Traditional mechanisms (33–37) of hyperglycemia-induced complications involve oxidative stress, polyol pathway activation, advanced glycation end products (AGEs) formation, protein kinase C pathway activation, and hexosamine pathway activation. However, these mechanisms inadequately explain the “metabolic memory” phenomenon. While previous research noted that AGEs accumulate in patients with long-term poor glycemic control and continuously exert pathological effects promoting vascular disease (33), this mechanism remains too generalized to explain the dynamic characteristics and individual variability of “metabolic memory”.

Clinical data from the Diabetes Control and Complications Trial (DCCT) demonstrated that hyperglycemic environments induce persistent epigenetic modifications in immune and tissue cells of type 1 diabetes (T1D) patients. Epigenetic markers such as H3K9Ac in monocytes significantly correlate with previous glycated haemoglobin (HbA_1c) levels (38, 39), suggesting “metabolic memory” is closely linked to long-term epigenetic regulation (40, 41). However, traditional epigenetic explanations fail to clarify why “metabolic memory” persists for decades despite the relatively short lifespan and constant renewal of peripheral effector cells.

The trained immunity paradigm provides a novel framework for understanding the “metabolic memory” phenomenon. Hyperglycemia, as a trained immunity inducer, affects not only mature circulating immune cells but also crucially induces persistent metabolic and epigenetic reprogramming in hematopoietic stem cells (HSCs) and myeloid progenitor cells (8, 42, 43). This progenitor-level “memory” ensures that newly generated immune cells maintain pro-inflammatory phenotypes even after glycemic control is achieved, thereby explaining the continued progression of long-term complications. The trained immunity theory’s key advantage over other explanations lies in its focus on progenitor cell-level mechanisms and their intergenerational transmission.

Notably, glycemic variability—characterized by unstable fluctuations between peak and nadir blood glucose levels—is a common phenomenon in diabetes management (44). Clinical studies have shown an association between glycemic variability and the development and progression of diabetic complications (44–46). This pattern of intermittent hyperglycemic stimulation exhibits similarities to the initial stimulus and re-stimulation model characteristic of trained immunity. Glycemic variability likely triggers epigenetic and metabolic reprogramming in immune cells, inducing trained immunity that contributes to the establishment and maintenance of “metabolic memory”.

Taken together, diabetes impacts immune cell function via complex metabolic and epigenetic network remodeling, creating a regulatory network spanning metabolism, immunity, tissue homeostasis, and hematopoiesis. Diabetic patients, particularly those with type 2 diabetes (T2D), frequently present with comorbid conditions including hypertension (47), hyperlipidemia (48, 49), and obesity (50). Therefore, the trained immunity caused by other abnormal factors in diabetes and glycemic variability should also be given attention.

In trained immunity, non-permanent histone modifications are closely associated with gene activation. In the β-glucan-induced trained immunity model of macrophages, H3K4 monomethylation (H3K4me1) and trimethylation (H3K4me3) are significantly enriched in the enhancer regions of pro-inflammatory genes. This activation mechanism depends on upregulated expression of Set7 lysine methyltransferase. In vitro experiments confirm that Set7 inhibitors suppress pro-inflammatory memory effects induced by β-glucan (51, 52).

Hyperglycemia induces similar epigenetic remodeling in trained immunity. Mechanistic analysis reveals that high glucose promotes H3K4me3 deposition in pro-inflammatory gene promoter regions by upregulating the glycolytic pathway in monocytes and the mixed lineage leukemia (MLL) family of H3K4 methyltransferases. Clinical data further support that glycolysis-related genes and MLL methyltransferases are significantly upregulated in CD14⁺ monocytes of patients with T1D and THP-1 cells cultured under hyperglycemic conditions (42).

