ACOT12, a novel factor in the pathogenesis of kidney fibrosis, modulates ACBD5

Introduction

Chronic kidney disease (CKD) is a progressive disorder characterized by changes in kidney structure, such as cysts, tumors and atrophy and/or loss of kidney function, over a period of months to years. Kidney dysfunction can change the output or quality of urine and is most often recognized by increased serum levels of creatinine, cystatin C or blood urea nitrogen (BUN). Patients with CKD have an increased risk of developing other complications, such as cardiovascular diseases, hypertension and bone disorders. Although the impact of CKD on morbidity and mortality is increasing worldwide¹, there is no effective treatment for CKD other than dialysis or kidney transplantation.

Kidney fibrosis is the final common pathway in CKD and is characterized by glomerulosclerosis, tubulointerstitial fibrosis, inflammatory infiltration and kidney parenchyma loss. Excessive accumulation of the extracellular matrix, which is primarily composed of collagen, results in loss of kidney function^2,3, and impeding or reversing this process is one of the current therapeutic approaches for controlling CKD⁴. Recently, peroxisome proliferator-activated receptors (PPARs) have been implicated in the regulation of kidney inflammation and fibrosis and are considered potential therapeutic targets⁵. PPARα regulates age-associated kidney fibrosis⁶. PPARα deficiency leads to the accumulation of lipids in kidney tubules and exposure to PPARα agonists reduces the production of transforming growth factor-beta (TGFβ), IL-1β, IL-6 and TNF, leading to a reduction in tubulointerstitial fibrosis and inflammation^7,8. Moreover, PPARγ, which is known to be expressed in glomerular mesangial cells, podocytes and proximal epithelial cells, has a protective role against acute kidney injury by inhibiting inflammation, and pioglitazone, a PPARγ agonist, prevents mesangial expansion, glomerulosclerosis, tubulointerstitial inflammation and fibrosis, as well as tubular dilation and atrophy during the pathogenesis of diabetic nephropathy^9,10.

Increasing evidence supports the hypothesis that lipid abnormalities contribute to the progression of kidney fibrosis¹¹. Fatty acid oxidation (FAO) is closely associated with the development and progression of fibrosis because the kidneys mostly rely on fatty acid (FA) β-oxidation as their energy source. Thus, FAO dysfunction can exacerbate kidney fibrosis. Carnitine palmitoyltransferase 1A (CPT1A), a rate-limiting enzyme in the FAO pathway, is expressed in kidney tubules and is significantly reduced in kidney fibrosis¹². FA transporters in the kidney, such as clusters of differentiation-36 (CD36), FA-binding proteins and FA-transport proteins, also play important roles in kidney disease. CD36, a FA translocase, is involved in Wnt/β-catenin activation and contributes to lipid accumulation and kidney fibrosis^13,14. CD36 is associated with the collagen I-discoidin domain receptor 1 pathway and causes kidney failure in a collagen 4-mutated lipotoxic injury model¹⁵. Inhibition of FABP4 attenuates endoplasmic reticulum stress, mitochondrial dysfunction and inflammation in kidney fibrosis^16,17. FATP2, encoded by Slc27a2, promotes kidney fibrosis by inhibiting FAO¹⁸. These data suggest that lipid metabolism accompanying CKD in kidney fibrosis could be a potential therapeutic target. Given the critical role of lipid metabolism in kidney fibrosis, we focused on acyl-CoA thioesterase 12 (ACOT12) in this study. ACOT12 is an enzyme that hydrolyzes acyl-CoA to free FAs and CoA, which is crucial in regulating the intracellular levels of acyl-CoA and free FAs. It is predominantly expressed in the kidney tubules, suggesting its potential involvement in renal lipid metabolism. However, the role of ACOT12 in kidney fibrosis has not been previously explored. We hypothesized that ACOT12 might be a key regulator linking lipid metabolism to kidney fibrosis. Our results demonstrate that Acot12 deficiency induces lipid accumulation and exacerbates kidney fibrosis independently of PPARα, and that restoring Acot12 effectively ameliorates kidney fibrosis.

