Targeting the HIF-1α/PPARγ axis with curcumin carbon dots to combat uranium nephrotoxicity

Abstract

Uranium-induced nephrotoxicity involves complex mechanisms and lacks therapeutic interventions. This study identifies the HIF-1α/PPARγ axis as a key driver of renal lipid reprogramming and injury following uranium exposure. Using molecular docking, co-immunoprecipitation (Co-IP), ChIP-qPCR, and dual-luciferase reporter assays, we confirmed that HIF-1α directly regulates PPARγ expression and downstream lipotoxicity. The causal role of this axis was further validated using HIF-1α knockdown/knockout models and PPARγ pharmacological modulators (GW9662 and Rosi). Curcumin-derived carbon dots (CCDs) were synthesized at temperatures ranging from 120 to 210 °C. As-prepared CCDs-210 exhibited uniform size (< 10 nm), and excellent biocompatibility. Carbonization temperature governed cellular uptake: CCDs-210 entered via active endocytosis, while CCDs-120 used passive diffusion. In uranium-exposed HK-2 cells, CCDs-210 scavenged ROS, stabilized mitochondrial membrane potential, and suppressed the HIF-1α/PPARγ axis, reversing lipid dysregulation and improving cell viability. In vivo, CCDs-210 attenuated renal histopathological damage, restored renal function, and conferred multi-organ protection. These findings establish the HIF-1α/PPARγ axis as a central mediator of uranium nephrotoxicity and demonstrate that CCDs-210, by targeting this axis, effectively mitigate oxidative stress, inflammation, and lipid metabolic disruption. This work provides both a mechanistic foundation and a translatable nanotherapeutic strategy for treating uranium-induced kidney injury.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04620-x.

Introduction

Uranium, a naturally occurring radioactive heavy metal, is ubiquitous within the environment. Following internalization via occupational exposure, environmental contamination, or dietary intake, uranium exhibits multisystem toxicity, with the kidney serving as the primary site of accumulation and pathological injury [1–3]. While previous studies have confirmed that uranium exposure induces diverse renal pathological changes, including impaired glomerular filtration, renal tubular injury, and interstitial fibrosis [4, 5], the underlying molecular mechanisms remain incompletely elucidated. Recently, dysregulated lipid metabolism has emerged as a focal point in the pathogenesis and progression of kidney diseases. Abnormal lipid accumulation can reportedly exacerbate renal injury by triggering oxidative stress, inflammatory responses, and mitochondrial dysfunction [6–9]. However, the precise mechanisms by which uranium exposure modulates renal lipid metabolic reprogramming and its consequent association with nephrotoxicity require further investigation.

Hypoxia-inducible factor 1α (HIF-1α), a master transcription factor mediating cellular adaptation to hypoxic microenvironments, functions beyond oxygen homeostasis to critically regulate lipid metabolism [10, 11]. Concurrently, peroxisome proliferator-activated receptor γ (PPARγ), a prominent member of the nuclear receptor superfamily, serves as a core regulator of lipid homeostasis; its aberrant activation induces lipid accumulation and subsequent cellular dysfunction [12–14]. Although previous studies have suggested potential crosstalk between HIF-1α and PPARγ in metabolism-related diseases[15, 16], their synergistic involvement and underlying molecular mechanisms in uranium-induced renal injury remain unexplored.

Curcumin (Cur) is recognized as a potential renoprotective agent due to its anti-inflammatory, antioxidant, and HIF-1α inhibitory properties [17–20]; It alleviates histopathological damage in various models, such as ischemia-reperfusion injury and diabetic nephropathy, through pleiotropic mechanisms [21]. however, its clinical translation is severely hindered by poor bioavailability. To address this limitation, curcumin-derived carbon dots (CCDs), an emerging class of nanomaterials, have been developed to effectively circumvent these pharmacokinetic bottlenecks. Compared with native Cur, CCDs retain their intrinsic pharmacological activity while exhibiting significantly enhanced in vivo stability and therapeutic efficacy, attributable to their superior water solubility and biocompatibility [22]. These attributes position CCDs as promising therapeutic candidates [23, 24]. While CCDs have demonstrated potent efficacy in mitigating inflammation, oxidative stress, and metabolic disorders [25, 26], their capacity to alleviate uranium-induced renal lipid reprogramming via modulation of the HIF-1α/PPARγ axis remains unclear.

This study sought to investigate whether uranium exposure triggers renal lipid metabolic reprogramming via the HIF-1α/PPARγ pathway and delineates the associated mechanisms of renal injury. Furthermore, the capacity of CCDs to mitigate these pathological effects was evaluated. Overall, our findings offer novel insights into the molecular basis of uranium toxicity, highlighting potential therapeutic targets for the management of uranium-induced nephropathy.

Methods

Cell culture and treatment

Human renal tubular epithelial HK-2 cells (Cell Bank of the Chinese Academy of Sciences) were maintained in MEM supplemented with 10% fetal bovine serum (FBS) at 37 °C in a humidified 5% CO₂ atmosphere. For mechanistic analyses (e.g., HIF-1α/PPARγ axis validation), cells were exposed to lower concentrations (0.5–2.5 mM) of uranyl nitrate (Macklin, China) for 12–48 h to induce sustained stress while preserving viability. To evaluate the therapeutic efficacy of CCDs, a reversible injury model was established via a short 4 h exposure to 5 mM uranyl nitrate.

Establishment of HK-2^shHIF−1α stable cell line

The PPIGi-shHIF-1α plasmid was constructed and packaged (Supplementary Information Table S1 and Fig. S1). Log-phase HK-2 cells were seeded into 6-well plates (5 × 10⁵ cells/well) and transduced with viral supernatant upon reaching 30–40% confluence. The medium was replaced with fresh complete medium 8–10 h post-transduction. At 48 h, flow cytometry was utilized to evaluate initial GFP positivity. Subsequent puromycin selection was applied until the GFP⁺ population exceeded 95%, after which the stable cells were cryopreserved at −80 °C.

Preparation of CCDs

Cur (200 mg) was pyrolyzed under nitrogen (N₂) atmosphere at 120 °C, 150 °C, 180 °C and 210 °C for 4 h. The resulting orange or brown products were ultrasonicated in deionized water (20 mL) and centrifuged (10,000 rpm, 30 min) to discard aggregates. Supernatants were then dialyzed (1 kDa MWCO) against deionized water for 36 h (water refreshed every 8 h) and lyophilized to yield the respective solid quantum dots (CCDs-120, CCDs-150, CCDs-180, and CCDs-210).

Characterization methods

The microscopic morphology of CCDs was characterized using High-resolution Transmission Electron Microscopy (TEM, FEI Talos F200x system, USA). The crystal structure of Cur and CCDs were conducted on Powder X-ray diffraction (PXRD, Bruker D2 Phaser diffractometer, USA). The functional groups, chemical composition, and surface structure of Cur and CCDs were performed using Fourier Transform Infrared spectrometer (FTIR, Thermo Fisher Scientific Nicolet iS20, USA), X-ray Photoelectron Spectroscopy (XPS, Thermo Scientific K-Alpha spectrometer, USA). The structural characteristics of Cur and CCDs were examined using Ultraviolet-visible Spectrophotometer (UV-vis, Shimadzu UV-3600, JPN) and Fluorescence spectroscopy (FL, Edinburgh FS5, UK).

Cell viability assay

Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8) assay (Beyotime, China). HK-2 and HK-2^shHIF−1α cells were seeded into 96-well plates and cultured for 24 h. The cells were then exposed to various concentrations of uranyl nitrate (0.5–7.5 mM; Macklin, China) for 24, 48, or 72 h. To determine the functional necessity of PPARγ, cells were pre-treated with GW9662 or Rosi prior to uranyl nitrate exposure. Post-treatment, cells were washed with PBS and incubated with a 10% CCK-8 solution for 1 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader (Synergy LX). Cell viability was calculated as the ratio of the absorbance of treated wells to that of untreated control wells.

To evaluate the therapeutic effect of CCDs, HK-2 cells were initially exposed to 5 mM uranyl nitrate for 4 h. Following injury, the medium was replaced with fresh medium containing Cur and CCDs (1.25–10 µg/mL) for an additional 24 h, after which cell viability was determined. For rescue experiments, cells were co-treated with CCDs-210 (5 µg/mL) and HIF-1α stabilizer dimethyloxalylglycine (DMOG, 1 mM) to evaluate whether HIF-1α stabilization compromises the cytoprotective effects of CCDs.

Apoptosis assay

HK-2 cells and HK-2^shHIF−1α cells were treated with uranyl nitrate (0.5, 1, 2.5 mM) for 12 h. Post-treatment, cells were harvested using EDTA-free trypsin, washed with pre‑cooled 1× PBS, and resuspended in 1× binding buffer. The cells were then dual-stained with Annexin V‑FITC and propidium iodide (PI) for 15 min in the dark at room temperature. Apoptotic cells were subsequently quantified via flow cytometry.

