Affinity, DF6010), Beclin1 Affinity (Affinity, AF5128), LC3 (Affinity, AF4650), p62 (Affinity, AF5384), GAPDH (Affinity, AF7021).

Immunofluorescence

The samples were fixed in 4% paraformaldehyde and blocked with 5% BSA containing 0.3% Triton X-100 for 1 h. After overnight incubation with primary antibodies at 4 °C, samples were incubated with FITC-labeled Goat Anti-Rabbit IgG (Beyotime, Shanghai, China) for 1 h. The nuclei were stained with DAPI (Beyotime, Shanghai, China). The immunofluorescence images were acquired using a fluorescence microscope (Olympus, Tokyo, Japan).

Transmission electron microscopy (TEM)

Samples were fixed in 2.5% glutaraldehyde and 1% osmic acid for 2 h, then post-fixed in 1% osmium tetroxide. Following staining with aqueous uranyl acetate, samples were embedded in epoxy resin. Ultrathin sections were stained using lead citrate and observed by transmission electron microscope (Tecnai Spirit Biotwin, USA).

Statistical analysis

Data are represented as mean ± standard deviation (SD) from three independent experiments. Group comparisons were analyzed by one-way analysis of variance (ANOVA), with p < 0.05 considered statistically significant.

Results

ULK1 and autophagy were involved in the pathogenesis of depression

The CUMS paradigm was employed to investigate the underlying mechanism of depression. Behavioral tests including SPT, TST, FST and OPT were performed to validate the CUMS model (Fig. 1A-F). In CUMS mice, the sucrose preference markedly reduced (Fig. 1A), indicating the emergence of anhedonia symptom. The prolonged immobility time in TST and FST (Fig. 1B and C) demonstrated the despair behavior in CUMS mice. The time spent in central area, number of rearing, and movement distance of CUMS mice significantly decreased (Fig. 1D-F), which reflected the appearance of anxiety behavior and the decline of locomotion activity. These results demonstrated the successful establishment of the CUMS model. To explore the potential targets and relevant pathways of depression, we collected the prefrontal cortex of non-stressed and CUMS mice for RNA sequencing and bioinformatics analysis (Fig. 1G-I). Among the top 30 differentially expressed mRNAs, Ulk1 expression markedly decreased in CUMS mice (Fig. 1G). GO enrichment analysis showed that autophagy, autophagosome and ubiquitin protein ligase binding were the top 3 GO processes (Fig. 1H). KEGG enrichment analysis indicated that the most associated signaling pathway of the differentially expressed mRNAs was autophagy. (Fig. 1I). Moreover, CUMS mice exhibited a markedly reduction in autophagosome numbers in prefrontal cortex (Fig. 1J). Immunofluorescence assay showed the decrease expression of ULK1 in CUMS mice (Fig. 1K). The above outcomes suggested that ULK1 and autophagy were involved in the pathogenesis of depression.

Fig. 1.

ULK1 and autophagy were associated with the pathogenesis of depression. (A-F) Sucrose preference, immobility time, time spent in central area, number of rearing, and movement distance of the mice in behavioral tests (n = 8). *p < 0.05, compared with Control group. (G) Clustered heat map of differentially expressed mRNAs. (H) GO enrichment analysis of differentially expressed mRNAs. (I) KEGG pathway enrichment analysis of differentially expressed mRNAs. (J) Autophagosomes (red arrows) observed by transmission electron microscopy in prefrontal cortex of the mice (scale bar in normal image: 2 μm, scale bar in magnified image: 1 μm). (K) ULK1 protein expression examined by immunofluorescence assay in prefrontal cortex of the mice (scale bar: 50 μm)

