2.1 The Expression of TRIM65 Is Increased in AKI

To investigate the role of TRIM proteins in AKI, two mouse models were established: an I/R and a rhabdomyolysis-induced AKI model. The expression levels of TRIM family members in renal tissues from both models were quantified by RT-qPCR. As illustrated in Figure 1A, numerous TRIM family members exhibited a striking upregulation, with TRIM65 expression being particularly pronounced. In the renal tissue of rhabdomyolysis-induced AKI, TRIM65 expression increased 23-fold, while in I/R-induced AKI renal tissue, it increased threefold. Histological examination by hematoxylin and eosin (H&E) staining revealed the development and progression of proximal tubular dilation, epithelial simplification, inflammatory cell infiltration, and cast formation in the renal tissue of the glycerol injection group in comparison to the saline group (Figure 1B). Immunohistochemical (IHC) staining using TRIM65-specific antibodies indicated a significant upregulation of TRIM65 in renal tissue of AKI mice induced by rhabdomyolysis, with a predominant distribution in renal tubular epithelial cells, and no significant difference observed in glomeruli (Figure 1B). Similarly, extensive injury to the renal tubular epithelium was observed in I/R-induced AKI mice, which was characterized by tubular dilation and swelling, cast formation, and interstitial inflammatory cell infiltration (Figure 1C). Moreover, IHC staining showed a significant increase in TRIM65 expression in tubular cells following I/R surgery when compared to the sham group (Figure 1C). Western blot analysis indicated a notable elevation in TRIM65 expression in the kidneys of mice subjected to rhabdomyolysis and I/R-induced AKI models (Figure 1D,E). Furthermore, to investigate whether TRIM65 expression is also elevated in human AKI samples, the expression matrix was downloaded from the GEO dataset (GSE139061), which includes data from nine reference nephrectomies and 39 human AKI samples. Following normalization using the edgeR package, the expression of TRIM65 was found to be significantly increased in AKI samples, as determined by the Wilcoxon test (p < 0.001, Figure 1F). These findings indicate that TRIM65 is induced in AKI.

2.2 Genetic Deletion of TRIM65 Protected AKI in Experimental Mice

To ascertain the function of TRIM65 in AKI, we initially assessed the impact of TRIM65 gene deletion on renal function and pathological alterations in rhabdomyolysis, I/R, and cisplatin-induced mouse models of AKI. Serum levels of urea nitrogen and creatinine (BUN and sCr) were markedly elevated in WT mice with AKI, whereas TRIM65 deletion resulted in a notable decline in these levels (p < 0.001, Figure 2A,B, see also Figures S1A,B and S2A,B). Western blot analysis showed that two of the most utilized markers of AKI, kidney injury molecule 1 (KIM-1) and neutrophil gelatinase-associated lipocalin (NGAL), were found to be significantly upregulated in the renal tissue of WT mice that bad been induced to develop AKI. In contrast, these markers were significantly inhibited in the renal tissues of Trim65^−/− mice (Figure 2C,D, see also Figures S1C,D, S2C, S3A). Histological examination of mouse renal tissues revealed that TRIM65 deficiency resulted in the attenuation of AKI (Figure 2E,F, see also Figures S1E,F and S2D). IHC staining suggests that the expression of KIM-1 and NGAL was elevated in AKI mice. Consequently, the elevated protein level of KIM-1 and NGAL observed after I/R was markedly decreased by TRIM65 deletion (Figure 2E,G, see also Figures S1E,G and S2D). Additionally, the mRNA expression levels of the inflammatory factors MCP-1, IL-6, and IL-10 in the kidney tissues were markedly elevated (Figure S3B). In comparison to WT mice, the mRNA expressions of pro-inflammatory factors MCP-1 and IL-6 were decreased, whereas the anti-inflammatory factor IL-10 was significantly increased in Trim65^−/− mice (Figure S3B). To evaluate renal tissue function, a GFR experiment was conducted utilizing a fluorescently labeled exogenous tracer, FITC-sinistrin. The results demonstrated that the clearance time of the tracer in the PBS group was approximately 70 min, while the clearance time in the cisplatin-induced WT mice group was significantly extended. In contrast, the tracer clearance time in Trim65^−/− mice was significantly shorter than that of WT mice (Figure S2E,F). Furthermore, Trim65^−/− mice exhibited greater tolerance to cisplatin-induced death than WT mice (Figure S2G). These findings indicate that TRIM65 deficiency protects against AKI in experimental mice.

