A total of 60 patients were enrolled in this study, with a male-to-female ratio of 2:3. The participants ranged in age from 18-65 years. Patient recruitment was conducted between March 2024 and March 2025. Renal biopsy specimens from patients with DKD were collected from the Department of Nephrology at the Second Affiliated Hospital of Guangdong Medical University using needle biopsy. Normal control renal tissues were obtained from histologically normal, non-tumorous regions of nephrectomy specimens via wedge biopsy, located at least 5 cm away from the tumor margin and were confirmed intraoperatively by pathological examination to be free of malignant cell infiltration (27). Written informed consent was obtained from all patients. All procedures were approved by the Medical Ethics Committee of the Second Affiliated Hospital of Guangdong Medical University (approval no. YSC202400101, 2024) and were performed in accordance with the Declaration of Helsinki.

Male BKS. Cg-Dock7m^+/+ Leprdb/J (db/db) mice, together with their age-matched non-diabetic lean db/m littermates, were purchased from the Model Animal Research Center of Nanjing University. A total of 100 male mice (db/m weighting 27±5 g, db/db weighting 41±10 g), were raised to 10 weeks old housed under specific pathogen-free (SPF) conditions at a constant temperature of (23±2°C), relative humidity of 50-60%, with a 12-h light/dark cycle. Animals had free access to standard chow and sterile acidified drinking water ad libitum. Experimental protocols were approved by the Animal Ethics Committee of Guangdong Medical University (approval no. GDY2402166, 2024). Mice in the treatment groups were administered FX-11 at a dose of 15 mg/kg via intraperitoneal injection once daily for 10 weeks.

For PFKFB3 knockdown, an adeno-associated virus (AAV9) encoding a short hairpin RNA targeting PFKFB3 (shPFKFB3 target sequence: GCCTCCAACATCATGGAAGTT; shNC target sequence: 5'-CCTAAGGTTAAGTCGCCCTCG-3', Shanghai GeneChem Co., Ltd.) was injected directly into the renal pelvis of db/db mice at 10 weeks of age.

Blood glucose concentrations were measured using a Roche Diabetes Care glucometer (Roche Diagnostics, Ltd.) from tail vein blood samples collected at 8 weeks of age. Mice exhibiting a fasting blood glucose level exceeding 16.7 mM were considered successfully diabetic. Urinary albumin, urinary creatinine, and serum creatinine levels were quantified using commercial kits from Roche Diagnostics, Ltd., following previously established protocols (28). Starting from 10 weeks of age, the db/db group received AAV-shPFKFB3 treatment and all groups were fasted for at least 8 h prior to blood glucose and urine tests, during which free access to water was permitted. Anesthesia was induced in an induction chamber with 5% isoflurane delivered in 100% oxygen at a flow rate of 1.5 l/min, and was continued until the loss of righting and pedal reflexes was confirmed. A blood volume of ~1.43 ml was obtained from each mouse via cardiac puncture at 20 weeks of age.

According to the AVMA Guidelines for the Euthanasia of Animals (2020 Edition) (29), at the end of the experimental protocol, all animals were humanely sacrificed. In accordance with the approved animal ethics protocol, mice were euthanized by overdose of isoflurane anesthesia. Briefly, mice were placed in an induction chamber and exposed to 5% isoflurane in oxygen until deep anesthesia was achieved, as confirmed by the absence of pedal withdrawal and corneal reflexes. Isoflurane exposure was continued for at least 1 min after respiratory arrest. Mortality was further confirmed by the absence of respiration, heartbeat and reflex responses, followed by vital organ harvest.

Human proximal tubular epithelial (HK-2) cells were cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum. To model diabetic conditions, cells were exposed to high glucose (HG; 30 mM), with normal glucose (NG; 5.5 mM) as control. PFKFB3 silencing was achieved by transfecting HK-2 cells with small interfering (si)RNA targeting PFKFB3 (siPFKFB3) at a final concentration of 50 nM using Lipofectamine^® RNAiMAX Transfection Reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Cells were incubated with the transfection complex at 37°C for 6 h, after which the medium was replaced with fresh complete medium. Cells were harvested for subsequent experiments 48 h after transfection. The siNC (5'-UUCUCCGAACGUGUCACGUTT-3') was used as a normal control. For lactate supplementation experiments, cells were treated with 10 mM sodium L-lactate (cat. no. S108838, Aladdin) in addition to HG conditions.