Materials and methods

Ethical approval

All animal studies were approved by the Institutional Animal Care and Use Committee of Wonkwang University and followed institutional guidelines (WKU20-60, WKU20-114, WKU21-36, WKU21-89 and WKU22-40). All the mice were housed at 23 ± 1 °C for 12 h. The light/dark cycles and relative humidity were 50 ± 5%, with food and water available ad libitum. Human kidney tissues were obtained from 12 patients with chronic obstructive nephropathy and 17 patients with clear cell renal cell carcinoma. The study protocol was approved by the clinical research ethics committee of Keimyung University Dongsan Hospital (institutional review board no. 2022-08-022), and all experiments were performed following approved guidelines.

Animal experiments

Global Acot12^−/− mice were generated via the germline transmission of an RGEN-induced mutant allele¹⁹. Acot12^−/−PPARα^−/− double knockout (KO) mice were generated by crossbreeding. Male 6–7-week-old C57BL/6N mice underwent unilateral ureteral obstruction (UUO) surgery. UUO surgery was performed under anesthesia, as previously described²⁰, and the kidneys were collected on the seventh day. Fenofibrate (125 mg/kg) was orally administered once daily from three days before UUO surgery to the day before killing. For the adenine diet model, the mice were fed a 0.25% adenine diet for 2 weeks.

Lentiviral packaging

Lenti-X 293T cells (Clontech) were transfected with pLenti-GIII-CMV-Acot12 or pLenti-GIII-CMV-PPARαpLenti-GIII-CMV-PPARα via a third-generation packaging system mix (Applied Biosystems) according to the manufacturer’s protocol. The lentiviral particles were concentrated via a Lenti-X concentrator (Takara) and stored at −80 °C. The lentiviruses were injected into four sites in the kidney parenchyma after UUO surgery.

Synthesis of CHI–DHCA

Catechol-modified chitosan (CHI–DHCA) was synthesized via standard carbodiimide chemistry. Briefly, chitosan (0.5 g) (molecular weight of 100 kDa, 70% deacetylated, Heppe Medical Chitosan, GmbH) was dissolved in distilled water (44.5 ml) containing 5 ml of 1 N HCl. After the complete dissolution of the chitosan, the pH was adjusted to 5 with 1 N NaOH. Then, 3,4-dihydroxyhydrocinnamic acid (DHCA, 0.59 g) (Sigma‒Aldrich) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (0.62 g) (TCI-SU) in ethanol (12.5 ml) were added slowly to the chitosan solution and allowed to react for 12 h. The pH was adjusted to 5.0 using 1 N HCl to prevent the undesirable oxidation of the catechol groups during the reactions. The product was purified via a dialysis membrane (molecular weight cutoff of 12–14 kDa; SpectraPor) against a pH 2 NaCl (10 mM) solution for 2 days, distilled water for 4 h and subsequently lyophilized. The degree of catechol substitution was confirmed via ¹H nuclear magnetic resonance (Bruker Avance, 500 MHz) and ultraviolet‒visible (UV–Vis) spectroscopy (UV-1900i, Shimadzu). The degree of DHCA substitution of CHI–DHCA was calculated from the absorbance at 280 nm with standard curves of DHCA monomer concentrations.

Generation of Acot12-conjugated CHI–DHCA patches

Chitosan-catechol (40 mg/ml) was mixed with 50 μg of pcDNA-RFP-Acot12, lyophilized for 1 day and used for the delivery of Acot12 into the kidney.

PTEC culture

Kidney cortical fragments isolated from 3-week-old male C57BL/6 mice were digested with collagenase type I (Thermo Fisher Scientific) and filtered through a 100 μm strainer¹². After red blood cells were lysed with RBC lysis buffer (Roche), the primary tubule epithelial cells (PTECs) were centrifuged, washed twice with RPMI 1640 and grown in RPMI 1640 (Gibco) supplemented with 20 ng/ml recombinant human epidermal growth factor (Sigma) and 10% fetal bovine serum.