Western blot analysis

Total proteins were extracted from tissues and cells using RIPA lysis buffer. Equal amounts of protein were resolved in 10% SDS-PAGE gels and transferred to PVDF membranes. After blocking with 5% non-fat dry milk in TBST for 2 h, the membranes were probed overnight at 4 °C with primary antibodies against HIF-1α, PPARγ and key lipid metabolism enzymes (ACC1, FASN, FABP4, DGAT2, LPL, and CPT-1 A) (Abcam, Cambridge, UK). Following incubation with secondary antibodies for 2 h at room temperature, protein bands were visualized using an ECL substrate (Bio-Rad, USA) and a chemiluminescence imaging system. GAPDH and β-actin served as loading controls.

Quantitative real-time polymerase chain reaction (RT-qPCR)

Total RNA was extracted using TRIzol reagent and reverse-transcribed into cDNA using a Thermo Fisher transcription kit. RT-qPCR was performed using 2X SYBR Green Pro Taq HS Premix. Relative gene expression was calculated via the 2^−ΔΔCt method, with GAPDH serving as the endogenous reference for normalization. Primer sequences are listed in Supplementary Information Table S2 and Table S3.

Co-immunoprecipitation (Co-IP)

Lysates from log-phase HK-2 cells were incubated overnight with an anti-HIF-1α antibody (Abcam, Cambridge, UK) or a control mouse IgG (Proteintech, Chicago, USA). Immune complexes were captured using Protein A/G agarose beads and eluted by boiling in (1×) SDS-PAGE loading buffer at 100 °C for 10 min. The co-precipitated proteins were subsequently analyzed via Western blot using an anti-PPARγ antibody.

ChIP-qPCR

ChIP assays were performed using the EZ-ChIP^™ Kit. Briefly, formaldehyde-crosslinked chromatin from HK-2 cells was sheared via sonication into 200–1000 bp fragments. The chromatin was then immunoprecipitated overnight at 4 °C with anti-HIF-1α or control IgG antibodies (Abcam, Cambridge, UK) followed by capture with Protein A/G magnetic beads. After a series of stringent washes, the protein-DNA complexes were eluted, and cross-links were reversed at 65 °C. The purified DNA was then subjected to qPCR analysis using PPARγ promoter-specific primers. Data were calculated as a percentage of the starting input (1%) to quantify the enrichment of HIF-1α at the target locus.

Dual-luciferase reporter assay

Transcriptional activity was evaluated using the Dual-Luciferase Reporter Assay Kit (yeasen). HK-2 cells were co-transfected with a pcDNA3.1-Luc plasmid containing the human PPARγ promoter and a pRL-TK Renilla luciferase control vector via Lipofectamine 3000 (Invitrogen). At 48 h post-transfection, relative luciferase activity was determined by normalizing firefly signals to Renilla signals.

Cellular uptake

HK-2 cells (2 × 10⁵ cells/well) were seeded into 6-well plates and cultured for 24 h. Cells were then treated with Cur or CCDs (5 µg/mL) under three distinct conditions: (1) for time-dependent analysis, cells were incubated for 1–2 h at 37 °C; (2) to elucidate endocytic mechanisms, cells were pre-treated for 30 min with specific inhibitors including chlorpromazine (10 µg/mL), genistein (55 µg/mL), or cytochalasin D (10 µg/mL) (Macklin, China) prior to a 2 h incubation; (3) for temperature-dependent assays, cells were pre-incubated at 4 °C for 30 min, followed by a 2 h treatment at 4 °C. Post-treatment, all cells were washed with ice-cold PBS, harvested, and analyzed via flow cytometry.

Nile red staining

To evaluate lipid accumulation, HK-2 and HK-2 ^shHIF−1α cells were subjected to various treatments prior to Nile Red staining. For mechanistic analyses, cells were exposed to 1 mM uranyl nitrate for 24 h, with or without the PPARγ antagonist GW9662 or agonist Rosi. In rescue experiments, shHIF-1α stable cells were co-treated with 5 mM uranyl nitrate and Rosi to determine whether PPARγ activation mitigates the protective effects of HIF-1α knockdown. To assess the therapeutic efficacy of CCDs, cells were injured with 5 mM uranyl nitrate for 4 h, followed by a 24 h post-treatment with Cur or CCDs (5 µg/mL). Post-treatment, all cells were fixed with 10% paraformaldehyde (PFA), stained with 0.05 µg/mL Nile Red (Sigma-Aldrich), and counterstained with Hoechst (Beyotime, China). Fluorescent images were acquired using a confocal laser scanning microscope (CLSM, STELLARIS 5) and quantified via ImageJ software.

Oxidative stress and mitochondrial membrane potential (ΔΨ_m) assays

HK‑2 cells were exposed to 5 mM uranyl nitrate for 4 h, followed by 24 h post-treatment with Cur or CCDs (5 µg/mL). Cells were stained with 1 mL of DCF‑DA, DHE, or JC‑1 working solution at 37 °C for 30 min in the dark. Following PBS washes, fluorescence images were captured via CLSM and quantitatively analyzed with ImageJ software. For the JC‑1 assay, monomeric (green) and aggregated (red) fluorescence were detected simultaneously. The red/green fluorescence intensity ratio was used to reflect changes in mitochondrial membrane potential. Staining intensities of DCF-DA, DHE, and JC-1 were quantified via ImageJ, with mean fluorescence intensity (MFI) defined as the ratio of integrated density (IntDen) to the selected area. At least five random fields per group were captured for quantification.

Mouse breeding and genotyping

Male heterozygous knockout mice (C57BL/6J‑Hif1α^em1C/Cya, background C57BL/6J; Cyagen) were crossed with female wild‑type C57BL/6J mice (GemPharmatech) to generate F1 progeny. Following PCR-based genotyping, genotypically verified WT (HIF‑1α^+/+) and heterozygous (HIF‑1α^+/−) F1 littermates were rigorously selected for subsequent phenotypic analyses and comparative uranium exposure studies.

Animal experiments

Uranium-Induced Kidney Injury Model:

To establish the kidney injury model, mice were administered uranyl nitrate (5 mg/kg in saline). Twenty-four mice were randomly allocated into four groups based on genotype: WT, HIF‑1α^+/−, U, and U + HIF‑1α^+/− (n = 6 per group). Peripheral blood (250 µL per mouse) was collected via retro‑orbital puncture at 24 and 48 h post-administration. Serum (60–90 µL) was separated by centrifugation and stored at −80 °C. At 48 h, mice were euthanized by cervical dislocation. The right kidneys were fixed in 4% PFA; while the left kidneys were tri-sectioned, flash‑frozen in liquid nitrogen, and stored temporarily at −80 °C.

Carbon‑Dot Treatment Experiment:

Kidney injury was established via intravenous administration of uranyl nitrate (5 mg/kg). At 24 h post-injury, mice were randomly assigned to receive daily intravenous injections of saline, Cur, CCDs-120, or CCDs-210 (100 mg/kg) for three consecutive days. Upon completion of the treatment regimen, peripheral blood and kidney tissues were collected and processed according to the aforementioned protocols.

Renal function indicator assessment and pathological analysis

Serum urea (UREA) and creatinine (CREA) levels were quantified via a ChemRay automated biochemical analyzer with commercial kits. For histological assessment, 4% PFA-fixed kidney tissues were paraffin-embedded and sectioned at a thickness of 5 μm. Hematoxylin and eosin (H&E) staining was performed to assess structural morphology, and the Paller scoring system was applied to evaluate tubular damage based on brush border loss, tubular dilation/atrophy, vacuolization, and inflammatory infiltration. Lesions were graded from 0 to 4: 0 (normal), 1 (mild, damage < 5%), 2 (moderate, 5–25%), 3 (severe, 25–75%), and 4 (extensive, > 75%).

Cytokine assay

Serum concentrations of TNF-α, IL-6, and IL-1β were determined using ELISA kits (Ruixin BioTech, Fujian, China) per the manufacturer’s instructions. Optical density (OD) was measured at 450 nm using a microplate reader (Bio-Rad Laboratories, Inc.).

Measurement of oxidative stress in renal tissue

Right kidney sample (~0.1 g) was homogenized, and the resulting homogenate was analyzed to determine the levels of SOD, GSH, and MDA using the kits provided by the Nanjing Jiancheng Bioengineering Institute (Nanjing, China), as per the manufacture’s guidelines.

Oil red O staining

OCT-embedded renal tissue samples were cryosectioned into 10 μm thick slices. The sections were stained with Oil Red O (Solarbio, Beijing, China), differentiated in 60% isopropanol, and counterstained with hematoxylin. Representative images were captured using a light microscope (Leica, Germany). Lipid droplet areas were measured using Image‑Pro Plus 6.0 software, and the relative percentage was calculated using the following formula:

Relative lipid droplet area (%) = (Lipid droplet area/Tissue area) × 100%.

Transmission electron microscopy (TEM) analysis

Renal tissue was fixed using an electron microscope fixative and osmium tetroxide, dehydrated with graded ethanol, and permeabilized using an acetone-embedding agent. Samples were polymerized at 60 °C for 48 h. Ultrathin Sects. (60–80 nm) were cut using an ultramicrotome, stained with 2% uranyl nitrate-saturated ethanol solution and 2.6% lead citrate, and observed using TEM for pyroptosis features.

