STUB1-mediated PGC-1α ubiquitination promotes postoperative cognitive dysfunction in aged mice via modulating autophagy–lysosome machinery

Abstract

Objective

Postoperative cognitive dysfunction (POCD) is a common neurological complication in elderly patients after anesthesia and surgery, and autophagy plays a critical regulatory role in its pathogenesis. This study aimed to systematically clarify the molecular mechanism by which STUB1 mediates POCD in aged mice.

Methods

A POCD model in aged mice was established by internal fixation of tibial fracture under anesthesia. Cognitive function was evaluated via conditioned fear test and Morris water maze, and neuronal damage and apoptosis was evaluated through H&E and TUNEL staining. To identify molecular mechanisms, RNA-seq was conducted on the hippocampus, and STUB1/PGC-1α/TFEB axis was identified. STUB1 was knocked out and overexpressed in mouse hippocampal neuron cells (HT22), respectively. The expression of STUB1/PGC-1α/TFEB axis, and the activity of downstream mitophagy, autophagy–lysosome, and apoptosis were investigated both in the hippocampus and neuron cells through qRT-PCR, immunofluorescence, western blot, transmission electron microscope, flow cytometry, and TUNEL staining. PGC-1α ubiquitination was measured through Co-IP.

Results

Following surgery and anesthesia, the aged mice showed cognitive dysfunction, and neuronal apoptosis, with elevated expression of E3 ubiquitin ligase STUB1, decreased PGC-1α and TFEB expression, and increased PGC-1α ubiquitination in the hippocampus. Moreover, aberrant PINK1/PARKIN-mediated mitophagy, and LAMP-1/LC3/Cathepsin D autophagy–lysosome signaling were found in the hippocampus. In neuron cells, STUB1 overexpression reduced the expression of PGC-1α and TFEB, enhanced PGC-1α ubiquitination, facilitated neuron cell apoptosis, and attenuated mitophagy and autophagy–lysosome pathways.

Conclusion

STUB1 directly binds to PGC-1α and promotes its ubiquitination and degradation, thereby inhibiting the TFEB-dependent autophagy–lysosome pathway and inducing hippocampal neuronal apoptosis, ultimately contributing to POCD in aged mice. This study provides a novel theoretical basis for understanding the pathogenesis of POCD and identifies potential therapeutic targets.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40001-025-03771-1.

Introduction

Postoperative cognitive dysfunction (POCD) is a prevalent neurological complication in elderly patients following anesthesia and surgery, characterized by short-term or long-term impairments in learning, memory, and other cognitive functions. POCD can persist for months, seriously affecting patients’ postoperative quality of life and overall health. Clinical studies have reported that the incidence of POCD ranges from 10 to 54%, with a pooled incidence of 29% [1, 2]. To date, effective prevention and treatment strategies for POCD remain lacking, primarily due to the incomplete understanding of its underlying molecular mechanisms.

Hippocampal neuronal apoptosis is a key pathological feature of POCD induced by anesthesia and surgery [3–5]. Advanced age is recognized as the most prominent risk factor for POCD, and mitochondrial homeostasis imbalance is a hallmark of aging, which also serves as the pathophysiological basis for mitochondrial dysfunction and hippocampal neuronal apoptosis after anesthesia and surgery [6]. Mitochondrial homeostasis is maintained through a series of processes including mitochondrial biogenesis, fission, fusion, and mitophagy. Regulation of mitochondrial homeostasis has been shown to alleviate neuronal apoptosis and cognitive impairment [7, 8]. Mitophagy, a process that eliminates damaged mitochondria via autophagy, is essential for maintaining mitochondrial homeostasis [9–11]. Mitophagy, a process that eliminates damaged mitochondria via autophagy, is essential for maintaining mitochondrial homeostasis [12]. Peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) is a core regulator of mitochondrial function and has been implicated in cognitive impairment [13–15]. In addition, interaction between PGC-1α and PINK1/PARKIN (a classical mitophagy pathway) is involved in maintenance of mitochondrial homeostasis [16]. Therefore, PGC-1α plays a central role in several processes that maintain mitochondrial homeostasis. Transcription factor EB (TFEB) functions as an important driver for enhancing the autophagy–lysosome machinery through transcriptionally regulating genes of autophagy and lysosome biogenesis [17, 18]. Although PGC-1α is reported to interact with TFEB, the specific regulatory mechanism between them in the context of POCD remains unclear [19–21].