2.3 TRIM65 Physically Binds to VDAC1

To elucidate the mechanism of action of TRIM65 in AKI, we employed yeast two-hybridization to identify potential substrates [14]. Among these, VDAC1 emerged as a particularly promising candidate, given that VDAC1 has recently been reported to play a pivotal role in AKI [16, 17]. First, a yeast two-hybrid system was employed to ascertain the interaction between TRIM65 and VDAC1 (Figure 3A). To further substantiate these findings, a co-immunoprecipitation (co-IP) experiment was conducted in HEK293T cells. As shown in Figure 3B,C, TRIM65 can bind with VDAC1 and vice versa. Furthermore, the interaction between the two entities was confirmed in HK-2 cells (Figure 3D). Another TRIM family protein, TRIM47, which is in the adjacent position of TRIM65 in human chromosome 17, was observed to fail to bind VDAC1 (Figure S4A). In contrast, Bax2, a known VDAC1 binding protein, was found to detect their interaction in HK-2 cells (Figure S4B). Given the interaction between TRIM65 and VDAC1, we sought to determine whether they co-localize at the subcellular level in HK-2 cells. Immunofluorescence experiments demonstrated that TRIM65 was predominantly cytoplasmic, with partial nuclear expression, while VDAC1 was exclusively cytoplasmic and exhibited strong co-localization with TRIM65 (Figure 3E). To ascertain whether there is a physical direct interaction between TRIM65 and VDAC1, a GST pull-down experiment was conducted. The results showed that GST-TRIM65 could pull down VDAC1 in vitro, whereas none of the VDAC1 protein interacted with the beads coated with GST (Figure 3F), which suggests that TRIM65 may directly interact with VDAC1.

TRIM65, a member of the TRIM protein family, is composed of four domains: the Really Interesting New Gene (RING), B-Box-type Zinc Finger (BBOX), Coiled Coil (CC), and SPla domain of the Ryanodine Receptor (SPRY) [14]. To identify which domain of TRIM65 was responsible for the interaction with VDAC1, plasmid constructs were generated, encoding each domain, to perform a co-IP assay (Figure 3G,H). The results revealed that the CC and SPRY domain of TRIM65 were indispensable for the interaction with VDAC1 (Figure 3G,H). Conversely, our findings indicated that the amino acids sequence 1–131 of VDAC1, rather than the amino acids sequence 132–283, was sufficient for binding to TRIM65 (Figure 3I,J).

2.4 TRIM65 Mediates Ubiquitination of VDAC1

Given that TRIM65 has been demonstrated to possess E3 ubiquitin ligase activity, the aim of this study was to investigate the impact of TRIM65 on VDAC1 ubiquitination. As shown in Figure 4A,B, TRIM65 markedly augmented the ubiquitination of VDAC1, whereas the enzymatically inactive mutant of TRIM65 did not result in an increase in the ubiquitination modification of VDAC1. Additionally, the overexpression of TRIM65 was observed to markedly elevate the ubiquitination level of VDAC1 in HK-2 cells (Figure 4C). It has been demonstrated that E3 ubiquitin ligases can mediate diverse types of ubiquitin chain modification to the substrate, thereby exerting distinct functions [18]. To determine the type of ubiquitin chains added by TRIM65 to VDAC1, a series of ubiquitin mutants were utilized in ubiquitination assays. The results demonstrated that both K48- and K63-linked ubiquitin modifications of VDAC1 appeared in the presence of TRIM65, while other types of ubiquitin modifications were not detected (Figure 4D). As illustrated in Figure 4E, the overexpression of TRIM65 resulted in a notable elevation in the K48- and K63-linked ubiquitination of VDAC1. Moreover, ubiquitination experiments were conducted using ubiquitin molecules that had been mutated from lysine to arginine at positions 48 and 63 (K48R and K63R), thereby confirming that these two lysine residues are pivotal sites on VDAC1 polyubiquitination modified ubiquitin (Figure 4F). Therefore, TRIM65-conjugated ubiquitin chains on VDAC1 were predominantly K48 and K63 ubiquitin linkages. This was evidenced by the efficient conjugation of ubiquitin containing K48 and K63 as the sole lysine residue, whereas ubiquitin with K48R or K63R mutation was not.