GFP-LC3/RFP-SKL transfection

PTECs were transfected with plasmids encoding GFP-LC3 and RFP-SKL via Lipofectamine 3000 (Invitrogen) in the presence of 5 ng/ml TGFβ and 10 nM bafilomycin A for 24 h, and fluorescence images were acquired via EVOS FL Auto software v.1.7.

Histology and immunohistochemistry

The kidneys were fixed in 4% paraformaldehyde (pH 7.4), dehydrated and embedded in paraffin. After deparaffinization and rehydration, sectioned kidneys were stained with Masson’s trichrome. For immunohistochemistry, endogenous peroxidase was blocked with 3% hydrogen peroxide, and antigen retrieval was performed with 0.01 M sodium citrate buffer (pH 6.0). Tris–EDTA buffer (pH 9.0) was used for ACOT12 staining. After blocking with normal horse serum, the sections were incubated overnight with the following primary antibodies: ACOT12 (1:1,000; AbFrontier), PLIN2 (1:1,000; Abcam), Collagen I (1:400; Southern Biotech), αSMA (1:1,000; Abcam), PPARα (1:100; Santa Cruz) and ACBD5 (1:100; Novus). After incubation with peroxidase-conjugated secondary antibodies for 1 h, the antigen was detected via ImmPACT DAB Substrate (Vector Laboratories). Images were acquired via EVOS FL Auto and quantified via Image-Pro Plus 4.5 (Cybernetics) or ImageJ software (ImageJ, National Institutes of Health; https://imagej.nih.gov/ij/).

mRNA sequencing analysis

For mRNA sequencing analysis, total RNA was isolated from the contralateral or UUO kidneys of Acot12^+/+ or Acot12^−/− mice using RNAiso Plus (Takara). The RNA quality was verified via a Bioanalyzer 2100 system (Agilent), a cDNA library was constructed using the NEBNext Ultra II Directional RNA-Seq kit (NEB) and the library was sequenced via a NovaSeq 6000 instrument (Illumina). Trimming, mapping, normalization and initial analyses were performed by E-Biogen, Inc. Further analysis was performed via ExDEGA (E-Biogen) and the Database for Annotation, Visualization and Integrated Discovery (DAVID)²¹. Networks were analyzed via the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING; https://string-db.org).

qRT–PCR

Total RNA was isolated using TRIzol reagent (Invitrogen) or RNAiso Plus (Takara), according to the manufacturer’s instructions. cDNA was synthesized via a high-capacity cDNA reverse transcription kit (Applied Biosystems) or 5× RT Master Mix (Takara). Quantitative PCR with reverse transcription (qRT–PCR) was performed using SYBR Green PCR Master Mix (Applied Biosystems) or AMPIGENE qPCR Green Mix Hi-ROX (Enzo) on an Applied Biosystems 7300 or ABI StepOnePlus instrument (Applied Biosystems). The relative expression level of each gene was normalized to the 18S rRNA expression level. The sequences of primers used in this study are listed in Supplementary Table 1.

Western blot analysis

Kidney proteins were extracted via RIPA buffer (Cell Signaling Technology) with 2 mM phenylmethylsulfonyl fluoride, separated via sodium dodecyl sulfate‒polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride or nitrocellulose membranes (GE Healthcare BioSciences). The membranes were blocked with 5% skimmed milk in TBS–Tween 20 and incubated with the following primary antibodies: ACOT12 (1:2,000; AbFrontier), PPARα (1ː1,000; Santa Cruz) and β-actin (1:3,000; Sigma). The blots were developed with horseradish peroxidase-conjugated secondary antibodies and detected via an enhanced chemiluminescence method. The bands were quantified via ImageJ software and normalized to β-actin.

Plasma analysis via enzyme-linked immunosorbent assay

Blood samples were collected via heparinized syringes, and the plasma was collected via centrifugation at 3,000 rpm for 20 min at 4 °C. BUN (Arbor Assays), plasma cystatin C (R&D Systems) and plasma neutrophil gelatinase-associated lipocalin (NGAL; Immunology Consultant Laboratory) were measured according to the manufacturer’s instructions.