CircADAM9 played a critical role in the pathogenesis of depression

To explore the potential role of circRNAs in depression, we performed circRNA sequencing on prefrontal cortex samples from the brains of non-stressed and CUMS mice. A clustered heatmap revealed the top 30 differentially expressed circRNAs (Fig. 2A), among which circ_0001665 showed significant downregulation in CUMS mice. Subsequent RT-qPCR analysis on the top 5 differentially expressed circRNAs confirmed this result, with circ_0001665 showing the most pronounced differential expression (Fig. 2B), suggesting its potential importance in depression. According to UCSC Date genome browser, circ_0001665 derived from Adam9 exons 12–14 with a length of 461 bp. Sagner sequencing confirmed that the amplified product was consistent with the spliced junction site of circ_0001665 (Fig. 2C). Divergent primer amplification confirmed the circular nature of circ_0001665, as evidenced by cDNA-specific amplification without gDNA contamination (Fig. 2D). The stability of circ_0001665 was assessed by RNase R digestion (Fig. 2E) and actinomycin D treatment (Fig. 2F). After RNase R digestion, circ_0001665 exhibited stronger stability than linear mRNA Adam9. Concurrently, actinomycin D treatment showed minimal impact on the stability of circADAM9. The subcellular location and expression of circ_0001665 were investigated by FISH assay, as shown in Fig. 2G. Circ_0001665 predominantly located in the cytoplasm, and exhibited a lower expression in CUMS group compared with non-stressed mice. These results confirmed the existence and crucial involvement of circ_0001665 in depression pathophysiology. Given that circ_0001665 originated from back-splicing of the Adam9 gene, we named it circADAM9.

Fig. 2.

CircADAM9 (circ_0001665) played a crucial role in the pathogenesis of depression. (A) Clustered heat map of differentially expressed circRNAs. (B) Relative expression of the top 5 differentially expressed circRNAs (n = 5). *p < 0.05, **p < 0.01, compared with Control group. (C) Schematic diagram showing that circADAM9 originated from the circulation of Adam9 exons 12–14. The junction site of circADAM9 were verified by Sanger sequencing. (D) Existence of circADAM9 validated by agarose gel electrophoresis. (E) Relative expression of Adam9 and circADAM9 after RNase R digestion (n = 3). *p < 0.05, compared with Mock group. (F) Relative expression of Adam9 and circADAM9 after Actinomycin D treatment (n = 3). *p < 0.05, compared with Adam9 group. (G) Location and expression of circADAM9 detected by FISH assay in prefrontal cortex of the mice (scale bar: 50 μm)

CircADAM9 relieved depressive-like behaviors in CUMS mice

To investigate the influence of circADAM9 on depressive-like behaviors, LV-circADAM9 (for overexpression) and SH-circADAM9 (for knockdown) lentivirus were microinjected into the prefrontal cortex of the mice before CUMS procedure (Fig. 3A). The overexpression and knockdown of circADAM9 in mice were validated by RT-qPCR (Fig. S1). The depressive-like behaviors were assessed by behavioral tests, with experimental schematics illustrated in Fig. 3B. Compared with CUMS group, circADAM9 overexpression increased the sucrose preference in SPT (Fig. 3C), while shortened immobility time in TST (Fig. 3D) and FST (Fig. 3E), indicating that circADAM9 alleviate anhedonia and despair induced by CUMS. In OPT test (Fig. 3F-H), the time spent in central area, number of rearing, and movement distance in LV-circADAM9 group showed statistically higher values than CUMS group, demonstrating the ameliorative effects of circADAM9 on anxiety and locomotion activity. CircADAM9 knockdown exerted an opposite impact on depressive-like behaviors compared to its overexpression. These findings demonstrated that circADAM9 ameliorated depressive-like behaviors in CUMS mice, and provided further evidence for the close association between circADAM9 and depression.

Fig. 3.

CircADAM9 improved depressive-like behaviors in CUMS mice. (A) Schematic diagram of the experiment procedures. (B) Schematic diagrams of behavioral tests including SPT, TST, FST and OFT. (C) Sucrose preference of the mice in SPT. (D) Immobility time of the mice in TST. (E) Immobility time of the mice in FST. (F-H) Time spent in central area, number of rearing, and movement distance of the mice in OFT. *p < 0.05, compared with Control group, ^#p < 0.05, compared with CUMS group (n = 8)

CircADAM9 promoted ULK1-mediated autophagy in the prefrontal cortex of CUMS mice

Our results have demonstrated the involvement of circADAM9 and ULK1-mediated autophagy in depression. Moreover, overexpression of circADAM9 led to an upregulation of Ulk1 gene expression, while its knockdown resulted in a corresponding downregulation in mice (Fig. S1). Thus, we hypothesized an interaction between circADAM9 and ULK1-mediated autophagy and probed their intrinsic relationship in prefrontal cortex of CUMS mice. Western blot and immunofluorescence assay (Fig. 4A-C) revealed significant downregulation of ULK1, ATG5, Beclin1 and LC3, alongside p62 accumulation in CUMS mice, indicating the inhibition of autophagy. Conversely, circADAM9 overexpression normalized these alterations. Furthermore, TEM analysis (Fig. 4D) showed that circADAM9 overexpression significantly increased the number of autophagosomes in the prefrontal cortex of CUMS mice, suggesting the enhanced autophagosome biogenesis. The impacts of circADAM9 knockdown on autophagy related proteins and autophagosomes were contrary to circADAM9 overexpression. The above results confirmed that circADAM9 acted as an activator of ULK1-mediated autophagy in CUMS mice.