To identify the precise site of ubiquitination by TRIM65, we generated VDAC1 mutants bearing single lysine-to-arginine substitutions at each potential ubiquitylation site. As illustrated in Figure 4G, the extent of ubiquitin conjugation to the VDAC1 mutant with lysine 161 to arginine (K161R) and lysine 200 to arginine (K200R) was substantially diminished in comparison to that of either the WT VDAC1 or any mutant bearing single lysine to arginine substitutions on each of the other ubiquitylation sites. These findings suggest that lysine 161 and 200 are the primary sites for TRIM65-mediated VDAC1 ubiquitination.

2.5 TRIM65 Maintains VDAC1 Protein Stability

Given that protein ubiquitination affects a multitude of cellular processes by regulating protein degradation and protein-protein interactions [19], we proceeded to determine the effect of TRIM65 on VDAC1 stability. The results showed that TRIM65 increased VDAC1 expression concentration dependently (Figure 5A). However, the overexpression of TRIM65 did not result in an increase in protein level following the mutation of the site where TRIM65 ubiquitinates VDAC1 (Figure S5). In HK-2 cells, the knockdown of TRIM65 was observed to significantly reduce VDAC1 protein levels and vice versa (Figure 5B). Polyubiquitin linkages via lysine 48 (K48) or 63 (K63) can result in the differential targeting of proteins for 26S proteasomal degradation or endosome trafficking to the lysosome [20]. To investigate how TRIM65 regulates VDAC1 protein levels in renal cells, it was necessary to determine which of the major VDAC1 degradation pathways was affected by TRIM65. Inhibitors for autophagosome (3-methyladenine [3-MA]), lysosome (chloroquine [CQ]), and proteasome (MG132) were administered to control or TRIM65 knockdown HK-2 cell; both CQ and 3-MA abolished the negative regulation of sh-TRIM65 on VDAC1 levels (Figure 5C), indicating that TRIM65 blocks the degradation of VDAC1 through the autophagy pathway. The impact of TRIM65 on the stability of VDAC1 was subsequently verified. A cycloheximide chase assay demonstrated that the overexpression of TRIM65 and 3-MA markedly impeded the degradation of VDAC1 protein in HK-2 cells (p < 0.001, Figure 5D). IHC and western blot analysis of renal tissues from I/R mice revealed that VDAC1 expression was increased in AKI, whereas it was lower in TRIM65 knockout mice in both sham-operated and I/R groups (Figure 5E–G). The correlation analysis revealed that the protein level of VDAC1 in renal tissue exhibited a significant positive correlation with the expression of TRIM65 in the I/R-induced AKI mouse model (p < 0.05, r = 0.8907, Figure 5H). Further validation in the rhabdomyolysis-induced AKI model demonstrated that the expression of VDAC1 was consistent with that of TRIM65 (Figure 5I). Additionally, correlation analysis revealed a significant positive correlation between the expression of TRIM65 and VDAC1 (p < 0.05, r = 0.8525, Figure 5J). In conclusion, TRIM65 plays a role in stabilizing VDAC1 and regulating its expression in AKI.