Statistical analyses

All the results are expressed as the means ± s.e.m. of at least three independent experiments. Representative experiments and images are shown unless otherwise stated. Two-tailed Student’s t-tests were used to compare two groups. One-way or two-way analysis of variance was used to assess the differences between multiple groups, followed by Fisher’s least significant difference test. Statistical tests with P < 0.05 were considered statistically significant.

Results

ACOT12 is a key factor linking lipid metabolism to kidney fibrosis

UUO, which mimics chronic obstructive nephropathy in humans, results in kidney fibrosis as well as lipid and extracellular matrix accumulation²² (Supplementary Fig. 1a). Perilipin 2 (PLIN2), which coats intracellular lipid droplets, was observed as a marker of lipid metabolism activation²³. In addition, treatment with losartan, an angiotensin II type 1a receptor antagonist, attenuated fibrosis²⁴ and decreased lipid deposition in UUO-induced mice (Supplementary Fig. 1a). These results suggest that dysregulated lipid metabolism is closely associated with the development and progression of kidney fibrosis. Next, genome-wide transcript levels in the UUO model were analyzed to identify the key player responsible for lipid accumulation and fibrosis (Supplementary Fig. 1b). Gene Ontology analysis revealed that three metabolic processes, the FA metabolic process (n = 26), lipid metabolic process (n = 44) and acyl-CoA metabolic process (n = 12), were significantly altered in UUO kidneys compared with control kidneys (Supplementary Fig. 1c). Moreover, acyl-CoA synthetase medium-chain family members (Acsm) 1/2/3/5, Acot 11/12, enoyl-CoA hydratase and 3-hydroxyacyl CoA dehydrogenase (Ehhadh) were common genes in all three groups. Acyl-CoA synthetase medium-chain (ACSM)s generate acyl-CoA from free FAs and CoA, whereas ACOTs hydrolyze acyl-CoA to FAs and CoA. EHHADH is a peroxisomal β-oxidation enzyme. The protein–protein interaction network constructed via the STRING database suggested that ACOT12 is a key player in lipid accumulation and fibrosis (Supplementary Fig. 1d). The transcriptional and translational levels of ACOT12 were significantly lower in UUO kidneys than in control kidneys (Supplementary Fig. 1e). Furthermore, losartan treatment was associated with an increase in ACOT12 expression, which correlated with reduced kidney injury (Supplementary Fig. 1f). This observation is consistent with various models, including diabetes (GSE246576), an adenine-rich diet (GSE150641) and folic acid exposure (GSE222570), all of which are associated with decreased levels of ACOT12 expression (Fig. 1a). The adenine-rich diet models revealed common lipid-related pathways that matched those found in our UUO mRNA sequencing data, highlighting the consistent enrichment of signaling pathways across these various conditions (Fig. 1b). Diminished expression of ACOT12 in kidneys subjected to an adenine-rich diet was further confirmed through immunohistochemistry (Fig. 1c). In human kidney biopsies, we noted a significant reduction in ACOT12 compared with that in normal kidneys (Fig. 1d). In silico analysis using Gene Expression Omnibus (GEO) (GSE7869) confirmed that ACOT12 expression was lower in patients with autosomal dominant polycystic kidney disease (ADPKD) than in normal patients (Fig. 1e). Analysis of the tubulointerstitial transcriptome from European Renal cDNA Bank (ERCB) patients with CKD identified in the NephroSeq v5 online database revealed a significant positive correlation between ACOT12 and the estimated glomerular filtration rate (eGFR). Consistent with these findings, we also observed a significant decrease in ACOT12 in UUO kidneys compared with that in contralateral kidneys (Fig. 1f). A negative correlation between PLIN2 expression/fibrosis and ACOT12 expression was observed not only via immunohistochemistry images, but also via Pearson correlation analysis of those images (Fig. 1g).