Accumulating evidence indicates that ubiquitination is involved in mitochondrial protein metabolism and mitophagy, and participates in the regulation of POCD. STIP1 homology and U-box-containing protein 1 (STUB1), an E3 ubiquitin ligase, has been shown to be associated with cognitive impairment in cerebellar ataxias, and can regulate TFEB activity by promoting the ubiquitination and degradation of phosphorylated TFEB [19–21]. However, whether STUB1 is involved in POCD by regulating PGC-1α ubiquitination and the TFEB-dependent autophagy–lysosome pathway remains unreported. In this study, we hypothesized that STUB1 mediates PGC-1α ubiquitination and degradation, thereby inhibiting the TFEB-dependent autophagy–lysosome pathway and inducing hippocampal neuronal apoptosis, ultimately contributing to POCD in aged mice. This study aims to verify this hypothesis and provide new insights into the pathogenesis of POCD.

Materials and methods

Animals

Twelve week-old C57BL/6 male mice were purchased from Cavins, China. The mice were kept at 24 ± 1 °C, 50–70% relative humidity, and a 12 h light/dark cycle, with free to food and water. The animal experiments complied with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH) and the ARRIVE guidelines, and were approved by laboratory animal ethics committee of JXMC.

Model establishment

Mice were placed on a heating blanket to maintain normal body temperature in a customized transparent glass small animal anesthesia box. They were anesthetized by continuous inhalation of 1.5% iso-ether + 50% oxygen with an oxygen flow rate of 3L/min. EtCO₂, MAC value and oxygen concentration were continuously monitored by a multifunctional gas monitor (Drager, Germany). Four hours after inhaled anesthesia, open tibial fracture internal fixation was performed. In detail, the mice were immobilized in supine position and disinfected with towel. A longitudinal incision (about 1 cm long) was made in front of the tibial side to separate tibial fascia and muscle in turn, exposing tibial trochanter. A small hole was drilled here with a trocar, and a 0.7 cm sterile 22G trocar core was inserted into the tibial bone marrow cavity. A surgical clamp was then applied to the middle and lower third of the tibia resulting in an artificial fracture. A small amount of gentamicin was used to flush the wound. After the mice woke up, the movement of the hind limb of the surgical side was observed, and the success of the operation was judged on the basis that the joint movement was not affected. The oxygen saturation of toe pulse was monitored continuously during the operation. The wound was smeared with 2.5% lidocaine + 2.5% propoxycaine at 8 h intervals for 48 h after surgery. The control group inhaled 50% air + 50% oxygen for 4 h. To prevent interference of spatial density effect on subsequent behavioral tests, the original squirrel cage was used as the anesthesia box for control mice, and SP02 of control mice was not monitored to avoid stress.

Conditioned fear test

Conditioned fear test was conducted after successful modeling. This experiment was divided into the establishment of conditioned fear and the test of conditioned fear. The hardware adopted a 67 cm × 65 cm × 80 cm Skinner box, and the software adopted Visu Track rodent behavior video analysis software (version 3.0.0.1) from Shanghai Xinsoft Information Technology Co., LTD., China. Conditioned fear was established 48 h after surgery. The animal was placed in the Skinner box for 30 s, a single frequency sound signal (4.5KHZ, 60 dB, 30 s) was given 4 times. Then, 6 times of inescapable plantar electric shock (0.5 mA, 5 s) prompted by sound signal was given. Sound and electric shock ended at the same time, and the interval between the two sounds was 30 s. The conditioned fear test was performed 24 h after the conditioned fear was established. The conditioned fear animal was placed in the Skinner box again for 30 s, and was only given the sound signal (4.5KHZ, 60 dB, 30 s) for 10 times, each time at an interval of 30 s.

Morris water maze

This experiment included navigation test and space exploration test. For navigation test, the mice were trained for five consecutive days, four times a day, each time at one hour interval. The time it took the mice to get from four entry points to the water and from the water to find the platform was determined as the escape platform period. The average scores of the four incubation periods were counted as the final results of the day. For space exploration test, after six days, the platform was of removal while the animals were put into the water from the farthest end of the platform. The swimming track in 60 s was recorded, and the staying time of the mice in the target quadrant and the number of times it crossed the platform were observed and recorded.

Hematoxylin–eosin (H&E) staining

Mouse hippocampus tissues were dehydrated, transparent, and waxed in the dehydrator (JT-12 J, WHJJ, China), followed by paraffin embedding, slicing, and baking. After rinsing the sections with distilled water, hematoxylin (H9627, Sigma, USA) was added and stained for 5 min. The sections were washed with distilled water, and 1% ethanol hydrochloride (10011018, Sinopharm, China) was added to differentiate for 2 s. After rinsing, the sections were dyed utilizing eosin (71014544, Sinopharm) for 2–8 s. Next, they were rinsed and dehydrated with anhydrous ethanol (10009218, Sinopharm). Neutral gum (10004160, Sinopharm) was added and sealed. Utilizing an IX71 light microscope (Olympus, Japan), images were observed and photographed.

