Bioenergetic failure in diabetic peripheral neuropathy: from glucotoxicity to multidimensional metabolic imbalance Authors: Wang Tao, Yunfeng Yu, Xuerong Wu, Jiacheng He, Xiu Liu, Juan Huang, Fan Xiao, Rong Yu Published: 2026 ABSTRACT Diabetic peripheral neuropathy (DPN), particularly distal symmetric polyneuropathy, characterized by length‑dependent axonal damage, is a common chronic complication of type 2 diabetes mellitus. The pathogenesis of DPN is complicated, yet one thing is clear: long axons require a great deal of energy. When axonal transport is compromised by impaired energy metabolism, neuronal somata fall into an energy deficit that triggers neurodegeneration. While earlier work centered on hyperglycemia-induced cytotoxicity, recent studies have increasingly implicated dysregulation of glucose, lipid, and amino acid metabolism as key contributors to DPN. In this review, we integrate the anatomical organization of peripheral nerves, bioenergetic pathways, and axon–Schwann cell interactions to establish a framework for understanding how glucose, lipid, and amino acid dysregulation converge to induce bioenergetic failure in DPN. Based on these mechanisms, we further discuss novel strategies aimed at restoring metabolic homeostasis in neurons and Schwann cells. Importantly, correcting a single metabolic pathway is unlikely to halt or reverse DPN. Instead, restoring global energy homeostasis to rebalance axonal energy supply and demand may be essential for preserving peripheral nerve function. Introduction Diabetic complications are becoming an increasingly substantial burden on public health due to the increasing prevalence of diabetes worldwide. Nearly 50% of people with diabetes may eventually develop diabetic neuropathy, according to epidemiological data [1]. The clinical manifestations of diabetic neuropathy vary depending on the type of nerve fibers involved (small, large, or mixed) and the distribution of lesions (diffuse, mononeuropathy, or radiculopathy/plexopathy). The most prevalent and distinctive diffuse form of diabetic peripheral neuropathy (DPN) among these is distal symmetric polyneuropathy [2]. DPN is highly prevalent in the diabetic population, yet it remains clinically under-recognized. The prevalence of DPN in patients with type 2 diabetes mellitus (T2DM) has been reported to be 26.71% [3], whereas in Chinese patients with T2DM, it was 35.34% and exceeded 50% in those with a disease duration of more than 10 years. Of particular concern, approximately 15% of patients with newly diagnosed T2DM already exhibit overt peripheral nerve damage at the time of diagnosis [4]. Moreover, depending on the diagnostic criteria and assessment tools used, varying degrees of peripheral nerve impairment, predominantly small‑fiber involvement, can be detected in individuals with prediabetes [5]. Currently, DPN screening and diagnosis still rely heavily on nerve conduction studies, which primarily assess large‑fiber (Aα and Aβ) dysfunction. In contrast, skin biopsy for intraepidermal nerve fiber density (IENFD) and corneal confocal microscopy are objective morphological tools for evaluating small‑fiber (thinly myelinated Aδ and unmyelinated C fibers) involvement, but they are not yet widely used in clinical practice, thereby limiting early detection and diagnostic accuracy [6]. Small fibers primarily mediate pain, temperature sensation, and autonomic function; accordingly, their involvement can present early with pain, burning, and paresthesias. As the disease advances, patients may develop typical distal numbness, pain, and sensory disturbances in a “stocking-glove” distribution. In severe cases, the condition can further lead to foot ulceration and even non-traumatic amputation, markedly reducing quality of life. Current clinical management of DPN relies primarily on strict glycemic control, lifestyle modification, neurotrophic therapy, and conventional symptomatic analgesic therapy. However, accumulating clinical evidence indicates that intensified glycemic control alone is insufficient to reverse the progression of DPN, and its effect on neuropathic pain relief is limited, particularly in patients with T2DM [7]. This therapeutic impasse strongly suggests that, beyond hyperglycemia, other metabolic and bioenergetic abnormalities jointly contribute to the development and progression of DPN. In recent years, the unique anatomical and bioenergetic features of the peripheral nervous system (PNS) have offered a novel viewpoint on this enduring issue. The PNS is composed of neurons (cell bodies and axons), supporting glial cells such as Schwann cells, a microvascular network, and connective tissue. Neuronal somata are only micrometers in diameter, but their axons can extend over a meter in length. Long axons require a continuous supply of adenosine triphosphate (ATP) to sustain action potential propagation, membrane potential stability, axonal transport, and distal regenerative capacity [8]. Distal axons rely on the anterograde transport of mitochondria to maintain energy homeostasis; however, impaired mitochondrial trafficking, disrupted fission/fusion dynamics, and reduced local biogenesis can further compromise energy homeostasis. Consequently, under metabolic stress, distal axons are particularly vulnerable to length-dependent energy failure and axonal degeneration, a pattern that closely parallels the “stocking-glove” distribution characteristic of DPN [9]. At the same time, Schwann cells, which tightly ensheath axons, require highly active lipid synthesis and metabolism to maintain myelin structure and insulating function [10]. Thus, axons and Schwann cells are structurally and functionally tightly coupled, forming a metabolically interdependent “unit” with high energetic demands [11]. However, energy homeostasis within this unit is especially vulnerable in T2DM, rendering it a critical determinant of peripheral nerve injury. In light of these considerations, this review aims to integrate the latest evidence in the field to systematically elucidate the mechanisms of bioenergetic dysregulation in the PNS under diabetic conditions. We first summarize the unique energy demands and metabolic networks of the PNS. We then discuss in detail how disturbances in glucose, lipid, and amino acid metabolism disrupt axonal and Schwann cell homeostasis. Finally, we explore emerging therapeutic strategies aimed at restoring neuronal and Schwann cell metabolic homeostasis, with the goal of providing an updated theoretical framework and translational perspective for the development of potential therapies to ameliorate DPN. Anatomical and physiological basis and energy metabolism of DPN Anatomical and physiological basis The PNS comprises all nerves outside the brain and spinal cord, including 12 pairs of cranial nerves, 31 pairs of spinal nerves, and the autonomic nervous system (sympathetic and parasympathetic divisions). These nerve fibers are organized into ganglia, nerve trunks, and plexuses, forming the essential bidirectional link between the central nervous system and peripheral receptors and effectors. The PNS contains sensory, motor, and autonomic nerve fibers, each composed of axons, myelin sheaths (where present), and connective tissue layers (endoneurium, perineurium, and epineurium), together forming a complex and highly ordered conduction network. Based on the presence of myelin, nerve fibers are broadly classified into myelinated fibers (large fibers) and unmyelinated fibers (small fibers). Large fibers are formed by Schwann cells that wrap axons with multiple layers of compact myelin, and the nodes of Ranvier enable rapid saltatory conduction. Small fibers, by contrast, are embedded within the basal lamina of Schwann cells, forming Remak bundles in which multiple unmyelinated axons are ensheathed together, allowing continuous and slow signal transmission (Fig. 1) [12].Fig. 1Core anatomical organization of the PNS and metabolic derangements in DPN. A Cross-sectional anatomy of a peripheral nerve. The entire nerve is enclosed by the epineurium, individual fascicles are surrounded by the perineurium, and single nerve fibers are embedded within the endoneurium, together with the accompanying vasa nervorum. B Classification of nerve fibers. Myelinated nerve fibers feature compact myelin sheaths formed by Schwann cells, enabling rapid saltatory conduction at the nodes of Ranvier. Unmyelinated nerve fibers consist of multiple small axons embedded within a single Schwann cell to form Remak bundles, which conduct slowly and continuously. C Metabolic abnormalities. Dysregulation of glucose, lipid, and amino acid metabolism impairs glucose transport, disrupts lactate shuttling, and induces mitochondrial dysfunction. These mechanisms collectively trigger segmental demyelination, axonal degeneration, and damage to Remak bundles, ultimately leading to reduced nerve conduction velocity, axonal degeneration, and neuropathic pain Core anatomical organization of the PNS and metabolic derangements in DPN. A Cross-sectional anatomy of a peripheral nerve. The entire nerve is enclosed by the epineurium, individual fascicles are surrounded by the perineurium, and single nerve fibers are embedded within the endoneurium, together with the accompanying vasa nervorum. B Classification of nerve fibers. Myelinated nerve fibers feature compact myelin sheaths formed by Schwann cells, enabling rapid saltatory conduction at the nodes of Ranvier. Unmyelinated nerve fibers consist of multiple small axons embedded within a single Schwann cell to form Remak bundles, which conduct slowly and continuously. C Metabolic abnormalities. Dysregulation of glucose, lipid, and amino acid metabolism impairs glucose transport, disrupts lactate shuttling, and induces mitochondrial dysfunction. These mechanisms collectively trigger segmental demyelination, axonal degeneration, and damage to Remak bundles, ultimately leading to reduced nerve conduction velocity, axonal degeneration, and neuropathic pain Schwann cells are the principal glial cells of the PNS. Immature SCs derived from the neural crest can further differentiate into distinct functional states during development and after injury, including myelinating Schwann cells, non‑myelinating Schwann cells, repair Schwann cells, and perisynaptic Schwann cells [13]. Different Schwann cell subtypes exhibit marked differences in metabolic demand, functional specialization, and vulnerability, thereby accounting for their distinct patterns of involvement in DPN [14, 15]. Myelinating Schwann cells undergo extensive plasma membrane expansion to form and maintain myelin sheaths; their membrane surface area increases approximately 6,500-fold, and internodal length and myelin thickness are precisely scaled to axonal diameter [16]. Myelin is a unique membrane structure in which lipids account for 70%–80% of the dry weight of compact myelin, with cholesterol (~ 26%), glycolipids (including galactocerebroside, ~ 31%), and sphingomyelin (10%–35%) as the major lipid components [17]. Therefore, myelinating Schwann cells are particularly