Periodic acid schiff (PAS) staining

Paraffin-embedded renal Sect. (3 μm) were deparaffinized, rehydrated, and oxidized in 0.5% periodic acid for 10 min. After incubation with Schiff reagent for 20 min in the dark and a tap water rinse for 10 min, the slides were counterstained with hematoxylin, and mounted.

Immunofluorescence staining

Renal paraffin sections were dewaxed, rehydrated, and treated with EDTA antigen retrieval solution by microwave heating. After serum blocking, the sections were incubated overnight at 4 °C with anti-HIF-1α and anti-PPARγ antibodies, with or without co-staining for AQP1 (a proximal tubule marker) or Podocin (a glomerular marker). Following a 1 h dark incubation with fluorescent secondary antibodies and DAPI counterstaining, a minimum of seven random high-power fields per sample were imaged using fluorescence microscopy. Positive signal area was quantified using ImageJ software.

Isothermal titration calorimetry (ITC)

Binding thermodynamics between CCDs-210 and HIF-1α were determined using a MicroCal ITC system. Titrations were performed at 25 °C to record the heat of reaction. The dissociation constant (K_D), enthalpy change (ΔH), and entropy change (ΔS) were calculated by fitting the raw data to a one-site binding model. These parameters elucidated the primary non-covalent driving forces (e.g., hydrogen bonding, van der Waals interactions) underlying the binding event.

Cellular thermal shift assay (CETSA)

To verify intracellular target engagement, intact HK-2 cells treated with CCDs-210 were subjected to a temperature gradient (37–70 °C) for 3 min. Following cell lysis via freeze-thaw cycles, the thermal stability of HIF-1α was analyzed by Western blot. The resulting melting curves (Tm) were compared between the control and CCDs-210 treated groups to assess the ligand-induced stabilization of the target protein.

Lipidomic and metabolomic profiling

Sample Preparation for Cell Lipidomics: Cryopreserved cell samples were thawed on ice and resuspended in 500 µL of pre-cooled extraction solvent (80% aqueous methanol) containing internal standards. Following a 3 min vortex, complete lysis was achieved via three freeze-thaw cycles (liquid nitrogen and ice bath, 5 min each). The lysates were centrifuged (12,000 rpm, 3 min, 4 °C), and a 200 µL aliquot of the supernatant was transferred to a low-binding vial for lipidomic analysis.

Sample Preparation for Tissue Metabolomics: Tissue samples (~100 mg) were ultrasonically homogenized. The homogenates were extracted using a cold solvent mixture comprising 500 µL of 80% methanol and 250 µL of 20% acetonitrile in methanol (v/v). After vortexing for 3 min, the mixture was centrifuged (12,000 rpm, 10 min, 4 °C). The supernatant was incubated at −20 °C for 30 min to further precipitate proteins and subjected to a second identical centrifugation. Finally, 180 µL of the clarified supernatant was filtered through a 0.22 μm plate for analysis.

Instrumental Analysis: All prepared samples were analyzed using an ultra-performance liquid chromatography system (ExionLC^™AD, SCIEX) coupled with a tandem mass spectrometer (QTRAP^®6500+, SCIEX).

Statistical analysis

Data normality was assessed via the Shapiro-Wilk test. Normally distributed data are presented as mean ± standard deviation (SD). Comparisons between two groups were performed using an unpaired Student’s t-test (parametric) or a Mann-Whitney U test (non-parametric), as appropriate. For multi-group comparisons, one-way or two-way analysis of variance (ANOVA) was conducted, followed by Tukey’s or Dunnett’s post hoc tests for multiple comparisons. n represents the number of biological replicates. All statistical analyses were executed using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA). Statistical significance was defined as P < 0.05, with significance levels denoted as *P < 0.05, **P < 0.01, and ***P < 0.001.

For lipidomics and metabolomics, data were processed using the Metware Cloud platform. Raw intensities were normalized to internal standards and total peak area to minimize technical variation. To eliminate technical variations, raw data were corrected using internal standards and normalized to the total peak area. Prior to multivariate analysis, peak intensities were log-transformed and Pareto-scaled. Quality control (QC) samples were interspersed throughout the analytical sequence to monitor instrument stability; features with a coefficient of variation (CV) > 30% in the QC samples were discarded. Multiple comparisons were adjusted using the false discovery rate (FDR), with an FDR < 0.05 established as the threshold for identifying significantly differential lipids or metabolites.

Results

HIF-1α deficiency attenuates uranium-induced renal injury

To elucidate the role of HIF-1α in uranium-induced renal injury, we established a stable HIF-1α-knockdown HK-2 cell line (HK-2^shHIF−1α) and exposed both parental HK-2 and HK-2^shHIF−1α cells to varying concentrations of uranium (0–7.5 mM) for 24 h, 48 h, and 72 h (Fig. 1A). Uranium inhibited HK-2 cell viability in a dose- and time-dependent manner. Notably, growth inhibition was significantly attenuated in HK-2^shHIF−1α cells at equivalent uranium concentrations (P < 0.05), indicating that HIF-1α deficiency confers protection against uranium cytotoxicity. Apoptosis assays at 0.5, 1, and 2.5 mM uranium confirmed a concentration-dependent increase in apoptosis in parental HK-2 cells, which was significantly mitigated in HK-2^shHIF−1α cells (P < 0.05) (Fig. 1B).

Fig. 1.

HIF-1α mediates uranium-induced renal injury and lipid accumulation. (A) Viability of HK-2 and HK-2^shHIF-1α cells treated with uranium (0–7.5 mM) for 24, 48, and 72 h. (B) Apoptosis percentage of HK-2 and HK-2^shHIF-1α cells exposed to 0.5, 1, and 2.5 mM uranium. (C) Schematic of the HIF-1α knockout mouse model generation. (D, E) Representative H&E staining of renal tissues and corresponding injury scores. (F) Serum creatinine (Crea) and blood urea nitrogen (Urea) levels. (G) Representative Nile Red staining of HK-2 and HK-2^shHIF-1α cells following uranium exposure. (H) Representative Oil Red O staining of renal tissues from wild-type (WT) and HIF-1α^+/- mice following uranium exposure. (I) Quantification of Nile Red and Oil Red O staining intensity. Data are presented as mean ± SD. **P < 0.01, ***P < 0.001

In vivo validation was performed using a HIF-1α heterozygous knockout (HIF-1α ^+/−) mouse model (Fig. 1C). Histopathological assessment via H&E staining demonstrated that uranium exposure precipitated severe renal architectural distortion, characterized by diffuse tubular epithelial swelling, vacuolar degeneration, and leukocytic infiltration into the interstitium (Fig. 1D). These changes were reflected in significantly elevated renal injury scores (Fig. 1E). Uranium exposure significantly elevated serum Crea and Urea levels (P < 0.05), whereas HIF-1α deficiency remarkably attenuated renal impairment, maintaining these parameters at significantly lower levels than the wild-type group (Fig. 1F). Despite the potential for compensatory activation of the remaining allele, the haploinsufficiency-driven reduction in HIF-1α expression offered significant protection against nephrotoxicity. This indicates that the severity of uranium-induced renal damage is highly dependent on HIF-1α protein levels.

Given the established link between renal injury and ectopic lipid deposition, we assessed lipid accumulation by Nile Red staining in HK-2^shHIF−1α cells (Fig. 1G) and Oil Red O staining in HIF-1α^+/− mouse kidney tissues (Fig. 1I). Uranium exposure markedly triggered lipid accumulation in both parental HK-2 cells and wild-type mouse kidneys, whereas HIF-1α deficiency significantly attenuated this deposition (P < 0.05) (Fig. 1H), suggesting that HIF-1α mediates uranium-induced lipid dysregulation.

HIF-1α/PPARγ axis dictates uranium-induced renal injury, with PPARγ acting as the functional pivot

Although the HIF-1α/PPARγ axis has been implicated in the pathogenesis of lipid metabolic disorders [27, 28], the direct crosstalk between these two factors in the context of uranium-induced renal injury remains to be elucidated. RT-qPCR and Western blot analyses were initially used to evaluate the effects of uranium exposure on the expression levels of HIF-1α and PPARγ in HK-2 cells. The results demonstrated that uranium treatment (0.5, 1, and 2.5 mM) elicited a significant, time- and dose-dependent upregulation of both mRNA and protein levels of HIF-1α and PPARγ at 24 h and 48 h (P < 0.05). Interestingly, the uranium-induced overexpression of these two factors was markedly abrogated in the stable HK-2^shHIF−1α cell line (P < 0.05) (Figs. 2A-B). In HIF-1α^+/− mice, genetic ablation synchronously reduced uranium-induced HIF-1α and PPARγ expression at mRNA and protein levels in renal tissues (P < 0.05) (Figs. 2C-E). Immunofluorescence confirmed that the augmented HIF-1α and PPARγ signals in uranium-exposed kidneys were effectively suppressed by HIF-1α deficiency (Fig. 2F).

Fig. 2.