TUNEL staining

Following the TUNEL staining kit (C1088, Beyotime, China) manufacturer’s instruction, TUNEL staining was conducted. The sections were washed with PBS, and 100 μL protease K reaction solution containing 2000U was dropped on the sections and treated at 37 ℃ for 20 min. After rinsing the sections with PBS (3 times × 5 min), 50 μL TUNEL test solution was added and reacted at 37 ℃ for 1 h without light. After rinsing the sections with PBS (3 times × 5 min), Hoechst staining (C1017, Beyotime) was conducted at room temperature for 5 min. The slices were rinsed by PBS (3 times × 5 min) as well as sealed by anti-fluorescence quencher. Results were analyzed and photographed utilizing an IX71 light microscope (Olympus).

RNA sequencing (RNA-seq)

Total RNA was extracted, and cDNA library was prepared. The paired-end reads were generated on Illumina Novaseq 6000. Differentially expressed transcripts between POCD and control groups were screened in accordance with |log2fold change|> 1 & adjusted p < 0.05 [22, 23]. Based on them, functional enrichment analysis was conducted.

qRT-PCR

Total RNA extraction was conducted by Trizol (15596018, Invitrogen, USA). RNA concentration was measured, and cDNA was synthesized by reverse transcription synthesis kit. Primers were synthesized according to the primer sequences: mouse Stub1, 5′-GATAAGAGCCCGAGTGC-3’ (forward (F)), 5′-AGTGGGTTCCGAGTGAT-3′ (reverse (R)); mouse Pgc-1α, 5′-GGTGTAGCGACCAATC-3′ (F), 5′-ATCAAATGAGGGCAAT-3′ (R); mouse Tfeb, 5′-AGGAGCGGCAGAAGAA-3′ (F), 5′-AGGATGGTGCCTTTGTT-3′ (R); mouse Pink1, 5′-CAAGTCCGACAACATCCT-3′ (F), 5′-GCCACCACGCTCTACAC-3′ (R); mouse Parkin, 5′-ACAAGGACACGTCGGTAG-3′ (F), 5′-TAGCCAAGTTGAGCATCG-3′ (R); mouse Lamp-1, 5′-CAAGACGGTGACCAGA-3′ (F), 5′-GCATTCATCCCAAACT-3′ (R); mouse Cathepsin D, 5′-GCAGTGCCTCTTATCC-3′ (F), 5′-TGTCTGGGTGTAGTTCAT-3′ (R); mouse Caspase3, 5’-TGGAAAGCCGAAACTC-3′ (F), 5′-GGATGAACCACGACCC-3′ (R); mouse Gapdh, 5′-GCCTTCCGTGTTCCTA-3′ (F), and 5′-AGACAACCTGGTCCTCA-3′ (R). Real-time PCR reaction system was prepared, and PCR amplification was performed. Real-time fluorescence quantitative PCR analyzer (7500, ABI, USA) was used to automatically analyze the results. Relative expression levels were calculated with 2^−△△CT method.

Immunofluorescence

The tissue sections were dewaxed and repaired by antigen. The cells were fixed in 4% paraformaldehyde for 15 min, and exposed to 0.1% Triton for 10 min. Under the serum closure, primary antibodies were added and incubated at 4 °C overnight. The slices were rinsed by PBS (3 times × 5 min), with subsequent incubation with secondary antibodies at room temperature for 50 min away from light. Hoechst staining (C1017, Beyotime) was performed at room temperature for 5 min, with subsequent PBS rinsing (3 times × 5 min). Subsequently, the sections were sealed with anti-fluorescence quencher, followed by image acquisition under an IX71 light microscope (Olympus). Primary antibodies included STUB1 (A11751, ABclonal, China), PGC-1α (1:100, 66369-1-Ig, Proteintech, China), TFEB (1:100, 13372-1-AP, Proteintech), LAMP-1 (1:100, ab62562, Abcam, USA), and LC3 (1:100, CL488-14600, Proteintech).