sensitive to abnormalities in glucose and lipid metabolism. Hyperglycemia induces Schwann cell dedifferentiation and downregulates myelin‑associated protein expression through pathways such as sorbitol accumulation, IGF-1 downregulation, and mTOR signaling inhibition, thereby promoting demyelination [18, 19]. Non‑myelinating Schwann cells primarily support the survival and metabolic homeostasis of unmyelinated small fibers and are susceptible to insulin resistance and metabolic syndrome in prediabetes, leading to small‑fiber injury and sensory abnormalities [20]. Perisynaptic Schwann cells are a specialized subset of non‑myelinating Schwann cells located at the neuromuscular junction; they support axonal health, maintain the synaptic microenvironment, and secrete neurotrophic factors, and are likewise vulnerable to metabolic and nutritional disturbances [21]. Repair Schwann cells are involved in axonal regeneration, myelin debris clearance, and remyelination; their formation requires injury-induced metabolic reprogramming and is regulated by transcription factors such as c-Jun [22]. Hyperglycemia‑induced endoplasmic reticulum stress, oxidative stress, and disturbances in glycolysis, the tricarboxylic acid (TCA) cycle, and lipid metabolism may all impede the establishment and maintenance of the repair Schwann cell phenotype [23]. During this process, some myelinating Schwann cells can dedifferentiate into an immature-like state under metabolic stress, underscoring the close link between Schwann cell phenotype switching and metabolic dysregulation [20]. It is noteworthy that, although myelin is essential for rapid nerve impulse conduction, the axon is the core structure responsible for signal propagation in the PNS, and its integrity is essential for nerve function. Existing clinical, pathological, and electrophysiological evidence indicates that DPN is fundamentally a length-dependent neuropathy characterized predominantly by axonal degeneration. Electrophysiological studies have shown that a reduction in action potential amplitude typically precedes slowing of nerve conduction velocity, suggesting that axonal dysfunction is the primary early event in DPN, whereas demyelination is largely a secondary consequence of axonal injury [24, 25]. This view is further supported by nerve biopsy findings. Multiple studies by Said, Llewelyn, Behse, and colleagues [26–28] have shown that peripheral nerve biopsies from diabetic patients commonly demonstrate decreased nerve fiber density, distal axonal atrophy, loss of small fibers, and axonal degeneration, while primary segmental demyelination is relatively rare. Clinically, some patients already exhibit reduced IENFD on skin biopsy and reduced corneal nerve fiber density before any overt abnormality is detected on conventional nerve conduction studies, indicating that small-fiber injury can occur earlier than large-fiber dysfunction [29, 30]. Therefore, DPN should be recognized as a length-dependent neuropathy characterized predominantly by distal axonal and small-fiber damage, with demyelination representing a secondary change. Energy metabolic dysfunction DPN is characterized by length‑dependent axonal degeneration, largely reflecting the high energy demand of axons that can extend over a meter in length. Even at rest, the Na⁺/K⁺ pump continuously hydrolyzes ATP to maintain the transmembrane sodium‑potassium gradient and the resting membrane potential. During excitation, action potential generation and propagation, neurotransmitter release and reuptake, and axonal transport further increase energy demand. Hyperglycemia, however, impairs Na⁺/K⁺ pump function through the polyol pathway and via specific fatty acids (e.g., C20:4 and C22:4) [31]. Pump dysfunction reduces the transmembrane sodium and decreases sodium currents, thereby compromising action potential conduction. This is consistent with altered nerve excitability observed in patients with DPN, manifesting as pain, paresthesia, muscle cramps, and fasciculations [32]. In long distal axons, anterograde mitochondrial transport and local energy supply are particularly critical, because the distal terminal must rely on mitochondria derived from the cell body to supply ATP and metabolic intermediates and maintain bioenergetic stability [33]. When substrate availability is limited or mitochondrial transport is impaired, the distal axon is the first to deplete its energy reserves and undergo degeneration. Mitochondrial dysfunction is a key amplifier of axonal injury in DPN. Beyond transport efficiency, mitochondrial biogenesis, fission/fusion dynamics, and oxidative phosphorylation capacity are all essential for maintaining axonal energy homeostasis [34]. Diabetes‑related lipid metabolic abnormalities lead to the accumulation of lipotoxic molecules such as saturated fatty acids, an altered phosphatidylethanolamine/phosphatidylcholine (PE/PC) ratio, and ceramides. These metabolic disturbances directly damage mitochondrial membrane structure and dynamics, impair ATP synthesis, and exacerbate distal axonal bioenergetic failure [35]. In axonal homeostasis, the sterile alpha and TIR motif-containing 1 (SARM1)/nicotinamide mononucleotide adenylyl transferase 2 (NMNAT2) axis is a central regulator of axonal degeneration after injury. NMNAT2 is a short-lived NAD⁺-synthesizing enzyme that relies on continuous axonal transport to replenish the distal axon, thereby maintaining local NAD⁺ levels and preventing axonal degeneration [36]. When hyperglycemia and related metabolic stress impair axonal transport, insufficient NMNAT2 supply leads to accumulation of nicotinamide mononucleotide (NMN), which in turn activates SARM1. Moreover, SIRT3-mediated deacetylation of SARM1 further enhances its NADase activity, ultimately triggering NAD⁺ depletion and axonal degeneration [37–39]. Thus, the SARM1/NMNAT2 axis may be regarded as a key molecular hub that transduces metabolic stress into structural axonal injury in DPN. In addition to axonal mitochondria, the metabolic coupling between Schwann cells and the axon is also fundamental to peripheral nerve homeostasis. Under physiological conditions, Schwann cells provide an alternative energy substrate to the axon through the “lactate shuttle”: Schwann cells take up glucose via GLUT1/3, generate pyruvate via glycolysis, convert it to lactate, and release lactate into the extracellular space via MCT1/4, where it is taken up by the axon via MCT2 [40]. Inside the axon, lactate is converted back to pyruvate, enters the TCA cycle, and fuels oxidative phosphorylation to produce ATP, thereby supporting axonal function when axonal mitochondrial capacity is limited or energy demand is high [41]. However, the metabolic homeostasis of Schwann cells is highly susceptible to insulin resistance, which in turn affects their energy support for axons. Insulin receptors (INSR) and insulin‑like growth factor‑1 receptors (IGF1R) are expressed in both peripheral neurons and Schwann cells, and they regulate axonal survival, neurite outgrowth, myelination, and metabolic homeostasis through signaling pathways such as PI3K/Akt/mTOR [42, 43]. Using Schwann cell‑specific INSR/IGF1R double‑knockout mice, Hackett et al. [44] demonstrated that loss of insulin signaling in Schwann cells leads to axonal damage and myelin thinning, accompanied by downregulation of genes involved in fatty acid and cholesterol synthesis. In a normoglycemic high-fat diet-induced insulin resistance model, Elzinga et al. [45] observed nerve conduction deficits and decreased intraepidermal nerve fiber density, indicating that insulin resistance can directly damage peripheral nerves even in the absence of overt hyperglycemia. Multi-omics studies provide direct support for this systemic metabolic derangement. Compared with controls, DPN patients and animal models show marked downregulation of genes involved in glucose and lipid metabolism in sural or sciatic nerves, including peroxisome proliferator‑activated receptor γ (PPARγ), apolipoprotein E, perilipin 1, and adiponectin. Such downregulation exacerbates energy dyshomeostasis and promotes triglyceride accumulation in the PNS [46, 47]. In db/db mice, the sciatic nerve and dorsal root ganglia (DRG) exhibit reduced expression of enzymes governing glycolysis, the TCA cycle, and lipid synthesis, alongside upregulation of lipid degradation enzymes [48]. In streptozotocin-induced diabetic rats, the sciatic nerve shows impaired glycolysis and oxidative phosphorylation as early as 12 weeks, accompanied by accumulation of branched‑chain amino acids. By 22 weeks, key TCA cycle intermediates (citrate and α‑ketoglutarate) are further depleted, while sorbitol and lactate accumulate abnormally, causing intracellular osmotic imbalance, cellular damage, and energy metabolism failure [49]. Clinical studies have similarly shown that, even with relatively good glycemic control, metabolic syndrome components such as abdominal obesity, hypertriglyceridemia, higher BMI, and an increased waist-to-hip ratio remain significantly associated with DPN risk [50–52]. In a population‑based study, Ziegler et al. [53] reported a polyneuropathy prevalence of 28.0% in patients with diabetes, compared with 13.0% in those with impaired fasting glucose and 11.3% in those with impaired glucose tolerance, indicating that metabolic abnormalities are already linked to nerve injury before the diagnostic threshold for diabetes is reached. Moreover, low HDL cholesterol and elevated triglycerides are associated with reduced sensory action potential amplitude [53, 54], and Davalos et al. [55] found that metabolic syndrome, abdominal obesity, and dyslipidemia are significantly associated with neuropathic pain, neuropathy severity, and mixed-fiber involvement. Of note, combined diet and exercise interventions significantly improved both distal and proximal IENFD in patients with impaired glucose tolerance, and the improvement in proximal IENFD correlated with reduced neuropathic pain and increased sural sensory amplitude, suggesting partial reversibility with early metabolic intervention [54]. Taken together, DPN should not be viewed simply as a complication of hyperglycemia, but rather as a systemic metabolic disorder of energy homeostasis driven by the interplay of hyperglycemia, insulin resistance, lipotoxicity, and mitochondrial dysfunction. Accordingly, the following sections focus on how disturbances in glucose, lipid, and amino acid metabolism intersect to disrupt peripheral nerve energy homeostasis and drive the pathological progression of DPN (Fig. 2).Fig. 2Schematic illustration of metabolic disturbances and axon-Schwann cell injury in DPN. Under physiological conditions, Schwann cells take up glucose via GLUT1/GLUT3, convert it to lactate through glycolysis, and release lactate via MCT1. Axons take up lactate through MCT2, convert it to pyruvate, and channel it into the mitochondrial tricarboxylic acid (TCA) cycle to efficiently generate ATP required for normal function. Under diabetic conditions, hyperglycemia impairs regulated glucose handling by Schwann cells, activates