Uranium exposure promotes renal injury via activation the HIF‑1α/PPARγ axis. (C) RT-qPCR analysis of HIF-1α and PPARγ mRNA levels and (D) representative Western blot images and densitometric quantification of HIF-1α and PPARγ protein levels in HK-2 and HK‑2^shHIF‑1α cells following uranium treatment. (E) RT-qPCR analysis of HIF-1α and PPARγ mRNA levels and (F-G) representative Western blot images and densitometric quantification of HIF-1α and PPARγ protein levels in renal tissues of wild-type (WT) and HIF-1α^+/- mice exposed to uranium. (H) Immunofluorescence staining of HIF-1α and PPARγ (red) in renal tissues. Nuclei were stained with DAPI (blue). Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Molecular docking simulations were employed to predict the binding affinity and specific interaction sites between HIF-1α and PPARγ (Fig. 3A). Following the screening of optimal docking conformations via Zdock and Zrank, the top-ranked complex exhibited an electrostatic potential energy of −120.07 kcal/mol, an optimized van der Waals non-bonded interaction energy of −0.059 kcal/mol, and a desolvation energy of −17.10 kcal/mol. The resulting E-rdock score of −9.5 indicated a strong binding potential between the two proteins. Previous reports have indicated that the amino acid sequence within the 232–500 residue region of PPARγ serves as a pivotal domain for molecular interactions [29]. Consistently, the two-dimensional interaction map generated via LigPlot (Supplementary Information Figure S2) revealed that residues including Gln286 and Thr459 in PPARγ form hydrogen bond interactions with HIF-1α, further supporting the structural stability of the complex [30]. To experimentally validate this interaction, Co-IP was performed using an anti-HIF-1α antibody for pulldown, followed by Western blot analysis with an anti-PPARγ antibody. Distinct bands were detected at approximately 58 kDa in both the Input and IP fractions, confirming a direct physical interaction between HIF-1α and PPARγ (Fig. 3B). To further determine whether HIF-1α directly transactivates PPARG at the transcriptional level, we performed chromatin immunoprecipitation (ChIP) and luciferase reporter assays. As shown in Fig. 3C, ChIP assays revealed a distinct amplification band at the target fragment (approximately 150–200 bp) in the IP group compared to the IgG negative control. HIF-1α showed significant enrichment at the PPARG promoter region, with an enrichment efficiency of approximately 20%–35% relative to the input, confirming a direct physical interaction between them. Furthermore, dual-luciferase reporter assays indicated that exogenous overexpression of HIF-1α markedly enhanced the transcriptional activity of the luciferase reporter by 4- to 6-fold compared with the empty vector group (Fig. 3D). Taken together, these data establish HIF-1α as a direct transcriptional activator of PPARG, exerting its influence through both site-specific physical binding and functional modulation.

Fig. 3.

HIF‑1α-driven PPARγ activation mediates uranium-induced renal lipotoxicity. (A) Molecular docking and (B) Co-immunoprecipitation (Co-IP) assay confirming the physical interaction between HIF-1α and PPARγ. (C) Chromatin immunoprecipitation (ChIP) assay showing the enrichment of HIF-1α at the PPARG promoter region. (D) Dual-luciferase reporter assay demonstrating the transcriptional activation of the PPARG promoter by HIF-1α. (E) Impact of PPARγ antagonist (GW9662) and agonist (Rosi) on cell viability and lipid accumulation (F, G) in uranium-treated HK-2 cells. (H) Rescue-reversal experiments in HK-2^shHIF−1α cells: Rosi treatment reversed the protective effects of HIF-1α knockdown, as evidenced by cell viability and (I, J) Nile Red staining. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

To evaluate the functional necessity of PPARγ in uranium-induced nephrotoxicity, we employed the specific PPARγ antagonist GW9662 and the high-affinity agonist Rosi. CCK-8 assays (Fig. 3E) and Nile Red staining results (Figs. 3F-G) showed that GW9662 significantly attenuated uranium-induced cellular damage and lipid deposition, whereas Rosi exacerbated these pathological changes. To definitively establish HIF-1α as the upstream driver, we performed rescue-reversal experiments. In HK-2^shHIF−1α cells treated with 5 mM uranium, the cytoprotection and reduction in lipid accumulation conferred by HIF-1α inhibition were significantly reversed by the addition of the PPARγ agonist Rosi (manifested as decreased cell viability and aggravated lipid accumulation) (Figs. 3H-J). This finding confirms that the protective effect of HIF-1α inhibition is mediated by suppression of its downstream target, PPARγ, thereby establishing the hierarchical causal relationship within this signaling axis.

Collectively, these findings suggest that uranium exposure modulates PPARγ expression by activating HIF-1α, thereby establishing a functional HIF-1α/PPARγ axis that mediates the progression of uranium-induced renal injury.

HIF-1α deficiency ameliorates uranium-induced lipid metabolic reprogramming via downstream pparγ-mediated enzymatic networks

To elucidate the regulatory role of HIF-1α in uranium-induced lipid metabolism, lipidomic profiling was conducted. Principal Component Analysis (PCA) demonstrated distinct clustering and clear separation among the Control, U, HK-2^shHIF−1α, and U + HK-2^shHIF−1α groups, indicating significant divergence in their lipid metabolic profiles (Fig. 4A). A volcano plot was generated, revealing that uranium exposure significantly upregulated 234 lipid species (P < 0.05, FC > 2), while HIF-1α deficiency markedly downregulated 207 lipid species compared to the U group (P < 0.05, FC < 0.5) (Fig. 4B). K-means clustering confirmed that HIF-1α suppression shifted the lipidome toward the baseline profile (Figs. 4C-E). Heatmap analysis showed that triglycerides (TG), diglycerides (DG), and free fatty acids (FFA) were most profoundly suppressed by HIF-1α deficiency (Fig. 4F). KEGG enrichment identified uranium-perturbed pathways, including ether lipid metabolism, steroid biosynthesis, and linoleic, arachidonic, and α-linolenic acid metabolism. HIF-1α deficiency further modulated glycerophospholipid metabolism (Fig. 4G).

Fig. 4.

Lipidomic profiling reveals that HIF-1α knockdown attenuates uranium-induced lipid dysregulation in HK-2 cells. (A) PCA score plot of lipidomic profiles across four groups: NC, U, HK‑2^shHIF‑1α, and U + HK‑2^shHIF‑1α. (B) Volcano plots identifying differentially abundant lipids in the U vs. NC (left) and U + HK‑2^shHIF‑1α vs. U (right) comparisons. Orange and green dots represent significantly upregulated (P < 0.05, FC > 2) and downregulated (P < 0.05, FC < 0.5) lipids. (C) Integrated volcano plot comparing differential lipids across three comparisons: U vs. NC (part 1), U + HK‑2^shHIF‑1α vs. U (part 2), and U + HK‑2^shHIF‑1α vs. NC (part 3). (D) Classification scatter plot illustrating the alteration patterns of five major lipid classes-fatty acyls (FA), glycerolipids (GL), glycerophospholipids (GP), sphingolipids (SP), and sterol lipids (ST)-between the indicated comparison groups. (E) K-means clustering analysis of differentially expressed lipids. (F) KEGG pathway enrichment analysis. Bubble plots show significantly enriched pathways in the U vs. NC group (left) and U + HK‑2^shHIF‑1α vs. U group (right). (G) Heatmap displaying the relative abundance and P-values of representative differential lipids across the U vs. NC, U + HK‑2^shHIF‑1α vs. U, and U + HK‑2^shHIF‑1α vs. NC comparisons

To correlate these lipidomic shifts with downstream molecular events, we investigated the expression of key metabolic enzymes regulated by the major lipid-metabolism regulator, PPARγ (Figs. 5A-C). Uranium exposure significantly induced the mRNA and protein expression of lipogenic enzymes (ACC1, FASN, LPL), the fatty acid transporter (FABP4), and the rate-limiting enzyme for triglyceride synthesis (DGAT2), while concurrently suppressing the fatty acid oxidation catalyst (CPT-1 A). Importantly, these enzymatic alterations were markedly reversed by the PPARγ-specific antagonist GW9662, confirming the substantial role of the HIF-1α/PPARγ axis in governing the expression of these metabolic enzymes.

Fig. 5.

The HIF-1α/PPARγ axis drives renal lipid remodeling by regulating key metabolic enzymes. (A-C) RT-qPCR and Western blot analysis of lipogenic enzymes (ACC1, FASN, LPL), fatty acid transporters (FABP4), triglyceride synthesis enzymes (DGAT2), and fatty acid oxidation catalysts (CPT-1 A) in HK-2 cells. Cells were exposed to uranium with or without pharmacological modulation of PPARγ via the antagonist GW9662 or agonist Rosi. (D) Schematic diagram illustrating the mechanistic link between HIF-1α/PPARγ-mediated enzymatic shifts and specific lipidomic alterations. Coordinated upregulation of ACC1, FASN, LPL, and DGAT2 facilitates FFA accumulation and TG/DG synthesis, while FABP4 induction and CPT-1 A suppression further exacerbate lipid deposition and impair mitochondrial β-oxidation. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

These enzymatic shifts provide a direct mechanistic explanation for the specific lipidomic alterations observed (Fig. 5D). Coordinated upregulation of ACC1, FASN, and LPL reflects enhanced PPARγ-driven de novo fatty acid synthesis and uptake, leading to the accumulation of FFA. The simultaneous induction of DGAT2, which catalyzes the final step of TG synthesis, directly accounts for the elevated levels of TG and DG. The upregulation of FABP4 facilitates intracellular fatty acid transport and lipid droplet formation, whereas the downregulation of CPT-1 A indicates impaired mitochondrial fatty acid oxidation. This metabolic imbalance creates a “feed-forward” loop that further exacerbates lipid deposition. Collectively, these findings substantiate that HIF-1α acts as a critical regulatory node in uranium-induced nephrotoxicity by driving PPARγ-dependent metabolic remodeling.