2.6 TRIM65 Aggravated Mitochondrial Dysfunction

Given that TRIM65 may interact with the mitochondrial outer membrane protein VDAC1 to affect I/R-induced AKI, we postulated that TRIM65 may exert a certain impact on mitochondrial function. Transmission electron microscopy (TEM) was employed to ascertain whether TRIM65 affects the ultrastructure of mitochondria in AKI. As illustrated in Figure 6A, the mitochondrial cristae are well developed in normal renal cells, with the cristae connected to the inner membrane. Furthermore, TRIM65 knockdown does not appear to affect mitochondrial structure under normal conditions. However, this organization was lost during the progression of AKI, and the mitochondrial cristae were fewer in number and unorganized in AKI renal tissue cells. Additionally, there was a notable presence of swollen vacuoles and enlarged individual organelles, which did not appear to be connected in a mitochondrial network (Figure 6A). TRIM65 deletion ameliorated the disruption and damage to mitochondrial structure in AKI renal cells (Figure S6A). These findings indicate that AKI results in a reduction in the structural integrity of mitochondria, and that TRIM65 knockdown protects against AKI by preventing the destruction of mitochondrial ultrastructure. It is well established that the mitochondrial transmembrane potential (ΔΨm) serves as a crucial indicator of mitochondrial function [21]. A decline in this parameter is frequently considered an indicator of mitochondrial dysfunction [21]. JC-1 is a ratiometric probe that monitors ΔΨm by determining the relative dual emission of the mitochondrial J aggregates (red fluorescent) versus JC-1 monomers (green fluorescent). In normal cells, JC-1 is predominantly located in the mitochondria, resulting in the accumulation of red J aggregates with a slight green fluorescent monomeric form remaining in the cytoplasm (Figure 6B). In contrast, following hypoxia-reoxygenation treatment, the green fluorescence of J monomers in the cytoplasm of HK-2 cells gradually increased, while the red fluorescence of J aggregation proteins (normal mitochondria) was reduced, indicating a collapse of ΔΨm (Figure 6B). This indicated that the mitochondria had been severely damaged. However, overexpression of TRIM65 significantly inhibited J aggregates (Figure S6B), suggesting that TRIM65 may act as a negative regulator of mitochondrial function in AKI.

2.7 Protective Effect of TRIM65 Deficiency in AKI Reversed by VDAC1 Overexpression

To ascertain the role of TRIM65 in AKI by targeting VDAC1, a serotype 9 adeno-associated virus vector packaged in a serotype 9 capsid (AAV2/9) was employed. This vector has been demonstrated to exhibit excellent renal tissue specificity and efficacy in transducing the carried gene [22]. In brief, AAV2/9 constructs carrying either the VDAC1 gene or EGFP were injected into renal tissues via intrapelvic injection in WT and Trim65 knockout mice. Twenty-one days later, I/R-induced AKI was performed, and high green fluorescence in renal tissues of all groups was observed by fluorescence microscopy. This suggests that AAV2/9 successfully mediated either VDAC1 or EGFP into the kidneys (Figure 7A). The degree of AKI in WT and Trim65^−/− mice induced by I/R was not altered by the empty AAV vector. In contrast, TRIM65 knockdown showed a significant protective effect (Figure 7A,B). However, renal overexpression of VDAC1 almost completely abolished the protective effect of TRIM65 deletion and also evidently exacerbated AKI in WT mice (Figure 7A,B). It is noteworthy that VDAC1 did not affect the expression of TRIM65 in renal tissues of WT mice (Figure 7A,C). Consequently, serum levels of BUN, sCr, and Kim-1 were markedly diminished in AAV-control-treated Trim65^−/− mice in comparison to AAV-control-treated WT mice. However, AAV-VDAC1 treatment yielded no statistical difference between the two groups in WT and Trim65^−/− AKI mice (Figure 7D). Western blot analysis confirmed that the injection of AAV-VDAC1 resulted in the overexpression of VDAC1 in the kidney tissues of both WT mice and Trim65 knockout mice (Figure 7E,F). The VDAC1 significantly increased the protein levels of KIM-1 and NGAL in the kidney tissues of Trim65^−/− AKI mice (Figure 7A,C,E,F). Indeed, the overexpression of VDAC1 was observed to markedly elevate the protein levels of KIM-1 and NGAL in the kidney tissues of WT AKI mice (Figure 7A,C,E,F). Collectively, the overexpression of VDAC1 in kidney tissues effectively counteracted the protective effect of TRIM65 knockdown on AKI, thereby suggesting that TRIM65 may potentially target VDAC1 to regulate AKI.