The spatial transcriptomics (ST) data of human DKD kidney samples have been deposited in the Gene Expression Omnibus (GEO) database at the National Center for Biotechnology Information (accession number: GSE261545; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE261545). For the present study, the ST datasets were retrieved from this GEO repository (https://www.ncbi.nlm.nih.gov/geo/). Dimensionality reduction was conducted using the Seurat package (version 4.2.2; https://satijalab.org/seurat/) within the R software environment (version 4.3.2; https://cran.r-project.org/bin/windows/base/old/4.3.2/). Previously established morphology-based clustering annotations were incorporated into the metadata. SCTransform was applied for data normalization and scaling using Seurat (version 4.2.2) in the R environment (version 4.3.2). The first 20 principal components were then used for Uniform Manifold Approximation and Projection to achieve dimensionality reduction, generating a comparable number of clusters for assessment against the morphology-based clusters. Subsequently, high-resolution spatial plots were generated to visualize the spatial distribution of these clusters alongside the expression levels of feature genes. The Loupe file produced by the 10X Space Ranger pipeline (10x Genomics; https://www.10xgenomics.com) was initially opened in the Loupe Browser (10x Genomics; version 6.5.0; https://www.10xgenomics.com/products/loupe-browser). A renal pathologist (P.I.) used hematoxylin and eosin (H&E)-stained (hematoxylin, 5 min; eosin, 2 min at room temperature) bright-field mosaics (ST slide) as anatomical references to perform morphology-based clustering for each spatially resolved spot on the Visium slide (Wuhan Servicebio Technology Co., Ltd.). This process involved manual delineation of renal morphological structures and lesions to identify the clusters.

GO enrichment analysis was per for med using the clusterProfiler R package (version 4.3.2; https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) based on the GO database. Differentially expressed genes were subjected to enrichment analysis, and terms with P<0.05 and FDR <0.05 were considered significantly enriched. Results were visualized using R software.

Kidney tissues were fixed in 10% neutral-buffered formalin at room temperature for 24 h, dehydrated through a graded ethanol series, cleared in xylene, and embedded in paraffin. Paraffin-embedded tissues were sectioned at 3 μm. For Masson's trichrome staining, sections were stained using a Masson's Trichrome Staining kit (Wuhan Servicebio Technology Co., Ltd.) at room temperature for 10 min according to the manufacturer's instructions. For Sirius Red staining, sections were stained using a Picro Sirius Red Staining kit (Wuhan Servicebio Technology Co., Ltd.) at room temperature for 60 min. For immunohistochemistry, sections were deparaffinized and rehydrated, followed by permeabilization with 0.1% Triton X-100 for 10 min. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 10 min. Sections were blocked with 5% bovine serum albumin (BSA) at room temperature for 30 min. Sections were incubated overnight at 4°C with primary antibodies against PFKFB3 (Proteintech Group, Inc.; cat. no. 13763-1-AP. 1:100), fibronectin (Abcam; cat. no. ab2413, 1:200), or α-smooth muscle actin (α-SMA; Abcam; cat. no. ab5694, 1:200). After washing, sections were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies (1:500; Wuhan Servicebio Technology Co., Ltd.) at room temperature for 1 h. Immunoreactivity was visualized using diaminobenzidine (DAB), followed by hematoxylin counterstaining for 3 min at room temperature. Images were acquired using a light microscope at ×200 and ×400 magnification.

Lactate levels in kidney tissue homogenates and HK-2 cell culture supernatants were measured using a colorimetric assay kit (Biovision; Abcam Bio X Cell; cat. no. K627-100) according to the manufacturer's protocol. Absorbance was read at 450 nm using a microplate reader, and lactate concentrations were calculated from a standard curve.

For immunofluorescence staining, frozen kidney sections (3 μm) or cultured HK-2 cells were fixed in ice-cold methanol/acetone for 10 min and washed with phosphate-buffered saline (PBS). Samples were blocked with 5 % BSA (Beyotime Biotechnology; cat. no. ST027) in PBS for 1 h at room temperature. Cells or sections were then incubated overnight at 4°C with primary antibodies against E-cadherin (Santa Cruz Biotechnology, Inc.; cat. no. sc-59778), PFKFB3 (Abcam, ab181861), H4K12la (PTM Bio, PTM-1411RM,), or HIPK2 (Abcam; cat. no. ab108543), all the primary antibody were diluted to 1:500. After washing with PBS, samples were incubated with fluorophore-conjugated secondary antibodies (1:1,000; Invitrogen; Thermo Fisher Scientific, Inc.) for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI (1 μg/ml) for 5 min at room temperature. Fluorescence images were captured using a fluorescence microscope.