the polyol pathway, protein kinase C (PKC) signaling, and advanced glycation end-product (AGE) formation, and induces endoplasmic reticulum stress and inflammatory cytokine release. Dysregulated lipid metabolism directly damages mitochondrial function and disrupts myelin synthesis, whereas abnormal amino acid metabolism alters neuronal excitability by affecting Kv1.2 channel function, thereby exacerbating neuropathic pain. These three major metabolic disturbances converge on mitochondria, triggering reactive oxygen species (ROS) overproduction, Ca2⁺ dyshomeostasis, and apoptotic cascades, ultimately leading to axonal degeneration and segmental demyelination, the pathological hallmarks of DPN Schematic illustration of metabolic disturbances and axon-Schwann cell injury in DPN. Under physiological conditions, Schwann cells take up glucose via GLUT1/GLUT3, convert it to lactate through glycolysis, and release lactate via MCT1. Axons take up lactate through MCT2, convert it to pyruvate, and channel it into the mitochondrial tricarboxylic acid (TCA) cycle to efficiently generate ATP required for normal function. Under diabetic conditions, hyperglycemia impairs regulated glucose handling by Schwann cells, activates the polyol pathway, protein kinase C (PKC) signaling, and advanced glycation end-product (AGE) formation, and induces endoplasmic reticulum stress and inflammatory cytokine release. Dysregulated lipid metabolism directly damages mitochondrial function and disrupts myelin synthesis, whereas abnormal amino acid metabolism alters neuronal excitability by affecting Kv1.2 channel function, thereby exacerbating neuropathic pain. These three major metabolic disturbances converge on mitochondria, triggering reactive oxygen species (ROS) overproduction, Ca2⁺ dyshomeostasis, and apoptotic cascades, ultimately leading to axonal degeneration and segmental demyelination, the pathological hallmarks of DPN DPN and glucose metabolism The PNS is metabolically highly active, requiring a continuous energy supply to sustain neuronal excitability, axonal transport, and chemical signaling. Glucose serves as the primary fuel for these processes. Zhao et al. [55] reported marked alterations in key serum metabolites, including glucose‑6‑phosphate, isocitrate, and glycerate, in both patients with DPN and rat models. Among the affected pathways, the TCA cycle showed the most pronounced changes, followed by the pentose phosphate pathway and glycolysis. Collectively, these abnormalities in glucose metabolism lead to insufficient ATP production and energy supply to the distal axon, while excessive accumulation of metabolic byproducts further provokes oxidative stress and cellular injury. In the following sections, we discuss how abnormalities in glucose metabolism mechanistically affect the energetic demands of axons, Schwann cells, and myelin maintenance (Fig. 3).Fig. 3Overview of the mechanisms by which dysregulated glucose metabolism contributes to DPN. In Schwann cells, excess glucose is diverted into pathogenic side pathways, including the polyol, PKC, and AGE pathways, thereby triggering endoplasmic reticulum (ER) stress and inflammation. Meanwhile, reduced GLUT1/3 expression limits glucose uptake, and impairment of MCT1-mediated lactate shuttling deprives axons of critical metabolic support. Within axons, hyperglycemia suppresses mitochondrial biogenesis through the AMPK/SIRT1/PGC-1α axis. Pathological activation of SIRT3 promotes SARM1 deacetylation, leading to NAD⁺ and ATP depletion. This bioenergetic crisis, together with ROS overproduction, ultimately drives demyelination and axonal degeneration. AGEs advanced glycation end-products, AMPK AMP-activated protein kinase, ER endoplasmic reticulum, GLUT1/3 glucose transporter 1/3, MCT1/2 monocarboxylate transporter 1/2, NAD⁺ oxidized nicotinamide adenine dinucleotide, PGC-1α peroxisome proliferator-activated receptor gamma coactivator 1-alpha, PKC protein kinase C, ROS reactive oxygen species, SCs Schwann cells, SIRT1/3 sirtuin 1/3, TRPV1 transient receptor potential vanilloid 1 Overview of the mechanisms by which dysregulated glucose metabolism contributes to DPN. In Schwann cells, excess glucose is diverted into pathogenic side pathways, including the polyol, PKC, and AGE pathways, thereby triggering endoplasmic reticulum (ER) stress and inflammation. Meanwhile, reduced GLUT1/3 expression limits glucose uptake, and impairment of MCT1-mediated lactate shuttling deprives axons of critical metabolic support. Within axons, hyperglycemia suppresses mitochondrial biogenesis through the AMPK/SIRT1/PGC-1α axis. Pathological activation of SIRT3 promotes SARM1 deacetylation, leading to NAD⁺ and ATP depletion. This bioenergetic crisis, together with ROS overproduction, ultimately drives demyelination and axonal degeneration. AGEs advanced glycation end-products, AMPK AMP-activated protein kinase, ER endoplasmic reticulum, GLUT1/3 glucose transporter 1/3, MCT1/2 monocarboxylate transporter 1/2, NAD⁺ oxidized nicotinamide adenine dinucleotide, PGC-1α peroxisome proliferator-activated receptor gamma coactivator 1-alpha, PKC protein kinase C, ROS reactive oxygen species, SCs Schwann cells, SIRT1/3 sirtuin 1/3, TRPV1 transient receptor potential vanilloid 1 Glucose uptake Normal function of the PNS relies on a continuous and efficient supply of glucose. This process requires the coordinated action of multiple membrane transporters, including GLUTs and MCTs. In myelinated nerve fibers, GLUT1 is enriched in metabolically active regions, particularly the nodes of Ranvier, as well as the plasma membrane and cytoplasm of Schwann cells surrounding the Schmidt-Lanterman incisures (SLIs). In contrast, GLUT3 is predominantly localized in Schwann cells and the paranodal axolemma [56]. SLIs are specialized structures formed during the wrapping of axons by myelinating Schwann cells in myelinated fibers. By maintaining cytoplasmic continuity between the inner and outer layers of the myelin sheath, they contribute to metabolic exchange and support saltatory conduction [57]. Given their high metabolic demand, these regions are particularly rich in transport proteins. Likewise, MCT1, which mediates the shuttling of lactate, acetoacetate, and pyruvate, is highly expressed in both myelin and SLIs [58]. Different MCT isoforms display distinct affinities for lactate: MCT1 has intermediate affinity and can both import and export lactate under physiological conditions, whereas MCT2 (high affinity) and MCT4 (low affinity) primarily mediate lactate uptake and efflux, respectively. This isoform specialization ensures that axonal metabolic demands for lactate are met [58]. Hyperglycemia and insulin resistance alter the expression or activity of these transporters. Consistent findings from clinical and animal model studies indicate that GLUT1, GLUT3, and MCT1 are downregulated in multiple cell types, including erythrocytes, cardiomyocytes, skeletal muscle fibers, and PNS cells, in diabetic patients and rat models [59–61]. Szabó et al. [60] reported that elevated HbA1c levels were significantly associated with reduced membrane expression of GLUT1, GLUT3, and MCT1 in erythrocytes. However, the mechanisms underlying this downregulation remain unclear. One plausible explanation is that this downregulation represents a compensatory response to hyperglycemia, limiting excessive glucose influx into neurons and Schwann cells and consequently attenuating downstream glucotoxic effects [62]. In parallel, chronic inflammation and insulin resistance, such as elevated interferon‑γ levels, may further suppress transporter expression, thereby impairing the reparative capacity of Schwann cells by reducing glucose uptake [63]. Of particular importance, the reduction of MCT1 directly disrupts the lactate shuttle-mediated metabolic coupling between axons and myelinating SCs, forcing the axon, especially the distal axon, to rely more heavily on its own limited energy supply and thereby increasing its susceptibility to degeneration. In streptozotocin‑induced diabetic mice, MCT1 expression in PNS and DRG was significantly reduced as early as 2 weeks after disease induction. More importantly, mice with heterozygous MCT1 knockout developed more severe neuropathy than wild-type diabetic mice, characterized by greater axonal loss and demyelination, slower nerve conduction velocities, and more profound loss of mechanical sensitivity [64, 65]. Therefore, GLUT1, GLUT3, and MCT1 may serve as potential biomarkers of early metabolic uncoupling in DPN, and restoring the axon-Schwann cell lactate shuttle represents a novel strategy to improve distal axonal bioenergetic support. Glycolytic metabolism Glycolysis, the central pathway of glucose catabolism, is essential for PNS energy homeostasis. Under physiological conditions, pyruvate derived from glycolysis enters mitochondria to fuel the TCA cycle, supporting efficient ATP production and normal nerve function. In diabetes, persistent hyperglycemia and insulin resistance disrupt this process, making glycolytic dysregulation a key pathogenic driver of DPN [66]. First, diabetic complications often involve reduced pyruvate availability or impaired pyruvate utilization, which may represent an early link between hyperglycemia and neuronal bioenergetic failure [67]. In vitro studies show that pyruvate deprivation leads to SC and neuronal death, whereas pyruvate supplementation enhances TCA cycle flux and promotes ATP production in diabetic sciatic nerves [68]. Second, excessive glucose uptake in peripheral nerve tissues activates deleterious glycolytic branch pathways, including the polyol pathway, protein kinase C signaling, and AGE formation. These pathways converge to trigger ER stress, oxidative stress, and chronic inflammation [69]. In addition, when glycolytic flux exceeds the TCA cycle capacity in T2DM, oxidative phosphorylation is impaired, resulting in inadequate ATP production and abnormal accumulation of metabolic intermediates such as ketone bodies, lactate, and acylcarnitines [70]. In distal axons, excessive accumulation of these intermediates can damage mitochondria, induce excessive fission, trigger ROS overproduction, and disrupt calcium homeostasis, ultimately compromising axonal structure and function [71, 72]. Among these processes, the role of lactate is particularly complex and has garnered considerable attention. Physiologically, lactate serves as an energy substrate shuttled from Schwann cells to axons [73]. For example, conditional knockout of the glycolytic enzyme PKM2 in myelinating Schwann cells reduces lactate production, leading to insufficient axonal lactate supply after repetitive electrical stimulation [74]. Jia et al. [75] further reported that knockout of Rheb, a regulator of energy metabolism in Schwann cells, resulted in increased lactate production, which in turn led to elevated axonal ROS, mitochondrial damage, and decreased axonal stability. Notably, inhibition of glycolysis with 2-deoxy-D-glucose or blockade of lactate transport into axons with an MCT1 inhibitor reversed these changes. On the