**Fig. 1: The expression level of Acot12 in a CKD model.**

Acot12
^−/− mice develop age-related renal fibrosis

Here, we observed an age-dependent loss of ACOT12 expression in the renal tubules (Fig. 2a). In the kidneys of 20-month-old Acot12^+/+ mice, the expression level of ACOT12 dramatically decreased, with increased kidney dysfunction and injury observed via plasma cystatin C and NGAL analysis (Fig. 2b). Acot12 KO (Acot12^−/−) mice previously generated via RGEN-induced mutation of the Acot12 gene (Supplementary Fig. 2) were used to understand the role of ACOT12 in kidney disease pathogenesis. In 20-month-old Acot12^−/− mice, the average kidney size was markedly larger than that in Acot12^+/+ mice (Fig. 2c and Supplementary Fig. 3a). Significant increases in three types of kidney dysfunction and injury markers in the plasma, namely, BUN, cystatin C and NGAL, were also observed (Fig. 2d). Masson’s trichrome staining revealed a significant increase in the number of polycysts in the kidneys of 20-month-old Acot12^−/− mice compared with Acot12^+/+ mice (Fig. 2e). Age-dependent increases in kidney fibrosis (Fig. 2e and Supplementary Fig. 3b) were observed, with the most significant increase observed in 20-month-old Acot12^−/− kidneys. Significant lipid accumulation was observed in the kidneys of 6- and 12-month-old Acot12^−/− mice compared with those of Acot12^+/+ mice (Fig. 2f). This early lipid deposition suggests that dysregulated lipid metabolism due to ACOT12 deficiency may contribute to the initiation and progression of kidney fibrosis, highlighting the role of lipid accumulation as an early pathogenic event in kidney disease.

**Fig. 2: Association of ACOT12 deficiency with aging-dependent loss of kidney function.**

ACOT12 restoration attenuates lipid accumulation and kidney fibrosis

To fully understand the role of ACOT12 in the pathogenesis of kidney disease, Acot12^−/− mice subjected to UUO surgery were used. The kidney weight to body weight ratio (KW/BW) was significantly lower in Acot12^−/− kidneys than in Acot12^+/+ kidneys (Fig. 3a). Kidney fibrotic lesions, as assessed by Masson’s trichrome and type I collagen staining, were significantly greater in Acot12^−/− kidneys than in Acot12^+/+ kidneys (Fig. 3b). The expression levels of genes associated with inflammation, such as Mcp1, or fibrosis, such as Tgfβ, Col1a1 and αsma, were increased in Acot12^−/− kidneys, with the most significant increase observed in UUO-induced Acot12^−/− kidneys (Fig. 3c). The expression levels of genes involved in lipid synthesis, such as Acly, Acaca, Fasn and Scd1, were also increased in Acot12^−/− kidneys and UUO-induced Acot12^−/− kidneys (Fig. 3d). However, PLIN2 staining revealed no difference in lipid accumulation between Acot12^−/− and Acot12^+/+ kidneys.