Preparation and morphological characterization of CCDs

Having established that the HIF-1α/PPARγ axis is a central driver of uranium-induced renal lipid reprogramming and subsequent injury, we engineered a nanomedicine platform to target this pathological axis as a potential therapeutic intervention. Given that Cur possesses inherent regulatory effects on lipid metabolism yet exhibits poor bioavailability, CCDs were fabricated via a simplified pyrolysis method to preserve its bioactive groups while forming a stable sp²-hybridized carbon core [23], thereby significantly improving the water solubility and therapeutic potential of the precursor.

CCDs were synthesized by pyrolyzing Cur under N₂ at 120, 150, 180, and 210 °C for 4 h (Fig. 6A, Supporting Information Figure S3). With increasing temperature, the powder color deepened from orange-yellow to dark brown through sequential dehydration, polymerization, and carbonization, yielding CCDs-120, CCDs-150, CCDs-180, and CCDs-210. High-resolution transmission electron microscopy (HRTEM) (Figs. 6B-I) revealed uniform morphologies with particle sizes of 6.1 ± 1.8, 9.9 ± 2.4, 7.0 ± 1.1, and 5.6 ± 1.2 nm for CCDs-120, −150, −180, and −210, respectively. This nanoscale size is conducive to cellular internalization and in vivo bioavailability, both of which are critical for subsequent renal-protective applications. Furthermore, HRTEM images revealed distinct lattice fringes with a spacing of 0.21–0.22 nm, corresponding to the (100) crystal plane of graphite [31, 32]. This result confirms that during pyrolysis, Cur undergoes sequential dehydration, polymerization into a polymer intermediate, and subsequent carbonization, ultimately forming a stable sp²-hybridized carbon core.

Fig. 6.

Preparation and characterization of Cur and CCDs. (A) Schematic illustration of the synthesis of CCDs via the pyrolysis method. (B-I) TEM and HRTEM images of CCDs; Insets show the particle size distribution histograms. (J) XRD pattern of CCDs. (K) FTIR spectrum of CCDs. (L-O) High-resolution XPS spectra of C 1 s (L, M) and O 1 s (N, O) for Cur and CCDs-210. (P) UV-vis absorption spectrum of Cur and CCDs. (Q) FL spectrum of Cur and CCDs. (R) Photographs of Cur and CCDs-210 solutions

PXRD patterns were used to analyze the crystal structure of Cur and CCDs (Fig. 6J). As the pyrolysis temperature increased, the intensity of Cur’s characteristic diffraction peaks gradually decreased, indicating disruption of its crystalline structure during dehydration and carbonization. At pyrolysis temperatures of 180 and 210 °C, the characteristic peaks related to Cur completely disappeared, and a broad diffraction peak appeared at around 20°, which corresponds to the (002) crystal plane of graphitized carbon [31]. The higher peak intensity of CCD-210 than that of CCD-180 suggests that the degree of graphitization is positively correlated with the carbonization temperature.

FTIR, XPS, UV-vis, and FL were next used to characterize the functional groups, chemical composition, and surface structure of Cur and CCDs. FTIR spectra further confirmed the structural changes in the CCDs (Fig. 6K). The broad peak around 3400 cm⁻¹ was attributed to the stretching vibration of -OH, and the peaks of 1635 cm^− 1 and 1266 cm^− 1 corresponded to C = O and C-O bonds, respectively. Although the intensity of C = O peaks decreased slightly at higher temperatures, their persistence confirms that CCDs retained significant bioactive functional groups, which are pivotal for anti-inflammatory and antioxidant activities.

XPS analysis of the C 1 s spectra (Figs. 6L-M) revealed that the sp²-C peak of 284.3 eV in CCDs-210 was significantly enhanced, suggesting higher carbonization [33, 34]. In the O 1 s spectra (Figs. 6N-O), the C = O peak of 532.5 eV increased at higher temperature, which may be attributed to the oxidation of amorphous carbon by residual oxygen, leading to the formation of -OH and -COOH groups [35].

UV-vis was used to analyze the optical properties and electronic transitions of CCDs (Fig. 6P). The curcumin-associated absorption peak at 264 nm (attributed to C = C bonds) weakened with increasing temperature, indicating structural modification. For CCDs-180 and CCDs-210, a new absorption peak emerged at 288 nm, which was attributed to the π-π* transition of conjugated C = C bonds in the sp²-C with a graphite-like structure. Besides, CCDs-210 exhibited a distinct peak at 364 nm, corresponding to the core n→π* transition of C = O groups, attributed to the breakage and recombination of the original conjugated structure and the formation of nanosized carbon cores.

The photoluminescent properties of the CCDs were characterized via fluorescence spectroscopy (Fig. 6Q). The fluorescence emission peak of Cur to CCDs-210 gradually red-shifted from 575 nm to 612 nm, mainly due to the destruction of Cur’s original structure at high temperature, which promoted the formation of a sp²-conjugated system.

CCDs-210 formed a clear, stable aqueous solution in contrast to the turbid Cur suspension (Fig. 6R), confirming that high-temperature carbonization yields smaller, highly dispersible nanoparticles with markedly improved water solubility.

Cellular uptake behavior and mechanism analysis of CCDs

To characterize the cellular uptake profiles of CCDs, we systematically investigated their internalization behavior and underlying mechanisms. CLSM (Fig. 7A) and flow cytometry (Fig. 7B) revealed that at 1 h and 2 h, the uptake intensity of CCDs-210 was significantly lower than that of both free Cur and CCDs-210 (P < 0.05). Notably, no significant difference was observed between free Cur and CCDs-120 (P > 0.05). The internalization of both free Cur and all CCDs exhibited a distinct time-dependent progression (Fig. 7C). Mechanistic investigations demonstrated that the uptake of free Cur and CCDs-120 in HK-2 cells was refractory to low-temperature incubation (4 °C, an inhibitor of energy-dependent transport[36), whereas the internalization of CCDs-210 was significantly impeded under these conditions (P < 0.05). Furthermore, the uptake of both free Cur and CCDs was significantly inhibited by clathrin-mediated endocytosis inhibition (10 µg/mL chlorpromazine), caveolae-mediated endocytosis inhibition (55 µg/mL genistein), and macropinocytosis inhibition (10 µg/mL cytochalasin D) (P < 0.05) (Figs. 7D-E). These findings collectively indicate that CCDs-120 largely retains the passive diffusion behavior of free Cur, attributable to the modest cross-linking induced at 120 °C, whereas the well-defined graphitic carbon core of CCDs-210, formed at 210 °C, necessitates energy-dependent active endocytosis.

Fig. 7.

Cellular uptake characteristics and internalization mechanisms of CCDs in HK-2 cells. (A) Representative confocal laser scanning microscopy images showing the time-dependent intracellular distribution of free Cur and CCDs (green fluorescence) at 1 and 2 h post-incubation. (B-C) Time-dependent uptake analysis via flow cytometry histograms and corresponding mean fluorescence intensity quantification. (D-E) Mechanistic insights into endocytosis: Fluorescence intensity quantification and representative flow cytometry profiles of HK-2 cells treated with free Cur and CCDs at 4 °C or in the presence of specific endocytosis inhibitors. Data are presented as mean ± SD. ** P < 0.01, *** P < 0.001

Protective effects of CCDs against uranium-induced HK-2 cell injury in vitro

The structural transition from “molecule-like” passive diffusion (CCDs-120) to “nanoparticle-like” active endocytosis (CCDs-210) observed in our uptake studies provided the theoretical basis for evaluating their biological performance. Having established that 210 °C carbonization yields nanoparticles with optimized water solubility and distinct internalization kinetics, we next investigated whether these physicochemical advantages translate into superior protection against the HIF-1α/PPARγ-mediated injury identified earlier.

The therapeutic efficacy of Cur and CCDs synthesized at different temperatures was evaluated in vitro using a uranium-induced cell injury model. Following 4 h of exposure to 5 mM uranium, HK-2 cells were treated with Cur or various CCDs to assess cellular recovery. The results demonstrated that 5 µg/mL CCDs-210 exhibited the most potent reparative effect, restoring cell viability to approximately 100%, which was significantly superior to the Cur group (P < 0.05) (Fig. 8A). RT-qPCR analysis revealed that uranium exposure significantly upregulated the mRNA expression of HIF-1α and PPARγ compared to the control group (P < 0.05). Conversely, intervention with 5 µg/mL Cur, CCDs-120, or CCDs-210 significantly attenuated the expression of both factors (P < 0.05) (Fig. 8B). Among the treatment groups, CCDs-210 demonstrated the most pronounced inhibitory capacity, suggesting superior therapeutic potential. Consistent with the transcriptomic data, Western blot analysis confirmed that uranium-induced overexpression of HIF-1α and PPARγ proteins was significantly suppressed by all intervention groups (P < 0.05) (Fig. 8C).

