Mitochondrial metabolic reprogramming drives diabetic kidney disease progression: cell-specific mechanisms, metabolic memory, and targeted strategies Authors: Lingfei Zhang, Yue Zhang, Liu Han, Jialei Wang, Lijie Hou, Songbo Fu Published: 2026 ABSTRACT AbstractDiabetic Kidney Disease (DKD), as a major microvascular complication of diabetes, is the leading cause of chronic kidney disease and end-stage kidney disease. Metabolic reprogramming is an important concept in cancer research, and the most classic example is the Warburg effect. This phenomenon has also been observed in DKD, suggesting that metabolic reprogramming also plays a significant role in DKD. Mitochondria are crucial in both catabolism and anabolism, making them the center of metabolic reprogramming research. Additionally, mitochondria are involved in methylation and acetylation modifications, potentially playing a vital role in “metabolic memory.” Starting from the metabolic landscape of mitochondria in cells, this review discusses mitochondrial metabolic reprogramming in different types of renal and immune cells, providing insights into the mechanisms underlying “metabolic memory.” In addition, we comprehensively summarize therapeutic strategies targeting mitochondrial metabolic reprogramming from both preclinical and clinical perspectives. Unlike previous studies that primarily focused on individual pathophysiological processes, this review integrates the intrinsic link between mitochondrial metabolic reprogramming and epigenetic modifications associated with “metabolic memory,” thereby offering a novel interdisciplinary perspective for overcoming current therapeutic bottlenecks in DKD.Keywords: Diabetic kidney disease, Metabolic reprogramming, Mitochondria, Metabolic memory Introduction With the prevalence of high-calorie diets and sedentary lifestyles, the global burden of diabetes mellitus (DM) is increasingly heavy. Based on the prediction of the International Diabetes Federation, the global number of people with DM will reach 853 million in 2050 (Duncan et al. 2025). As the most severe microvascular complication of diabetes, diabetic kidney disease (DKD) has a prevalence rate as high as 40% in type 2 DM patients and 30% in type 1 DM patients (Gupta et al. 2023), and has become the leading cause of global chronic kidney disease and end-stage kidney disease (ESKD) (Hoogeveen 2022). At present, renin-angiotensin system (RAS) blockers, sodium-glucose cotransporter 2 (SGLT2) inhibitors, glucagon-like peptide-1 receptor agonists (GLP-1RAs), and non-steroidal mineralocorticoid receptor antagonists (MRAs) are widely used in clinical practice to delay the progression of DKD (Martinez Leon et al. 2025). However, existing therapies still fail to completely halt its ultimate deterioration trajectory towards ESKD. Therefore, deeply exploring the potential pathogenic mechanisms and intervention targets of DKD has become an urgent need to break through the current treatment bottleneck. Metabolic reprogramming was first systematically defined in the field of oncology. Its landmark feature “Warburg effect” describes the metabolic preference of cells to rely primarily on glycolysis for energy even in aerobic conditions (Fendt 2024). This phenomenon not only meets the energy demand of rapid proliferation of cancer cells (Du et al. 2025), but also has been widely observed in DKD tissues in recent years (Zhang et al. 2024; Kanasaki 2023; Lu et al. 2025), suggesting that metabolic network remodeling is also a deep mechanism driving DKD pathogenesis (Li et al. 2023). The key to dissecting metabolic reprogramming lies in focusing on mitochondria. As the core hub of cellular metabolism, mitochondria not only efficiently produce ATP through the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS), but also play a pivotal role in the synthesis and homeostasis regulation of macromolecules such as fatty acids, amino acids, heme, and iron-sulfur clusters (Nunnari and Suomalainen 2012). What is more groundbreaking is the direct involvement of mitochondrial metabolites in epigenetic modifications in the nucleus, providing important mechanistic support for their role in the “metabolic memory” effect in DKD (Sarwar et al. 2023). Clinical observations have revealed that even after intensive hypoglycemic therapy, some DM patients continue to experience progressive DKD lesions (Mimura et al. 2025). The academic community summarizes this phenomenon as “metabolic memory” (Natarajan 2021), and hypothesizes that its underlying logic is that the dramatic fluctuations of metabolites induced by early hyperglycemia leave an epigenetic “imprint” on the cellular genome, and continue to transmit during cell division (VanderJagt et al. 2015). Therefore, further research on the pathogenesis of DKD from the perspective of mitochondrial metabolic reprogramming holds the potential to become a key breakthrough in overcoming its treatment dilemma. In this review, we first outline the metabolic landscape of mitochondria within cells, then systematically summarize mitochondrial metabolic reprogramming in different types of renal and immune cells, and further discuss its role in “metabolic memory.” Finally, we synthesize pharmacological agents targeting mitochondrial metabolic reprogramming from both preclinical and clinical perspectives. Innovatively, this review establishes a logical chain encompassing “mitochondrial metabolism-epigenetic modifications-metabolic memory,” which overcomes the limitations of conventional studies that discuss these mechanisms in isolation. This integration offers novel insights into the pathogenesis and targeted therapy of DKD. Metabolic landscape of mitochondria within cells Mitochondrial metabolism consists of two major aspects: mitochondrial energy metabolism and mitochondrial quality control. The latter serves as a guarantee for the former, and any disorder in the former can also affect the latter (Fan et al. 2024; Zhang et al. 2023). Mitochondrial energy metabolism and quality control are integrated through mitochondrial structure. Within mitochondria, different compartments and components are responsible for regulating energy metabolism and quality control. A structural perspective of mitochondria can help to better delineate the overall relationship between the two, thereby facilitating a more comprehensive understanding of the cellular metabolic landscape of mitochondria (Fig. 1). The metabolic landscape of mitochondria in cells Mitochondrial structure Mitochondria are composed of the outer membrane, inner membrane, intermembrane space and matrix. The outer mitochondrial membrane (OMM) has relatively high permeability, and it mediates the entry and exit of metabolic substrates (such as pyruvate, fatty acids, etc.) and ions into and out of the mitochondria (Chen et al. 2023). OMM proteins Dynamin-related protein 1 (DRP1) and Mitofusin 1/2 (Mfn1/2) are closely related to mitochondrial dynamics (Tábara et al. 2025). The OMM also contains autophagy receptors such as BCL2 (B-cell lymphoma 2) interacting protein 3 (BNIP3), which mediate the mitochondrial non-ubiquitin autophagy pathway (Li et al. 2024). The inner mitochondrial membrane (IMM) is highly folded into cristae structures, significantly increasing its surface area and providing more space for the components of the electron transport chain (ETC) located on it (Stephan et al. 2020). The strict impermeability of IMM to ions and small molecules is crucial for maintaining the proton gradient required for ATP synthesis (Prinz et al. 2020). IMM fusion is mainly mediated by the IMM protein optical atrophy 1 protein (OPA1) (Tábara et al. 2025). Meanwhile, the generation of reactive oxygen species (ROS) is also one of the important functions of IMM (2023). The mitochondrial intermembrane space located between the two membranes serves as a storage reservoir for proton pumps during the electron transfer process. This proton gradient is crucial for ATP synthesis (Goyani et al. 2024). The space enclosed by IMM is called the mitochondrial matrix, which has highly active biochemical activity. It contains a series of enzymes that drive processes such as the TCA cycle and fatty acid β-oxidation, and is responsible for the oxidative breakdown of substrates and the production of NADH and FADH2 (Eniafe and Jiang 2021). The mitochondrial matrix also harbors various enzymes that scavenge ROS, including superoxide dismutase (SOD) and glutathione peroxidase (Napolitano et al. 2021). In addition, protein channels on IMM and OMM also play a role in mitochondrial biogenesis (Emma et al. 2016). Mitochondrial energy metabolism In terms of energy metabolism, mitochondria are the main productivity organelles, which are mainly responsible for catabolism, but also affect anabolism.Some catabolites will be transported out of mitochondria and participate in cytoplasmic anabolism, such as lipid synthesis. In addition, mitochondria are involved in methylation modification and acetylation modification, which are also closely related to mitochondrial energy metabolism (Du et al. 2025). Cytosolic glucose is metabolized via glycolysis to pyruvate, which can either be converted into lactate by lactate dehydrogenase (LDH) or transported into mitochondria via the mitochondrial pyruvate carrier (MPC) to generate acetyl-CoA, which subsequently enters the TCA cycle (Liang et al. 2025). Mitochondria