other hand, whether pathologically excessive lactate production is harmful to axons remains debated. Rahman et al. [76] found that lactate exposure in cultured DRG neurons upregulated pain‑related ion channels and reduces neuronal viability. Moreover, knockout of pyruvate dehydrogenase kinase 2/4 inhibited diabetes-induced lactate accumulation, reduced expression of pain-related ion channels and macrophage infiltration in the DRG, and alleviated DPN-associated pain. However, Beirowski et al. [77] observed that simply increasing lactate concentration in DRG culture systems did not induce overt axonal degeneration, suggesting that the primary injury is not lactate per se but the broader metabolic dysregulation accompanying its elevation. Therefore, whether lactate accumulation independently damages the PNS remains to be fully determined. Oxidative phosphorylation Mitochondrial oxidative phosphorylation is the principal bioenergetic pathway in neurons, generating large amounts of ATP via the electron transport chain to support electrophysiological activity, axonal transport, and other energy-demanding processes [78]. In diabetes, impaired insulin/IGF-1 signaling drives an aberrant metabolic phenotype that suppresses oxidative phosphorylation and shifts metabolism toward glycolysis, thereby limiting ATP supply to distal axons and promoting distal axonal degeneration [79]. Mitochondrial dysfunction, characterized by inadequate ATP production and excessive ROS generation, is a central pathological mechanism in the onset and progression of DPN [80]. Classical studies have shown that hyperglycemia inhibits the AMPK/SIRT/PGC-1α signaling axis. Downregulation of this axis reduces mitochondrial biogenesis, decreases the catalytic activity of respiratory chain complexes, and destabilizes the mitochondrial membrane potential, ultimately impairing oxidative phosphorylation efficiency [79]. When intracellular glucose overload coexists with excessive fatty acid flux, surplus NADH and FADH₂ overload the electron transport chain, creating a state of metabolic congestion that elevates the proton gradient and further suppresses ATP synthesis [81]. At the same time, an overloaded electron transport chain becomes prone to electron leakage, promoting excessive ROS production and triggering pronounced oxidative stress [82]. Excessive ROS not only directly damage mitochondrial DNA, respiratory chain protein complexes, and membrane potential, but also impair structural proteins and lipid membranes in SCs and axons, thereby aggravating neural tissue damage [83, 84]. These changes significantly compromise energy supply to distal axons. Conversely, enhancing mitochondrial biogenesis, activating ROS‑scavenging systems, or improving mitochondrial transport capacity may alleviate injury and promote neural repair [83]. Xu et al. [85] demonstrated that satellite glial cells transfer mitochondria to the DRG via myosin‑10 (MYO10)‑mediated tunneling nanotubes in vitro, ex vivo, and in vivo models, thereby alleviating neuropathic pain and suggesting a novel therapeutic strategy for DPN. DPN and lipid metabolism Lipids are not only essential structural components of neuronal and Schwann cell membranes and myelin, but also serve as key metabolic substrates for maintaining axonal mitochondrial function and peripheral nerve homeostasis. Compared with abnormalities in glucose metabolism, the impact of lipid metabolic disturbances in DPN is more directly reflected in Schwann cell dysfunction and aberrant myelin lipid remodeling, which in turn compromise Schwann cell metabolic support for axons and ultimately lead to axonal degeneration [86]. Cohort studies have shown that the lipidomic profiles of patients with T2DM and DPN differ significantly from those of patients with diabetes alone, primarily characterized by reduced acylcarnitines and sphingolipids and increased free fatty acids and ceramides [87, 88]. In individuals with obesity, specific lipid subclasses such as diacylglycerols, sphingomyelins, and ceramides could effectively distinguish between those with and without DPN [89]. In addition, 1‑deoxydihydroceramide levels are inversely correlated with intraepidermal nerve fiber density in the legs of patients with type 2 diabetes, suggesting that specific lipotoxic molecules may be closely associated with small‑fiber injury [90]. These clinical findings indicate that lipid metabolic abnormalities are closely linked to the development and progression of DPN and may also serve as potential biomarkers. To determine whether lipid metabolism contributes to nerve injury independently of hyperglycemia, Obrosova et al. [91] first established a high-fat diet-induced model of peripheral neuropathy. This model exhibited features consistent with STZ-induced DPN, including sensory nerve conduction deficits, tactile allodynia, and thermal hypoalgesia. However, pathological features such as epidermal nerve fiber loss and axonal atrophy were relatively mild. Notably, switching these mice to a normal diet for six weeks reversed thermal hypoalgesia and restored sensory nerve conduction velocity, although motor nerve conduction velocity remained unaffected [91]. These observations suggest that lipid metabolic dysregulation predominantly damages peripheral nerves by impairing Schwann cell lipid remodeling and metabolic support to axons, thereby leading to distal axonal energy failure and degeneration (Fig. 4).Fig. 4Summary of lipid metabolic dysregulation in Schwann cells and axons in DPN. In Schwann cells, downregulation of Maf and p75NTR suppresses the expression of the rate-limiting enzyme HMGCR via the SCAP/SREBP2 pathway, thereby reducing myelin cholesterol synthesis and ultimately triggering demyelination. In the axonal microenvironment, exogenous long-chain SFAs are transported via CD36, activate the SREBP1/SCAP/FASN pathway to promote de novo fatty acid synthesis, and are esterified by DGAT2 into triglycerides for storage in lipid droplets. MUFAs can likewise be sequestered into lipid droplets through FATP/FABP-mediated handling, whereas excess free SFAs that exceed lipid droplet buffering capacity induce lipotoxicity and mitochondrial injury. In parallel, SFAs promote ceramide accumulation through the SPT pathway. The combined mitochondrial toxicity caused by ceramide accumulation, PC/PE imbalance, and impaired β-oxidation is characterized by downregulation of the mitochondrial fusion factors Mfn1/2, upregulation of the fission factors Mff and Drp1, and activation of Bax/Bak-mediated MOMP, ultimately culminating in axonal degeneration. Bax/Bak Bcl-2-associated X protein/Bcl-2 antagonist killer. CD36 cluster of differentiation 36, DGAT2 diacylglycerol O-acyltransferase 2, Drp1 dynamin-related protein 1, FABP fatty acid-binding protein, FASN fatty acid synthase, FATP fatty acid transport protein, HMGCR 3-hydroxy-3-methylglutaryl-CoA reductase, MUFAs monounsaturated fatty acids, Mff mitochondrial fission factor, Mfn1/2 mitofusin ½, MOMP mitochondrial outer membrane permeabilization, p75NTR p75 neurotrophin receptor, PC phosphatidylcholine, PE phosphatidylethanolamine, SREBP1/2 sterol regulatory element-binding protein ½, SCAP SREBP cleavage-activating protein, SCs Schwann cells, SFAs saturated fatty acids, SPT serine palmitoyltransferase, TG triglyceride Summary of lipid metabolic dysregulation in Schwann cells and axons in DPN. In Schwann cells, downregulation of Maf and p75NTR suppresses the expression of the rate-limiting enzyme HMGCR via the SCAP/SREBP2 pathway, thereby reducing myelin cholesterol synthesis and ultimately triggering demyelination. In the axonal microenvironment, exogenous long-chain SFAs are transported via CD36, activate the SREBP1/SCAP/FASN pathway to promote de novo fatty acid synthesis, and are esterified by DGAT2 into triglycerides for storage in lipid droplets. MUFAs can likewise be sequestered into lipid droplets through FATP/FABP-mediated handling, whereas excess free SFAs that exceed lipid droplet buffering capacity induce lipotoxicity and mitochondrial injury. In parallel, SFAs promote ceramide accumulation through the SPT pathway. The combined mitochondrial toxicity caused by ceramide accumulation, PC/PE imbalance, and impaired β-oxidation is characterized by downregulation of the mitochondrial fusion factors Mfn1/2, upregulation of the fission factors Mff and Drp1, and activation of Bax/Bak-mediated MOMP, ultimately culminating in axonal degeneration. Bax/Bak Bcl-2-associated X protein/Bcl-2 antagonist killer. CD36 cluster of differentiation 36, DGAT2 diacylglycerol O-acyltransferase 2, Drp1 dynamin-related protein 1, FABP fatty acid-binding protein, FASN fatty acid synthase, FATP fatty acid transport protein, HMGCR 3-hydroxy-3-methylglutaryl-CoA reductase, MUFAs monounsaturated fatty acids, Mff mitochondrial fission factor, Mfn1/2 mitofusin ½, MOMP mitochondrial outer membrane permeabilization, p75NTR p75 neurotrophin receptor, PC phosphatidylcholine, PE phosphatidylethanolamine, SREBP1/2 sterol regulatory element-binding protein ½, SCAP SREBP cleavage-activating protein, SCs Schwann cells, SFAs saturated fatty acids, SPT serine palmitoyltransferase, TG triglyceride Cholesterol metabolism Cholesterol is a critical structural lipid of the peripheral nerve myelin sheath, and its homeostasis is essential for axonal insulation, saltatory conduction, and nerve impulse propagation. A long-term follow-up study showed that, even after adjustment for baseline glycated hemoglobin (HbA1c) and diabetes duration, serum total cholesterol (TC; OR = 1.26, 95% CI 1.10–1.45), low-density lipoprotein cholesterol (LDL-C; OR = 1.22, 95% CI 1.03–1.45), and triglycerides (TG; OR = 1.35, 95% CI 1.16–1.57) remained positively associated with DPN over a 7.3-year follow-up period [92]. However, whether serum TC is an independent risk factor for DPN remains controversial, as different meta-analyses have yielded inconsistent results [93, 94]. The underlying pathophysiology of this association is multifactorial. Cholesterol, a key component of myelin, is synthesized de novo by Schwann cells rather than relying on uptake from circulation. Disruption in this synthetic process is a core mechanism underlying myelin damage in DPN. The rate-limiting enzyme of cholesterol biosynthesis, 3-hydroxy-3-methylglutaryl-CoA reductase, and the neurotrophin receptor p75, which regulates cholesterol metabolism in myelinating Schwann cells, directly impair Schwann cell cholesterol synthesis. This in turn leads to thinner and loosely compacted myelin, severely disrupting saltatory conduction and impairing nerve impulse propagation [95, 96]. Rescue experiments in mice further validated the critical role of this mechanism. Restoration of the cholesterol synthesis-related protein DHCR7 in p75-deficient Schwann cells significantly attenuated DRG neuron loss [97]. In the setting of Schwann cell insulin resistance, the PI3K/Akt/mTOR-SREBP signaling axis downregulates key enzymes involved in fatty acid and cholesterol biosynthesis, leading to myelin thinning and altered myelin lipid composition [44, 98]. In