Fig. 8.

Restorative effects of Cur and CCDs on uranium-induced injury in HK-2 cells. (A) Cell viability of HK-2 cells treated with free Cur or CCDs following uranium exposure. (B) RT-qPCR analysis of HIF-1α and PPARγ mRNA levels. (C) Representative Western blot images and densitometric analysis of HIF-1α and PPARγ protein levels. (D) Representative Nile Red staining images and relative quantification of fluorescence intensities. (E) Evaluation of oxidative stress and mitochondrial function: Representative fluorescence images and relative quantification of total ROS (DCFH-DA), superoxide anions (DHE), and mitochondrial membrane potential (ΔΨm, JC-1). Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Nile Red staining confirmed that uranium exposure significantly induced the aberrant accumulation of lipid droplets in HK-2 cells. While both Cur and various CCDs effectively reduced lipid droplet density, the CCD groups exhibited superior efficacy compared to the Cur group (P < 0.05) (Fig. 8D). Subsequently, DCF, DHE, and JC-1 fluorescent probes were used to evaluate reactive oxygen species (ROS) scavenging capacity and mitochondrial integrity following CCD intervention. DCF assay results confirmed that CCDs-210 effectively neutralized uranium-induced oxidative stress by significantly diminishing the green fluorescence signal, confirming its potent ROS-scavenging capability. Consistently, DHE fluorescence analysis revealed that uranium-induced injury significantly augmented superoxide levels, as indicated by a marked increase in fluorescence intensity compared with the control group (P < 0.05). Following treatment with Cur, CCDs-120, or CCDs-210, this intensity was significantly attenuated (P < 0.05), with CCDs-210 exhibiting the most pronounced reduction. Assessment of mitochondrial membrane potential (ΔΨm) via JC-1 staining demonstrated that uranium exposure triggered mitochondrial depolarization, evidenced by the prevalence of JC-1 monomers and a significant increase in green fluorescence. Notably, treatment with CCDs-210 promoted the formation of red-fluorescent JC-1 aggregates and simultaneously weakened the green monomeric signal. This shift indicates a significant reduction in ΔΨm loss and a restoration of mitochondrial functional integrity (Fig. 8E).

In vivo protective effect of CCDs in a mouse model of uranium-induced renal injury

Given that in vitro results confirmed CCDs-210 most effectively inhibited the HIF-1α/PPARγ axis and alleviated oxidative stress, we next evaluated its systemic bioactivity in a murine model of acute uranium poisoning.

Animal experimental results demonstrated that mice subjected to uranium exposure exhibited profound renal dysfunction compared with the control group. Notably, serum Crea and Urea levels were elevated by approximately 21-fold and 7-fold, respectively (P < 0.05) (Fig. 9A). Concurrently, the expression levels of pro-inflammatory cytokines, including TNF-α, IL-6, and IL-1β, were significantly upregulated in renal tissues following uranium exposure (P < 0.05). Conversely, treatment with either Cur or CCDs significantly suppressed the production of these inflammatory mediators, effectively restoring them to near-basal levels (P < 0.05). Among the treatment groups, the CCDs-210 group demonstrated the most potent therapeutic efficacy (Fig. 9A).

Fig. 9.

In vivo protective effects of Cur and CCDs against uranium-induced nephrotoxicity. (A) Assessment of renal function (serum urea and crea), inflammatory cytokines (TNF-α, IL-6, and IL-1β), and oxidative stress markers (SOD, GSH, and MDA). (B) Representative images of renal histopathology (H&E and PAS staining), and ultrastructural morphology via TEM. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Assessment of oxidative stress in renal tissues revealed that uranium exposure compromised the antioxidant defense system. Specifically, SOD activity and GSH levels were significantly diminished (P < 0.05), while lipid peroxidation was exacerbated, as evidenced by a significant increase in MDA levels (P < 0.05). Following Cur or CCDs administration, these parameters were significantly ameliorated (P < 0.05). Notably, the CCDs-210 group exhibited the most pronounced recovery (Fig. 9A, Supplementary Information Table S4).

H&E staining demonstrated partial glomerular atrophy and mild mesangial expansion. Renal tubular epithelial cells exhibited diffuse or focal swelling, characterized by vacuolar degeneration (loose, translucent cytoplasm). PAS staining indicated thickening and rupture of the renal tubular basement membrane (TBM), the formation of PAS-positive casts in the lumen, and a mild accumulation of PAS-positive matrix in the glomerular mesangial area. Administration of Cur and CCDs alleviated these uranium-induced pathological injuries to varying degrees. Notably, the CCD groups demonstrated superior efficacy compared to the Cur group, with CCDs-210 exhibiting significantly greater improvement than CCDs-120 (Fig. 9B). Furthermore, TEM analysis confirmed that the uranium-exposed group exhibited typical ultrastructural damage, including mitochondrial swelling, cristae disruption or loss, and vacuolization. Following treatment with Cur or CCDs, mitochondrial structural integrity was significantly ameliorated, most notably in the CCDs-210 group (Fig. 9B).

Analysis of HIF-1α and PPARγ mRNA and protein expression in murine kidneys revealed significant upregulation following uranium exposure (P < 0.05), whereas treatments significantly downregulated these markers (P < 0.05). CCDs-210 restored HIF-1α mRNA expression to levels comparable to the control group and exerted the most potent inhibitory effect on PPARγ (Figs. 10A-B). Consistent with these findings, immunofluorescence assays demonstrated that the fluorescence intensity of HIF-1α and PPARγ was significantly higher in the uranium-exposed group compared to the blank control (P < 0.05). Following treatment, the intensities were significantly attenuated (P < 0.05). Notably, the fluorescence intensity in the CCDs-210 group was comparable to that of the control group, confirming its efficacy in alleviating uranium-induced renal injury (Fig. 10C). Multiplex immunofluorescence co-staining with the proximal tubule marker AQP1 and the podocyte marker Podocin further localized the uranium-induced upregulation of HIF-1α and PPARγ predominantly to the proximal tubular epithelial cells (Fig. 10D). Importantly, CCDs-210 exhibited the most robust suppression of these markers specifically within AQP1-positive tubules. This cell-type-specific inhibition highlights the potent regulatory efficacy of CCDs-210 on the HIF-1α/PPARγ axis within its primary pathological niche, further underscoring its therapeutic potential as a targeted intervention for renal tubular injury.

Fig. 10.

CCDs-210 targets the HIF-1α/PPARγ axis within the proximal tubular pathological niche. (A) RT-qPCR and (B) Western blot analysis showing the mRNA and protein expression of HIF-1α and PPARγ in murine kidneys across treatment groups. (C) Representative immunofluorescence images of HIF-1α and PPARγ in renal tissues. (D) Multiplex immunofluorescence co-staining of HIF-1α and PPARγ with AQP1 (proximal tubule marker) and Podocin (podocyte marker), demonstrating the cell-type-specific inhibitory efficacy of CCDs-210 within the proximal tubular niche. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

Moreover, our findings revealed that CCDs conferred broad multi-organ protection (Fig. 11). Within the myocardium, CCDs attenuated cardiomyocyte edema, inflammatory infiltration, and myofiber disarray. Hepatic analysis revealed that CCDs normalized hepatocyte morphology while significantly reducing necrosis and portal inflammation. Furthermore, CCDs treatment restored splenic tissue architecture and red/white pulp demarcation. Pulmonary evaluations indicated an alleviation of alveolar epithelial damage and interstitial inflammation. Across all assessed organs, CCDs demonstrated superior efficacy compared to free Cur, underscoring their therapeutic potential for the systemic management of uranium-induced multi-organ toxicity.

Fig. 11.

Systemic biosafety and ameliorative effects of CCDs-210 across major extra-renal organs. Representative H&E staining images of the heart, liver, spleen, and lungs from each group

CCDs-210 directly targets the HIF-1α/PPARγ axis to mediate its protective effects

To determine whether CCDs-210 modulates the HIF-1α/PPARγ axis through direct molecular engagement, we performed structural, biophysical, and functional validation.

Molecular docking simulations revealed that Cur forms a stable complex with HIF-1α protein, characterized by a steady-state RMSD of ~1.05 nm and a high binding affinity (Fig. 12A). During pyrolysis at 210 °C, CCDs-210 partially retains the core active functional groups of Cur. This provides a potential structural basis for the physical interaction between CCDs-210 and the target protein, thereby supporting its specific regulation of the signaling axis.

Fig. 12.