Western blot

Protein extraction was conducted utilizing RIPA lysis buffer (P0013B, Beyotime), with subsequent protein concentration measurement by BCA kit (BL521A, Biosharp, China). The extracted protein was separated utilizing PAGE, with subsequent transference to PVDF membrane (IPVH00010, Millipore, Germany). The membrane was rinsed by TBST (3 times × 5 min), and then slowly shaken with 5% BSA at 37 ℃ for 2 h. Incubation with primary antibodies of STUB1 (1:1000, 68407-1-Ig, Proteintech), PGC-1α (1:1000, 66369-1-Ig, Proteintech), TFEB (1:1000, 13372-1-AP, Proteintech), PINK1 (1:1000, ab300623, Abcam), PARKIN (1:1000, 14060-1-AP, Proteintech), LAMP-1 (1:1000, 21997-1-AP, Proteintech), Cathepsin D (1:1000, 21327-1-AP, Proteintech), Caspase3 (1:1000, 19677-1-AP, Proteintech), ubiquitin (1:1000, 10201-2-AP, Proteintech), and GAPDH (1:5000, 60004-1-Ig, Proteintech) was conducted at 4 ℃ overnight. After rinsing by TBST (3 times × 1 min), incubation with HRP-labeled goat anti-rabbit or anti-mouse IgG (1:5000, ZB2301 or ZB2305, ZSGB-BIO) was conducted at 37 ℃ for 1 h. After rinsing with TBST (3 times × 5 min), ECL luminescent solution (ECL-03-250, 7Sea Biotech, China) was added, and images were acquired via an integrated chemiluminescence apparatus (ChemiScope 5300 Pro, CLINX, China).

Transmission electron microscope

Mouse hippocampus tissues were fixed with 1% osmic acid in 0.1 M phosphoric acid buffer at room temperature for 2 h out of the light, and rinsed with 0.1 M phosphate buffer (3 times × 15 min). After dehydration at room temperature, the specimens were exposed to acetone (10000418, Sinopharm): 812 embedding agent (90529-77-4, SPI) (1:1) at 37 ℃ for 3 h, acetone: 812 embedding agent (1:2) at 37 ℃ overnight as well as pure 812 embedding agent at 37 ℃ for 6 h. After pouring pure 812 embedding agent into the embedding plate, they were inserted into the oven at 37 ℃ behind the embedding plate overnight. The embedded plates were polymerized in an oven at 60 ℃ for 48 h, and the resin blocks were removed for use. The resin blocks were sliced 60–80 nm in an ultra-thin microtome, with 150 mesh copper mesh. The copper mesh was dyed by 2% uranium acetate saturated alcohol solution (02624-AB, SPI) without light for 8 min, and cleaned with 70% alcohol for 3 times and ultra-pure water for 3 times. 2.6% lead citrate solution was dyed for 8 min with carbon dioxide avoidance, washed with ultra-pure water for 3 times, and slightly dried by filter paper. After putting into the copper mesh box, the copper mesh sections were dried overnight at room temperature.

Cell fixation was conducted by 2.5% glutaraldehyde for 4 h. Then, the cells were collected by cell scraping, and centrifuged at 1000 rpm. Fresh 2.5% glutaraldehyde was replaced and the cells were fixed at 4 ℃ for 2–4 h. After rinsing with 0.1 M phosphate buffer (3 times × 15 min), the cells were fixed by 1% osmic acid in 0.1 M phosphate buffer at room temperature for 2 h, and washed by 0.1 M phosphate buffer (3 times × 15 min). After dehydration, the cells were exposed to acetone: 812 embedding agent (1:1) overnight as well as pure 812 embedding agent overnight. The cells were polymerized at 60℃ for 2 days, with subsequent preparation of 60–80 nm ultra-thin sections. The sections were dyed by 2% uranium acetate saturated solution as well as lead citrate for 15 min each. After drying, the images were investigated utilizing a HT7800/HT770 transmission electron microscope (HITACHI, Japan).

Cell culture and transfection

Mouse hippocampal neuron cells HT22 (CTCC-551, CTCC, Zhejiang, China) were maintained in DMEM (SH30243.01, Hyclone, China) + 10% fetal bovine serum (FBS) (A511-001, Lonsera, China) + 1% penicillin–streptomycin (BL505A, Biosharp). HT22 cells were cultured in a 5% CO₂ incubator at 37℃.

The cells were inoculated into a 6-well plate (5 × 10⁵ cells /well). On the next day, each well was exchanged by 2 mL fresh medium. 125 μL of antibiotic- and serum-free medium, 100pmoL of small interfering RNA (siRNA) of Stub1, 2.5 μg of DNA, and 4μL Lipo8000™ transfection reagent (C0533, Beyotime) were mixed and stored at room temperature for 6 h. Each well was added with 125μL mixture. The full-length of Stub1 sequences were amplified by PCR and cloned into pcDNA3.1 vector (V012531, NovoPro, China). Through Lipo8000™ transfection reagent, the cells were transfected with pcDNA3.1 vector. After 48 h culture, transfection efficacy was evaluated.

Flow cytometry

Following the Annexin V-FITC/PI kit (CTCC-M009, CTCC) manufacturer’s instructions, apoptosis was detected. The cells were digested and centrifuged at 1500 rpm at room temperature for 5 min. They were gathered and suspended by precooling 1 × PBS. After centrifuging at 1500 rpm for 5 min, the cells were washed and suspended by 300 µL 1 × binding buffer, and exposed to 5 µL Annexin V-FITC at room temperature for 15 min, away from light. Then, 10 µL of PI was utilized to stain the cells at room temperature for 10 min away from light. Apoptotic levels were tested utilizing flow cytometry (FACSVerse, BD, USA), and assessed utilizing FlowJo v7.6 software.