Fig. 4.

CircADAM9 activated ULK1-mediated autophagy in the prefrontal cortex of CUMS mice. (A) Relative protein expression of ULK1, ATG5, Beclin1, LC3 and p62 measured by Western blot (n = 3). *p < 0.05, compared with Control group, ^#p < 0.05, compared with CUMS group. (B-C) Protein expression of ULK1 and LC3 detected by immunofluorescence assay (scale bar: 50 μm). (D) Autophagosomes (red arrows) observed by TEM in prefrontal cortex of the mice (scale bar in normal image: 2 μm, scale bar in magnified image: 1 μm)

CircADAM9 activated ULK1-mediated autophagy in astrocytes

To ensure the reliability of our results, we verified the regulatory effect of circADAM9 on ULK1-mediated autophagy in vitro. Given the close association between astrocytic autophagy and depression [28–30], astrocytes were employed for in vitro studies of depression. The astrocytes were stimulated with LPS (1 µg/mL) to model autophagy disruption caused by inflammatory stress and cellular damage in depression [31, 32]. The overexpression and knockdown of circADAM9 in astrocytes were validated by RT-qPCR (Fig. S2). As shown in Fig. 5A-C, LPS treatment successfully induced autophagy impairment, as evidenced by significant downregulation of ULK1, ATG5, Beclin1 and LC3-II, coupled with p62 upregulation. CircADAM9 overexpression reversed these LPS-caused impacts, restoring autophagic proteins expression to normal levels. Ultrastructural analysis (Fig. 5D) showed that circADAM9 overexpression rescued the reduce of autophagosomes in LPS-stimulated astrocytes, demonstrating the recuperative autophagosome formation. Inversely, circADAM9 knockdown exacerbated autophagic suppression caused by LPS. These findings in vitro consistently mirrored our observations in vivo, demonstrating the role of circADAM9 in promoting ULK1-mediated autophagy.

Fig. 5.

CircADAM9 promoted ULK1-mediated autophagy in astrocytes. (A) Relative protein expression of ULK1, ATG5, Beclin1, LC3 and p62 determined by Western blot (n = 3). *p < 0.05, compared with Control group, ^#p < 0.05, compared with LPS group. (B-C) Protein expression of ULK1 and LC3 determined by immunofluorescence assay (scale bar: 50 μm). (D) Autophagosomes (red arrows) observed by TEM in astrocytes (scale bar in normal image: 2 μm, scale bar in magnified image: 1 μm)

CircADAM9 regulated Ulk1 via sponging miR-302b-3p

Given that circRNAs can regulate gene expression through competitive endogenous RNA (ceRNA) mechanisms [21, 22], we investigated whether circADAM9 regulated Ulk1 gene expression by sponging specific miRNAs. Bioinformatics analysis identified 5 candidate miRNAs interacting with both circADAM9 and Ulk1 (Fig. 6A and B). Using a biotinylated circADAM9 probe (validated by RT-qPCR, Fig. 6C), RNA pull-down assay (Fig. 6D) revealed that miR-302b-3p was more prominently enriched by circADAM9 compared with the other 4 miRNAs. CircADAM9 overexpression and knockdown most significantly affected miR-302b-3p expression among the candidates (Fig. 6E). The expression of miR-302b-3p negatively correlated with circADAM9. MiR-302b-3p mimics or inhibitor transfection showed that miR-302b-3p negatively regulated Ulk1 expression (Fig. S3). FISH assay (Fig. 6F) confirmed the cytoplasmic localization of miR-302b-3p and its elevated expression in CUMS mice. Dual-luciferase reporter assays showed that miR-302b-3p mimics or inhibitor significantly changed the luciferase activity of circADAM9 WT and Ulk1 WT reporters, but not the MUT reporters (Fig. 6G-J), providing conclusive evidence for the specific binding between circADAM9 and miR-302b-3p, as well as between miR-302b-3p and Ulk1. These results proved that circADAM9 acted as a sponge of miR-302b-3p to regulate Ulk1.