Under steady-state conditions, immune cells display relatively low biosynthetic activity and predominantly rely on oxidative phosphorylation (OXPHOS) and fatty acid oxidation (FAO) for energy requirements. However, upon activation, innate immune cells undergo a substantial surge in energy demands. Consequently, aerobic glycolysis, glutaminolysis, cholesterol metabolism, and fatty acid synthesis become pivotal pathways to meet these elevated needs. This increased requirement for glucose and shift toward aerobic glycolysis resembles the “Warburg effect” observed in cancer cells (53, 54).

Elevated glucose levels drive macrophages toward glycolysis while reducing OXPHOS, a shift linked to protein kinase B (AKT) activation within the mechanistic target of rapamycin (mTOR) pathway (55). This metabolic reprogramming enables macrophages to secrete pro-inflammatory cytokines like tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and interleukin-1β (IL-1β), perpetuating chronic inflammation.

Beyond macrophages, neutrophils also undergo metabolic reprogramming in diabetes, enhancing glycolysis via the pentose phosphate pathway and FAO. This leads to acetyl-coenzyme A accumulation, which, mediated by ATP-citrate lyase, promotes histone acetylation. Consequently, neutrophils form excessive neutrophil extracellular traps (NETs), thereby impairing wound healing in diabetic patients (28).

Atherosclerosis (AS), a chronic inflammatory vascular disease driven by genetic susceptibility, lifestyle factors, and systemic inflammation, contributes to significant global morbidity and mortality (57). Macrophages, central to atherosclerotic plaque pathogenesis, exhibit enhanced glycolysis, disrupted tricarboxylic acid cycle, and epigenetic alterations under hyperglycemic conditions (43, 58, 59). This cellular adaptation redirects their polarization toward a pro-inflammatory M1 phenotype while suppressing reparative M2 functions (60).

Animal studies show that even transient hyperglycemia accelerates AS through enhanced myelopoiesis (61). In 2021, a study conducted by the team led by Robin P. Choudhury provided robust evidence that hyperglycemia promotes trained immunity in HSCs and macrophages, significantly exacerbating AS (8). Central to this process is the transcription factor RUNX1, which orchestrates HSC differentiation, macrophage survival, and inflammatory programming (8).

Macrophages from hyperglycemic mice maintain enhanced cytokine production even after being cultured in normal glucose for 7 days. This inflammatory priming has significant in vivo consequences: bone marrow transplantation from hyperglycemic mice accelerates plaque formation in normoglycemic LDL-knockout mice (8). These plaques show H3K4me3 enrichment in macrophage-rich regions—a trained immunity marker absent in controls. Similar epigenetic and functional alterations are observed in leukocytes from T2D patients, confirming the clinical relevance of this phenomenon (8).

Notably, in hyperglycemic environments, all exposed tissue cells are affected, potentially inducing reprogramming in multiple cell types related to vascular health. Trained immunity characteristics have been documented in various immune and non-immune cells critical to AS, including dendritic cells (62), neutrophils (63), natural killer cells (64), vascular smooth muscle cells, (65) and endothelial cells (66). These findings suggest that trained immunity extends beyond innate immune cells, with hyperglycemia potentially inducing long-term vascular endothelial dysfunction through epigenetic reprogramming mechanisms across multiple cell types critical for vascular health.

The pathogenesis of myocardial infarction (MI) is characterized by coronary artery obstruction, which subsequently leads to myocardial cell death due to ischemia and hypoxia. Diabetic patients face higher mortality and increased complications (reinfarction, heart failure, shock, arrhythmias), demonstrating hyperglycemia’s synergistic amplification of cardiac injury (67–72).

The post-MI inflammatory cascade is a tightly regulated yet complex process involving systemic and local immune activation. Bone marrow-derived immune cells are rapidly mobilized alongside resident cardiac immune responses, triggering the recruitment of circulating inflammatory cells critical for injury and repair. Neutrophils dominate the early phase (peaking at 24–48 hours post-MI), followed by macrophages, T/B cells, and dendritic cells, with macrophages playing dual roles in inflammation and tissue repair (53, 73).