**Fig. 3: ACOT12 deficiency exacerbates lipid accumulation and kidney fibrosis.**

Restoration of ACOT12 expression using a lentivirus (LV-Acot12) via parenchymal injection decreased kidney fibrosis and lipid accumulation (Fig. 4a and Supplementary Fig. 4a) as well as the expression levels of fibrosis-related genes by day 7 post-UUO (Fig. 4b). Additionally, on day 3 post-UUO, kidney fibrosis was reduced and ACOT12 was restored (Supplementary Fig. 4b,c). We also delivered Acot12 via a chitosan patch. To prepare the DNA-loaded adhesive patches, CHI–DHCA was first synthesized via standard carbodiimide chemistry. As shown in Fig. 4c, the primary amines of chitosan could react with the carboxylic acid groups of DHCA, resulting in the synthesis of CHI–DHCA. The DHCA conjugation of CHI–DHCA was confirmed by UV‒Vis spectroscopy. As previously reported, the catechol groups of DHCA exhibit an absorption maximum at a wavelength of 280 nm (ref. ²⁵). The typical absorption of catechol groups at 280 nm was detected in both DHCA and CHI–DHCA (Fig. 4d). The degree of catechol substitution of CHI–DHCA was 8.7%. Next, the CHI–DHCA patches were prepared via a simple freeze-drying method. CHI–DHCA was subsequently dissolved in ddH₂O and freeze dried. Uniform sponge-like structures were found in the CHI–DHCA patches (Fig. 4e). The application of the chitosan patch with the Acot12 expression vector (patch-ACOT12) to the kidney significantly reduced kidney fibrosis (Fig. 4f and Supplementary Fig. 4d,e). Furthermore, the levels of inflammatory and fibrotic genes were also simultaneously lower in UUO kidneys attached with patch-ACOT12 than in control kidneys (patch-control) (Fig. 4g).

**Fig. 4: Restoration of ACOT12 ameliorates lipid accumulation and kidney fibrosis in *Acot12*^−/− UUO kidneys.**

ACOT12 deficiency exacerbates kidney fibrosis independently of PPARα

Previous research has demonstrated that PPAR plays a key role in protecting the kidney from CKD as well as acute injury⁵. As shown in Fig. 5a, the PPAR pathway emerged as a critical biological pathway that was significantly altered in the pathogenesis of kidney fibrosis. Notably, ACOT12 expression seems to be regulated by PPARα due to the presence of a predicted PPAR-binding motif. To further test whether the activation of PPARα could overcome kidney fibrosis induced by ACOT12 deficiency, fenofibrate, a known PPARα agonist, was used. In UUO-induced Acot12^+/+ mice, fenofibrate treatment significantly reduced lipid accumulation and kidney fibrosis and increased ACOT12 expression (Fig. 5b). However, fenofibrate treatment did not restore kidney fibrosis in UUO-induced Acot12^−/− mice, although fenofibrate activated PPARα (Fig. 5c and Supplementary Fig. 5a). Consistent with this finding, a significant decrease in plasma NGAL levels was observed in the fenofibrate-treated Acot12^+/+ mice, whereas plasma NGAL levels were not altered in the fenofibrate-treated Acot12^−/− mice (Fig. 5d). These data suggest that kidney fibrosis induced by ACOT12 deficiency is caused by a signaling pathway independent of PPARα signaling. To verify this, we used Pparα^−/− mice to generate Acot12 and Pparα double-KO (Acot12^−/− Pparα^−/−) mice (Supplementary Fig. 2). Lentiviral introduction of ACOT12 into UUO-induced Pparα^−/− mice via parenchymal injection significantly reduced lipid accumulation and kidney fibrosis (Fig. 6a), as did the expression levels of fibrosis genes (Fig. 6b). Moreover, the restoration of ACOT12 in Acot12^−/− Pparα^−/− mice significantly reduced lipid accumulation and kidney fibrosis (Fig. 6c). Surprisingly, however, restoring PPARα in Acot12^−/− Pparα^−/− mice exacerbated lipid accumulation and kidney fibrosis.

**Fig. 5: ACOT12 deficiency may stimulate kidney fibrosis independently of PPARα.**

**Fig. 6: Restoration of Acot12 into *Acot12*^−/−*Pparα*^−/− kidneys reduces kidney fibrosis but restoration of PPARα does not.**