Fig. 6.

CircADAM9 sponged miR-302b-3p to regulate Ulk1. (A) Venn diagram showing the overlap of miRNAs interacted with circADAM9 and Ulk1. (B) Interaction network of the 5 candidate miRNAs with circADAM9 and Ulk1. (C) Efficiency of RNA probe detected by RT-qPCR (n = 5). **p < 0.01, compared with oligo probe group. (D) Relative expression of the 5 candidate miRNAs after pulling down by corresponding probe (n = 5). *p < 0.05, **p < 0.01, compared with oligo probe group. (E) Relative expression of the 5 candidate miRNAs after circADAM9 overexpression and knockdown (n = 5). *p < 0.05, **p < 0.01, compared with Control group. (F) Location and expression of miR-302b-3p measured by FISH assay in prefrontal cortex of the mice. (G) Putative binding sites of miR-302b-3p in circADAM9. (H) Luciferase activity of circADAM9 WT and MUT after miR-302b-3p mimics or inhibitor transfection (n = 5). *p < 0.05, compared with Control group. (I) Putative binding sites of miR-302b-3p in 3’ UTR of Ulk1. (J) Luciferase activity of Ulk1 WT and MUT after miR-302b-3p mimics or inhibitor transfection (n = 5). *p < 0.05, compared with Control group

CircADAM9 promoted autophagy via miR-302b-3p/ULK1 axis in astrocytes

Having demonstrated that circADAM9 modulated Ulk1 through sponging miR-302b-3p, we proceeded to explore whether circADAM9 activated autophagy via miR-302b-3p/ULK1 signaling cascade in astrocytes. The astrocytes were transfected with LV-circADAM9 and miR-302b-3p mimics. The overexpression of circADAM9 and miR-302b-3p in astrocytes were validated by RT-qPCR (Fig. S4). The autophagic proteins expression were determined by western blot and immunofluorescence assay (Fig. 7A-C). As consistently observed in our experiments, LPS inhibited the expression of ULK1, ATG5, Beclin1, LC3-II and increased p62 expression, all of which were rectified by circADAM9 overexpression. MiR-302b-3p mimics aggravated the abnormal expression of autophagic proteins induced by LPS, and neutralized the restorative effects of circADAM9 overexpression. Ultrastructural analysis by TEM (Fig. 7D) further revealed that miR-302b-3p mimics counteracted circADAM9-induced autophagosome biogenesis. These data confirmed that circADAM9 activated ULK1-mediated autophagy via miR-302b-3p in vitro, and suggested the pivotal role of circADAM9/miR-302b-3p/ULK1 axis in promoting autophagy.

Fig. 7.

CircADAM9 activated autophagy through miR-302b-3p/ULK1 axis in astrocytes. (A) Relative protein expression of ULK1, ATG5, Beclin1, LC3 and p62 detected by Western blot (n = 3). *p < 0.05, compared with Control group, ^#p < 0.05, compared with LPS group, ^Δp < 0.05, compared with LPS + LV-circ group. (B-C) Protein expression of ULK1 and LC3 measured by immunofluorescence assay (scale bar: 50 μm). (D) Autophagosomes (red arrows) observed by TEM in astrocytes (scale bar in normal image: 2 μm, scale bar in magnified image: 1 μm)

CircADAM9 improved depressive-like behaviors and activated autophagy through miR-302b-3p/ULK1 axis in CUMS mice

The functional involvement of circADAM9/miR-302b-3p/ULK1 axis in depressive-like behaviors and autophagy was investigated in CUMS mice. The overexpression of circADAM9 and miR-302b-3p in mice were validated by RT-qPCR (Fig. S5). Behavioral tests (Fig. 8A-F) showed that circADAM9 overexpression alleviated CUMS-induced anhedonia, despair, anxiety and declining locomotion activity. MiR-302b-3p mimics not only exacerbated these symptoms, but also reversed the ameliorative effects of circADAM9 overexpression, indicating that circADAM9 relieved depressive-like behaviors via miR-302b-3p. The western blot and immunofluorescence assay (Fig. 8G-I) and ultrastructural analysis (Fig. 8J) revealed that miR-302b-3p mimics diminished the promotive effect of circADAM9 on ULK1-mediated autophagy. These findings in vivo were consistent with the results in vitro, demonstrating that circADAM9 activated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors in vivo.