addition, the transcription factor Maf is a key regulator coordinating Schwann cell myelination and cholesterol biosynthesis. Maf is highly expressed during the critical phase of myelination initiation, and its expression is tightly coupled to the progression of myelin formation. Accordingly, Maf-deficient mice exhibit marked hypomyelination, largely attributable to impaired cholesterol biosynthetic capacity in Schwann cells [99]. These findings underscore the pivotal role of transcriptional regulation of cholesterol metabolism in DPN-associated myelin injury. Triglyceride metabolism Multiple studies have established a positive correlation between elevated triglyceride levels in serum and sciatic nerve tissues and decreased nerve conduction velocity, reduced myelinated fiber density, and other objective neuropathic markers [100–102]. Notably, the neurotoxic effects of triglycerides depend strongly on the carbon chain length and degree of saturation of the fatty acids they contain. Research by O’Brien et al. [103] revealed that hypertriglyceridemia induced by high-fat diets is associated with elevated levels of SFAs, such as palmitic acid (C16:0) and stearic acid (C18:0), as well as unsaturated fatty acids like oleic acid (C18:1) and linoleic acid (C18:2). These lipid changes are closely linked to the degree of myelinated fiber loss. Mechanistically, these observations have been validated at the cellular level. Rumora et al. [104, 105] showed that treating DRG neurons with long-chain SFAs, including palmitate (C16:0) and stearate (C18:0), led to a decline in mitochondrial membrane potential, decreased ATP production, abnormal mitochondrial morphology, and ultimately activation of caspase-3/7-dependent apoptosis. In contrast, shorter-chain SFAs, such as laurate (C12:0) and myristate (C14:0), exerted no appreciable effect on axonal mitochondrial transport. Under high-fat diet conditions, non-myelinated Sewan cells are more susceptible to lipid metabolic disorders, which is why lipotoxicity affects the metabolic support of amyloid small fibers earlier [20]. Interestingly, monounsaturated fatty acids such as oleic acid confer neuroprotection by promoting lipid droplet formation within axons, thereby preserving mitochondrial function and transport while counteracting the detrimental effects of SFAs [105]. These differences are attributed to the structural properties of long-chain SFAs, which lack double bonds and exhibit lower mitochondrial metabolic efficiency. This inefficiency leads to the accumulation of metabolic intermediates, disruption of the electron transport chain, and oxidative stress [106]. Clinically, fatty acid profiling combined with routine lipid testing could be used for DPN risk stratification to identify individuals with abnormally elevated long‑chain saturated fatty acids. Activation of long-chain fatty acid metabolism requires upregulation of long-chain acyl-CoA synthetase 1 (ACSL1). In Schwann cells, ACSL1 upregulation under conditions of elevated long-chain fatty acids alleviates oxidative stress, mitochondrial dysfunction, incomplete β-oxidation, and Schwann cell injury [107]. Triglyceride metabolism is also subject to coordinated regulation by multiple enzymes. Among them, diacylglycerol acyltransferase 2 (DGAT2) catalyzes the final and irreversible step of triglyceride synthesis. Its protein expression across different strata of metabolic disease severity has been linked to the pathogenesis of peripheral neuropathy, suggesting that DGAT2 may represent a potential therapeutic target [103]. Additionally, both de novo fatty acid synthesis and fatty acid uptake profoundly influence the homeostasis of neural lipids. Fatty acid synthase, the key enzymatic complex responsible for the de novo synthesis of medium- and long-chain fatty acids up to 16 carbons in length, is regulated by sterol regulatory element-binding protein 1 and its cleavage-activating protein and is closely associated with developmental myelination [17]. In Schwann cells, uptake of circulating fatty acids is mediated primarily by cluster of differentiation 36 (CD36), fatty acid transport proteins (FATPs), and fatty acid-binding proteins (FABPs). By incorporating fatty acids into myelin, these transporters profoundly influence myelin lipid composition and functional stability. Recent studies suggest that fatty acid transport mediated by CD36, FATPs, and FABPs is dysregulated in peripheral neuropathy associated with prediabetes and T2DM, highlighting these pathways as potential molecular targets [108, 109]. Interestingly, the latest evidence indicates that inhibition of FABP5 and FABP7 does not alter hyperglycemia-induced metabolic changes in Schwann cells, suggesting that these proteins may promote oxidative stress independently of metabolic-state-driven signaling [110]. Nevertheless, direct evidence that targeted modulation of these transporters can improve or reverse DPN remains lacking, making this an important area for future investigation. Glycerophospholipid metabolism Glycerophospholipids are crucial components of peripheral nerve myelin and mitochondrial membranes. According to the hydrophilic head group linked to the phosphate moiety, glycerophospholipids are classified into PC, PE, cardiolipin (CL), and related subclasses. Among them, PC is one of the most abundant phospholipids in myelin. A Mendelian randomization study suggested that specific PC species, such as PC (16:0_20:2) and PC (16:1_18:1), may protect against diabetic neuropathy [111]. In a cohort study, Rumora et al. [88] found that patients with T2DM and DPN had elevated PE levels but reduced PC and CL levels. Preclinical studies have demonstrated that specific lipid species in the sciatic nerve are lost at an early stage of the disease, including significant reductions in cardiolipin, galactocerebroside, sulfatide, sphingomyelin, the two major plasmalogen forms of phosphatidylethanolamine (PE; P18:0‑18:1 and P18:1‑18:1), lysophospholipids (LPC and LPE), and plasmenyl-PE; these changes occur prior to structural damage to mitochondria, myelin, and axons [35, 112]. Given that PC and PE are the predominant phospholipids in mitochondrial membranes, particularly in the inner mitochondrial membrane, even subtle changes in their relative abundance may compromise membrane integrity and fluidity. This, in turn, destabilizes respiratory chain complexes, impairs oxidative phosphorylation, and ultimately contributes to peripheral nerve injury [113]. CL, a signature phospholipid of the inner mitochondrial membrane, is similarly indispensable for the assembly and function of electron transport chain supercomplexes; thus, its depletion directly disrupts mitochondrial bioenergetics [114]. Therefore, an elevated PE/PC ratio and decreased cardiolipin levels may serve as potential lipid metabolic biomarkers for DPN. In a Trx2-deficient model of mitochondria-driven lipid dysregulation, Schwann cells exhibited particularly pronounced disturbances in glycerophospholipid and sphingolipid homeostasis, supporting the notion that intrinsic lipid metabolic defects in glial support cells may also contribute to nerve injury [115]. However, direct functional studies demonstrating a causal role for this pathway in peripheral nerve tissue or cell-based models remain lacking. How phospholipid dysregulation specifically affects mitochondrial dynamics, respiratory chain activity, and energy supply in axons and Schwann cells, and whether these alterations constitute an independent driving force in DPN, remain to be further clarified experimentally. Sphingolipid metabolism Sphingolipids are a class of complex lipids characterized by a sphingoid base backbone, including ceramides, sphingomyelins, and glycosphingolipids, and are essential components of nerve cell membranes. Ceramide sits at the metabolic hub of sphingolipid metabolism, and its dysregulation has been closely linked to DPN [116]. Multiple clinical studies have shown that patients with DPN have higher plasma levels of deoxysphingolipids (e.g., deoxy‑C24‑ceramide), C24‑ and C26‑ceramides, and cysteine than patients without neuropathy; moreover, 1‑deoxydihydroceramide levels are negatively correlated with intraepidermal nerve fiber density in the leg [117, 118]. These findings suggest that specific ceramide subspecies have potential as diagnostic and prognostic biomarkers for DPN. Mechanistically, sphingolipid dysregulation impairs axonal mitochondrial transport and local energy supply mainly through disruption of mitochondrial dynamics, induction of oxidative stress, and activation of apoptotic pathways, thereby contributing to peripheral nerve injury. Ceramides and diacylglycerol downregulate mitochondrial fusion proteins and upregulate fission-associated proteins, promoting mitochondrial fragmentation [119]. In particular, C16:0 ceramide synthesized by ceramide synthase 6 (CerS6) and C18:0 ceramide generated by CerS1 drive mitochondrial fission through interaction with mitochondrial fission factor [120]. By contrast, the sphingosine kinase 2 (SphK2)/S1P axis appears to play a protective role in maintaining mitochondrial dynamic balance. S1P generated by SphK2 binds to prohibitin 2, stabilizing respiratory chain complexes and improving oxidative phosphorylation efficiency [121]. In addition, ceramides increase mitochondrial outer membrane permeabilization (MOMP), trigger cytochrome c release, and activate the caspase cascade[122]. This process not only leads to neuronal cell death but also induces apoptosis in both non‑myelinating and myelinating Schwann cells, thereby compromising their metabolic and structural support for axons [123]. Notably, ceramides with different acyl‑chain lengths, synthesized by distinct CerS isoforms, exhibit functional heterogeneity. C16:0 ceramide synthesized by CerS5 promotes neuronal apoptosis by increasing MOMP sensitivity, whereas C24:1/C24:0 ceramides produced by CerS2 may exert prosurvival effects through stabilization of mitochondrial membranes [124]. However, excessive CerS2 expression leads to the accumulation of very‑long‑chain ceramides, mitochondrial dysfunction, oxidative stress, and compensatory mitophagy [125]. Meanwhile, ceramide and diacylglycerol activate NADPH oxidase and damage respiratory chain complexes, leading to excessive ROS generation and lipid peroxidation, producing neurotoxic products such as 4‑hydroxynonenal and malondialdehyde, thereby disrupting membrane fluidity [126]. Animal studies have shown that CerS2 knockout mice, due to blockade of very‑long‑chain ceramide synthesis, accumulate sphingoid precursors, resulting in an increased proportion of C16:0 ceramide, inhibition of mitochondrial complex IV activity, loss of membrane potential, and increased ROS production [127]. Conversely, inhibiting serine palmitoyltransferase, the rate-limiting enzyme in de novo ceramide synthesis, or protein phosphatase 2A reduces mitochondrial ROS levels and apoptosis [128]. Overall, sphingolipid imbalance, by continuously impairing mitochondrial homeostasis and reducing axonal energy supply, is an important mechanism in