Mechanistic verification of the HIF-1α-dependent protective effects of CCDs-210. (A) Molecular docking simulation showing the binding affinity and stable complex formation between Cur and the HIF-1α protein. (B) Isothermal titration calorimetry (ITC) analysis of the interaction between CCDs-210 and HIF-1α protein. (C) Cellular thermal shift assay (CETSA) in HK-2 cells demonstrating the ligand-induced thermal stabilization of HIF-1α by CCDs-210. (D) Cell viability analysis and (E, F) fluorescence quantification of ROS (DCF) and superoxide (DHE), revealing that the cytoprotective efficacy of CCDs-210 is functionally dependent on HIF-1α inhibition rather than solely on ROS scavenging. (G) Schematic diagram of the pleiotropic mechanism of CCDs-210 in mitigating uranium-induced renal injury. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001

To provide definitive evidence of direct engagement, Isothermal Titration Calorimetry (ITC) and Cellular Thermal Shift Assay (CETSA) were utilized. The in vitro interaction between CCDs-210 and recombinant HIF-1α protein exhibited a characteristic exothermic heat-release curve (Figs. 12B-C). Quantitative analysis yielded a dissociation constant K_D = 75.6 µM, with enthalpy and entropy changes of ΔH = −34.15 kJ/mol and ΔS = −35.69 J/mol·K, respectively. These thermodynamic parameters indicated an enthalpy-driven, non-covalent interaction (predominantly van der Waals forces and/or hydrogen bonding), confirming the direct molecular recognition capability of CCDs-210. To verify this binding within the complex intracellular milieu, CETSA was performed using HK-2 cells. CCDs-210 treatment significantly enhanced the thermal stability of the HIF-1α protein, as evidenced by a distinct rightward shift in the thermal melting curve. This result directly confirms that CCDs-210 physically interacts with HIF-1α in a live-cell context.

Finally, rescue experiments were conducted using DMOG (a HIF-1α stabilizer) to distinguish between targeted pathway intervention and non-specific stress mitigation. Although CCDs-210 retained a degree of broad-spectrum antioxidant capacity, its ability to restore cell viability was markedly attenuated upon artificial sustainment of HIF-1α activity by DMOG (Fig. 12D). Given that non-specific ROS scavenging alone accounted for the protective effects of CCDs-210, the recovery of cell viability would have remained largely unaffected (Figs. 12E-F). Conversely, the pronounced reduction in protection following HIF-1α stabilization indicated that the therapeutic efficacy of CCDs-210 was functionally contingent upon direct inhibition of the HIF-1α/PPARγ signaling axis (Fig. 12G).

Metabolomic insights into the protective mechanism of CCDs against uranium-induced renal injury

To gain system-level insights into CCDs-210-mediated renoprotection, we performed comprehensive metabolomic profiling of renal tissues. Volcano plots showed that uranium was markedly elevated in most renal metabolites; all CCDs and Cur treatments reversed this trend toward a normal metabolic profile (Fig. 13A). Heatmap analysis revealed that CCDs-120 and CCDs-210 significantly downregulated multiple lipid classes, including acylcarnitines, free fatty acids, lysoPCs, oxidized lipids, PCs, and sphingosines, relative to the U group (Fig. 13B). CCDs-210 exerted the most comprehensive restorative effect across all uranium-perturbed metabolites (Fig. 13C). KEGG pathway enrichment analysis identified sphingolipid signaling, sphingolipid metabolism, and glycerophospholipid metabolism as the primary hubs modulated by CCDs-210 (Fig. 13D). From a mechanistic standpoint, CCDs-210 exerts its effect on lipid homeostasis by regulating the downstream enzymatic network linked to this signaling axis. By suppressing uranium-induced activation of HIF-1α/PPARγ, CCDs-210 effectively modulates the expression of FASN and DGAT2. This enzymatic inhibition restricts the expansion of the fatty acid and diacylglycerol pools, thereby limiting the substrate supply required for pathological remodeling of glycerophospholipid metabolism. Furthermore, although direct enzymatic validation of the sphingolipid pathway was not performed, the marked reduction in sphingolipid accumulation following CCDs-210 treatment, coupled with the known substantial crosstalk between sphingolipids and fatty acid/glycerophospholipid metabolism, strongly suggests that CCD-210 mitigates renal sphingolipid dysregulation by disrupting the upstream HIF-1α/PPARγ signaling axis. In summary, these findings substantiate the role of CCDs-210 as a potent metabolic modulator that mitigates uranium-induced renal injury by precisely blocking the HIF-1α/PPARγ-dependent lipid reprogramming cascade.

Fig. 13.

CCDs-210 modulate the renal metabolic profile in mice with uranium-induced kidney injury. (A) Volcano plots displaying differentially abundant metabolites in renal tissues across groups. Orange and green dots represent significantly upregulated (P < 0.05, FC > 2) and downregulated (P < 0.05, FC < 0.5) metabolites. (B) Heatmap of differentially abundant metabolites. The color scale ranges from green (downregulation) to red (upregulation). (C) Integrated volcano plots highlighting metabolic shifts across multiple comparisons: U vs. NC (part1); U + Cur vs. U (part2); U+CCDs-120 vs. U (part3); U+CCDs-210 vs. U (part4). (D) Bubble plots of KEGG pathway enrichment analysis. The color gradient represents the P-value (redder colors indicate smaller P-values), and the bubble size represents the number of metabolites enriched in each pathway

In addition to these core pathways, alpha-linolenic, linoleic, and arachidonic acid metabolism were significantly implicated in the therapeutic response, reflecting a broad-spectrum anti-inflammatory and pro-resolving lipidomic shift. Collectively, these findings substantiate that CCDs-210 alleviates renal injury by blocking the HIF-1α/PPARγ axis, which in turn suppresses a cascade of lipid-mediated injuries.

Discussion

This work provides hitherto undocumented evidence that uranium exposure triggers renal injury via HIF-1α/PPARγ axis activation, which in turn mediates lipid reprogramming in the kidney. It further demonstrates that CCDs intervene in this pathogenic cascade, thereby mitigating the resulting damage. These findings establish a novel conceptual foundation and point to CCDs as a potential therapeutic avenue for uranium-associated nephropathy.

Importantly, the HIF-1α/PPARγ axis was identified as a central orchestrator of uranium-induced nephrotoxicity. While our initial transcriptomic profiling noted the synchronous upregulation of HIF-1α and PPARγ (Supplementary Information Figure S4), subsequent investigations revealed a sophisticated dual-level regulatory mechanism: HIF-1α acts as not only a direct transcriptional activator of the PPARG promoter (evidenced by ChIP and luciferase assays), but also a physical binding partner of the PPARγ protein (confirmed by Co-IP). This dual-track regulation ensures robust, synergistic activation of the axis under uranium-induced stress, transcending mere correlation to establish a definitive biochemical link. To determine whether the resulting lipid dysregulation is a primary pathogenic driver, we established a rigorous causal chain through multi-layered validation. At the molecular level, uranium triggered a “synthesis-enhanced, oxidation-impaired” enzymatic shift, characterized by the upregulation of lipogenic pivots (ACC1, FASN, DGAT2, LPL) and the suppression of the rate-limiting enzyme for fatty acid oxidation, CPT-1 A. Functional validation of this imbalance was achieved through a rescue-reversal strategy: PPARγ inhibition by GW9662 effectively bypassed uranium-induced damage, whereas the agonist Rosi worsened both lipid accumulation and cytotoxicity. These findings, corroborated by the consistent protection observed in HK-2^shHIF−1α and HIF-1α^+/− mice, consolidate the role of HIF-1α/PPARγ-driven lipid reprogramming as a bona fide cause of renal injury. Meanwhile, the accumulation of lipotoxic intermediates likely fuels a vicious cycle of oxidative stress and mitochondrial dysfunction, thereby further amplifying the primary insult [37, 38] Crucially, this study expands the functional repertoire of HIF-1α beyond its canonical role in hypoxia. By uncovering this novel HIF-1α/PPARγ-mediated lipid reprogramming mechanism, we provide a new paradigm for understanding the molecular pathogenesis of heavy metal-induced nephrotoxicity and identify potential therapeutic targets to mitigate uranium-induced damage.

As a natural polyphenolic compound, Cur shows significant therapeutic potential in managing renal diseases [18, 39]. It mitigates histopathological damage across various models, including ischemia-reperfusion injury and diabetic nephropathy, through pleiotropic mechanisms. These include modulating oxidative stress pathways (e.g., activating the Nrf2/HO-1 system), suppressing pro-inflammatory cytokines (e.g., TNF-α and IL-6), and enhancing mitochondrial function[40, 41]. However, the presence of multiple phenolic hydroxyl groups results in poor aqueous solubility, severely limiting bioavailability and clinical translation [23]. Biomass-derived CDs not only exhibit excellent biocompatibility and low toxicity but also retain the intrinsic pharmacological activities of their precursor molecules, demonstrating substantial potential for potential therapeutic candidate [42, 43].