Fig. 8.

CircADAM9 ameliorated depressive-like behaviors and promoted autophagy via miR-302b-3p/ULK1 axis in CUMS mice. (A-F) Sucrose preference, immobility time, time spent in central area, number of rearing, and movement distance of the mice in behavioral tests (n = 8). (G) Relative protein expression of ULK1, ATG5, Beclin1, LC3 and p62 measured by Western blot (n = 3). (H-I) Protein expression of ULK1 and LC3 determined by immunofluorescence assay (scale bar: 50 μm). (J) Autophagosomes (red arrows) observed by TEM in prefrontal cortex of the mice (scale bar in normal image: 2 μm, scale bar in magnified image: 1 μm). *p < 0.05, compared with Control group, ^#p < 0.05, compared with CUMS group, ^Δp < 0.05, compared with CUMS + LV-circ group

Discussion

The present study demonstrated that circADAM9 ameliorated depressive-like behaviors, and uncovered the role of circADAM9/miR-302b-3p/ULK1 axis in regulating autophagy in the pathophysiology of depression. Our investigations began with RNA sequencing and bioinformatics analysis, which revealed the critical involvement of circADAM9 and ULK1-mediated autophagy in depression. The relationship of circADAM9 and ULK1-mediated autophagy were further explored in vivo and in vitro. Interestingly, our results indicated that circADAM9 overexpression relieved the depressive symptoms induced by CUMS, and activated ULK1-mediated autophagy. Furthermore, we found that circADAM9 could sponge miR-302b-3p to modulate Ulk1 gene expression. In CUMS mice and astrocytes, miR-302b-3p mimics abolished the behavioral improvements and pro-autophagic effect modulated by circADAM9. These findings comprehensively suggested that circADAM9 activated autophagy via miR-302b-3p/ULK1 axis to alleviate depressive-like behaviors, as depicted in Fig. 9.

Fig. 9.

CircADAM9 activated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors

Autophagy is a self-protective intracellular degradation process, which serves as a pivotal homeostatic regulator in the pathophysiology of depression. During the onset of depression, stress triggers the release of damage-associated molecular patterns (DAMPs), which activate pattern recognition receptors (PRRs) and initiate inflammatory responses. The inflammatory signals disrupt monoamine neurotransmitter synthesis, glutamate metabolism, and neuroplasticity, contributing to the development of depression [14]. Importantly, autophagy prevents this process by degrading stress-induced DAMPs and PRRs (such as mitochondrial DNA and NLRP3 inflammasome), thereby attenuating inflammatory cascades and suppressing depression progression [15]. ULK1 is a primary activator of autophagy, which coordinates the entire autophagosome biogenesis process by activating distinct molecular complexes [10–13]. The suppression of ULK1-mediated autophagy in depression has been reported in some preclinical studies [4–7]. Consistent with these findings, we found the involvement of ULK1 and autophagy in depression through RNA sequencing and bioinformatics analysis. In CUMS mice, the downregulation of ULK1 was also observed, concomitant with the impaired autophagy. Emerging evidence indicates that astrocytic autophagy plays an important role in the pathophysiology of depression [28–30], leading to its increasing utilization in depression research [31, 32]. Thus, we employed astrocytes as an in vitro model for our investigation. The astrocytes were stimulated with LPS to induce inflammatory stress and cellular damage, and the results showed that ULK1-mediated autophagy was disrupted after LPS treatment. This outcome in vitro was accordant with the observation in vivo, providing compelling justification for utilizing astrocytes in depression research.