Co-immunoprecipitation (Co-IP)

The cells were lysed by Western/IP lysis buffer supplemented with protease and phosphatase inhibitors. After centrifuging at 12,000 × g, the supernatant samples were precleared with protein A/G-magnetic beads (MedChemExpress, USA), and incubated with primary antibodies at 4 °C overnight, followed by incubation with protein A/G-magnetic beads at 4 °C for 2 h. The beads were washed by Western/IP lysis buffer, and the indicated proteins were measured utilizing western blot.

Statistical analysis

All data are presented as the mean ± standard deviation (SD) from at least three independent experiments. Sample size was determined based on statistical power analysis, with a power of 0.8 and α = 0.05. Statistical analysis was performed using R v3.6.1 and GraphPad Prism v9.0.0 software. Differences between two groups were analyzed by unpaired Student’s t test. Multiple comparisons were performed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. P < 0.05 was considered statistically significant. Detailed statistical methods, n values, and error bar definitions are specified in each figure legend.

Results

Cognitive dysfunction and hippocampal neuronal apoptosis in aged mice with anesthesia and surgery

The present study constructed a POCD model in aged mice via internal fixation of tibial fracture under anesthesia. To evaluate cognitive function, conditioned fear test and Morris water maze were carried out. For the conditioned fear test, the ratio of freezing time was declined in POCD mice than controls (Fig. 1A, B). In addition, as shown in the Morris water maze test results, the number of shuttles in the target quadrant and the escape platform distance were declined in POCD mice than controls (P < 0.05), without difference in the latency and total distance (Fig. 1C–G). This was indicative of impaired spatial learning and memory of aged mice following internal fixation of tibial fracture and anesthesia. To assess the neuronal damage, H&E staining was conducted on mouse hippocampus tissues. Consequently, the neurons of controls were tightly arranged, while those in POCD mice were disordered, loosely arranged and the nuclei dissolved (Fig. 1H). As depicted in TUNEL staining, hippocampal neuronal apoptosis was notably increased following anesthesia and surgery (Fig. 1I, J). These results confirm that anesthesia and surgery induce cognitive dysfunction and hippocampal neuronal apoptosis in aged mice.

Fig. 1.

Cognitive dysfunction and neuronal apoptosis in the hippocampus in aged mice after anesthesia and surgery. A, B Conditioned fear test and analysis of ratio of freezing time in the POCD model and control groups (n = 3 per group). The POCD model was constructed by internal fixation of tibial fracture under anesthesia. C–G Morris water maze test of the latent period, total distance, number of shuttles in the target quadrant, and escape platform distance. H H&E staining of mouse hippocampus tissues. Scale bar, 50 μm; n = 6 per group. I, J TUNEL staining of mouse hippocampus tissues and analysis of TUNEL-positive cell percentage. Scale bar, 50 μm; n = 6 per group. ns, p > 0.05; **p < 0.01; ***p < 0.001 (Student’s t test)

Aberrant STUB1/PGC-1α/TFEB axis in the hippocampus of aged mice with anesthesia and surgery

To investigate the molecular mechanisms underlying POCD, RNA-seq was conducted on the hippocampus of POCD and control mice. Consequently, 1044 transcripts were upregulated and 529 transcripts were downregulated after anesthesia and surgery (Fig. 2A, B), which might be involved in POCD. Several key pathways (e.g., apoptosis, mitophagy, lysosome, ubiquitin-mediated proteolysis, and proteasome) were remarkably enriched by the differentially expressed transcripts (Fig. 2C–G). Among the differentially expressed transcripts, PGC-1α (also known as PPARGC1A), a transcriptional coactivator mediating mitochondrial biogenesis and autophagy, presented significant downregulation in the hippocampus of aged mice with anesthesia and surgery (P < 0.05). Through the Ubibrowser (http://ubibrowser.bio-it.cn/ubibrowser/), STUB1 was inferred as a potential E3 ubiquitin ligase of PGC-1α (Supplementary Fig. 1). Prior studies have suggested that TFEB may be a direct target of PGC-1α, thus regulating mitochondrial biogenesis and autophagy [21, 24, 25]. The STRING predicted that mitophagy-, autophagy–lysosome-, and apoptosis-related markers were interacted with PGC-1α (Supplementary Fig. 2). Altogether, we identified the STUB1/PGC-1α/TFEB axis that might be involved in POCD. qRT-PCR, multiple immunofluorescence, and western blot demonstrated that STUB1 expression was elevated, and /PGC-1α and TFEB expression was declined in the hippocampus of aged mice with anesthesia and surgery (Figs. 2H–N and 3A–D). These results indicate that the STUB1/PGC-1α/TFEB axis is abnormally regulated in POCD mice.