the development and progression of DPN. DPN and amino acid metabolism As an integral component of the bioenergetic network, amino acid metabolism has attracted increasing attention in the context of DPN. Emerging evidence suggests that alterations in circulating amino acid profiles are associated not only with the risk of developing DPN, but also with the direct regulation of nerve injury and pain through multiple mechanisms (Fig. 5).Fig. 5Amino acid metabolic dysregulation drives neuropathic pain through mitochondrial bioenergetic failure and ion channel remodeling. In the spinal dorsal horn, excessive extracellular glutamate accumulation leads to overactivation of postsynaptic NMDA and AMPA receptors, thereby amplifying nociceptive signaling. At the same time, depletion of essential amino acids, including arginine, tyrosine, glycine, and serine, deprives neurons of key metabolic substrates. This shortage results in mitochondrial dysfunction, reduced ATP production, and cellular bioenergetic failure, further disrupting neuronal homeostasis. In DRG neurons, deficiency of branched-chain amino acids (BCAAs; leucine, isoleucine, and valine) triggers a stress response characterized by upregulation of ATF4. Increased ATF4 expression enhances the expression of the amino acid transporter LAT1. This maladaptive response, potentially accompanied by accumulation of specific metabolites such as 4-methyl-2-oxopentanediol, may lead to downregulation or dysfunction of the voltage-gated potassium channel Kv1.2. Loss of Kv1.2 lowers the threshold for action potential generation, resulting in neuronal hyperexcitability and thermal/mechanical allodynia. AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, ATP adenosine triphosphate, ATF4 activating transcription factor 4, BCAA branched-chain amino acid, DRG dorsal root ganglion, Kv1.2 voltage-gated potassium channel 1.2, LAT1 L-type amino acid transporter 1, NMDA N-methyl-D-aspartate Amino acid metabolic dysregulation drives neuropathic pain through mitochondrial bioenergetic failure and ion channel remodeling. In the spinal dorsal horn, excessive extracellular glutamate accumulation leads to overactivation of postsynaptic NMDA and AMPA receptors, thereby amplifying nociceptive signaling. At the same time, depletion of essential amino acids, including arginine, tyrosine, glycine, and serine, deprives neurons of key metabolic substrates. This shortage results in mitochondrial dysfunction, reduced ATP production, and cellular bioenergetic failure, further disrupting neuronal homeostasis. In DRG neurons, deficiency of branched-chain amino acids (BCAAs; leucine, isoleucine, and valine) triggers a stress response characterized by upregulation of ATF4. Increased ATF4 expression enhances the expression of the amino acid transporter LAT1. This maladaptive response, potentially accompanied by accumulation of specific metabolites such as 4-methyl-2-oxopentanediol, may lead to downregulation or dysfunction of the voltage-gated potassium channel Kv1.2. Loss of Kv1.2 lowers the threshold for action potential generation, resulting in neuronal hyperexcitability and thermal/mechanical allodynia. AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, ATP adenosine triphosphate, ATF4 activating transcription factor 4, BCAA branched-chain amino acid, DRG dorsal root ganglion, Kv1.2 voltage-gated potassium channel 1.2, LAT1 L-type amino acid transporter 1, NMDA N-methyl-D-aspartate Non-branched-chain amino acids Xu et al. [129] found that levels of arginine and tyrosine are positively correlated with nerve conduction velocity in patients with DPN, whereas glutamate shows a negative correlation. This observation aligns with findings from Ganz et al. [130], who reported that low circulating arginine levels could predict DPN risk and that L-arginine supplementation effectively prevents mechanical hypersensitivity and thermal hyperalgesia in STZ-induced DPN models [131]. Moreover, metabolic profiling of DRG neurons from animal models of DPN highlights significant enrichment in pathways such as arginine and proline metabolism, aspartate metabolism, glycine and serine metabolism, and glutamate metabolism, with marked reductions in levels of glycine and non-proteinogenic amino acids [132]. Although direct evidence linking amino acid depletion to neuronal injury and pathological pain remains limited, disturbances in amino acid availability are known to be important contributors to mitochondrial dysfunction, bioenergetic crisis, and impaired cellular repair after injury [133]. In contrast, glutamate, the principal excitatory neurotransmitter in the CNS, may exacerbate neuropathic pain when it accumulates excessively in peripheral tissues by activating glutamate receptors on dorsal horn neurons and amplifying nociceptive transmission [134]. Branched-chain amino acids Elevated circulating levels of branched-chain amino acids (BCAAs), including leucine, isoleucine, and valine, are associated with an approximately fivefold increased risk of diabetes compared with that in metabolically healthy individuals [135]. Interestingly, among individuals already diagnosed with T2DM, those with DPN have significantly lower circulating BCAA levels than those without neuropathy [136]. Zhou et al. [136] demonstrated that BCAA deficiency exacerbates DPN phenotypes in animal models, a mechanism linked to increased expression of the L-type amino acid transporter (LAT1) via activating transcription factor 4. Abnormally elevated LAT1 interacts with the voltage-gated potassium channel 1.2 (Kv1.2), inhibiting its membrane localization, resulting in a reduced number of functional membrane-localized Kv1.2 channels, heightened neuronal excitability, and ultimately enhanced neuropathic pain. Consistent with this observation, levels of 4-methyl-2-oxopentanoate, a leucine-derived metabolite involved in BCAA metabolism, are also significantly reduced in patients with T2DM complicated by DPN [88]. Regulation of metabolism in DPN Dysregulated energy metabolism drives oxidative stress, mitochondrial dysfunction, and peripheral nerve injury through multiple targets and pathways. However, effective interventions for neuropathy associated with metabolic abnormalities remain limited. At present, therapeutic strategies fall into two main categories: pharmacological interventions and lifestyle-based interventions (as shown in Table 1). Table 1Summary of metabolic intervention strategies for DPNIntervention categorySpecific interventionPrimary mechanisms of actionKey effects on DPNPharmacological interventionsAldose reductase inhibitors (epalrestat [137], compound 15c [182])1. Reduced sorbitol accumulation, and inhibited polyol pathway2. Improved axonal sodium currents, restored Na⁺/K⁺ pump function and reversed the dedifferentiation state of myelinating Schwann cells under high glucose conditions1. Repaired structure of myelin and axonal structure2. Improved nerve conduction velocityAdiponectin receptor agonist (AdipoRon) [139]1. Reduced systemic oxidative stress and insulin resistance2. Upregulated MCT1/2/4 and enhances lactate and ATP transport to axons1. Restored energy supply between myelinating Schwann cells and axons2. Reduced neuronal apoptosisHMG-CoA reductase inhibitors (atorvastatin, ezetimibe/simvastatin, rosuvastatin [145, 147, 148, 183])1. Inhibited HMG-CoA reductase, and lowered serum cholesterol2. Reduced total ceramide, sphingomyelin and other sphingolipid metabolites3 Activated glutathione antioxidant system, and reduced lipid peroxidation1. Ameliorated oxidative stress in sciatic nerve2. Alleviated DPN severity and symptoms3. Improved nerve conduction velocityPPARγ agonists (rosiglitazone, pioglitazone [140, 141])1. Reduced oxidative stress and prevented thermal hypoalgesia2. Improved evoked pain behavior3.Decreased TNF-α and IL-1β levels in sciatic nerve4.Downregulated Nav1.7 channel expression1. Improved morphology and function of unmyelinated nerve fibers2. Shortened thermal latency, and increased nerve conduction velocityPPARγ agonist compounds (gentiopicroside [184], magnolol [185], lycorine [186])1. Modulated PPAR-γ/AMPK/ACC, PPAR-γ/AMPK/PGC-1α, and PPARγ/NF-κB signaling pathways2. Regulated lipid metabolism, improved mitochondrial function, and promoted neural blood flowImproved myelin sheath structure in sciatic nerve and intraepidermal nerve fiber density in hind paw, and alleviated hyperalgesiaPPARα agonist (fenofibrate [143, 144])1. Activated PPARα-AMPK-PGC-1α-eNOS pathway2. Reduced 1-deoxysphingolipids and other atypical sphingoid bases, with no effect on typical sphingolipidsProtected endothelial cells, and reduced apoptosisGLP-1R agonists(liraglutide [157, 187], exenatide [188])1. Shortened large intestinal transit time, reduced motility index2. Decreased inflammatory response and MAD levels, increased GSH-Px and SOD levels1. Improved enteric nervous system.2. Increased MNCV and SNCV of median and common peroneal nerves.3. Improved nerve excitability parametersGLP-1R agonists(liraglutide [154], exenatide [155, 156])1. Reduced Interleukin-6 levels2. Promoted corneal regeneration1. No effect on electrophysiological or small‑fiber neuropathy parameters2. No improvement in NCV or neurological symptomsSGLT‑2 inhibitors (Dapagliflozin [159], Empagliflozin [160])Increased glutathione peroxidase, and enhanced antioxidant capacity1. Increased corneal nerve fiber density, corneal nerve branch density, corneal nerve fiber length, and intraepidermal nerve fiber density2. Improved motor and sensory nerve conduction velocity3. Reduced neuropathy riskSGLT-2 inhibitors (canagliflozin [163])–1. No effect on sensorimotor olyneuropathy, diabetic neuropathy, or non-autonomic neuropathy events2. Increased risk of lower limb amputationDPP4 inhibitors(sitagliptin [166], teneligliptin [165])Elevated GLP-1 levels, and restored insulin signaling1. Reduced triglyceride levels and improved nerve conduction2. Improved sympathetic dysfunction, sudomotor function, and heart rate response to standingMolecules that modulate SARM1 acetylation or catalytic activity [167, 168] [169]1. Attenuated ROS accumulation, ATP depletion, and NAD⁺ decline2. Modulated the SARM1/NMNAT2 axis and restored axonal mitochondrial transportReduce intraepidermal nerve fiber loss in paw pad skin and axonal degenerationNAD⁺ precursor supplementation or NMNAT2 overexpression (e.g., epigallocatechin gallate) [170, 171]Derived neuronal NAD⁺ production, restored neuronal glycolysis, and restored axonal mitochondrial transportReduced axonal damage and neurodegenerationDiet and exercise interventionsEndurance training combined with sensorimotor training [173, 174], or aerobic exercise combined with resistance training [175]1. Regulated expression of nerve growth factor and brain-derived neurotrophic factor2. Modulated polyol-sorbitol pathway3. Enhanced antioxidant capacity1. Promoted improvement in peroneal nerve conduction velocity2. Improved balance functionCaloric restriction combined with high-intensity interval training [189]Activated AMPK, and regulated glucose and lipid metabolic homeostasisImproved