In diabetic patients, hyperglycemia-induced trained immunity disrupts this balance, exacerbating post-MI inflammation. Experimental models demonstrate that Ly6C^Hi monocytes exhibit a pathological “second wave” of infiltration into ischemic myocardium, mirroring their delayed polarization to Ly6C^Lo phenotypes observed in diabetic wound healing (74). Some researchers speculate that this mechanism is likely to be related to the healing process of MI as well (75). Hyperglycemia further entrenches a pro-inflammatory macrophage phenotype, increasing their infiltration into ischemic tissue and suppressing reparative functions. The resulting inflammatory milieu not only delays healing but also heightens risks of adverse remodeling and heart failure.

Interestingly, a recent study has revealed that MI can act as a priming factor for monocytes to enhance trained immunity, thereby promoting the progression of AS (76). In patients with acute coronary syndrome (ACS), the expression of spleen tyrosine kinase (SYK) in monocytes may serve as a potential biomarker for predicting the risk of recurrent ischemic events. In this context, MI can be regarded as the “first hit,” while hyperlipidemia represents the “second hit.” Both conditions exert their effects through epigenetic modifications within the bone marrow and monocytes, jointly leading to increased SYK expression and maintenance of a persistent pro-inflammatory phenotype (76).

Based on these observations, a bidirectional interaction has been established between trained immunity and cardiovascular injury. Cardiovascular damage itself may initiate immune training via persistent inflammatory signaling. AS increases the risk of MI, while MI-induced immune priming subsequently exacerbates residual atherosclerotic lesions.

The kidney serves as a key target organ for microvascular damage in diabetes. Diabetic kidney disease (DKD) pathogenesis is complex, arising from the interplay of multiple factors, including genetic predisposition, environmental influences, metabolic disorders, hemodynamic abnormalities, and immune responses. This pathological process is characterized by persistent hyperglycemia, immune complex deposition in the glomeruli, increased chemokine production, and macrophage recruitment (77, 78). These events trigger complex crosstalk between macrophages, non-myeloid cells, and adaptive immune cells. The inflammatory cascade is tightly linked to dysregulated macrophage function.

Notably, a growing body of evidence indicates that innate immune cells in the kidneys exhibit phenotypes consistent with trained immunity. Patients with chronic kidney disease (CKD) exhibit elevated CD14⁺⁺CD16⁺ pro-inflammatory monocytes in bone marrow alongside heightened systemic levels of IL-6, IL-1β, and TNF-α, suggesting persistent innate immune activation (79). Monocytes stimulated by Ox-LDL and subsequently exposed to Toll-like receptor (TLR) 2 and TLR4 agonists demonstrate enhanced production of IL-6 and TNF-α, with upregulated H3K4me3 modification levels at inflammatory mediator gene promoters. This epigenetic modification is reversible by histone methyltransferase inhibition (24).

Environmental stressors further potentiate renal immune memory: high-salt diets exacerbate macrophage-mediated inflammation during secondary pathogen challenges, characterized by CD45⁺F4/80⁺ macrophage infiltration and cytokine surges that accelerate renal fibrosis (80). Uremic toxin accumulation in CKD, particularly indoxyl sulfate, activates trained immunity via aryl hydrocarbon receptor (AhR)-dependent arachidonic acid pathways, perpetuating inflammatory cascades (81). In experimental high-fat diet (HFD)+CKD models, synergistic lipid metabolism disturbances and caspase-11/LPS interactions upregulate 998 cytoplasmic genes linked to vascular inflammation via trained immunity mechanisms (82).

Although current research on hyperglycemia-induced trained immunity primarily focuses on the cardiovascular system, its specific manifestations and mechanisms in renal pathophysiology remain poorly understood. As a hallmark microvascular complication of diabetes, DKD pathogenesis likely involves trained immunity as a pivotal link connecting hyperglycemic memory to renal inflammatory injury.