Non-coding RNAs have exhibited multifaceted functions in psychiatric disorders [33]. CircRNAs, characterized by a stable closed-loop structure, are increasingly recognized as significant regulator in depression. A prior study reported that circSTAG1 could regulate astrocyte dysfunction and depressive-like behaviors via N⁶-methyladenosine modification [16], while microglial circUBE2K was found to aggravated depression by modulating parental gene UBE2K via targeting HNRNPU [17]. Distinct from previous investigations, our research focused on discovering circRNAs acting as upstream regulators governing ULK1-mediated autophagy. We conducted circRNA sequencing in the prefrontal cortex of non-stressed and CUMS mice and identified circADAM9 from differentially expressed circRNAs. The subsequent in vivo and in vitro results demonstrated that circADAM9 ameliorates depressive-like behaviors, upregulated ULK1 expression, and rescued the autophagy impairment induced by CUMS or LPS. These findings presented the first evidence for the mechanistic involvement of circADAM9 in the pathogenesis of depression, and established a definitive functional interplay between circADAM9 and ULK1-mediated autophagy.

Our findings had established a connection between circADAM9 and ULK1-mediated autophagy, yet the underlying molecular mechanisms required further exploration. CircRNAs can regulate gene expression via the ceRNA mechanism to mediate diverse biological processes [21, 22]. We had observed that circADAM9 overexpression or knockdown induced corresponding changes in Ulk1 gene expression. This prompted us to postulate that circADAM9 regulated Ulk1 by sponging specific miRNAs. We identified five candidate miRNAs through bioinformatic analysis, among which miR-302b-3p possessed the most pronounced binding affinity to circADAM9. The expression of miR-302b-3p negatively correlated with circADAM9, while miR-302b-3p inversely regulated Ulk1 gene expression, inferring a potential ceRNA network. Subcellular localization analysis indicated cytoplasmic enrichment for both circADAM9 and miR-302b-3p, implying spatial compatibility for the interaction between them. Dual-luciferase reporter assays provided compelling evidence for the direct binding between circADAM9 and miR-302b-3p, as well as between miR-302b-3p and Ulk1. Ultimately, miR-302b-3p mimics diminished the improvement of depressive-like behaviors and the enhancement of ULK1-mediated autophagy modulated by circADAM9, suggesting that circADAM9 activated autophagy via miR-302b-3p/ULK1 axis to alleviate depressive-like behaviors. To our knowledge, no prior studies have implicated miR-302b-3p in the pathogenesis of depression. Our findings uncovered the important role of circADAM9/miR-302b-3p/ULK1 axis in regulating autophagy in depression.

In summary, the present work provided the first evidence that circADAM9 activated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors. Our study advances the mechanistic understanding of depression and proposes a new potential therapeutic target for the treatment of the disorder. However, our current research remains limitations. The regulatory landscape of circRNAs in diseases is an intricate network, whereas we specifically elucidated the circADAM9/miR-302b-3p/ULK1 axis in depression in this study. Additionally, novel therapeutic approaches targeting circADAM9 are urgently required. In the future, we will dedicate to map the comprehensive regulatory network of circRNAs in depression, and develop novel therapeutic strategies targeting circADAM9.

Conclusion

In conclusion, the present work provided the first evidence that circADAM9 activated autophagy via miR-302b-3p/ULK1 axis to improve depressive-like behaviors. Our study advances the mechanistic understanding of depression and proposes a new potential therapeutic target for the treatment of the disorder.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ANOVA

analysis of variance

ceRNA

competitive endogenous RNA

circRNA

circular RNA

CUMS

chronic unpredictable mild stress

DAMP

damage-associated molecular pattern

FISH

fluorescence in situ hybridization

FST

forced swimming test

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

LPS

lipopolysaccharide

miRNA

microRNA

OPT

open field test

PRR

pattern recognition receptor

RT-qPCR

real-Time quantitative PCR

SD

standard deviation

SPT

sucrose preference test

TEM

transmission electron microscopy

TST

tail suspension test

ULK1

Unc-51-like kinase 1

Author contributions

Shan Liu: Methodology, Formal analysis, Investigation, Writing-Original Draft. Shangmin Yu: Conceptualization, Validation, Supervision, Funding acquisition, Writing-Review & Editing.

Funding

This study was supported by the Foundation of University Natural Science Research Project of Anhui Province (No. 2024AH051224) and the Foundation of Bengbu Medical University (No. 2023byzd046).

Data availability

The RNA sequencing data in the study have been deposited in the NCBI Sequence Read Archive (SRA) repository under BioProject accession number PRJNA1304014. All the data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (publication No. 85 − 23, revised 1996) and were approved by the Laboratory Animal Welfare and Ethics Committee of Bengbu Medical University (Approval No. 2024 − 525). Clinical trial number: not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.