• Peer-reviewed original articles or reviews;

• Studies focusing on mouse models of diabetes or diabetic complications;

• Articles examining pathophysiological mechanisms, functional assessments, or therapeutic interventions in diabetic nephropathy, neuropathy, or retinopathy.

• Studies published in non-English languages;

• Conference abstracts, letters to the editor, or reports lacking primary data;

• Articles not directly related to mouse models of diabetic complications;

• Duplicate publications or studies lacking a clear methodology.

NF-$\kappa$B signaling plays a central role in the inflammatory processes underlying diabetic complications. Hyperglycemia and AGEs serve as primary activators, promoting leukocyte adhesion, leukostasis, and endothelial injury that ultimately lead to vascular leakage [12, 13]. Beyond NF-$\kappa$B, additional pathways, such as receptor-interacting protein kinase 3 (RIPK3)–mediated NOD-, LRR-, and pyrin domain-containing protein 3 (NLRP3) inflammasome activation and Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling, contribute to the complex inflammatory network [14, 15].

NF-$\kappa$B-mediated inflammation follows a shared framework yet presents with tissue-specific outcomes. For instance, upregulation of intercellular adhesion molecule 1 (ICAM-1) and cluster of differentiation 18 (CD18) is observed in affected tissues, and genetic deletion of NF-$\kappa$B components leads to reduced inflammatory lesions [16]. Tissue selectivity is also evident, as toll-like receptor 4 (TLR4) and interleukin-6 (IL-6) upregulation is more prominent in organs experiencing progressive damage, even under comparable glycemic conditions [17]. Mendelian randomization studies further support the causal role of genetically elevated inflammatory mediators—such as stem cell growth factor beta (SCGF-$\beta$), interleukin-8 (IL-8), and growth-regulated oncogene alpha (GRO-$\alpha$)—in disease risk and progression [18]. Sustained inflammatory activation results in tissue remodeling and functional decline.

Targeting inflammation yields promising therapeutic outcomes. Pharmacologic inhibition of NF-$\kappa$B reduces both inflammation and oxidative stress [17], while photobiomodulation therapy decreases NF-$\kappa$B and receptor for advanced glycation end products (RAGE) expression and shifts cytokine profiles toward resolution [19]. Suppressor of cytokine signaling (SOCS) mimetic peptides, targeting the JAK/STAT pathway, also confer anti-inflammatory and metabolic benefits [15]. Collectively, these interventions highlight inflammation as a unifying and actionable target across diabetic complications.

AGE–RAGE signaling is a pivotal mechanism connecting chronic hyperglycemia to oxidative stress and inflammation [20]. AGEs interact with RAGE on target cells, activating downstream Mitogen-activated protein kinase (MAPK) pathways and inducing pro-inflammatory cytokine production [21, 22]. RAGE expression is markedly upregulated in diabetic lesions, and genetic polymorphisms in RAGE influence susceptibility to complications [23, 24].

This pathway contributes to both systemic and tissue-specific pathology. RAGE activation promotes expression of vascular endothelial growth factor (VEGF) and recruits inflammatory cells [25], sustaining chronic inflammation and driving pathological tissue remodeling. It is also linked to autonomic dysfunction and peripheral nerve injury. Specific AGE profiles—reflecting glycolytic dysfunction, lipid peroxidation, and glucotoxicity—correlate with impaired function in multiple organ systems [26].

Therapeutic targeting of AGE–RAGE interactions is multifaceted. Inhibitors of AGE formation, such as pyridoxamine, reduce tissue damage; combination therapies further enhance efficacy [27]. Natural compounds and herbal formulations have shown potential in lowering AGE levels, boosting endogenous antioxidant defenses, and modulating apoptosis-related pathways [28]. RAGE-deficient models exhibit preserved tissue integrity and reduced inflammation, reinforcing the clinical relevance of this pathway [29].