ACOT12 deficiency inhibits pexophagy via the downregulation of ACBD5

To investigate the underlying mechanism involved in kidney fibrosis induced by ACOT12 deficiency, we extracted genes differentially expressed in Acot12^−/− UUO kidneys or commonly overlapping genes and analyzed them via the DAVID bioinformatics resource²⁶ (Fig. 7a). The lysosome, peroxisome, phagosome, autophagy, cellular senescence, mTOR and p53 signaling pathways were enriched in genes differentially expressed in Acot12^−/− UUO kidneys. Among these, significant attention was given to peroxisomal biogenesis factor (PEX) genes, such as Pex1 and Pex2, which were found to be downregulated, indicating disrupted peroxisome biosynthesis in the absence of ACOT12 (Fig. 7b). In addition to a reduction in the expression of peroxisomal biosynthesis genes (Fig. 7b and Supplementary Fig. 5b), the expression of acyl-CoA-binding domain containing 5 (Acbd5), a gene encoding the peroxisome receptor for pexophagy, was also significantly decreased in Acot12^−/− UUO kidneys. Given the critical role of ACBD5 in facilitating the interaction between peroxisomes and autophagosomes, its downregulation suggests a potential mechanism through which impaired pexophagy contributes to the fibrotic phenotype observed in these kidneys. The colocalization of peroxisomes with the autophagosome marker LC3 was more prominent in Acot12^+/+ primary murine tubular epithelial cells than in Acot12^−/− cells, indicating the activation of pexophagy (Fig. 7c and Supplementary Fig. 5c). However, in Acot12^−/− primary murine tubular epithelial cells, a significantly lower colocalization level was observed, which was rescued by the restoration of Acot12. Immunohistochemical staining revealed that ACBD5 expression was significantly lower in Acot12^−/− kidneys than in Acot12^+/+ kidneys and in Acot12^+/+ UUO kidneys than in contralateral kidneys (Fig. 7d,e). The decreased level of ACBD5 in Acot12^+/+ UUO kidneys was reversed by treatment with losartan and fenofibrate (Supplementary Fig. 6a,b), which was previously shown to induce ACOT12 expression. However, fenofibrate treatment did not lead to an increase in ACBD5 expression in Acot12^−/− UUO kidneys. Consistent with ACOT12 expression, an age-dependent loss of ACBD5 expression was observed (Supplementary Fig. 6c), suggesting that the fibrosis induced by ACOT12 may be linked to pexophagy. The overexpression of ACBD5 in Acot12^−/− UUO kidneys also ameliorated kidney fibrosis and lipid accumulation (Fig. 7f). Although the introduction of ACOT12 or ACBD5 into 23-month-old Acot12^+/+ kidneys did not achieve as high a level of overexpression as that observed after earlier injections, the areas where ACOT12 or ACBD5 was overexpressed presented notable improvements in kidney fibrosis and lipid accumulation (Fig. 7g and Supplementary Fig. 7).

**Fig. 7: ACOT12 deficiency may inhibit pexophagy via the downregulation of ACBD5.**

Discussion

Accumulating evidence has revealed that defective FAO in the kidneys can induce kidney fibrosis^11,27. Tubular epithelial cells in the kidney use FA instead of glucose as the major energy source for ATP generation, and the importance of mitochondrial β-oxidation has been suggested^28,29,30. The downregulation of the CPT family proteins CPT1 and CPT2, which are involved in transporting FAs into the mitochondria, induces FA accumulation and contributes to CKD and kidney fibrosis^12,31. In this study, we found that ACOT12 KO exacerbated lipid accumulation and kidney fibrosis in kidney tubules.