are also the central hub for fatty acid oxidation (FAO). Long-chain fatty acids are transported into mitochondria via the carnitine palmitoyltransferase (CPT) complex and undergo β-oxidation to produce acetyl-CoA. This acetyl-CoA pool not only fuels the TCA cycle for energy production but also serves as a substrate for acetylation modifications (Du et al. 2025). Additionally, cytosolic amino acids can be transported into mitochondria to participate in the TCA cycle through transamination and decarboxylation reactions. Serine metabolism provides one-carbon units such as CH₂-THF to one-carbon metabolism, thereby contributing to DNA, RNA, and histone methylation (2022). The catabolism of glucose, lipids, and amino acids ultimately converges in the TCA cycle, which subsequently drives robust ATP synthesis through OXPHOS (Bartman et al. 2023). In DKD, a hyperglycemic environment promotes increased renal glucose uptake and significantly enhances glycolytic flux (Fan et al. 2024). Studies have shown that mRNA expression of key glycolytic enzymes, including hexokinase 1 (HK1), phosphofructokinase (PFK), and pyruvate kinase (PK), is significantly upregulated in the renal cortex of diabetic mice. In addition, metabolomic analyses have revealed that intermediates of glycolysis (hexose-6-phosphate and pyruvate) and the TCA cycle (fumarate and malate) are both elevated (Sas et al. 2016). A plausible explanation is that DKD is associated with a hypoxic microenvironment, which drives reliance on anaerobic glycolysis for rapid ATP production to meet the high metabolic demands of the kidney (Kanasaki 2023). Patients with DKD also exhibit dyslipidemia characterized by elevated cholesterol and triglyceride levels, as well as impaired lipoprotein metabolism (Mitrofanova et al. 2021). Lipid catabolism is also suppressed in DKD, which may result from an imbalance between increased lipid uptake and reduced lipid oxidation and efflux. Studies have demonstrated that transporters and receptors mediating fatty acid and cholesterol uptake (e.g., CD36, SR-A1, LOX-1, and LDLR) are overexpressed, whereas enzymes involved in FAO (e.g., ACO1, CPT1) and cholesterol efflux transporter ATP-binding cassette A1 (ABCA1) are downregulated, leading to significant lipid accumulation in renal tissues of DKD patients (Herman-Edelstein et al. 2014). Mitochondrial quality control Mitochondrial quality control primarily includes mitochondrial oxidative/antioxidant balance, mitochondrial biogenesis, mitochondrial dynamics, and mitophagy (Zhang et al. 2023). In DKD, impaired mitochondrial quality control leads to reductions in mitochondrial membrane potential (MMP), ETC complex activity, and oxygen consumption rate, ultimately resulting in defective OXPHOS and decreased ATP production (Sung et al. 2023; Liu et al. 2024). Mitochondrial oxidative/antioxidant balance Mitochondria are the principal source of ROS, with ETC complexes Ⅰ and Ⅲ serving as the primary sites of mitochondrial ROS (mtROS) generation (Zorov et al. 2014). Under physiological conditions, antioxidant enzymes within the mitochondrial matrix efficiently scavenge mtROS, maintaining them at low levels (Napolitano et al. 2021; Kowalczyk et al. 2021). When mtROS levels increase to a moderate extent without causing cellular dysfunction, they induce retrograde signaling that activates the expression of mitochondrial antioxidant proteins (Kops et al. 2002). This adaptive process is primarily mediated by the binding of transcription factors, such as nuclear factor erythroid 2-related factor 2 (Nrf2), to antioxidant response elements (AREs). Under unstressed conditions, Nrf2 is mainly degraded via the Kelch-like ECH-associated protein 1 (Keap1)-mediated proteasomal pathway. Upon mitochondrial oxidative stress, the affinity between Keap1 and Nrf2 diminishes, which impedes Nrf2 degradation. Stabilized Nrf2 subsequently translocates into the nucleus, where it binds to AREs and upregulates the expression of a series of antioxidant-related genes (Bryan et al. 2013). However, under pathological conditions, excessive accumulation of mtROS triggers severe mitochondrial oxidative stress, ultimately impairing normal mitochondrial metabolism (Zhao et al. 2021). In DKD, RAS promotes renal tubular injury by driving proximal tubular mitochondrial oxidative stress via NADPH oxidase 4 activation, as well as inducing distal tubular cellular senescence through oxidative DNA damage (Yang et al. 2026). Mitochondrial biogenesis To counteract oxidative stress-induced damage or replenish aged mitochondria, cells initiate mitochondrial biogenesis (Harrington et al. 2023). The vast majority of mitochondrial structural and functional proteins are encoded by nuclear DNA, transcribed in the nucleus, translated in the cytoplasm, and subsequently imported into mitochondria (Emma et al. 2016). Consequently, mitochondrial biogenesis is highly dependent on bidirectional communication between the nucleus and mitochondria. On the one hand, the nucleus regulates mitochondrial activity through anterograde signaling to facilitate adaptation to environmental changes. On the other hand, mitochondria transmit retrograde signals that modulate nuclear gene expression, thereby reshaping cellular energy metabolism (Quirós et al. 2016). This finely tuned regulatory network is mediated by nuclear transcription factors (NRFs, PPARs, ERRs, etc.), transcriptional co-regulators (PGC-1, RIP140, etc.), and upstream energy sensors (AMPK, SIRT, etc.) (Rackham et al. 2025; Malik et al. 2023; Abu Shelbayeh et al. 2023). In DKD, the secreted protein progranulin (PGRN) mediates mitochondrial biogenesis via the PGRN-Sirtuin 1 (SIRT1)-Peroxisome proliferator activated receptor gamma coactivator 1 alpha (PGC-1α)/Forkhead box protein O1 (FoxO1) signaling pathway to maintain mitochondrial homeostasis. However, PGRN levels are significantly diminished in the kidneys of both streptozotocin (STZ)-induced diabetic mice and patients with DKD, thereby exacerbating mitochondrial injury (Zhou et al. 2019). Mitochondrial dynamics Mitochondria continuously undergo fission and fusion to dynamically regulate their structure and maintain normal function. DRP1 regulates mitochondrial fission, whereas MFN1/2 and OPA1 mediate mitochondrial fusion (Tábara et al. 2025). Mitochondrial fission serves as a prerequisite for autophagic degradation (Burman et al. 2017). In DKD, a hyperglycemic environment triggers cyclase-associated protein 1 (CAP1)-induced actin depolymerization, which subsequently drives the translocation and enrichment of inverted formin 2 (INF2) from the endoplasmic reticulum (ER) to the mitochondria-associated ER membrane (MAM). At the MAM interface, enhanced protein–protein interactions between CAP1 and enriched INF2 further facilitate DRP1-mediated mitochondrial fission (Yao et al. 2026). Under hyperglycemic conditions, CD4 + T cells release small extracellular vesicles (EVs), leading to dysregulation of mitochondrial dynamics in renal tubular epithelial cells, characterized by decreased membrane potential and an imbalance in DRP1/MFN1/MFN2 expression (Han et al. 2026). IL-15 activates IL-15Rα in podocytes and induces histone H3 acetylation at the OPA1 promoter, thereby upregulating OPA1 expression, promoting mitochondrial fusion, and reversing hyperglycemia-induced mitochondrial fragmentation. Conversely, plasma IL-15 levels are significantly reduced in patients with DKD and are negatively correlated with the urine albumin-to-creatinine ratio (Li et al. 2026). Mitophagy Severe mitochondrial damage that is beyond repair triggers mitophagy, a selective autophagic process that removes damaged mitochondrial components or entire mitochondria to maintain mitochondrial homeostasis (Song et al. 2021). Currently, the major mechanisms of mitophagy include ubiquitin-dependent and ubiquitin-independent pathways. The ubiquitin-dependent pathway is primarily mediated by PTEN-induced kinase 1 (PINK1) and Parkin (Narendra and Youle 2024), whereas the ubiquitin-independent pathway is regulated by mitophagy receptors located on the OMM, such as BNIP3 and FUN14 domain-containing protein 1 (FUNDC1), whose LC3-interacting regions directly bind autophagy-related proteins to initiate mitophagy (Li et al. 2024). In DKD, the expression of the nuclear receptor estrogen-related receptor alpha (ESRRA) is markedly downregulated, which transcriptionally represses autophagy-related 5 (ATG5) and thereby blocks PINK1/Parkin-dependent mitophagy. Impaired mitophagy fails to effectively degrade mitochondria-localized arginase 2 (ARG2), shifting L-arginine metabolism toward the urea cycle rather than nitric oxide (NO) synthesis, ultimately resulting in reduced NO production and endothelial dysfunction (Hu et al. 2026). In addition, the upregulation of X-linked inhibitor of apoptosis (XIAP) in the renal tissues of DKD patients drives the degradation of unc-51-like autophagy activating kinase 1 (ULK1) via K48-linked polyubiquitination, thereby impairing mitophagy and disrupting carnitine metabolism (Hu et al. 2026). Under physiological conditions, cells maintain normal mitochondrial metabolism through the mechanisms described above. However, under pathological conditions, cells alter their original metabolic strategies and undergo mitochondrial metabolic reprogramming to adapt to the altered environment, thereby sustaining survival and