The Nrf2 pathway is a critical regulator of antioxidant responses and cellular redox balance. In DM, this pathway becomes dysfunctional, impairing the transcription of cytoprotective genes and exacerbating oxidative damage [30, 31]. Under normal conditions, oxidative stress induces Nrf2 translocation to the nucleus; however, chronic hyperglycemia disrupts this process [32].

Under high-glucose conditions, Nrf2 activity undergoes dynamic regulation. Antioxidant gene expression is transiently suppressed, followed by partial compensation despite ongoing accumulation of reactive oxygen species (ROS). At the same time, elevated NF-$\kappa$B p65 expression indicates antagonistic crosstalk between antioxidant and inflammatory pathways [33]. In diabetic tissues, the Nrf2 pathway is further compromised. Enhanced Kelch-like ECH-associated protein 1 (Keap1)-Nrf2 interaction impairs Nrf2 nuclear translocation, leading to glutathione depletion and weakened antioxidant defense [34, 35]. Nevertheless, the pathway remains a viable therapeutic target, given that Nrf2 activity is inducible and its deficiency exacerbates oxidative stress and inflammation [36]. Accordingly, pharmacologic activation of Nrf2—through agents such as oltipraz, resveratrol, and tBHQ—has been shown to confer protection against diabetic complications [37, 38, 39]. These findings highlight Nrf2 as a convergent therapeutic target across complications.

Mitochondrial dysfunction is a unifying pathogenic feature in diabetic complications, linking hyperglycemia to cellular injury through impaired energy metabolism, oxidative stress, and disrupted organelle dynamics [40, 41]. Fragmented mitochondria, an imbalance in adenosine triphosphate/adenosine diphosphate (ATP/ADP) ratio, and defective mitophagy are observed across affected tissues. Reduced expression of NAD⁺-dependent deacetylase sirtuin-3 (SIRT3) further exacerbates mitochondrial damage [42].

Mitochondrial stress generates ROS and releases mitochondrial deoxyribonucleic acid (mtDNA), which activates innate immune responses and contributes to fibrosis [43]. Synergistic effects of hyperglycemia and hypertension further compromise mitochondrial integrity, increasing superoxide and depleting glutathione [44]. Matrix metalloproteinases (MMP)-2 and MMP-9 aggravate this dysfunction via increased membrane permeability [45, 46]. Downregulation of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1$\alpha$) disrupts mitochondrial biogenesis and homeostasis [47].

Targeting mitochondrial quality control offers therapeutic promise. SIRT3 overexpression restores mitophagy through the Forkhead box O3a (FoxO3a)–PTEN-induced kinase 1 (PINK1)–Parkin pathway [48], while AICAR, an AMP analog, activates PGC-1$\alpha$ to improve mitochondrial dynamics and reduce oxidative injury [49]. Overexpression of mitochondrial transcription factor A (TFAM) also helps preserve mitochondrial integrity and tissue function [50]. These findings support mitochondrial dysfunction as both a central mechanism and a viable therapeutic target in diabetic complications.

Purinergic signaling has recently been recognized as an important contributor to the pathogenesis of DM and its complications, but it remains underexplored in most preclinical models. Extracellular nucleotides (e.g., ATP) and nucleosides (e.g., adenosine) act through P2 and P1 receptors to regulate inflammation, oxidative stress, and fibrosis, thereby exacerbating microvascular and macrovascular injury [51, 52]. In metabolic regulation, purinergic pathways modulate pancreatic $\beta$-cell function, insulin resistance, and hepatic glucose metabolism [53]. They also contribute to neuroinflammation and microvascular damage relevant to diabetic neuropathy [54]. In the eye, excessive extracellular ATP and overactivation of the purinergic receptor P2X7 (P2X7R) have been linked to diabetic retinopathy, where P2X7R stimulation promotes VEGF release and pathological angiogenesis [55]. In the kidney, activation of the ATP–P2X7R axis promotes NLRP3 inflammasome activation and accelerates injury [56]. Although this has been primarily demonstrated in models of acute kidney injury, the same pathway is increasingly implicated in the chronic inflammation and fibrosis that underlie diabetic nephropathy. Collectively, these findings underscore purinergic signaling as a common mechanistic link across diabetic complications and a promising therapeutic target for hyperglycemia-associated tissue injury.