5.1 Animals

Female C57BL/6 WT mice (20–25 g) were obtained from Hunan Shrek Jingda Experimental Animal Co., Ltd. The Trim65^−/− mice were generated in the C57BL/6 background by backcrossing the C57BL/6 WT mice with Trim65^+/− mice (Cyagen Biosciences, Inc.) as previously described [14, 15]. The genotype was identified by polymerase chain reaction (PCR). The mice were housed at a temperature of 22 ± 2°C, with a 12-h light/dark cycle and 50% humidity, and were provided with food and water ad libitum. All animal protocols employed in this study were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Nanchang University (approval number: CDYFY-IACUC-202301QR010). Eight-week-old female mice were randomly allocated to each group of six, with 12 mice in each group, for the survival experiment. The number of animals used and the associated suffering were kept to a minimum throughout the study.

5.2 Rhabdomyolysis-Induced AKI Model

Following an overnight fast, all animals were anaesthetized using isoflurane (2%) and oxygen flow (2 L/min) to induce and maintain anesthesia. Once the mice had reached the appropriate depth of anesthesia, they were weighed and then intramuscularly injected with 50% glycerol (8 mL/kg, with a solvent of normal saline) into the hind limbs. The mice in the control group were weighed and injected with an equal volume of normal saline into the muscles. Observation of the hind limbs of mice that had just been injected with 50% glycerol revealed that they were in a straight and dragging state. Euthanasia was performed 24 h later, after which eyeball blood samples and kidney tissue samples were collected.

5.3 Ischemia/Reperfusion Induced-AKI Model

The mice were fasted overnight prior to anesthesia with 2% pentobarbital sodium (70 µg/kg). The mice were positioned in a supine position on the preheating pad. A 1- to 2-cm incision was made at the abdominal midline, following the requisite sterilization and skin preparation. Subsequently, the intestine was extracted with the aid of two cotton swabs, thereby preventing damage to the surrounding blood vessels and nerves. Subsequently, the renal pedicle was exposed. The kidney artery and vein were then separated and clamped with precision using a non-traumatic microsurgical vascular clip. The color of the kidney altered from bright red to dark red, indicating successful ischemia. Following a period of 40 min, the clip was released with great care to prevent the rupture and subsequent bleeding of the renal pedicle. The color of the kidney changed from dark red to bright red, indicating successful reperfusion. The sham group returned the kidney and intestine to the abdominal cavity after surgical exploration and kidney exposure without clipping the renal pedicle. The incision was then closed with 4/0 silk thread layer by layer. After 24 h, the mice were anesthetized and euthanized with isoflurane (2%) for the collection of eyeball blood samples and kidney tissue samples.