function. This phenomenon has been extensively studied in cancer cells, with the most classical example being the Warburg effect (Fendt 2024). Interestingly, similar metabolic alterations have also been observed in DKD tissues (Zhang et al. 2024; Kanasaki 2023; Lu et al. 2025), suggesting that mitochondrial metabolic reprogramming may likewise play a critical role in the pathogenesis of DKD. The role of mitochondrial metabolic reprogramming in the pathogenesis and progression of DKD The kidney is a highly heterogeneous organ composed of multiple distinct cell types, including tubular epithelial cells, podocytes, glomerular endothelial cells, and glomerular mesangial cells. The intrinsic renal cells and infiltrating immune cells exhibit markedly diverse metabolic characteristics and physiological functions. Emerging evidence indicates that stress-induced metabolic reprogramming in different cell types can exert divergent and even opposing effects on tissue injury and repair (Kobayashi et al. 2016; Bueno et al. 2015; Larson-Casey et al. 2016). Therefore, dissecting the metabolic network of DKD from a cell type- specific perspective is of critical importance. Mitochondrial metabolic reprogramming in proximal renal tubular epithelial cells Glucose filtered by the glomerulus is reabsorbed through SGLTs located on the apical membrane of proximal tubular epithelial cells (PTECs). This process is coupled with Na+ reabsorption (Yang and Xu 2022). To sustain this Na+-dependent glucose reabsorption, basolateral Na+/K+-ATPase continuously exports the reabsorbed Na+ out of PTECs. Under physiological conditions, this energetically demanding process relies heavily on FAO to generate substantial amounts of ATP (Hou et al. 2024; Console et al. 2020). Under high glucose (HG) conditions, increased glomerular glucose filtration markedly elevates ATP demand. Several studies have demonstrated that HG exposure upregulates key glycolytic enzymes (e.g., PKM2, HK2, and LDHA) and enhances lactate production in PTECs. However, expression of FAO-related enzymes, such as CPT1, CPT2, and ACOX1, is downregulated. This shift indicates that PTECs undergo metabolic reprogramming characterized by enhanced glycolysis and suppressed FAO (Zhang et al. 2024; Lee et al. 2020). This process is associated with multiple metabolic disturbances and the subsequent formation of vicious pathogenic cycles. Suppression of FAO directly leads to lipid accumulation (Gupta et al. 2024), which in turn induces the overexpression of the nuclear receptor transcription factor retinoid X receptor α (RXRα). This overexpression subsequently activates the RXRα- mineralocorticoid receptor (MR) pathway, further suppressing FAO and OXPHOS by transcriptionally inhibiting the rate-limiting FAO enzyme CPT1, upregulating fatty acid synthesis genes (ACACA, ACSL5, ELOVL5/7, etc.) and ATP hydrolase genes (ATP2A2, ATP6V0C, etc.), while downregulating ATP synthase genes (ATP5MC2, ATP5PB, etc.) (Feng et al. 2024). On the other hand, when hyperglycemia exceeds the glycolytic capacity of cells, excess carbon flux is redirected into stress-related pathways (Wu et al. 2023), leading to excessive ROS production and mitochondrial oxidative stress (Ma et al. 2023). This oxidative stress not only directly damages mitochondria (Shadel and Horvath 2015), but also promotes lipid droplet formation and sphingosine-1-phosphate accumulation. This further exacerbates lipid deposition (Hou et al. 2024) and establishes a self-reinforcing positive feedback loop that progressively amplifies mitochondrial oxidative injury (Mitrofanova et al. 2023). Collectively, mitochondrial metabolic reprogramming and its downstream vicious cycles severely compromise mitochondrial bioenergetic capacity, culminating in defective ATP generation. Under HG conditions, the capacity of PTECs to utilize alternative energy substrates is also impaired. Studies have shown that under HG conditions, PTECs exhibit a reduced reabsorption of ketone bodies from the urine. Given that ketone bodies serve as a vital alternative fuel for the kidney under stress conditions (Guo et al. 2023), this restricted supply further exacerbates the deficit in mitochondrial ATP production. Decreased ATP levels drive an increase in cellular AMP levels. This elevated AMP/ATP ratio promotes protein synthesis and stimulates PTEC hypertrophy (Huynh et al. 2023). In turn, hypertrophied PTECs demand greater energetic output, which further aggravates mitochondrial workload and subsequent injury (Hoogstraten et al. 2024). Impaired mitochondrial clearance and biogenesis lead to the accumulation of damaged mitochondria. In diabetic mice, PTEC-specific deletion of Yes-Associated Protein 1 (YAP1) exacerbates renal injury, highlighting the protective role of YAP1 in PTECs (Ye et al. 2024). YAP1 is involved in multiple biological processes, including cell death, oxidative stress, and inflammation (Fu et al. 2022). Its inactivation downregulates PGC-1α- TFAM/TOMM20-mediated mitochondrial biogenesis and impairs mitophagy by suppressing the PINK1-Parkin pathway. This dual inhibition impedes the timely clearance of damaged mitochondria. In addition, YAP1 promotes the secretion of the pro-inflammatory chemokine C-X-C motif chemokine ligand 1 (CXCL1) from PTECs, facilitating macrophage M1 polarization and exacerbating inflammation (Ye et al. 2024). The protective mechanisms of mitochondrial biogenesis are disrupted under HG conditions. HG significantly inhibits the expression of methyltransferase 3 (METTL3) and global RNA m6A modification levels. This results in decreased m⁶A modification at key sites of lncRNA TUG1. This subsequently compromises the recognition and binding affinity of insulin-like growth factor 2 binding protein 2 (IGF2BP2) toward TUG1, ultimately decreasing TUG1 stability and downregulating its expression. Reduced TUG1 levels fail to effectively activate the PPARGC1A promoter, leading to suppression of PGC-1α transcription and activation. This suppression downregulates downstream mitochondrial regulators (e.g., Nrf1, Nrf2, and TFAM), culminating in tubular mitochondrial dysfunction characterized by arrested mitochondrial biogenesis, insufficient ATP production, and diminished respiratory chain complex I/III activity (Chen et al. 2025). Furthermore, mitochondrial proteostasis and structural integrity are also compromised. Under HG conditions, the expression of the mitochondrial protease caseinolytic peptidase P (ClpP) is significantly upregulated in PTECs. As a core component of the proteolytic system in the mitochondrial unfolded protein response (UPRmt), its overexpression causes hyperactivation of the UPRmt, thereby worsening oxidative stress, cell apoptosis, and tubulointerstitial fibrosis (Wang et al. 2025). Additionally, elevated expression of the peptide hormone asprosin upregulates the SUMO E3 ligase PIAS1 and downregulates the deSUMOylating protease SENP1. This imbalance drives the excessive SUMOylation of Drp1, which prevents its degradation and accelerates mitochondrial fragmentation, thereby disrupting mitochondrial structural homeostasis (Huang et al. 2025). In conclusion, metabolic abnormalities and quality control defects synergistically contribute to mitochondrial dysfunction, driving the initiation and progression of tubular injury in DKD (Fig. 2). Mitochondrial metabolic reprogramming in proximal tubular epithelial cells in diabetic kidney disease In addition to PTECs, one study also revealed potential metabolic reprogramming in collecting duct principal cells (CD-PCs) and the descending limb of Henle (DLH). Specifically, ssGSEA scores for FAO, glycolysis, OXPHOS, and the TCA cycle were all significantly increased in CD-PCs and DLH cells. This may represent a compensatory adaptation during the early stage of DKD to sustain their high-demand transport functions (Wu et al. 2023). Mitochondrial metabolic reprogramming in podocytes Podocytes send out foot processes to form a filtration barrier with glomerular capillary endothelial cells and basement membrane to maintain normal filtration function (Wongmekiat et al. 2021). The morphology and function of foot processes depend on the dynamic regulation of actin microfilaments. Therefore, a substantial supply of ATP is essential for maintaining the integrity of the glomerular filtration barrier (Reynolds 2020). Podocytes primarily rely on glucose metabolism for energy production (Yuan et al. 2020). Before differentiation, they predominantly depend on glycolysis, whereas after differentiation both glycolysis and OXPHOS are enhanced, with OXPHOS serving as the major energy source to meet their high energy demands (Yuan et al. 2020). Podocytes in DKD exhibit multifaceted mitochondrial metabolic reprogramming. Regarding glucose metabolism, the expression of Mothers Against Decapentaplegic Homolog 4 (Smad4) is upregulated in podocytes from DKD patients and mice, leading to reductions in glycolysis and OXPHOS respectively through suppression of pyruvate kinase M2 (PKM2) activity and inhibition of the degradation of ATPase inhibitory factor 1 (ATPIF1) (Li et al. 2020). With respect to lipid metabolism, ABCA1 regulates cholesterol and phospholipid efflux through an ATP-dependent mechanism (Wang et al. 2001). Its downregulation induces the accumulation of mitochondrial cardiolipin and impairs FAO in podocytes (Ducasa et al. 2019). In terms of amino acid