motor nerve conduction velocityOmega-3 polyunsaturated fatty acids [177], or menhaden and flaxseed oil [190–192]1. Activated the nuclear factor erythroid 2-related factor 2 antioxidant pathway in Schwann cells2. Promoted vasodilationImproved motor and sensory nerve conduction velocity, thermal sensitivity, and intraepidermal nerve fiber densityKetogenic diet [180, 181]Promoted neurite outgrowth of sensory neuronsPrevented and reversed mechanical allodynia and thermal hyperalgesiaKetogenic diet [70]Induced metabolic dysregulation, mitochondrial damage, and homeostasis imbalance in Schwann cellsExacerbated DPN phenotypeLow-carbohydrate diet, caloric restriction, or alternate-day fasting[193]Improved body weight and glycemic control, with no effect on fat massNo significant improvement in DPN symptoms Table 1Summary of metabolic intervention strategies for DPNIntervention categorySpecific interventionPrimary mechanisms of actionKey effects on DPNPharmacological interventionsAldose reductase inhibitors (epalrestat [137], compound 15c [182])1. Reduced sorbitol accumulation, and inhibited polyol pathway2. Improved axonal sodium currents, restored Na⁺/K⁺ pump function and reversed the dedifferentiation state of myelinating Schwann cells under high glucose conditions1. Repaired structure of myelin and axonal structure2. Improved nerve conduction velocityAdiponectin receptor agonist (AdipoRon) [139]1. Reduced systemic oxidative stress and insulin resistance2. Upregulated MCT1/2/4 and enhances lactate and ATP transport to axons1. Restored energy supply between myelinating Schwann cells and axons2. Reduced neuronal apoptosisHMG-CoA reductase inhibitors (atorvastatin, ezetimibe/simvastatin, rosuvastatin [145, 147, 148, 183])1. Inhibited HMG-CoA reductase, and lowered serum cholesterol2. Reduced total ceramide, sphingomyelin and other sphingolipid metabolites3 Activated glutathione antioxidant system, and reduced lipid peroxidation1. Ameliorated oxidative stress in sciatic nerve2. Alleviated DPN severity and symptoms3. Improved nerve conduction velocityPPARγ agonists (rosiglitazone, pioglitazone [140, 141])1. Reduced oxidative stress and prevented thermal hypoalgesia2. Improved evoked pain behavior3.Decreased TNF-α and IL-1β levels in sciatic nerve4.Downregulated Nav1.7 channel expression1. Improved morphology and function of unmyelinated nerve fibers2. Shortened thermal latency, and increased nerve conduction velocityPPARγ agonist compounds (gentiopicroside [184], magnolol [185], lycorine [186])1. Modulated PPAR-γ/AMPK/ACC, PPAR-γ/AMPK/PGC-1α, and PPARγ/NF-κB signaling pathways2. Regulated lipid metabolism, improved mitochondrial function, and promoted neural blood flowImproved myelin sheath structure in sciatic nerve and intraepidermal nerve fiber density in hind paw, and alleviated hyperalgesiaPPARα agonist (fenofibrate [143, 144])1. Activated PPARα-AMPK-PGC-1α-eNOS pathway2. Reduced 1-deoxysphingolipids and other atypical sphingoid bases, with no effect on typical sphingolipidsProtected endothelial cells, and reduced apoptosisGLP-1R agonists(liraglutide [157, 187], exenatide [188])1. Shortened large intestinal transit time, reduced motility index2. Decreased inflammatory response and MAD levels, increased GSH-Px and SOD levels1. Improved enteric nervous system.2. Increased MNCV and SNCV of median and common peroneal nerves.3. Improved nerve excitability parametersGLP-1R agonists(liraglutide [154], exenatide [155, 156])1. Reduced Interleukin-6 levels2. Promoted corneal regeneration1. No effect on electrophysiological or small‑fiber neuropathy parameters2. No improvement in NCV or neurological symptomsSGLT‑2 inhibitors (Dapagliflozin [159], Empagliflozin [160])Increased glutathione peroxidase, and enhanced antioxidant capacity1. Increased corneal nerve fiber density, corneal nerve branch density, corneal nerve fiber length, and intraepidermal nerve fiber density2. Improved motor and sensory nerve conduction velocity3. Reduced neuropathy riskSGLT-2 inhibitors (canagliflozin [163])–1. No effect on sensorimotor olyneuropathy, diabetic neuropathy, or non-autonomic neuropathy events2. Increased risk of lower limb amputationDPP4 inhibitors(sitagliptin [166], teneligliptin [165])Elevated GLP-1 levels, and restored insulin signaling1. Reduced triglyceride levels and improved nerve conduction2. Improved sympathetic dysfunction, sudomotor function, and heart rate response to standingMolecules that modulate SARM1 acetylation or catalytic activity [167, 168] [169]1. Attenuated ROS accumulation, ATP depletion, and NAD⁺ decline2. Modulated the SARM1/NMNAT2 axis and restored axonal mitochondrial transportReduce intraepidermal nerve fiber loss in paw pad skin and axonal degenerationNAD⁺ precursor supplementation or NMNAT2 overexpression (e.g., epigallocatechin gallate) [170, 171]Derived neuronal NAD⁺ production, restored neuronal glycolysis, and restored axonal mitochondrial transportReduced axonal damage and neurodegenerationDiet and exercise interventionsEndurance training combined with sensorimotor training [173, 174], or aerobic exercise combined with resistance training [175]1. Regulated expression of nerve growth factor and brain-derived neurotrophic factor2. Modulated polyol-sorbitol pathway3. Enhanced antioxidant capacity1. Promoted improvement in peroneal nerve conduction velocity2. Improved balance functionCaloric restriction combined with high-intensity interval training [189]Activated AMPK, and regulated glucose and lipid metabolic homeostasisImproved motor nerve conduction velocityOmega-3 polyunsaturated fatty acids [177], or menhaden and flaxseed oil [190–192]1. Activated the nuclear factor erythroid 2-related factor 2 antioxidant pathway in Schwann cells2. Promoted vasodilationImproved motor and sensory nerve conduction velocity, thermal sensitivity, and intraepidermal nerve fiber densityKetogenic diet [180, 181]Promoted neurite outgrowth of sensory neuronsPrevented and reversed mechanical allodynia and thermal hyperalgesiaKetogenic diet [70]Induced metabolic dysregulation, mitochondrial damage, and homeostasis imbalance in Schwann cellsExacerbated DPN phenotypeLow-carbohydrate diet, caloric restriction, or alternate-day fasting[193]Improved body weight and glycemic control, with no effect on fat massNo significant improvement in DPN symptoms Summary of metabolic intervention strategies for DPN 1. Reduced sorbitol accumulation, and inhibited polyol pathway 2. Improved axonal sodium currents, restored Na⁺/K⁺ pump function and reversed the dedifferentiation state of myelinating Schwann cells under high glucose conditions 1. Repaired structure of myelin and axonal structure 2. Improved nerve conduction velocity 1. Reduced systemic oxidative stress and insulin resistance 2. Upregulated MCT1/2/4 and enhances lactate and ATP transport to axons 1. Restored energy supply between myelinating Schwann cells and axons 2. Reduced neuronal apoptosis 1. Inhibited HMG-CoA reductase, and lowered serum cholesterol 2. Reduced total ceramide, sphingomyelin and other sphingolipid metabolites 3 Activated glutathione antioxidant system, and reduced lipid peroxidation 1. Ameliorated oxidative stress in sciatic nerve 2. Alleviated DPN severity and symptoms 3. Improved nerve conduction velocity 1. Reduced oxidative stress and prevented thermal hypoalgesia 2. Improved evoked pain behavior 3.Decreased TNF-α and IL-1β levels in sciatic nerve 4.Downregulated Nav1.7 channel expression 1. Improved morphology and function of unmyelinated nerve fibers 2. Shortened thermal latency, and increased nerve conduction velocity 1. Modulated PPAR-γ/AMPK/ACC, PPAR-γ/AMPK/PGC-1α, and PPARγ/NF-κB signaling pathways 2. Regulated lipid metabolism, improved mitochondrial function, and promoted neural blood flow 1. Activated PPARα-AMPK-PGC-1α-eNOS pathway 2. Reduced 1-deoxysphingolipids and other atypical sphingoid bases, with no effect on typical sphingolipids 1. Shortened large intestinal transit time, reduced motility index 2. Decreased inflammatory response and MAD levels, increased GSH-Px and SOD levels 1. Improved enteric nervous system. 2. Increased MNCV and SNCV of median and common peroneal nerves. 3. Improved nerve excitability parameters 1. Reduced Interleukin-6 levels 2. Promoted corneal regeneration 1. No effect on electrophysiological or small‑fiber neuropathy parameters 2. No improvement in NCV or neurological symptoms 1. Increased corneal nerve fiber density, corneal nerve branch density, corneal nerve fiber length, and intraepidermal nerve fiber density 2. Improved motor and sensory nerve conduction velocity 3. Reduced neuropathy risk 1. No effect on sensorimotor olyneuropathy, diabetic neuropathy, or non-autonomic neuropathy events 2. Increased risk of lower limb amputation 1. Reduced triglyceride levels and improved nerve conduction 2. Improved sympathetic dysfunction, sudomotor function, and heart rate response to standing 1. Attenuated ROS accumulation, ATP depletion, and NAD⁺ decline 2. Modulated the SARM1/NMNAT2 axis and restored axonal mitochondrial transport 1. Regulated expression of nerve growth factor and brain-derived neurotrophic factor 2. Modulated polyol-sorbitol pathway 3. Enhanced antioxidant capacity 1. Promoted improvement in peroneal nerve conduction velocity 2. Improved balance function 1. Activated the nuclear factor erythroid 2-related factor 2 antioxidant pathway in Schwann cells 2. Promoted vasodilation Pharmacological interventions Aldose reductase inhibitors, such as epalrestat, reduce sorbitol accumulation, restore myo-inositol levels and Na⁺/K⁺ pump function, and reverse hyperglycemia-induced dedifferentiation of myelinating Schwann cells, thereby repairing axonal and myelin structures and improving nerve conduction velocity [137, 138]. The adiponectin receptor agonist AdipoRon upregulates MCT1/2/4, enhances lactate and ATP transport to the axon, and restores energy supply between myelinating Schwann cells and the axon [139]. Improving systemic insulin resistance is an important direction for early DPN intervention. PPARγ agonists such as rosiglitazone and pioglitazone [140, 141] and PPARα agonist fenofibrate [142–144], improve unmyelinated nerve fiber morphology and function, shorten thermal withdrawal latency in diabetic mice, and restore nerve conduction velocity. These beneficial effects are mediated through multiple mechanisms, including inhibition of NADPH oxidase expression, reduction of oxidative stress and inflammatory cytokine release, and activation of the AMPK/PGC‑1α pathway to improve insulin resistance and to reduce 1‑deoxysphingolipids levels. In addition to lowering serum cholesterol, statins also significantly reduce levels of sphingolipid metabolites, such as total ceramides (29%), sphingomyelins (15%), hexosylceramides (30%), and lactosylceramides (34%), by inhibiting hydroxymethylglutaryl-CoA reductase. Furthermore, they enhance antioxidant capacity through glutathione activation and attenuate oxidative stress injury in the sciatic nerve [145, 146]. However, evidence from randomized controlled trials (RCTs) regarding statin