5.4 H&E and IHC

The slides derived mouse kidney tissues were immersed in xylene for 20 min. Subsequently, the slices were incubated in ethanol with decreasing concentration gradients for a period of 5 min, followed by a further 5 min in double-distilled water and PBS. The H&E solution was added in a stepwise manner. The slices were dehydrated in ethanol with increasing concentration gradients and placed in xylene. After sealing with neutral gum, renal tubular damage images were acquired with a pathological section scanner (Sunny Optical Technology 193 (Group) Company Limited, China) in a blinded manner by a semiquantitative method as previously described [41]. The images were scored in accordance with the following criteria: A score of 0 was indicative of normal tissue, while a score of 1 indicated minor damage, defined as less than 5% of the tissue affected, A score of 2 was represented as mild damage, occurring in tissue with 5%–25% damage, while a score of 3 signified moderate damage, affecting tissue with 25%–75% damage. Finally, a score of 4 was assigned to cases of severe damage, occurring in tissue with more than 75% damage. The IHC procedure was conducted by the previously described methodology [14]. The primary antibodies employed in this study included anti-TRIM65 (Atlas Antibodies, Sweden; HPA021578, 1:400), anti-VDAC1 (Proteintech, China; 66345-1-Ig, 1:50), anti-KIM-1 (R&D Systems, USA; AF1817, 1:400), and anti-NGAL (R&D Systems, USA; AF1857, 1:400). The antibodies were incubated at 4°C overnight. On the following day, the slides were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies at room temperature. Following a washing step, the positive stain was detected with 3,3’-diaminobenzidine (DAB). The final evaluation was performed under a pathological section scanner and analyzed using the ImageScope system with a minimum of six sections.

5.5 Transmission Electron Microscope (TEM)

The extraction of kidney tissue from the mice was completed within 3 min of the mice being euthanized. The mouse renal cortex sample was isolated and fixed in 2.5% glutaraldehyde (Servicebio, China). The cortex tissues were embedded in resin. Ultrathin sections of 80 nm were obtained by an ultramicrotome (Leica, Germany) and examined with a TEM (HITACHI, Japan).

5.6 Cell Culture and Hypoxia/Reoxygenation (H/R) Procedures

The HK-2 cells were procured from Wuhan Procell Life Science & Technology Co., Ltd. The human embryonic kidney (HEK293T) cells were obtained from the American Type Culture Collection (ATCC). All cell lines were authenticated via short tandem repeat (STR) analysis and tested for mycoplasma contamination. The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Servicebio, China) supplemented with 10% fetal bovine serum (ExCell Bio, China) and 1% penicillin/streptomycin (NCM Biotech, China) in an incubator (Thermo Scientific, USA) set to 5% CO₂ at 37°C. The HK-2 cells were exposed to a well-controlled three-gas cell culture incubator (BioSpherix, USA) under nutrient-deprived conditions (1% O₂, 94% N₂, and 5% CO₂) for 12 h, then transferred into complete medium for 4 h of reoxygenation at 37°C and 5% CO₂. The control group was maintained in an adequate nutrition and normoxic environment throughout the experiment.

5.7 JC-1 Dye

HK-2 cells were plated in a six-well plate and transfected with Flag-TRIM65 or an empty vector using Hieff Trans Liposomal Transfection Reagent (YEASEN, Shanghai, China) for 48 h. Subsequently, the cells were subjected to a H/R procedure. The cells were then rinsed with PBS and incubated with the JC-1 fluorescent probe (Beyotime, China) at 37°C for 20 min, followed by washing twice with pre-cooled JC-1 staining buffer. The inverted fluorescence microscope was employed to detect the fluorescence. When the mitochondrial membrane potential was normal, JC-1 aggregated in the mitochondrial matrix to form polymers, which exhibited red fluorescence. Conversely, when the mitochondrial membrane potential was compromised, JC-1 was unable to aggregate in the mitochondrial matrix. Consequently, JC-1 existed as monomers, which emitted green fluorescence. The ratio of red and green fluorescence intensity was found to be negatively correlated with early cell apoptosis.

5.8 Statistical Analysis

All statistical analyses were conducted using GraphPad Prism version 8.0. The data were expressed as mean ± SEM. For data that met the normality test, a Student's t-test was employed for the comparison of two groups. Comparisons between multiple groups were made using one-way ANOVA followed by Tukey's post hoc test. Comparisons between different time points were made using repeated-measures ANOVA. Nonparametric rank-sum tests were used for data that did not meet the normality test. A p-value of less than 0.05 was considered statistically significant.