metabolism, podocytes in DKD exhibit defects in branched-chain amino acid (BCAA) catabolism, characterized by reduced activity of BCAA-degrading enzymes and the accumulation of BCAAs and their metabolic intermediates (Zhao et al. 2025). This disturbance promotes crosstalk with glucose metabolic pathways. In DKD, the expression of the mitochondria-localized, podocyte-specific protein phosphatase 2 C subtype m (PP2Cm) is downregulated, which increases the inhibitory phosphorylation of the rate-limiting BCAA catabolic enzyme branched-chain α-ketoacid dehydrogenase (BCKD). The accumulation of BCAAs induces dissociation and inactivation of the key glycolytic enzyme PKM2, resulting in suppression of OXPHOS and redirecting glucose metabolic flux toward serine biosynthesis and one-carbon metabolism, further exacerbating metabolic imbalance (Zhao et al. 2025). The ETC is also impaired in podocytes in DKD, as evidenced by reduced abundance of ETC complexes and decreased cristae density (Mise et al. 2024). To counteract these defects, podocyte-specific overexpression of NDUFS4 (NADH: ubiquinone oxidoreductase iron-sulfur protein 4, a subunit of ETC complex I) has been shown to preserve the assembly of respiratory supercomplexes (RSCs) and cristae integrity through interaction with the cristae-shaping protein stomatin-like protein 2 (STOML2). This intervention significantly improves mitochondrial respiration and cristae morphology and restores mitochondrial ATP production (Mise et al. 2024). These findings suggest that NDUFS4 represents a promising therapeutic target. In addition, the hyperglycemic environment activates stress-responsive metabolic pathways, leading to increased production of advanced glycation end-products (AGEs) (Wu et al. 2023). Binding of AGEs to receptor for advanced glycation end-products (RAGE) induces the expression of inducible nitric oxide synthase (iNOS), which increases NO production. NO subsequently reacts with superoxide anions (O₂⁻) generated by NADPH oxidase to form peroxynitrite (ONOO⁻) to damage podocytes (Lee et al. 2013). Moreover, AGEs promote the production of serum amyloid A (SAA), which further triggers RAGE activation, establishing a positive feedback loop that exacerbates podocyte injury (Pichler et al. 2017). Healthy podocytes can transfer mitochondria to AGEs-injured podocytes through tunneling nanotubes (TNTs) in a cytoplasmic protein tumor necrosis factor α induced protein 2 (TNFAIP2)-dependent manner. This process is accompanied by the exchange of autophagosomes and lysosomes, thereby improving autophagic flux and lysosomal function in the injured podocytes and alleviating podocyte apoptosis (Barutta et al. 2023). These findings suggest that the TNFAIP2-TNTs system may represent a novel therapeutic target. Dysregulation of mitochondrial quality control in podocytes is mainly manifested as mitochondrial oxidative stress and abnormal mitophagy, which together lead to podocyte dysfunction and apoptosis and drive DKD progression (Fig. 3). Regarding oxidative stress, AGEs, generated during metabolic reprogramming, can induce mitochondrial oxidative stress in podocytes (Ma et al. 2023). In addition to AGEs, abnormalities in endogenous signaling molecules within podocytes can also mediate mitochondrial oxidative damage. Insulin-like growth factor-binding protein 2 (IGFBP2), a ligand protein predominantly expressed by podocytes in the adult kidney, shows elevated expression in patients with type 2 DM, which is closely associated with decreased estimated glomerular filtration rate (eGFR) and increased proteinuria (Narayanan et al. 2012). Studies have shown that increased IGFBP2 promotes podocyte apoptosis by binding to integrin α5 (ITGA5) on podocytes and subsequently activating focal adhesion kinase (FAK) phosphorylation-mediated mitochondrial oxidative stress (Wang et al. 2024). Sirtuin 3 (SIRT3) is a mitochondrial deacetylase that regulates mitochondrial homeostasis and is expressed at low levels in DKD. Its downregulation enhances the inhibitory acetylation of MPC2, which exacerbates ROS generation and subsequently drives podocyte apoptosis (Feng et al. 2025). In terms of mitophagy, alterations in podocyte autophage in DKD appear to depend on the duration of hyperglycemic exposure. In the early stage of DKD, short-term exposure to high glucose may induce autophagy to protect the kidney, whereas in the late stage of DKD, prolonged hyperglycemia may suppress podocyte autophagy and contribute to disease progression (Lenoir et al. 2015). In DKD, the transcriptional repressor ubiquitin-like with PHD and RING finger domains 1 (UHRF1) is downregulated, which promotes the expression of thioredoxin-interacting protein (TXNIP). This upregulation suppresses PINK1-mediated mitophagy and exacerbates podocyte ferroptosis, ultimately disrupting the glomerular filtration barrier (Ji et al. 2024). The key mitophagy molecule PINK1 influences podocyte permeability to albumin (Audzeyenka et al. 2022). Therefore, insufficient autophagy in podocytes plays an important role in the development of massive proteinuria in advanced DKD. Impaired mitophagy is also associated with insulin resistance in podocytes. Podocytes are the only insulin-sensitive cells in the glomerulus, capable of inducing the translocation of glucose transporters GLUT1 and GLUT4 to the cell membrane under insulin stimulation, thereby increasing glucose uptake (Coward et al. 2005). However, PINK1 deficiency inhibits the translocation of GLUT4 to the cell membrane, leading to podocyte insulin resistance and further exacerbating DKD (Audzeyenka et al. 2022). Insufficient autophagy is also an important driver of renal function decline. In DKD, complement accumulation impairs podocyte mitophagy through the complement component 3a (C3a)/complement component 3a receptor (C3aR)/phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/FoxO1 axis (Weng et al. 2025), leading to a rapid decline in eGFR (Jiang et al. 2022). Mitochondrial metabolic reprogramming in podocytes in diabetic kidney disease Mitochondrial metabolic reprogramming in glomerular endothelial cells Mitochondrial metabolic reprogramming in glomerular endothelial cells (GECs) is mainly characterized by enhanced glycolysis, reduced FAO, and cholesterol accumulation. In terms of glycolysis, insulin-like growth factor-binding protein 5 (IGFBP-5) is upregulated in GECs under HG conditions and promotes the expression of the key glycolytic enzyme fructose-2,6-bisphosphatase 3 (PFKFB3) by activating the transcription factor early growth response 1 (EGR1), thereby enhancing glycolysis (Song et al. 2022). Enhanced glycolysis not only provides energy but also activates the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling pathway through the metabolic byproduct lactate, promoting the release of inflammatory cytokines such as monocyte chemoattractant protein-1 (MCP-1) and aggravating inflammation in GECs (Song et al. 2022). In addition, the enhancement of glycolysis in GECs can be further amplified through crosstalk with tubular epithelial cells (TECs). On the one hand, transforming growth factor-β (TGF-β) secreted by TECs binds to receptors on the GEC membrane and promotes the expression of IGFBP-5 in GECs via the extracellular signal-regulated kinase (ERK) signaling pathway, further enhancing glycolysis and exacerbating inflammation (Wang et al. 2025). On the other hand, IGFBP-5 secreted by some GECs acts in a feedback manner on TECs to stimulate the production of more TGF-β through the AKT signaling pathway, which subsequently acts again on GECs, forming a self-amplifying IGFBP-5/TGF-β positive feedback loop (Wang et al. 2025). Therefore, targeting IGFBP-5 may help disrupt this vicious cycle and thereby prevent further progression of inflammation. With respect to lipid metabolism, GECs mainly exhibit suppression of FAO and accumulation of cholesterol. Regulators of mitochondrial homeostasis play an important role in this process. Downregulation of the mitochondrial deacetylase SIRT3 suppresses FAO and promotes endothelial-to-mesenchymal transition in GECs by activating the TGF-β/Mothers against decapentaplegic homolog 3 (Smad3) signaling pathway, thereby exacerbating renal fibrosis (Srivastava et al. 2021). Moreover, the capacity of GECs for lipid transport and clearance is also impaired. In human renal glomerular endothelial cells (HRGECs) cultured under high-glucose and high-cholesterol conditions, deficiency of the lipid transport protein ABCA1 exacerbates cholesterol deposition in endothelial cells, leading to inflammation and apoptosis and disrupting the endothelial glycocalyx barrier (Zhang et al. 2023). Mitochondrial oxidative stress in GECs is an important mechanism driving the progression of DKD. Xanthine oxidase (XO) serves as a major source of ROS. Its overactivation upregulates the expression of vascular endothelial growth factor (VEGF) and its receptors VEGFR1 and VEGFR3, thereby activating NADPH oxidase. This signaling cascade subsequently promotes the phosphorylation of the transcription factor forkhead box O3a (FoxO3a) and suppresses endothelial nitric oxide synthase (eNOS) activation, ultimately exacerbating mitochondrial oxidative stress and glomerular endothelial cell dysfunction (Yang et al. 2023). Furthermore, mitochondrial oxidative stress and lipid metabolic reprogramming exhibit reciprocal regulation in