In 1993, Loe first identified an elevated risk of periodontitis in diabetic patients, noting that it ranks as the sixth leading complication of diabetes (83). Subsequent epidemiological and intervention studies have demonstrated that individuals with diabetes are at a 3–4 times greater risk for developing periodontitis compared to those without diabetes (84–88).

Recent insights into trained immunity provide a novel framework for explaining the bidirectional relationship between diabetes and periodontitis. Systemic inflammation from periodontitis may activate trained immunity in peripheral immune cells and their precursors. Studies using 18F-fluorodeoxyglucose positron emission tomography-computed tomography (18F-FDG-PET/CT) imaging in patients with periodontitis support this hypothesis, showing an association between periodontitis and increased hematopoietic tissue activity (89, 90). Peripheral neutrophils in chronic periodontitis patients exhibit hyperresponsiveness with excessive reactive oxygen species (ROS) production (91) and increased pro-inflammatory cytokine release (92). Peripheral blood mononuclear cells from individuals with severe periodontitis also exhibit heightened IL-6 production (93). This cellular hyperreactivity persists even after successful periodontal treatment, aligning with characteristics of trained immunity.

Periodontitis and diabetes may reciprocally amplify inflammatory responses via trained immunity mechanisms. Both conditions induce sustained reprogramming in myeloid cells and their progenitors (94, 95). Specific bacterial products or inflammatory mediators can activate both peripheral myeloid cells and their bone marrow precursors, enhancing their responsiveness to subsequent challenges.

This bidirectional interaction creates a pathological feedback loop in which periodontal inflammation can exacerbate diabetes-associated immune responses, while hyperglycemia-primed cells exhibit heightened reactions to periodontal pathogens. This inflammatory interaction may contribute to the progression of both conditions. Current evidence indicates that trained immunity links oral and systemic inflammation, highlighting the need for integrated clinical management of these interconnected conditions.

Vaccine-mediated trained immunity can induce enhanced innate immune responses against unrelated pathogens, providing non-specific protection, known as heterologous effects (97). Utilizing these heterologous effects, vaccines in the context of trained immunity are applied not only for infection prevention but also for regulating immune dysregulation diseases. Animal studies have demonstrated that BCG vaccination prevents candidiasis in severe combined immunodeficiency mice (98). Human research has confirmed that BCG-induced trained immunity provides non-specific protection against controlled human malaria (99) and experimental viral infections (100). For immunologically dysregulated tumors, BCG has been approved for intravesical administration in the treatment of non-muscle invasive bladder cancer (101). These studies highlight the therapeutic potential of vaccines in the field of trained immunity.

T1D is an autoimmune disease characterized by progressive destruction of pancreatic β cells (102), with pathological features including immune dysregulation and loss of self-tolerance. Current immunotherapeutic strategies for T1D primarily focus on targeting specific T and B cells to prevent islet β cell destruction. Treatment approaches such as anti-CD3 monoclonal antibody (Teplizumab) (103, 104) and anti-CD20 monoclonal antibody (Rituximab) (105) have shown certain efficacy. However, research on preventing T1D through targeted immune approaches remains relatively limited. The innate immune system, as the “first line of defense “ (106) and a key regulator of immune responses, has therapeutic potential through trained immunity. This approach, which targets innate immune cells to regulate immune tolerance or correct immune dysregulation, also deserves attention.

Epidemiological studies report that vaccination with the inactivated influenza vaccine Pandemrix^® reduces T1D risk in specific populations, suggesting that vaccine-induced trained immunity may participate in autoimmune regulation (107–110). An 8-year randomized study reported that double-dose BCG treatment could normalize HbA_1c in T1D patients after three years (111). Additionally, patients receiving BCG treatment exhibited a systemic metabolic shift from OXPHOS to aerobic glycolysis, consistent with trained immunity characteristics as confirmed in mouse experiments (111). BCG also restores insulin secretion and regulates immunity by inducing regulatory T cells and reducing autoreactive T cells (112).