Total proteins were extracted from mice kidney tissues and HK-2 cells using RIPA lysis buffer containing protease and phosphatase inhibitors (Beyotime Biotechnology; cat. no. P0013B,). Protein concentrations were determined using a BCA Protein Assay Kit (Beyotime Biotechnology; cat. no. P0010). Equal amounts of protein (20 μg per lane) were separated by 10% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes. Following 1 h blocking with 5% non-fat milk at 25°C and the membranes incubated overnight at 4°C with primary antibodies against PFKFB3 (1:1,000; Abcam; cat. no. ab181861,), fibronectin (1:1,000; Abcam ab2413,), α-SMA (1:1,000; Abcam; cat. no. ab5694,), H4K12la (1:1,000; PTMBio, PTM-1411RM), HIPK2 (1:1,000; Abcam; cat. no. ab108543,), or β-actin (1:10,000, loading control; Abcam; cat. no. ab8226,) for 1 h at room temperature. After incubation with horseradish peroxidase-conjugated secondary antibodies (1:10,000; Novus Biologicals; Bio-Techne), protein bands were visualized using an Enhanced Chemiluminescence (ECL) Detection Kit (Beyotime Biotechnology; cat. no. P0018FS) and quantified using Image J (version 1.52a; National Institutes of Health).

HK-2 cells were seeded at a density of 3×10⁵ cells/well in 6-well plates and harvested for RNA extraction. Total RNA was isolated from HK-2 cells and mice kidney tissues using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. RNA concentration and purity were determined using a Nanodrop spectrophotometer. Reverse transcription was performed using a PrimeScript RT Reagent Kit (Takara Bio, Inc.) according to the manufacturer's protocol. Quantitative PCR was carried out using TB Green Premix Ex Taq II (Takara Bio, Inc.) on a LightCycler Real-Time PCR System (Roche Diagnostics Ltd.) according to the manufacturer's instructions. The PCR cycling conditions were as follows: Initial denaturation at 95°C for 30 sec, followed by 40 cycles of denaturation at 95°C for 5 sec and annealing/extension at 60°C for 30 sec. Gene expression levels were normalized to β-actin as an internal control, and relative fold changes were calculated using the 2^-ΔΔCq method (30). Primer sequences are provided in Table SI. All experiments were performed in triplicate and repeated at least three independent times.

HK-2 cells (1×10⁵ cells per sample) were harvested and subjected to CUT&Tag assay using the Hyperactive CUT&Tag-seq kit (Vazyme Biotech Co., Ltd.; cat. no. 904). Briefly, cells were bound to concanavalin A-coated magnetic beads, followed by sequential incubation with anti-H4K12la antibody (PTMBio; cat. no. PTM-1411RM), a secondary antibody (goat anti-rabbit IgG H&L) (Abcam), and a hyperactive pA/G-TN5 transposase (Vazyme Biotech Co., Ltd.). Tagmented DNA fragments were extracted, amplified by PCR, and purified for library preparation. Library quality was assessed using a LabChip system, and sequencing was performed on an Illumina platform. The raw and processed sequencing data have been uploaded to the NCBI GEO database under accession number GSE328635 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE328635).

ChIP assays were performed using a commercial Chromatin Immunoprecipitation Kit (MilliporeSigma; cat. no. 17-371). HK-2 cells were seeded at a density of 2×10⁶ cells per 10 cm dish and cultured prior to experiments. Immunoprecipitation was carried out overnight at 4°C using either an anti-H4K12la antibody (1:100; PTMBio; cat. no. PTM-1411) or normal rabbit IgG as a negative control with 25-50 μg of sheared chromatin per immunoprecipitation reaction and 2-5 μg antibody per IP. A 1-5% aliquot of chromatin was saved as Input prior to immunoprecipitation. After reversing cross-links and purifying DNA, enrichment of specific genomic regions was analyzed by qPCR using promoter-specific primers (Table SII). PCR amplification was carried out using DreamTaq DNA polymerase (Thermo Fisher Scientific, Inc.; cat. no. 10342020). The cycling program was: 95°C for 3 min; 35 cycles of 95°C for 30 sec, 58°C for 30 sec, 72°C for 30 sec; and 72°C for 5 min. PCR products were electrophoresed on a 1.5% agarose gel, stained with ethidium bromide, and imaged using a ChemiDoc MP system (Bio-Rad Laboratories, Inc.). Washing steps were performed using the buffers provided in the kit, including low-salt, high-salt, LiCl, and TE wash buffers. Results are presented as percentage of input relative to IgG control. Quantification was performed using the Applied Biosystems 7500 Real-Time PCR System (Thermo Fisher Scientific, Inc.) and SDS software (v2.4; (Applied Biosystems; Thermo Fisher Scientific, Inc.).