GECs. Loss of the endothelial glucocorticoid receptor not only induces oxidative stress through activation of the wingless/integrated (Wnt) signaling pathway but also suppresses FAO (Srivastava et al. 2021). Treatment with Wnt inhibitors alleviates oxidative stress and restores FAO activity (Srivastava et al. 2021), indicating that Wnt signaling represents a key regulatory node linking mitochondrial metabolic reprogramming with oxidative stress in GECs (Fig. 4). Mitochondrial metabolic reprogramming in glomerular endothelial cells in diabetic kidney disease Mitochondrial metabolic reprogramming in glomerular mesangial cells Renal fibrosis is a core pathological feature of advanced DKD and refers to the excessive accumulation of extracellular matrix (ECM) in the kidney, manifested as glomerulosclerosis and/or tubulointerstitial fibrosis (Qian et al. 2008). Among these processes, injury to glomerular mesangial cells (GMCs) is considered an important factor driving renal fibrosis (Huang et al. 2023). Increasing evidence indicates that mitochondrial metabolic reprogramming plays a pivotal role in GMC injury. In the aspect of energy metabolism, FAO and glycolysis are two major energy sources for GMCs. Studies have shown that inhibition of the rate-limiting FAO enzyme CPT1 significantly reduces ATP production in GMCs, whereas inhibition of the glycolytic rate-limiting enzyme HK likewise diminishes ATP generation (Nagai et al. 2022). These findings suggest that both pathways jointly maintain cellular energy homeostasis. When hyperglycemia exceeds the capacity of glycolysis, the AGEs stress metabolic pathway is activated, which further drives lipid metabolic reprogramming and renal fibrosis in GMCs by increasing TGF-β levels (Wang et al. 2021). This process primarily depends on two parallel downstream signaling pathways. One is the TGF-β/Rho-associated coiled-coil-containing protein kinase 1(ROCK1)/Adenosine monophosphate-activated protein kinase (AMPK)/PGC-1α pathway. TGF-β activates the protein kinase ROCK1, which suppresses the energy sensor AMPK and its downstream transcriptional coactivator PGC-1α, thereby downregulating key FAO genes, reducing energy production, and mediating lipid metabolic reprogramming in GMCs (Nagai et al. 2022). The other pathway is the TGF-β/TGF-βR/Smad3 pathway. TGF-β binds to its receptor TGF-βR and activates Smad3 phosphorylation, which subsequently stimulates fibroblast proliferation and extracellular matrix deposition, ultimately leading to renal fibrosis (Wang et al. 2014). Notably, although both pathways share TGF-β as the upstream molecule, the ROCK1-mediated FAO suppression pathway and the Smad3-mediated fibrotic pathway are independent of each other and do not exhibit direct cross-regulation (Nagai et al. 2022). Dysregulation of mitochondrial quality control also plays an important role in GMC injury and the progression of DKD. HG-induced upregulation of regulators of calcineurin 1.4 (RCAN1.4) promotes the translocation of DRP1 to mitochondria, leading to mitochondrial fragmentation and increased ECM production, thereby promoting GMC fibrosis [109]. HG also enhances mitochondrial oxidative stress through multiple signaling mechanisms, further impairing GMC function and promoting fibrosis. One mechanism involves upregulation of the non-receptor tyrosine kinase Fyn under HG conditions, which inhibits the SIRT1/FoxO3a pathway through the adaptor protein Casitas B-lineage lymphoma proto-oncogene (c-Cbl), thus increasing cellular sensitivity to oxidative damage (Li et al. 2023). The second mechanism involves microRNA-1225-3p-mediated suppression of Rho GTPase-activating protein 5 (ARHGAP5), which enhances the activity of Smad ubiquitin regulatory factor 2 (SMURF2) and promotes the ubiquitination and degradation of carbohydrate response element binding protein (ChREBP). This process weakens the adaptive response of GMCs to glucose metabolic stress (Zhang et al. 2025). Together, the above mechanisms exacerbate mitochondrial oxidative stress in GMCs and drive the progression of fibrosis. Mitochondrial metabolic reprogramming in macrophages Single-cell RNA sequencing (scRNA-seq) has revealed that macrophages are the predominant infiltrating immune cells in the kidneys during early DKD (Youssef et al. 2024). Furthermore, the macrophage abundance score in the renal tissues of patients with advanced DKD is significantly higher than that in early DKD patients and normal controls (Ye et al. 2022). Peptidylarginine deiminase 4 (PAD4) facilitates macrophage migration and infiltration. Data from the United Kingdom Biobank (UK Biobank) cohort indicate that most PADI4 variants result in loss of function, which correlates significantly with higher eGFR (Xiong et al. 2026). These findings suggest that macrophage infiltration is a key driver of DKD. Based on their activation states and functional characteristics, macrophages are broadly categorized into two major distinct subsets: M1 macrophages (pro-inflammatory phenotype) and M2 macrophages (anti-inflammatory phenotype) (2024). A pronounced immune-metabolic crosstalk exists in DKD, wherein macrophages undergo metabolic reprogramming in response to local microenvironmental cues. Conversely, their metabolic status reciprocally dictates their immunological functions (Jing et al. 2020), a key aspect of which involves the regulation of macrophage polarization (Vitale et al. 2019). As a pivotal hub linking energy metabolism and inflammatory regulation, enhanced glycolysis primarily drives M1 polarization. This hypermetabolic state accelerates DKD progression by upregulating a cascade of inflammatory mediators, such as IL-6 and TNF-α. Under HG conditions, EVs released by PTECs enhance the stability of the transcription factor hypoxia inducible factor-1α (HIF-1α) by inhibiting its hydroxylation (Jia et al. 2022). Concurrently, downregulation of the demethylase fat mass and obesity-associated protein (Fto) increases the m6A modification levels on neuronal PAS domain protein 2 (Npas2) mRNA, which enhances its stability, upregulates Npas2 expression, and ultimately stabilizes HIF-1α levels (Zhu et al. 2025). Stabilized HIF-1α functions as a transcription factor to upregulate the expression of key glycolytic enzymes, thereby reinforcing glycolytic flux. This process not only provides energy for macrophages but also promotes inflammatory cytokine expression via lactate accumulation, thus facilitating M1 polarization (Jia et al. 2022). In DKD, distinct macrophage phenotypes exhibit divergent fatty acid metabolism profiles, which underlie their disparate biological functionalities. Enhanced fatty acid synthesis is a critical metabolic requirement for sustaining M1 macrophage effector functions, whereas M2 macrophages rely predominantly on accelerated FAO to maintain their activities (Yu et al. 2025). The TCA cycle also regulates macrophage polarization through metabolic reprogramming. M1 polarization is characterized by disruptions in the TCA cycle at isocitrate dehydrogenase and succinate dehydrogenase nodes, leading to the accumulation of citrate and succinate. Conversely, an intact and fully functional TCA cycle serves as the metabolic foundation for M2 polarization (Russo et al. 2021). Among TCA metabolites, itaconate plays a key role by activating the Nrf2 pathway, which upregulates antioxidant genes such as GPX4 to promote M2 polarization (Shi and Cai 2022). Consequently, utilizing itaconate derivatives to activate the Nrf2 pathway presents a promising therapeutic strategy for improving DKD by modulating macrophage metabolism. OXPHOS also participates in the metabolic control of macrophage polarization, where its enhancement promotes M2 polarization while suppressing the M1 phenotype (Wang et al. 2021; Liu et al. 2023). In addition, studies have shown that mitochondria derived from mesenchymal stem cells (MSCs-Mito) can be transferred into macrophages, where they activate PGC-1α-mediated mitochondrial biogenesis and PGC-1α/transcription factor EB (TFEB)-mediated lysosomal autophagy. This transfer drives M2 macrophage polarization and alleviates inflammation (Yuan et al. 2021), suggesting that mitochondrial quality control is also closely linked to macrophage M2 polarization (Fig. 5). However, although macrophages can polarize toward the M2 phenotype to initiate tissue repair, prolonged HG exposure disrupts their anti-inflammatory and reparative functions, resulting in an imbalanced state in which different macrophage phenotypes coexist and counteract each other, ultimately driving the progression of DKD toward ESKD (Lin et al. 2024). Mitochondrial metabolic reprogramming in macrophages in diabetic kidney disease In summary, a hyperglycemic environment induces mitochondrial metabolic reprogramming across diverse renal cells. At the level of energy metabolism, renal cells exhibit a compensatory response by switching energy substrates and remodeling metabolic pathways to maintain normal cellular function. However, sustained hyperglycemic stimulation maintains cells in a chronic abnormal metabolic state, ultimately leading to persistent cellular injury. With respect to mitochondrial quality control, reprogramming mechanisms are also activated at early stages. In early DKD, mitophagy is increased in podocytes to remove damaged mitochondria and maintain metabolic homeostasis (Lenoir et al. 2015). Meanwhile, mitochondrial