Fig. 2.

Aberrant STUB1/PGC-1α/TFEB axis in the hippocampus of aged mice following anesthesia and surgery. A, B RNA-seq for screening differentially expressed transcripts in the hippocampus of aged mice following anesthesia and surgery. The screening criteria were |log2fold change|> 1 and adjusted p < 0.05. C–G Significant enrichment of (C) apoptosis, (D) mitophagy, (E) lysosome, (F) ubiquitin-mediated proteolysis, and (G) proteasome pathways by the differentially expressed transcripts. H–J qRT-PCR of Stub1, Pgc-1α, and Tfeb transcript levels in the hippocampus of POCD and control mice. K, L Multiple immunofluorescence of STUB1 and PGC-1α in the hippocampus of POCD and control mice. M, N Multiple immunofluorescence of PGC-1α and TFEB in the hippocampus of POCD and control mice. Scale bar, 50 μm; n = 6 per group. **p < 0.01; ***p < 0.001; ****p < 0.0001 (Student’s t test)

Fig. 3.

The STUB1/PGC-1α/TFEB axis governs autophagy–lysosome machinery in the hippocampus of aged mice following anesthesia and surgery. A–I Western blot of the expression of STUB1, PGC-1α, TFEB, PINK1, PARKIN, LAMP-1, Cathepsin D, Caspase3 in the hippocampus of POCD and control mice. J, K Multiple immunofluorescence of LAMP-1 and LC3 in the hippocampus of POCD and control mice. Scale bar, 50 μm. L Transmission electron microscope of the hippocampus in POCD and control mice. Scale bar, 500 nm; n = 6 per group. ***p < 0.001; ****p < 0.0001 (Student’s t test)

STUB1/PGC-1α/TFEB axis governs autophagy-lysosome machinery in the hippocampus of aged mice with anesthesia and surgery

Through western blot, mitophagy markers (PINK1/PARKIN), autophagy–lysosome markers (LAMP-1, Cathepsin D), apoptosis marker (Caspase3) were detected in the hippocampus of aged mice with anesthesia and surgery. It was proven that PINK1, PARKIN, LAMP-1, Cathepsin D, and Caspase3 expression presented notable elevation following anesthesia and surgery (Fig. 3A and E–I), indicating the activation of mitophagy and autophagy–lysosome pathways. Multiple-immunofluorescence also demonstrated the elevation of LAMP-1 and LC3 expression in the hippocampal tissues after anesthesia and surgery, (Fig. 3J, K), indicating autophagosome–lysosome fusion was enhanced. The transmission electron microscope results showed that autophagosome, lysosome, and autophagolysosome were increased in the hippocampus with anesthesia and surgery, with mitochondrial enlargement and blurred mitochondrial structure (Fig. 3L). These results suggest that the abnormal STUB1/PGC-1α/TFEB axis regulates autophagy–lysosome pathway dysfunction and mitophagy activation in POCD mice.

STUB1 overexpression declines PGC-1α and TFEB expression in mouse hippocampal neuron cells

To further verify the STUB1/PGC-1α/TFEB axis in POCD, STUB1 was knocked out and overexpressed in mouse hippocampal neuron cells (HT22), respectively (Fig. 4A, B). STUB1 knockdown elevated PGC-1α and TFEB expression, while STUB1 overexpression significantly (P < 0.05) declined PGC-1α and TFEB expression in HT22 cells in accordance with the results of qRT-PCR and multiple-immunofluorescence (Fig. 4C–H). These results indicate that STUB1 negatively regulates PGC-1α and TFEB expression in hippocampal neurons.

Fig. 4.

STUB1 overexpression declines the expression of PGC-1α and TFEB in mouse hippocampal neuron cells. A qRT-PCR of Stub1 transcript levels in mouse hippocampal neuron cells (HT22) transfected with Stub1 siRNA (si-Stub1). B–D qRT-PCR of Stub1, Pgc-1α, and Tfeb transcript levels in HT22 cells transfected with si-Stub1 or Stub1 overexpression (OE-Stub1) plasmid. E, F Multiple-immunofluorescence of STUB1 and PGC-1α in HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. G, H Multiple-immunofluorescence of PGC-1α and TFEB in HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. Scale bar, 50 μm; n = 6 per group. **p < 0.01; ***p < 0.001; ****p < 0.0001 (Student's t-test)

STUB1 overexpression facilitates apoptosis and regulates autophagy-lysosome machinery in mouse hippocampal neuron cells