Extracellular acidification rate (ECAR), a proxy for glycolytic activity, was measured using a Seahorse XF⁹⁶ extracellular flux analyzer (Agilent Technologies, Inc.). HK-2 cells were seeded into XF²⁴ V7 cell culture microplates and incubated for 2 h at 37°C in assay medium containing 10 mM glucose. Subsequently, oligomycin (an ATP synthase inhibitor) and 2-deoxyglucose (a glycolysis inhibitor) were sequentially injected, and ECAR was recorded at regular intervals to assess basal glycolysis, glycolytic capacity, and glycolytic reserve.

Data represent mean ± SD from three independent replicates. Statistical evaluation was carried out using SPSS 15.0 (SPSS, Inc.). Unpaired two-tailed t-tests were applied for pairwise comparisons, and one-way ANOVA with Bonferroni correction was used for multiple groups. P<0.05 was considered to indicate a statistically significant difference.

ST profiling was performed on H&E-stained renal sections from patients with DKD to visualize transcriptional landscapes within histological context. ST analysis identified a series of spatial 'spots' that were mapped onto the renal microanatomy, encompassing glomeruli, tubules, vascular regions and tubular casts. Reanalysis of a publicly available ST data enabled improved mapping resolution and the identification of rare cell populations contributing to local transcriptomic heterogeneity (Fig. 1A). Distinct clusters emerged, each displaying marked enrichment of canonical marker genes corresponding to specific renal structures (Fig. 1C). Further subclustering revealed that proximal tubular (PT) cells were divided into normal and injury-associated subtypes, the latter representing 53.9% of the total tubular cells (Fig. 1B and D). The expression levels of PFKFB3 in injured PT cells were markedly increased (Fig. 1E and F). GO enrichment analysis demonstrated that the injury-associated PT cluster was markedly enriched for glycolysis-related pathways, suggesting a metabolic shift contributing to DKD progression (Fig. 1G).

To further determine whether PFKFB3 expression is increased in renal tubules during DKD, its transcript and protein levels were assessed in db/db mice. RT-qPCR revealed a marked elevation of PFKFB3 mRNA levels in kidney derived from db/db mice, compared with those derived from db/m group (Fig. 2A). As shown in Fig. 2B, downregulation of E-cadherin expression, a TEC specific marker critical for maintaining epithelial integrity, together with elevated PFKFB3 expression, indicated that PFKFB3 was actively involved in tubular injury during DKD. In line with the aforementioned results, western blot and IHC analyses confirmed a significant elevation of PFKFB3 protein expression in the db/db group (Fig. 2C-E).

A kidney-specific PFKFB3 knockdown model was generated via in situ renal delivery of AAV9-shRNA to db/db mice to assess the role of PFKFB3 in DKD (Fig. 3A). Silencing of PFKFB3 expression markedly reduced blood glucose levels, serum creatinine and urinary albumin-to-creatinine ratio (UACR) at 20 weeks compared with those noted in the db/db group (Fig. S1A and B; Fig. 3B). Histological examination revealed significant renal fibrosis in db/db mice relative to the db/m group, which was effectively mitigated by shPFKFB3, as demonstrated by Masson's trichrome and Sirius Red staining (Fig. 3C). IHC staining further demonstrated a marked upregulation of fibronectin (FN) and α-SMA expression in the db/db mice, while these elevations were apparently reversed following knockdown of PFKFB3 expression (Fig. 3C). In line with these results, western blot analysis verified that the expression levels of FN and α-SMA were increased in db/db mice compared with those of the db/m mice, and were markedly reduced in the shPFKFB3-treated group (Fig. 3D and E). Conversely, PFKFB3 overexpression in db/db mice exacerbated renal fibrosis (Figs. S1, S2A and S2B). Increased extracellular matrix accumulation and interstitial expansion were observed in these mice, accompanied by upregulation of the expression levels of FN and α-SMA (Fig. S2B-D). Collectively, these results indicated that inhibition of PFKFB3 expression effectively attenuates fibrotic progression in DKD.