fragmentation is also observed, which further facilitates the mitophagy process (Kang et al. 2024). Current evidence indicates that autophagic activity in podocytes changes dynamically during DKD progression, characterized by early activation and later suppression. This dysregulation of autophagy is a key contributor to the development of albuminuria (Lenoir et al. 2015). This suggests that once the extent of injury exceeds the compensatory threshold of mitochondrial reprogramming, the mitochondrial quality control system becomes dysfunctional, which in turn further impairs energy metabolism. Injured renal cells secrete cytokines such as TGF-β (Wang et al. 2025) and IL-1β (Srivastava et al. 2021), which not only trigger renal inflammation and fibrosis but also mediate macrophage recruitment. Macrophage polarization is highly microenvironment-dependent, and local signal-driven mitochondrial metabolic reprogramming precisely regulates their immune phenotypes and functional differentiation, establishing a sophisticated immune-metabolic crosstalk (Jing et al. 2020). Furthermore, the diverse cytokines secreted by injured cells collectively constitute the cellular secretome, which mediates signaling crosstalk among cells and reciprocally modulates their respective mitochondrial metabolic reprogramming (Srivastava et al. 2021). EVs also serve as vital vehicles for intercellular communication (Yang et al. 2025). Together, the DKD microenvironment is jointly shaped by renal cells, immune cells, the secretome, and EVs (Fig. 6). Consequently, mitochondrial metabolic reprogramming represents a fundamental driver of the establishment and evolution of the DKD microenvironment, playing a pivotal role in the pathogenesis of DKD. Mitochondrial metabolic reprogramming and cellular crosstalk in the microenvironment of diabetic kidney disease Nevertheless, several limitations and unresolved questions remain. Taking the core mitochondrial regulator PINK1 as an example, its exact role in DKD remains highly controversial. On the one hand, PINK1 expression shows opposite trends across different renal cell types. HG exposure upregulates PINK1 levels in PTECs (Sung et al. 2023) but downregulates its expression in podocytes and GMCs (Li et al. 2017; Yi et al. 2022). In PTECs, PINK1 upregulation activates mitophagy, preserves mitochondrial structural and functional integrity, and suppresses excessive ROS accumulation. This protective axis prevents the transition of PTECs toward a pro-fibrotic phenotype and attenuates necroptosis, thereby delaying tubulointerstitial injury and the progression of renal fibrosis (Sung et al. 2023). Conversely, a reduction in PINK1 expression in podocytes and GMCs triggers overproduction of ROS, mitochondrial fragmentation, and autophagic dysfunction, ultimately culminating in apoptosis. However, restoring PINK1 expression can reverse HG-induced cellular stress and mitigate these adverse effects (Li et al. 2017; Yi et al. 2022). These inconsistencies are likely attributable to substantial differences in experimental design, methods of diabetes induction, timing of sample collection, and choice of animal models. On the other hand, the causal relationship between PINK1 alterations and sustained hyperglycemia remains unclear, making it difficult to determine whether PINK1 dysregulation is a cause or a consequence of metabolic disturbance (Zhan et al. 2015; Xiao et al. 2017). Moreover, conflicting findings have also been reported regarding lipid catabolism in DKD. Some studies suggest impaired lipid catabolism exacerbates lipotoxicity (Herman-Edelstein et al. 2014), whereas others propose a compensatory enhancement of lipid breakdown to cope with substrate overload and heightened energy demands (Sas et al. 2016). This divergence may be attributed to pathophysiological differences between early and late-stage DKD. Future studies should further delineate mitochondrial characteristics across different stages of DKD and elucidate cell type-specific differences in mitochondrial bioenergetics, quality control, and signaling pathways, which will be essential for developing personalized predictive tools and therapeutic strategies. Mitochondrial metabolic reprogramming: a key to deciphering “metabolic memory” in DKD “Metabolic memory” is defined as the phenomenon in which, even after strict glycemic control corrects hyperglycemia in later stages, the pathological “imprint” left by earlier metabolic disturbances in the kidney continues to drive the progression of DKD. The underlying logic of this phenomenon is deeply rooted in the physical linkage between mitochondrial metabolism and epigenetic modifications. Mitochondria are not only the center of energy production, but also the core source of methyl and acetyl donors required for epigenetic modifications. In the one-carbon metabolism network, mitochondria can generate one-carbon units, which shuttle to the cytoplasm in the form of formic acid and critically regulate the dual cycles of folate and methionine in the cytoplasm (Du et al. 2025). The methionine cycle generates S-adenosylmethionine (SAM), an indispensable donor for DNA and histone methylation (Blomgren et al. 2024), whereas the folate cycle maintains the continuous production of SAM (Green et al. 2023). TeSlaa et al. revealed the direct link between mitochondrial metabolism and methylation. The key TCA cycle metabolite α-ketoglutarate activates ten-eleven translocation methylcytosine dioxygenase (TET), thereby increasing DNA hydroxymethylation and reducing DNA methylation, while also stimulating histone demethylases. In contrast, succinate suppresses these demethylation processes (TeSlaa et al. 2016). On the other hand, cytosolic acetyl-CoA, the primary substrate for histone acetylation, is mainly derived from mitochondrial citrate. Citrate is exported from mitochondria via the citrate shuttle system and subsequently converted into acetyl-CoA in the cytosol (Du et al. 2025). Lozoya et al. demonstrated that mitochondrial acetyl-CoA levels directly determine histone acetyltransferase (HAT) activity, reversibly regulate site-specific histone acetylation, and precisely influence downstream gene transcription. Conversely, mitochondrial dysfunction markedly reduces the intracellular acetyl-CoA pool, leading to sustained hypoacetylation of nuclear histones (Lozoya et al. 2019). These findings indicate that mitochondrial metabolism is closely associated with methylation and acetylation modifications. Previous studies have shown that accumulation of S-adenosylhomocysteine (SAH) leads to reduced histone H3 lysine 27 methylation, which is closely related to the “metabolic memory” of DKD (Dai et al. 2021). It can therefore be inferred that mitochondrial metabolic reprogramming may underlie the “metabolic memory” observed in DKD. Recent studies have provided strong evidence supporting this hypothesis. Sarwar et al. found that the flux of the mitochondrial methionine cycle is increased in DKD, and subsequent RNA-Seq and Methyl-seq confirmed that this metabolic reprogramming does cause changes in DNA methylation modifications in db/db mice and negatively affects gene expression (Sarwar et al. 2023). These results suggest that the study of mitochondrial metabolic reprogramming may provide new ideas for solving the important clinical problem of “metabolic memory” in DKD patients. Therapeutic strategies targeting mitochondrial metabolic reprogramming in DKD: a comprehensive overview Preclinical studies A variety of agents targeting mechanisms of mitochondrial metabolic reprogramming have been demonstrated to alleviate DKD injury (Table 1). A classic drug targeting mitochondrial energy metabolism is the SGLT2 inhibitor. The most prominent metabolic reprogramming in PTECs is the shift from FAO to glycolysis, and SGLT2 inhibitors can correct this by inducing a fasting-like metabolic pattern. By blocking glucose reabsorption in PTECs, SGLT2 inhibitors increase urinary glucose excretion and produce a daily negative energy balance, which mimics the fasting state and forces cells to switch their energy source from glucose to FAO for ATP production (Shao et al. 2024). Further mechanistic studies have shown that this effect is achieved through pathways such as inhibition of HIF-1α and restoration of SIRT3 activity (Lu et al. 2022). In addition, the increased urinary glucose excretion caused by SGLT2 inhibitors is accompanied by osmotic diuresis and enhanced solute excretion, leading to fluid loss. This activates fluid-conservation mechanisms that require cells to produce and accumulate high levels of organic osmolytes to maintain water-electrolyte balance, thereby promoting the upregulation of amino acid metabolism (Shao et al. 2024). Stachyose has been described as the first SGLT2 destabilizer. Unlike traditional inhibitors, stachyose binds with high affinity to the Tyr270 site of SGLT2 and exerts its effects by promoting SGLT2 degradation (Wu et al. 2025). Table 1Agents targeting mitochondrial metabolic reprogramming in diabetic kidney diseaseDrug/TherapyTargeted Cell TypeDKD ModelMechanism of ActionFunctionReferenceSGLT2 InhibitorsProximal Tubular Epithelial Cellsdb/db mice and the human proximal tubular cell line HK-2Blocks glucose reabsorption, mimics fasting state, inhibits HIF-1α, restores SIRT3Corrects metabolic shift from glycolysis to fatty acid oxidation Ye et al. 