Although these studies establish connections between vaccines and T1D through trained immunity, in-depth elucidation of the relevant immunomodulatory mechanisms remains insufficient, while research applications in T2D are considerably more limited. For T1D, future strategies could integrate adaptive immune-targeted monoclonal antibodies with innate immune interventions to achieve synergistic effects, enhancing preventive and therapeutic outcomes.

Nanomedicine is a rapidly evolving field that integrates nanotechnology, biomedicine, and pharmaceutical sciences (113). Nanoparticles, the fundamental components of nanomedicine, are biocompatible and biodegradable spherical systems that encapsulate conventional or biological drugs. They function as drug delivery vehicles that protect therapeutic agents from degradation at the administration site, facilitate targeted transport to specific tissues or organs, and enable controlled drug release in response to environmental stimuli at the target location (114).

The persistent effects of trained immunity originate from metabolic and epigenetic reprogramming of bone marrow progenitor cells, generating myeloid cells with enhanced responsiveness, termed “trained” myeloid cells (115, 116). This requires technologies capable of directly targeting myeloid progenitor cells (96). Given that nanomaterials inherently interact with phagocytic myeloid cells, nanomedicine provides an ideal platform for modulating trained immunity (117), enabling precise and efficient targeting of cells and inflammatory signaling pathways associated with trained immunity.

For example, nanoformulations loaded with mTOR inhibitors (mTORi-NB) can inhibit the production of pro-inflammatory cytokines in human monocytes stimulated with Ox-LDL (118). In experimental models, one-week treatment with mTORi-NB in ApoE^-/- mice fed a Western diet for 12 weeks attenuated plaque inflammation (117). Such anti-inflammatory nanotherapeutic approaches may have potential applications in diabetes management.

Nanomedicine can modulate trained immunity at cellular, metabolic, and epigenetic levels through diverse material technologies, enabling precise immune modulation (117). However, translational applications in diabetes require further investigation of nanoparticle biocompatibility, cellular uptake, and drug release in diabetes-specific microenvironments.

Metabolic and epigenetic reprogramming interact in trained immunity, with metabolic intermediates functioning as substrates, cofactors, or signaling molecules that regulate chromatin-modifying enzymes, establishing immunological memory. Given the role of metabolic alterations in driving the epigenetic foundations of trained immunity, targeting key metabolic enzymes to inhibit excessive trained immune responses represents a promising anti-inflammatory strategy.

Enhanced glycolysis, a critical metabolic signature of trained immunity, can be modulated by hexokinase inhibitors such as 2-deoxy-D-glucose (119) or mTOR pathway inhibitors including rapamycin and metformin (55). Glutamine catabolism is also upregulated during trained immunity. Succinate derived from glutaminolysis promotes pro-inflammatory histone modifications by inhibiting histone demethylase (HDM) activity, a process blocked by the glutaminase inhibitor BPTES(bis-2-(5-phenylacetamido-1,3,4-thiadiazol-2-yl)ethyl sulfide) (120). In addition, statins inhibit HMG-CoA reductase and attenuate β-glucan-induced trained immunity by depleting mevalonate pathway intermediates essential for epigenetic-modifying enzymes (121). Itaconate, produced via decarboxylation of cis-aconitate catalyzed by immunoresponsive gene 1, drives macrophage polarization toward an anti-inflammatory phenotype by inhibiting histone demethylases (122, 123).

Beyond metabolic targets, emerging pharmacological strategies focus on modulating key epigenetic modifiers in specific cellular or pathological contexts, including DNA methyltransferases (DNMTs), lysine methyltransferases (KMTs), and histone deacetylases (HDACs). For instance, DNMT inhibitors and HDAC inhibitors can reverse the silencing of pro-inflammatory genes, whereas activators of specific KMTs may regulate anti-inflammatory signaling pathways via modulation of histone methylation patterns (124–130).

Strategies targeting the metabolic-epigenetic axis offer multi-level intervention points for regulating trained immunity, but their clinical translation requires addressing drug specificity, tissue selectivity, and long-term safety considerations.