In this study, CCDs were successfully synthesized via tube furnace pyrolysis. Compared with the conventional hydrothermal approach, this pyrolysis method offers several distinct advantages. First, precise temperature control within the furnace minimizes particle size deviations typically caused by localized overheating. Second, the implementation of an inert gas atmosphere prevents the oxidative degradation of Cur[44]. Furthermore, the reaction duration (1–4 h) is significantly shorter than that of hydrothermal synthesis (4–12 h). The high batch-processing capacity of the tube furnace, ranging from the gram to kilogram scale, further underscores its suitability for industrial-scale production[43]. CCK-8 and hemolysis assays confirmed that the synthesized CCDs exhibited negligible cytotoxicity and superior biocompatibility (Supplementary Information Figure S5).

Moreover, our study identified the carbonization temperature as a critical determinant of the cellular uptake pathways and kinetics of CCDs. Specifically, varying the carbonization temperature modulates the physicochemical properties of the materials, facilitating a fundamental transition from “molecule-like” passive diffusion to “nanoparticle-like” active endocytosis [45]. The CCDs-120 could largely retain the rapid transmembrane capabilities characteristic of free Cur. We hypothesized that the relatively mild thermal treatment at 120 °C primarily induced dehydration or moderate cross-linking, resulting in molecular aggregates whose structures remained dominated by Cur derivatives[46]. This structural retention allowed the materials to maintain the lipophilicity of free Cur, thereby enabling passive diffusion across the lipid bilayer. In contrast, high-temperature treatment at 210 °C triggered a more pronounced carbonization process, yielding carbon dots with a well-defined nanocarbon core and significantly altered surface characteristics, including modified hydrophilicity and surface charge [47, 48]. As discrete nanoparticles, they are internalized via endocytic pathways, which are often slower and capacity-limited, accounting for their lower initial uptake rate compared with the 120 °C variants [49, 50]. Collectively, these findings provide a rational framework for designing carbon dots with programmable cellular delivery behaviors. This study systematically elucidates the potent protective mechanisms of CCDs-210 against uranium-induced renal injury. In HK-2 cells, 5 µg/mL CCDs-210 outperformed Cur by fully restoring cell viability, likely due to its enhanced surface activity and biocompatibility derived from 210 °C carbonization. Mechanistically, DCF, DHE, and JC-1 assays confirmed that CCDs-210 possesses exceptional ROS-scavenging capacity and stabilizes mitochondrial membrane potential, thereby averting apoptosis and maintaining energy homeostasis [51–53]. A key finding of our study is the identification of the HIF-1α/PPARγ axis as a novel molecular target of CCDs-210. We clarified a multifaceted regulatory model by which CCDs-210 modulated HIF-1α. Biophysical evidence from ITC and CETSA assays confirmed a direct physical interaction (target engagement) between CCDs-210 and HIF-1α. This interaction likely imposed steric hindrance, interfering with its transcriptional activity or the recruitment of essential co-activators. Beyond direct binding, CCDs-210 facilitated HIF-1α degradation by disrupting the “oxidative stress-HIF-1α stability” feedback loop. By scavenging excess ROS and restoring mitochondrial function, CCDs-210 may reactivate prolyl hydroxylases (PHDs), thereby promoting the ubiquitination-mediated degradation of HIF-1α ^{54, 55}. This dual-action model, combining direct protein inhibition with indirect environmental restoration, effectively suppressed the pathological activation of the HIF-1α/PPARγ axis. Consequently, lipid homeostasis was restored by inhibiting PPARγ overexpression and reducing lipotoxic lipid droplet accumulation, as visualized by Nile Red staining. DMOG rescue experiments further confirmed that this protective effect relies on the HIF-1α pathway; artificially stabilizing HIF-1α markedly weakened the protective action of CCDs-210. Collectively, the renoprotective effect of CCDs-210 was the synergistic result of specific targeting of the HIF-1α/PPARγ axis and broad-spectrum mitochondrial support, a pleiotropic advantage characteristic of curcumin-derived nanomaterials, offering a promising potential therapeutic candidate for heavy metal-induced nephrotoxicity.

The in vivo efficacy of CCDs-210 was validated in a murine model, where it significantly reduced serum Urea and Crea while attenuating inflammatory infiltration and cytokine expression (TNF-α, IL-6, and IL-1β). By potentiating the endogenous antioxidant defense system (increasing SOD/GSH and decreasing MDA), CCDs-210 countered systemic oxidative damage. These molecular improvements culminated in the marked recovery of renal histopathology. Besides, the broad-spectrum protective effects observed in other vital organs (heart, liver, spleen, and lung) underscore the significant potential of CCDs-210 for clinical translation in treating systemic uranium poisoning. Notably, the efficacy of CCDs varied significantly with synthesis temperature; CCDs-210 consistently outperformed CCDs-120 across all evaluation metrics. This suggests that the carbonization temperature is a decisive factor in determining biological activity, likely by modulating the size, crystallinity, and surface physicochemical properties (e.g., functional group type and density) of the carbon dots [42, 43].

To elucidate the metabolic mechanisms underlying the renoprotective effects of CCDs, this study characterized the renal metabolic profiles following uranium exposure. Uranium triggered a widespread accumulation of diverse lipid metabolites, indicating profound metabolic stress and mitochondrial dysfunction [56]. Notably, CCDs-210 most effectively reversed these alterations, underscoring its pivotal role in restoring energy homeostasis. Mechanistically, the dysregulation of sphingolipids and glycerophospholipids-essential structural components of biological membranes-compromised the integrity of renal tubular epithelial cells, thereby exacerbating oxidative stress and inflammatory cascades [57]. Furthermore, aberrations in the metabolism of α-linolenic acid, linoleic acid, and arachidonic acid were closely linked to lipid peroxidation and the release of pro-inflammatory mediators [58]. KEGG enrichment analysis identified these lipid pathways as primary targets of CCDs-210, suggesting that it preserves membrane stability and mitigates peroxidative product accumulation through targeted metabolic modulation. These metabolic shifts, in coordination with the HIF-1α/PPARγ axis and inherent antioxidant activities, orchestrate a comprehensive defense against uranium-induced nephrotoxicity. Future research will involve targeted validation across a larger cohort to further elucidate the metabolic dynamics of specific biomarkers, providing a more robust statistical basis for the therapeutic prospects of CCDs-210 against uranium-induced nephrotoxicity.

Conclusion

This study confirms that HIF-1α/PPARγ-mediated lipid reprogramming is a key driver of uranium-induced nephrotoxicity, clarifying that uranium exposure induces renal injury by activating this signaling axis. To intervene in this pathological process, a series of CCDs were synthesized, and those prepared under higher carbonization conditions exhibited superior physicochemical properties and significant renoprotective effects. The renoprotective effect of CCDs is mainly achieved by directly targeting the HIF-1α/PPARγ signaling axis coupled with their intrinsic antioxidant properties, thereby improving uranium-induced lipid metabolic disorders, oxidative stress, inflammatory responses and mitochondrial dysfunction. Both in vitro and in vivo studies validated these protective effects. CCDs additionally preserve other major organs, suggesting wider therapeutic reach against multi-organ uranium toxicity. Yet, several drawbacks exist. Most notably, despite CCDs-210’s efficacy in renal injury, one must temper enthusiasm for its bedside application. Current research has primarily focused on early-stage mechanistic exploration and efficacy validation. Future investigations should prioritize systematic pharmacokinetic assessments, long-term safety evaluations, and the definitive confirmation of target engagement. The accumulation of these critical data will establish a solid foundation for developing CCDs-210 as a promising nanomedicine candidate for the prevention and treatment of uranium-induced nephrotoxicity. Second, the potential universal protective effects of CCDs against other heavy metal-induced toxicities remain to be explored. Future research should address these challenges and further optimize the CCDs synthesis process to facilitate its clinical translation.

Author contributions

Zhimin Jia: Investigation, Research, Drafting the original manuscript; Chang Wang: Research methods, Data management; Xiaowen Han: Development of concepts, Formal analysis; Jiawei Zeng: Investigation, Data curation; Chuandong He: Data curation, Methodology; Zhengguo Chen and Yan Chen: Conceptualization, Supervision, Writing review & editing; Bei Xu: Conceptualization, Supervision, Resources, Overseeing the project, Acquiring funding, Administering the project.

Funding

This study was funded by the Sichuan Medical Association Youth Innovation Project (Grant No. Q20250002), Sichuan Province Medical Research Project Program (Grant No. S23067), NHC Key Laboratory of Nuclear Technology Medical Transformation (MIANYANG CENTRAL HOSPITAL) (Grant No. 2022HYX007 and Grant No. 2024HYX021), Incubation Project of Mianyang Central Hospital (Grant No. 2024FH002 and Grant No. 2025FH002), Health Commission of Sichuan Province Medical Science and Technology Program (Grant No. 25QNMP055 and Grant No. 24QNMP035), Traditional Chinese Medicine Research Program of Sichuan Administration of Traditional Chinese Medicine (Grant No. 2024MS584).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experiments were conducted in strict accordance with Chinese national standards, and all followed the experimental protocols approved by the Ethics Committee of Laboratory Animals, Hubei Bainte Biotechnology Co., Ltd. (BNB-2024-003).

Consent for publication

All authors have approved the manuscript and agree for the submission.

Competing interests

The authors declare no competing interests.