2022; Xiong et al. 2026)StachyoseProximal Tubular Epithelial Cellsdb/db mice, the human proximal tubular cell line HK-2 and primary TECsBinds to the Tyr270 site of SGLT2 and promotes SGLT2 degradationSGLT2 destabilizer Lin et al. 2024)Schizonepeta C. Y. WuGlomerular Mesangial CellsSTZ-induced SD rats and the rat glomerular mesangial cell line HBZY-1Improves mitochondrial membrane potential and inhibits caspase signaling pathwayInhibits mitochondrial oxidative stress-induced apoptosis Vitale et al. 2019)BerberinePodocytesdb/db mice and the mouse podocyte cell line MPC-5PKM2/PGC-1α/Opa1 pathwayImproves mitochondrial dynamics and reverses apoptosis Jia et al. 2022)MyricetinPodocytesdb/db mice and the mouse podocyte cell line MPC-5Inhibits PIK3R1 to activate the PI3K/Akt pathway; promotes XBP1 to activate the PINK1/Parkin pathwayRestores mitochondrial autophagy Yu et al. 2025)Schisandra chinensis MixtureProximal Tubular Epithelial CellsSTZ-induced SD rats and the human renal tubular epithelial cell line HK-2VDAC1/Grp75/IP3R pathwayPromotes mitochondrial autophagy and inhibits apoptosis Russo et al. 2021)Syzygium cuminiGlomerular Endothelial CellsSTZ combined with nicotinamide-induced Wistar ratsPromotes nuclear translocation of Nrf2Upregulates antioxidant enzyme activity, alleviates oxidative stress Shi and Cai 2022)Ginsenoside Rg1Glomerular Mesangial CellsHigh-fat diet and STZ-induced C57BL6J mice and the human glomerular mesangial cellsBinds to CD36 and downregulates its expressionInhibits excessive free fatty acid uptake and blocks fibrosis Wang et al. 2021)SS-31PodocytesSTZ-induced C57BL/6 mice and the mouse glomerular primary podocytesInhibits OMA1-mediated OPA1 hydrolysisImproves mitochondrial dynamics Yuan et al. 2021)P110Proximal Tubular Epithelial CellsSTZ-induced C57BL/6 mice, db/db mice and the human renal tubular epithelial cell line HK-2Disrupts the interaction between Drp1 and Fis1; activates the AMPK/PGC-1α/TFAM pathwayInhibits mitochondrial fission, promotes mitochondrial biogenesis Kang et al. 2024)AdiponectinProximal Tubular Epithelial CellsSTZ-induced SD rats and the rat renal tubular epithelial cell line NRK-52ERegulates the AdipoR1/CREB/PGC-1α/TFAM pathwayPromotes mitochondrial biogenesis and function Li et al. 2017)SZC-6Proximal Tubular Epithelial CellsSTZ-induced C57BL/6 mice and the human renal tubular epithelial cell line HK-2Upregulates MnSOD transcription via the SIRT3-Foxo3a axisImproves mitochondrial oxidative stress Yi et al. 2022)Tilapia Skin PeptidesGlomerular Mesangial CellsSTZ-induced SD rats and the rat glomerular mesangial cellsActivates the Bnip3/Nix axisReverses excessive mtROS production, inhibits fibrosis-related protein expression Zhan et al. 2015)Abbreviations: DKD diabetic kidney disease, HIF-1α hypoxia-inducible factor-1α, SIRT3 sirtuin 3, SGLT2 sodium-glucose cotransporter 2, TECs tubular epithelial cells, STZ streptozotocin, SD rats Sprague–Dawley rats, Tyr Tyrosine, PKM2 pyruvate kinase M2, PGC-1α peroxisome proliferator-activated receptor gamma coactivator 1α, Opa1 optic atrophy 1, PIK3R1 phosphatidylinositol 3-kinase regulatory subunit 1, PI3K phosphatidylinositol 3-kinase, Akt protein kinase B, XBP1 X-box binding protein 1, PINK1 PTEN induced kinase 1, VDAC1 voltage-dependent anion channel 1, Grp75 glucose-regulated protein 75, IP3R inositol 1,4,5-trisphosphate receptor, Nrf2 nuclear factor erythroid 2-related factor 2, CD36 cluster of differentiation 36, OMA1 OMA1 zinc metallopeptidase, Drp1 dynamin-related protein 1, Fis1 fission 1, AMPK adenosine 5'-monophosphate-activated protein kinase, TFAM mitochondrial transcription factor A, AdipoR1 adiponectin receptor 1, CREB cAMP response element-binding protein, Foxo3a forkhead box O3a, MnSOD manganese superoxide dismutase, Bnip3 BCL2 interacting protein 3, Nix BCL2 interacting protein 3 like, mtROS mitochondrial reactive oxygen species Agents targeting mitochondrial metabolic reprogramming in diabetic kidney disease Disrupts the interaction between Drp1 and Fis1; activates the AMPK/PGC-1α/ TFAM pathway Regulates the AdipoR1/CREB/ PGC-1α/TFAM pathway Abbreviations: DKD diabetic kidney disease, HIF-1α hypoxia-inducible factor-1α, SIRT3 sirtuin 3, SGLT2 sodium-glucose cotransporter 2, TECs tubular epithelial cells, STZ streptozotocin, SD rats Sprague–Dawley rats, Tyr Tyrosine, PKM2 pyruvate kinase M2, PGC-1α peroxisome proliferator-activated receptor gamma coactivator 1α, Opa1 optic atrophy 1, PIK3R1 phosphatidylinositol 3-kinase regulatory subunit 1, PI3K phosphatidylinositol 3-kinase, Akt protein kinase B, XBP1 X-box binding protein 1, PINK1 PTEN induced kinase 1, VDAC1 voltage-dependent anion channel 1, Grp75 glucose-regulated protein 75, IP3R inositol 1,4,5-trisphosphate receptor, Nrf2 nuclear factor erythroid 2-related factor 2, CD36 cluster of differentiation 36, OMA1 OMA1 zinc metallopeptidase, Drp1 dynamin-related protein 1, Fis1 fission 1, AMPK adenosine 5'-monophosphate-activated protein kinase, TFAM mitochondrial transcription factor A, AdipoR1 adiponectin receptor 1, CREB cAMP response element-binding protein, Foxo3a forkhead box O3a, MnSOD manganese superoxide dismutase, Bnip3 BCL2 interacting protein 3, Nix BCL2 interacting protein 3 like, mtROS mitochondrial reactive oxygen species Recent studies have shown that active compounds derived from traditional Chinese medicine can regulate metabolic reprogramming in DKD with minimal adverse effects in clinical settings (Pan et al. 2022). Schizonepeta tenuifolia (C. Y. Wu) can inhibit mitochondrial oxidative stress-induced apoptosis in mesangial cells, possibly by improving MMP and suppressing the caspase signaling pathway (Huang et al. 2020). Berberine (BBR) is one of the active components of the traditional Chinese medicine formula modified Hu-lu-ba-wan (MHLBW) and exhibits strong affinity for PKM2. The improved MHLBW formulation ameliorates mitochondrial dynamics and reverses podocyte apoptosis through the PKM2/PGC-1α/OPA1 signaling pathway (Gong et al. 2024). Furthermore, berberine can upregulate the expression of FAO enzymes including CPT1, Acyl-CoA oxidase 1 (ACOX1) and peroxisome proliferator-activated receptor α (PPARα), restore FAO in PTECs, and reduce aberrant lipid accumulation (Rong et al. 2021). Myricetin can restore podocyte mitophagy by activating the PI3K/AKT pathway through inhibition of phosphatidylinositol 3-kinase regulatory subunit 1 (PIK3R1) and by promoting X-box binding protein 1 (XBP1) to activate the PINK1/Parkin pathway (Liu et al. 2026). The Schisandra compound promotes mitophagy in PTECs and suppresses apoptosis via the voltage-dependent anion channel 1 (VDAC1)/glucose-regulated protein 75 (Grp75)/inositol 1,4,5-trisphosphate receptor (IP3R) axis, thereby alleviating renal fibrosis (Li et al. 2025). Syzygium jambos enhances the nuclear translocation of Nrf2, thus upregulating antioxidant enzyme activity and alleviating oxidative stress in GECs (Bangar et al. 2025). Ginsenoside Rg1 directly binds to cluster of differentiation 36 (CD36) and downregulates its expression, in turn inhibiting excessive free fatty acid uptake by mesangial cells and blocking lipotoxicity-mediated fibrotic progression (Han et al. 2023). In recent years, many emerging therapeutic approaches have provided multidimensional perspectives for the precision treatment of DKD. Mesenchymal stem cell therapy can prevent GMC death, ROS accumulation, and MMP damage. It also restores the expression of genes associated with mitophagy and mitochondrial biogenesis in GMCs, indicating its therapeutic potential for DKD (Sávio-Silva et al. 2021). D-Arg-dimethyl-Tyr-Lys-Phe-NH₂ (SS31) is a mitochondria-targeted tetrapeptide that improves mitochondrial dynamics in podocytes by inhibiting OMA1-mediated OPA1 cleavage (Yang et al. 2021). P110 is a selective peptide inhibitor that suppresses mitochondrial fission in PTECs by disrupting the interaction between Drp1 and fission 1 (Fis1). It also promotes mitochondrial biogenesis through activation of the AMPK/PGC-1α/mitochondrial transcription factor A (TFAM) signaling pathway (Yue et al. 2025). A recent study has shown that exosomes derived from regulatory T cells (Tregs-Exos) deliver microRNA-218-5p by inhibiting the tenascin-C (TNC)/toll-like receptor 4 (TLR4)/proto-oncogene tyrosine-protein kinase Src (SRC)/FUNDC1 pathway, enhancing mitophagy and reducing apoptosis in podocytes, suggesting that engineered RGD-Treg-Exos represent a novel strategy for DKD treatment (Guo et al. 2025). In addition, various natural bioactive compounds and chemical derivatives have also demonstrated promising therapeutic potential for DKD. Adiponectin (APN) is a plasma hormone protein specifically secreted by adipocytes that promotes mitochondrial biogenesis and function in PTECs through regulation of the adiponectin receptor 1 (AdipoR1)/cAMP response element-binding protein(CREB)/PGC-1α/TFAM signaling pathway (Chen et al. 2022). The novel coumarin derivative SZC-6 is an improved allosteric activator of SIRT3 that enhances the transcription of the antioxidant enzyme manganese superoxide dismutase (MnSOD) through the SIRT3-FoxO3a axis, thereby alleviating mitochondrial oxidative stress in PTECs (Liu et al. 2025). Tilapia skin peptides (TSPs) are a mixture of small molecular peptides extracted from tilapia skin. They improve mitochondrial morphology by activating the BNIP3/Nix axis and reverse excessive