Flow cytometry showed that STUB1 overexpression significantly (P < 0.05) increased the apoptotic rate of HT22 cells and primary hippocampal neurons, while STUB1 knockdown had no significant effect on apoptosis (P > 0.05, Fig. 5A, B). Through multiple-immunofluorescence, LAMP-1 and LC3 expression was measured to characterize autophagosome–lysosome fusion. Consequently, STUB1 overexpression led to a notable decline in LAMP-1 and LC3 expression (Fig. 5C, D), indicating the decline in autophagosome fusion by STUB1 overexpression. The opposite effect was observed when STUB1 was knock out. The TEM also confirmed the results (Fig. 5E). These results indicated that STUB1 overexpression facilitated apoptosis and controlled autophagy–lysosome machinery in mouse hippocampal neuron cells.

Fig. 5.

STUB1 overexpression facilitates apoptosis and autophagy-lysosome machinery in mouse hippocampal neuron cells. A, B Flow cytometry for measuring apoptosis in mouse hippocampal neuron cells (HT22) transfected with Stub1 siRNA (si-Stub1) or Stub1 overexpression (OE-Stub1) plasmid. C, D Multiple-immunofluorescence of LAMP-1 and LC3 in HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. Scale bar, 50 μm; n = 6 per group. E Transmission electron microscope of HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. Scale bar, 1 μm. ns, p > 0.05; ****p < 0.0001 (Student’s t test)

STUB1 overexpression reinforces PGC-1α ubiquitination to modulate autophagy-lysosome machinery

As depicted in the qRT-PCR results, STUB1 overexpression decreased (P < 0.05) the expression of Pink1, Parkin, Lamp-1, and Cathepsin D as well as elevated (P < 0.05) the expression of Caspase3 in HT22 cells (Fig. 6A–E). Western blot also proven that STUB1 overexpression led to a remarkable decline (P < 0.05) in STUB1, PGC-1α, TFEB, PINK1, PARKIN, LAMP-1, and Cathepsin D as well as led to a remarkable decline (P < 0.05) in Caspase3 in HT22 cells (Fig. 6F–N). The opposite results were observed when STUB1 was knock out. Co-IP showed that ubiquitination levels of PGC-1α were notably elevated in the hippocampus of aged mice with anesthesia and surgery (Fig. 6O). STUB1 overexpression resulted in the elevation (P < 0.05) in PGC-1α ubiquitination in HT22 cells (Fig. 6P). We have validated the key targets (STUB1, PGC-1α, TFEB) in human clinical POCD samples by analyzing the public dataset GEO: GSE115446. The results show that their expression trends are consistent with those in our mouse model: STUB1 is upregulated in the POCD group, while PGC-1α and TFEB are downregulated (as shown in the Supplementary Figs. 3, 4 and 5). These results demonstrate that STUB1 directly binds to PGC-1α, promote its ubiquitination and degradation, and thereby regulate the autophagy–lysosome pathway and neuronal apoptosis.

Fig. 6.

STUB1 overexpression reinforces PGC-1α ubiquitination to modulate autophagy-lysosome machinery. A–E qRT-PCR of Pink1, Parkin, Lamp-1, and Cathepsin D, and Caspase3 transcript levels in HT22 cells transfected with Stub1 siRNA (si-Stub1) or Stub1 overexpression (OE-Stub1) plasmid. F–N Western blot of the expression of STUB1, PGC-1α, TFEB, PINK1, PARKIN, LAMP-1, Cathepsin D, Caspase3 in HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. O Co-IP of PGC-1α ubiquitination levels in the hippocampus of POCD and control mice. P Co-IP of PGC-1α ubiquitination levels in HT22 cells transfected with si-Stub1 or OE-Stub1 plasmid. ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 (Student's t-test)

Discussion

POCD is a recognized clinical phenomenon, manifesting as cognitive dysfunction in elderly patients following anesthesia and surgery. Herein, we constructed a POCD model in aged mice via internal fixation of tibial fracture under anesthesia. The aged mice presented cognitive impairment, further confirming anesthesia and surgery as a dominant cause of POCD and aging as a key risk factor. Cognitive impairment has also been observed in other POCD mouse models, such as exploratory laparotomy and anesthesia [26], and unilateral nephrectomy and anesthesia [27]. While numerous mechanisms have been reported to mediate POCD, the underlying mechanisms behind it remain poorly understood.