ACOT12, which is found in the cytosol and peroxisomes and regulates acetyl-CoA metabolism, is highly expressed in the liver, kidney and intestine³². ACOT12 deficiency is known to increase de novo lipogenesis (DNL) and leads to the development of nonalcoholic fatty liver disease^19,33. In hepatocellular carcinoma, the accumulation of acetyl-CoA induced by ACOT12 deficiency increases the acetylation of TWIST2 and stimulates epithelial-mesenchymal transition (EMT)³⁴. In addition, the activation of acetyl-CoA synthetases increases DNL by stimulating the acetylation of acetyl-CoA carboxylase and FA synthase in hepatocellular carcinoma under hypoxia³⁵. These data suggest various biological and pathological roles for ACOT12. However, the pathophysiological role and mechanism of action of ACOT12 in kidney diseases remain unknown. Here, we found that ACOT12 deficiency significantly exacerbated kidney fibrosis in both aged and UUO-induced mice. Severe polycystic kidney injury and early lipid deposition in aged ACOT12-deficient mice indicate that the dysregulation of lipid metabolism is a critical factor in the aging process of the kidneys⁶. A recent study demonstrated that deficiency of PPARα stimulates DNL by modulating ACOT12 (ref. ¹⁹). Moreover, the administration of MHY2013, a PPARα/β agonist, reduced accumulation and alleviated age-associated kidney fibrosis³⁶. As defective lipid metabolism, such as impaired FAO signaling, has recently been implicated in kidney fibrosis, treatment with a PPARα agonist or overexpression of PPARα could be expected to exert an antifibrotic effect in a kidney fibrosis model. Here, we observed that exposure of UUO kidneys to fenofibrate reduces lipid accumulation and increases ACOT12 expression in kidney tubules. However, fenofibrate introduction to Acot12^−/− UUO kidneys did not ameliorate lipid accumulation or reverse fibrosis. Interestingly, restoring PPARα in Pparα^−/−Acot12^−/− kidneys exacerbated lipid accumulation and kidney fibrosis, suggesting an important role of ACOT12 in kidney fibrosis independent of PPARα.

Several studies have reported the relationship between autophagy and kidney fibrosis^37,38,39. Autophagy deficiency increases the sensitivity of kidneys to damage, leading to the accumulation of damaged mitochondria, impaired kidney function and severe kidney fibrosis⁴⁰. Moreover, the accumulation of cytosolic acetyl-CoA, potentially resulting from ACOT12 deficiency, is known to lead to the activation of acetyltransferase EP300 and mTORC1, and results in the inhibitory acetylation of autophagy-related proteins, leading to the suppression of autophagy^41,42. In this study, we found that pexophagy, a type of macroautophagy that selectively degrades peroxisomes, was decreased in the kidney tubules of Acot12^−/− UUO kidneys. The accumulation of defective peroxisomes or failure to remove defective peroxisomes due to dysfunctional pexophagy could alter cellular redox homeostasis and/or impair peroxisome regeneration^43,44. Defective pexophagy in the liver reportedly increases oxidative stress and liver injury and contributes to the development of liver fibrosis⁴⁵. The accumulation of impaired peroxisomes caused by pexophagy inhibition results in redox imbalance and kidney damage in the acute kidney injury model⁴⁶. It has also been reported that signaling pathways involved in fibrosis, such as the TGFβ signaling pathway, can affect peroxisome proliferation/dynamics⁴⁷. In this study, we observed decreased levels of peroxisomal biogenic factors in Acot12^−/− UUO kidneys. Moreover, we detected significantly decreased expression levels of ACBD5, which is essential for pexophagy and is required for the delivery of peroxisomes to lysosomes in Acot12^−/− UUO kidneys^48,49. These data indicate an association between pexophagy and kidney fibrosis induced by ACOT12 deficiency. However, the detailed mechanisms underlying this association and the extent of its involvement in different fibrotic conditions need to be investigated. Although an association between ACOT12 deficiency and kidney fibrosis has been identified, the precise mechanisms underlying this relationship and its prevalence across various fibrotic states require further investigation. Moreover, additional studies using conditional KO models are needed to closely examine the kidney-specific effects of ACOT12 deficiency. Considering the differences in susceptibility to kidney injury and peroxisome dysfunction between sexes, the implications of ACOT12 deficiency in females also need to be explored^{50,51,52,53,54}.

In summary, we used transgenic mouse models to delineate the biological functions of ACOT12 in the pathogenesis of kidney diseases, particularly kidney fibrosis. Our results demonstrated the important in vivo function of ACOT12 as a key regulator of lipid metabolism and kidney fibrosis. ACOT12 deficiency exacerbates kidney fibrosis possibly through the downregulation of pexophagy.

Data and materials availability

The authors confirm that all the data are available in the main text or the Supplementary Materials. For general comments or material requests, please contact E.-J.J. ([email protected]). The sequencing data that support the findings of this study can be accessed under the Gene Expression accession number GSE261115.