mtROS production and the decline in MMP. Through these mechanisms, TSPs suppress the expression of fibronectin, type IV collagen, and intercellular adhesion molecule 1 (ICAM-1) in GMCs (Jin et al. 2020). Clinical trials While numerous preclinical studies have demonstrated promising results, clinical trials targeting mitochondrial metabolic reprogramming in DKD are also continuously emerging. Fenofibrate (NCT03869931), a PPARα agonist, has been shown to restore FAO and is commonly used clinically to treat hypertriglyceridemia. Studies by Annelie et al. have demonstrated that inhibition of the VEGF-B signaling pathway reduces renal lipid accumulation, alleviates lipotoxicity, and prevents renal dysfunction (Mehlem et al. 2016). Currently, the anti-VEGF-B monoclonal antibody 2H10 (NCT04419467) is undergoing clinical evaluation to assess its efficacy in DKD patients. Additionally, the novel glucose-lowering agent dorzagliatin (NCT06222476) functions as a glucokinase activator that enhances endogenous glycemic regulation. Pentoxifylline, a methylxanthine phosphodiesterase inhibitor, may exert protective effects against renal injury induced by type 2 DM and other chronic diseases (Shan et al. 2012). A clinical trial investigating pentoxifylline (NCT03625648) is currently ongoing to determine whether it can delay the progression of DKD. Nrf2 primarily suppresses mitochondrial oxidative stress through the Nrf2/Keap1/AREs signaling pathway (Bryan et al. 2013). To date, bardoxolone methyl represents the most clinically promising Nrf2 agonist under development (Pergola et al. 2011). An early-phase clinical trial (NCT01351675) assessing the efficacy and safety of bardoxolone methyl in patients with type 2 DM demonstrated a significant increase in eGFR. However, the trial was terminated prematurely due to cardiovascular and renal safety concerns. A subsequent randomized clinical trial TSUBAKI (NCT02316821) evaluating the effect of bardoxolone methyl on GFR in DKD patients also confirmed that bardoxolone methyl improved eGFR and reduced the incidence of ESKD. These two trials underscore the feasibility of modulating the Nrf2/Keap1/AREs pathway to ameliorate DKD. The ongoing Phase III AYAME trial (NCT03550443) is anticipated to provide further clarity on this therapeutic approach in the near future. Conclusions and perspectives Mitochondrial metabolic reprogramming serves not only as an initial adaptive response to HG stress in early DKD but also as a core driver of long-term renal deterioration through the establishment of “metabolic memory” via epigenetic modifications. Through remodeling of metabolic flux and quality control, mitochondria not only trigger deleterious cycles such as ROS accumulation and lipotoxicity, but also function as biosynthetic hubs that supply key metabolites, thereby engraving persistent pathological “imprints” on the renal cell genome through epigenetic modifications. This metabolic dysregulation further propagates through secretome- and EVs-mediated intercellular crosstalk, diffusing single-cell homeostatic imbalance into the broader renal microenvironment and ultimately driving uncontrolled fibrosis and inflammation. In conclusion, supported by rigorous experimental designs and an increasingly refined theoretical framework, the intricate relationship between DKD and mitochondrial metabolic reprogramming will be more comprehensively elucidated. Accordingly, targeting this process is expected to emerge as a key therapeutic direction for DKD. At present, the mechanistic understanding of how mitochondrial metabolic reprogramming contributes to DKD progression remains incomplete. Given the pathogenic heterogeneity between type 1 and type 2 DM, future studies should pay closer attention to their distinct patterns of mitochondrial dysregulation. In addition, current research on mitochondrial metabolic reprogramming in distal tubular epithelial cells (DTECs) during DKD progression remains limited. This may be attributed to the long-standing focus of traditional studies on the “diabetic glomerulopathy” paradigm, which centers on glomerular injury and regards DTEC damage as a secondary consequence. However, with the emergence and growing acceptance of the “diabetic tubulopathy” theory (Yao et al. 2022), together with recent evidence that distal tubule-specific urinary biomarkers such as EGF and UMOD are closely associated with the risk of DKD progression (Tamargo et al. 2025), it is evident that DTECs also possess important investigational value. Therapeutically, agents targeting mitochondrial metabolic reprogramming have made significant progress in preclinical and clinical research. However, further preclinical investigations are still required to elucidate their mechanisms of action and safety profiles, thereby providing a stronger theoretical basis for subsequent clinical translation. Given that mitochondrial metabolic reprogramming often mediates deleterious paracrine loops among distinct renal cell types, such as the IGFBP-5/TGF-β positive feedback loop established between GECs and TECs (Wang et al. 2025), it is imperative to further investigate intercellular crosstalk mechanisms in DKD to identify key therapeutic targets capable of disrupting such pathogenic feedback loops. In addition, studies have demonstrated that p16-positive senescent cells increase pyruvate dehydrogenase kinase 1 (PDK1) expression by inhibiting the AMPK- mammalian target of rapamycin (mTOR) pathway, thereby suppressing pyruvate dehydrogenase (PDH) activity and reducing OXPHOS (Lu et al. 2025). Moreover, upregulation of glycolysis and downregulation of FAO can induce G2/M cell cycle arrest in PTECs and increase senescence-associated secretory phenotype (SASP) factors, thereby exacerbating tubular injury in DKD (Feng et al. 2024). These findings suggest that targeting the “mitochondrial metabolic reprogramming-cellular senescence” regulatory axis may represent a promising therapeutic strategy for DKD. Acknowledgements Not applicable. Abbreviations Diabetic kidney disease Diabetes mellitus End-stage kidney disease Renin-angiotensin system Sodium-glucose cotransporter 2 Tricarboxylic acid Oxidative phosphorylation Outer mitochondrial membrane Dynamin-related protein 1 Mitofusin 1/2 B-cell lymphoma 2 BCL2 interacting protein 3 Inner mitochondrial membrane Electron transport chain Optical atrophy 1 protein Reactive oxygen species Superoxide dismutase Lactate dehydrogenase Mitochondrial pyruvate carrier Carnitine palmitoyltransferase ATP-binding cassette A1 Mitochondrial membrane potential Mitochondrial ROS Nuclear factor erythroid 2-related factor 2 Antioxidant response elements Kelch-like ECH-associated protein 1 Progranulin Sirtuin 1 Peroxisome proliferator-activated receptor gamma coactivator 1-alpha Forkhead box protein O1 Streptozotocin Cyclase-associated protein 1 Inverted formin 2 Endoplasmic reticulum Mitochondria-associated ER membrane Sextracellular vesicles PTEN-induced kinase 1 FUN14 domain-containing protein 1 Nitric oxide Sproximal tubular epithelial cells High glucose Retinoid X receptor α Yes-Associated Protein 1 Mitochondrial unfolded protein response Collecting duct principal cells Descending limb of Henle Pyruvate kinase M2 Branched-chain amino acid NADH: ubiquinone oxidoreductase iron-sulfur protein 4 Advanced glycation end-products Receptor for advanced glycation end-products Tunneling nanotubes Tumor necrosis factor α induced protein 2 Insulin-like growth factor-binding protein 2 Estimated glomerular filtration rate Sirtuin 3 Glucose transporter Phosphatidylinositol 3-kinase Protein kinase B Glomerular endothelial cells Insulin-like growth factor-binding protein 5 Tubular epithelial cells Transforming growth factor-β Mothers against decapentaplegic homolog 3 Vascular endothelial growth factor Forkhead box O3a Nitric oxide synthase Wingless/integrated Extracellular matrix Glomerular mesangial cells Rho-associated coiled-coil-containing protein kinase 1 Adenosine monophosphate-activated protein kinase Peptidylarginine deiminase 4 Hypoxia inducible factor-1α Neuronal PAS domain protein 2 S-adenosylmethionine Modified Hu-lu-ba-wan Peroxisome proliferator-activated receptor α Exosomes derived from regulatory T cells Mitochondrial transcription factor A Tilapia skin peptides Distal tubular epithelial cells Authors’ contributions Conceptualization, S.F.; administrative support, S.F.; writing—original draft preparation, L.Z.; writing—review and editing, Y.Z.; formal analysis, L.Z. and Y.Z.; investigation, L.Z. and L.H.; visualization, L.Z., Y.Z., L.H., J.W. and L.H.; funding acquisition, S.F. All authors have read and agreed to the published version of the manuscript. Funding Gansu Provincial Joint Research Fund Project (23JRRA1490); National Key Research and Development Program of China (2023YFC3503400); 14th Five-Year Plan Project (2023YF2508305); Integrated Traditional Chinese and Western Medicine Collaborative Chronic Disease Management Research Project (CXZH2024082); National Natural Science Foundation of China (82560163); Lanzhou University Innovation and Entrepreneurship training program (20260060119); Cuiying Scholars Research Cultivation Program (CYXZPT2025-31). Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Lingfei Zhang and Yue Zhang contributed equally to this study and share the first authorship. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.