Among all PFKFB isoforms, PFKFB3 exhibits the highest kinase-to-phosphatase activity ratio, which preferentially shunts glucose metabolism toward glycolysis (31,32). Renal lactate levels were initially assessed, revealing a significant increase in db/db mice (Fig. S3). Inhibition of PFKFB3 expression markedly decreased lactate production, while its overexpression further elevated renal lactate accumulation compared with that noted in db/db mice. To determine whether lactate reduction could mitigate fibrotic injury, the lactate dehydrogenase inhibitor FX-11 was administered to db/db mice (Fig. 4A). Treatment with FX-11 led to a substantial decline in blood glucose and UACR (Fig. 4B). Histological evaluation revealed that FX-11 markedly alleviated tubulointerstitial fibrosis, as demonstrated by Sirius Red staining (Fig. 4C). Moreover, FX-11 treatment markedly decreased FN expression in db/db mice, while IF analysis revealed lack of notable changes in the db/m group (Fig. 4C). Consistently, western blot assays confirmed that FX-11 substantially suppressed the protein levels of FN and α-SMA compared with those of db/db group (Fig. 4D and E). Collectively, these findings suggested that lactate accumulation contributes to the progression of renal fibrosis in DKD and that inhibiting lactate production confers a protective effect against renal injury.

Given that lactate acts as a crucial metabolic substrate facilitating histone lactylation (33,34), it was hypothesized that this epigenetic modification may be aberrantly elevated in the kidneys of the DKD mouse models. Western blot analysis indicated that global histone lactylation was notably increased in TEC following stimulation with HG conditions compared with those noted under NG conditions, with a distinct band at ~11 kDa likely corresponding to histone H4 (Fig. 5A and B). Subsequent immunoblotting further verified that H4K12la levels were markedly upregulated in HG-exposed TECs relative to the NG group (Fig. 5C). Collectively, these data suggested that lactate accumulation promoted H4K12 lactylation under diabetic conditions. To characterize the potential downstream targets of histone lactylation in DKD, CUT&Tag sequencing was conducted using an anti-H4K12la antibody in TECs cultured under NG or HG conditions. HG exposure induced a notable increase in H4K12la enrichment peaks across the genome (Fig. 5D and E). Functional annotation of these peaks revealed a strong positive association with the p53 signaling pathway (Fig. 5F). Integrative analysis identified HIPK2 and ATM as putative lactylation-regulated genes, with genome browser tracks revealing prominent H4K12la enrichment within their respective promoter regions (Fig. 5G and H). RT-qPCR and ChIP-qPCR were performed to confirm the CUT&Tag target genes, revealing their elevated expression, H4K12la enrichment was performed at the promoters of HIPK2 and ATM, with both genes exhibiting substantial H4K12la modification (Fig. 5I and J).

To elucidate the role of PFKFB3/H4K12la in TEC function during DKD, renal histone lactylation levels were initially assessed. Western blot analysis revealed that inhibition of PFKFB3 expression markedly decreased global histone lactylation in db/db mice (Fig. S4A and B; Fig. 6A and B). Knockdown of PFKFB3 expression in HK-2 cells led to reduced histone lactylation (Fig. 6C and D). Immunofluorescence co-staining of H4K12la (red) and E-cadherin (green) demonstrated markedly elevated H4K12la signals in the renal tubules of db/db mice, which were attenuated by inhibition of PFKFB3 expression (Fig. 6E). Moreover, transfection of HK-2 cells with small interfering (si)PFKFB3 suppressed the expression levels of H4K12la, α-SMA and FN (Fig. 6F). Extracellular flux analysis revealed that PFKFB3 silencing reduced glycolytic capacity and lactate accumulation compared with those noted in HG group (Fig. 7A and B). In HK-2 cells, treatment with siPFKFB3 led to a decrease in the expression levels of H4K12la, HIPK2, α-smooth muscle actin (α-SMA), and FN compared with those noted in the HG group. Notably, excess lactate production reversed this suppression, exacerbating fibrotic responses and increasing H4K12la and HIPK2 levels relative to siPFKFB3-treated cells (Fig. 7C and D). Double immunofluorescence staining further validated the lactate-induced alterations in the levels of H4K12la and HIPK2 in HK-2 cells (Fig. 7E). Collectively, these results indicated that PFKFB3-driven lactate accumulation promotes tubular fibrosis through H4K12la-mediated regulation of profibrotic signaling.