Hippocampal neuronal apoptosis, a key mechanism underlying POCD, was observed in the hippocampal tissues of aged mice with anesthesia and surgery. On the basis of RNA-seq, PGC-1α, an important modulator of mitochondrial function, was found and proven to present downregulation in the hippocampal tissues of aged mice with anesthesia and surgery, consistent with previous studies [28, 29]. Our study explored the regulatory mechanisms of PGC-1α. Protein ubiquitination degradation is a process of protein degradation through proteasome after ubiquitination modification, which is a common endogenous protein degradation mode in cells [30]. Three ubiquitin enzymes, E1, E2 and E3, are involved in this process, among which E3 ubiquitin is the most diverse, and has been deeply studied because of its extensive involvement in various cellular life activities and signal regulation [31]. Protein ubiquitination is involved in mitochondrial protein metabolism and exerts an indispensable role in maintaining mitochondrial function, intracellular environment and mitophagy [32–34]. Several studies have demonstrated that ubiquitin modulates POCD [35, 36]. STUB1 presented upregulation in the hippocampus following anesthesia and surgery, which was proven as an E3 ubiquitin ligase of PGC-1α to induce PGC-1α degradation through enhancing its ubiquitination.

Clinically, STUB1 variants are a common cause of cognitive impairment in cerebellar ataxias [37]. STUB1 suppresses TFEB activity through preferentially targeting inactive phosphorylated TFEB for proteasomal degradation via the ubiquitin–proteasome machinery [17]. TFEB functions as a master regulator for reinforcing the autophagy–lysosome machinery through transcriptional regulation of autophagy- and lysosome biogenesis-associated genes [17, 18]. The research uncovered the downregulation of TFEB in the hippocampus following anesthesia and surgery. As previously reported [19–21], there is an interaction between PGC-1α and TFEB. For example, PGC-1α alleviates Huntington’s disease proteotoxicity through suppressing oxidative stress as well as reinforcing TFEB activity [19]. TFEB nuclear translocation binds to PGC1α promoter to induce PGC1α transcription and facilitate lipid metabolism and inflammatory response of astrocytes in POCD [38]. TFEB-mediated autophagy-lysosome pathway participates in modulating POCD in aged mice [39].

Based on the above, we proposed the STUB1/PGC-1α/TFEB axis as a key mechanism of POCD. PGC-1α ubiquitination levels were elevated in the hippocampus of aged mice with anesthesia and surgery. In addition, in neuron cells, STUB1 overexpression declined PGC-1α expression via reinforcing its ubiquitination and its protein degradation. Subsequently, PGC-1α attenuated downstream TFEB expression, thus facilitating neuron cell apoptosis through attenuating autophagy–lysosome pathway. Loss of mitochondrial and protein homeostasis governs hippocampal neuronal apoptosis [40, 41], and clearing damaged proteins and mitochondria is crucial for alleviating apoptosis of hippocampal neurons after stress through autophagy–lysosome pathway [42, 43]. Our study showed that PINK1/PARKIN-mediated mitophagy, and LAMP-1/LC3/Cathepsin D autophagy–lysosome signaling were abnormal in the hippocampus of aged mice with internal fixation of tibial fracture under anesthesia. STUB1 overexpression facilitated neuron cell apoptosis, and attenuated mitophagy and autophagy-lysosome pathways. Therefore, our study revealed that STUB1-mediated PGC-1α ubiquitination resulted in POCD of aged mice through modulating TFEB-mediated autophagy–lysosome machinery.

Conclusion

In conclusion, our study indicated that after anesthesia and surgery in aged mice, the expression of E3 ubiquitin enzyme STUB1 in neuronal cells was increased, which declined the expression of its targeted protein PGC-1α through reinforcing PGC-1α ubiquitination and its protein degradation. Subsequently, PGC-1α directly decreased the expression of TFEB and inhibited the mitophagy–lysosome pathway, resulting in disruption of mitochondrial homeostasis and increased apoptosis, eventually driving POCD. Therefore, our findings provided novel theoretical basis for the pathogenesis of POCD.

Abbreviations

POCD

Postoperative cognitive dysfunction

PGC-1α

Peroxisome proliferator-activated receptor-γ coactivator-1α

TFEB

Transcription factor EB

STUB1

STIP1 homology and U-box-containing protein 1

H&E

Hematoxylin–eosin

RNA-seq

RNA sequencing

FBS

Fetal bovine serum

siRNA

Small interfering RNA

Co-IP

Co-immunoprecipitation

SD

Standard deviation

ANOVA

Analysis of variance

Author contributions

Q. H. S. and J. Y. X.: Study concept and design; S. S. X: Acquisition of the data and analysis and interpretation of the data; Y. M: Drafting of the manuscript and critical revision of the manuscript for important intellectual content.

Funding

This research was supported by Zhejiang Natural Science Exploration Project (LQ23H090002), Zhejiang Provincial Clinical Key Specialties-Anesthesiology (2023-ZJZK-001) and Jiaxing Supporting Disciplines—Anesthesiology (2023-ZC-001).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.