therapy remains mixed. Some RCTs have reported that statin therapy for 12–16 weeks reduces plasma lipid peroxidation levels and improves peroneal nerve conduction velocity [147, 148]. By contrast, large‑scale cohort studies and meta‑analyses have not found a significant association between statin use and DPN risk, suggesting that the efficacy of statins may depend on baseline lipid profiles and the stage of neuropathy [149, 150]. GLP-1 receptor agonists, SGLT-2 inhibitors, and DPP-4 inhibitors are three classes of novel glucose‑lowering agents that have garnered considerable attention in DPN research because of their neuroprotective effects mediated by metabolic modulation, attenuation of oxidative stress, improvement of mitochondrial function, and suppression of inflammatory responses[151, 152]. GLP‑1 receptors are expressed in neurons and Schwann cells of the DRG and sciatic nerve; their activation upregulates pERK1/2 and mitigates the reduction in motor nerve conduction velocity and plantar intraepidermal nerve fiber density in diabetic mice [153]. Some clinical studies have observed that GLP‑1 receptor agonists promote corneal nerve and small‑fiber regeneration, but their effects on neurophysiological parameters have been inconsistent [154–157]. A recent meta‑analysis suggested that GLP‑1 receptor agonists, particularly liraglutide, improve nerve conduction velocity in patients with type 2 diabetes and DPN; however, the analysis included only six studies with high heterogeneity, and therefore this conclusion requires further validation [158]. SGLT‑2 inhibitors also exhibit neuroprotective potential. RCTs and meta-analyses indicate that dapagliflozin and empagliflozin improve clinical symptoms and nerve conduction velocity in patients with DPN and promote small-fiber repair and regeneration [159–161]. A nationwide Danish cohort study found that patients treated with SGLT-2 inhibitors had a lower risk of diabetic foot disease than those receiving GLP-1 receptor agonists, suggesting that SGLT-2 inhibitors may confer a greater advantage in reducing DPN risk [162]. However, a post‑hoc analysis of the CREDENCE trial showed that canagliflozin did not significantly reduce the risk of sensorimotor polyneuropathy [163]; moreover, canagliflozin has been associated with an increased risk of lower‑limb amputation, so caution is warranted when using it in diabetic patients with foot ulcers [164]. DPP‑4 inhibitors exert their effects by prolonging the half‑life of endogenous GLP‑1 and other neurotrophic peptides. Studies by Syngle, Barros, and colleagues have demonstrated that DPP-4 inhibitors such as vildagliptin and sitagliptin promote neurite outgrowth, increase mechanical and thermal pain thresholds, and improve cardiac autonomic function and nerve conduction velocity [165, 166]. However, clinical evidence for DPP‑4 inhibitors in DPN treatment remains relatively limited. Overall, evidence for these three drug classes remains preliminary, and rigorous large-scale, multicenter RCTs with standardized endpoints, including pain assessment, small-fiber structure and function, nerve conduction, and autonomic measures, are needed for validation. Therapeutic strategies targeting the SARM1/NMNAT2 axis are also promising. Dysregulation of this axis is a core mechanism underlying the transport‑dependent energy vulnerability of axonal mitochondria. Several molecules that modulate SARM1 acetylation or catalytic activity have been developed, including the competitive inhibitory peptide Pep‑SP1 derived from amino acids 645–655 of the SARM1 TIR domain [167], small‑molecule compounds based on isoquinoline, isothiazole, pyridine, and tryptamine acrylamide moieties [168], and the rosemary‑derived antioxidant carnosol [169]. In preclinical models, these molecules reduce ROS accumulation, ATP depletion, and NAD⁺ decline, thereby attenuating intraepidermal nerve fiber loss in the plantar skin and axonal degeneration [167, 168]. Furthermore, NAD⁺ precursor supplementation or NMNAT2 overexpression, for example, using epigallocatechin gallate to drive neuronal NAD⁺ production, has shown potential to restore neuronal glycolysis, improve axonal transport and mitochondrial function, and prevent neuropathy [170, 171]. Although several clinical trials have begun to target the SARM1/NMNAT2 axis, the above results are all derived from animal models or preclinical studies and require further translational validation. Diet and exercise The American Diabetes Association recommends lifestyle interventions, including dietary adjustment and physical activity, for the management of metabolically associated DPN [172]. Such interventions improve systemic metabolic health in individuals with obesity, metabolic syndrome, and T2DM. Recent meta-analyses have shown that endurance training combined with sensorimotor training improves nerve conduction velocity and alleviates neuropathic pain and related symptoms [173, 174], whereas aerobic exercise combined with resistance training restores small-fiber sensory nerve function and improves muscle strength [175]. These benefits are likely attributable to exercise-induced reductions in hyperglycemia, dyslipidemia, insulin resistance, inflammatory mediators, and accumulation of AGEs [176]. Preclinical studies have shown that supplementation with omega‑3 polyunsaturated fatty acids (ω‑3 PUFAs) improves motor and sensory nerve conduction velocity, thermal sensitivity, and intraepidermal nerve fiber density by reducing oxidative stress injury in Schwann cells [177]. Clinical studies have also found that ω‑3 PUFA supplementation promotes small‑fiber nerve regeneration and increases corneal nerve fiber length, but has only limited effects on nerve conduction and sensory function [178]. The most recent meta‑analysis, which included only two clinical trials, reported that oral ω‑3 PUFA supplementation did not improve clinical symptoms or health‑related quality of life in patients with DPN [179]. Therefore, the clinical benefit of oral ω‑3 PUFAs in DPN remains uncertain. Whether the morphological improvements in small‑fiber nerve regeneration observed in clinical studies translate into functional benefits requires further validation through larger, well‑designed randomized controlled trials with standardized neurological endpoints. The ketogenic diet, which replaces glucose with ketone bodies as the primary circulating fuel, has been reported to improve metabolic indices in DPN, alleviate mechanical allodynia and thermal hyperalgesia, and promote regeneration of distal peripheral nerve fibers in animal models [180, 181]. Yet, the underlying mechanisms remain incompletely understood, and the therapeutic window has yet to be defined. Compounding this uncertainty, Li et al. [70] identified a pathogenic form of peripheral ketogenesis: aberrant ketone body production within Schwann cells, driven by the CB1R-Hmgcs2 axis, disrupts energy metabolism, damages mitochondria, and precipitates homeostatic decompensation, thereby contributing to DPN progression. In this context, pathological ketogenesis exacerbates DPN, whereas blocking this process reprograms Schwann cell metabolism and attenuates nerve injury. Collectively, these findings underscore a fundamental principle emerging in DPN research: globally reshaping energy metabolism, rather than targeting a single molecule, pathway, or dietary component, may represent a more promising therapeutic direction. Conclusions and perspectives In summary, the understanding of DPN pathogenesis has shifted from a hyperglycemia-centric model to one centered on systemic, multidimensional metabolic dysregulation involving glucose, lipids, and amino acids. Interactions among these metabolic networks, coupled with toxic intermediate accumulation, create a complex pathophysiological microenvironment in the PNS. As the hub of cellular energy conversion, mitochondria serve as the critical nodes linking glucose, lipid, and amino acid metabolism; disruption of mitochondrial function and dynamic homeostasis further support the notion that DPN is fundamentally a bioenergetic failure of peripheral nerve tissue. This review highlights key nodes of metabolic imbalance in DPN. First, transporter dysregulation (e.g., GLUTs, MCTs) and defects in glycolysis and oxidative phosphorylation disrupt axon‑Schwann cell metabolic coupling, leading to distal axonal energy deprivation and degeneration. Second, abnormal accumulations of saturated fatty acids, ceramides of specific chain lengths, and glutamate, along with their metabolic intermediates, further aggravate axonal and Schwann cell injury through oxidative stress, mitochondrial dysfunction, apoptosis, and ion channel activation. Based on these mechanisms, several candidate biomarkers have been preliminarily identified and may have potential for future clinical translation. The expression levels of GLUT1, GLUT3, and MCT1 may reflect early axon-Schwann cell metabolic uncoupling in DPN, whereas abnormal lipid profiles, particularly elevated long-chain saturated fatty acids, deoxysphingolipids (e.g., deoxy-C24 ceramide), specific ceramide subspecies (e.g., C24/C26 ceramides), 1-deoxydihydroceramide, an increased PE/PC ratio, and decreased cardiolipin levels, may help with risk stratification and disease progression assessment. In addition, alterations in amino acid metabolites (e.g., arginine, tyrosine, and glutamate) have been associated with nerve conduction velocity, but further validation is required before these markers can be translated into clinical use. In clinical practice, greater attention should be paid to early neuropathy screening in individuals with metabolic syndrome and prediabetes by incorporating sensitive small-fiber assessment tools such as IENFD, corneal confocal microscopy, and nerve excitability testing into the evaluation of at‑risk populations. For patients with different metabolic phenotypes, lifestyle intervention should serve as the foundation, supplemented by insulin sensitizers and other metabolism‑modifying agents as part of comprehensive glycemic control. Diet and exercise therapy have shown positive effects on improving small‑fiber sensory nerve function. Although agents such as GLP‑1 receptor agonists, SGLT‑2 inhibitors, PPAR agonists, and ω-3 PUFAs have demonstrated neuroprotective potential in clinical studies, their definitive benefit in DPN requires confirmation in high‑quality RCTs. The failure of single‑target strategies to sufficiently restore the complex energy homeostasis of peripheral nerves remains a major reason for the limited efficacy of current DPN interventions. Future studies should integrate multi-omics approaches, particularly single-cell and spatial metabolomics, to identify upstream master regulators of peripheral nerve energy metabolism and to provide a theoretical basis for multidimensional, multitarget synergistic interventions. Overall, this review emphasizes a paradigm shift from a hyperglycemia‑centric view to multidimensional energy metabolic dysregulation in DPN research, offering new perspectives for future studies and drug development.