DN is a major complication of DM and the leading cause of end-stage renal disease worldwide. It is characterized by progressive damage to glomerular, tubular, and interstitial compartments, ultimately resulting in renal failure [57]. DN pathogenesis involves a combination of fibrotic signaling and podocyte injury, with transforming growth factor-$\beta$ (TGF-$\beta$) playing a central regulatory role.

TGF-$\beta$ is a key driver of renal fibrosis in DN [58]. Chronic hyperglycemia induces sustained TGF-$\beta$1 expression in renal cells, activating the SMAD family member (Smad) 2/3 pathway and promoting extracellular matrix (ECM) accumulation. Fibroblast growth factor 21 (FGF21) mitigates this process by inhibiting TGF-$\beta$1-induced Smad2/3 nuclear translocation [59]. In diabetic mice, elevated TGF-$\beta$1 is accompanied by increased $\alpha$-smooth muscle actin ($\alpha$-SMA), fibronectin, and collagen deposition [14]. Smad3 deletion reduces mesangial expansion and proteinuria, reinforcing its pathological role [60]. Leucine-rich $\alpha$-2-glycoprotein 1 (LRG1), a TGF-$\beta$ effector, is upregulated in diabetic kidneys and exacerbates glomerular injury via Activin receptor-like kinase 1 (ALK1)–Smad1/5/8 signaling. LRG1 deletion confers renal protection, and elevated plasma LRG1 levels are associated with poor renal outcomes in type 2 DM [61, 62]. Beyond fibrosis, TGF-$\beta$ also induces podocyte-specific injury. In podocytes, it activates phosphoinositide 3-kinase (PI3K) signaling, upregulating monocyte chemoattractant protein-1 (MCP-1) and promoting C-C chemokine receptor type 2 (CCR2)-mediated podocyte migration, thereby disrupting the glomerular filtration barrier [63].

Podocyte injury is a hallmark of DN and contributes directly to proteinuria. The slit diaphragm—comprising nephrin, podocin, and other structural proteins—is compromised in DM, leading to foot process effacement and podocyte loss [64, 65]. Urinary mRNA levels of podocyte markers such as nephrin, podocin, synaptopodin, Wilms tumor 1 (WT-1), and alpha-actinin-4 are significantly elevated in patients with biopsy-confirmed DN. These elevations correlate with proteinuria severity and renal function decline, while WT-1 expression reflects the extent of tubulointerstitial fibrosis [66]. A key mechanism of podocyte injury involves dysregulation of the transient receptor potential canonical 6 (TRPC6) channel [67]. TGF-$\beta$1 enhances TRPC6 expression, resulting in reduced nephrin, elevated desmin and caspase-9 levels, and increased apoptosis. TRPC6 knockdown reverses these effects and preserves podocyte viability [68]. Hyperglycemia and angiotensin II further increase TRPC6 activity, triggering calcium influx, podocyte hypertrophy, process effacement, and apoptosis [69]. TRPC6 also impairs autophagy via calpain-mediated mechanisms, linking calcium dysregulation to podocyte dysfunction [70].

Together, these findings highlight TGF-$\beta$-driven fibrosis and TRPC6-mediated podocyte injury as central contributors to diabetic nephropathy progression.

DPN is one of the most common and debilitating complications of diabetes, affecting nearly half of individuals with type 1 and type 2 DM [71]. Its prevalence rises with disease duration, exceeding 50% in patients with DM for over 10 years [72]. DPN contributes significantly to morbidity, including lower-extremity amputations and chronic neuropathic pain [73]. Following amputation, life expectancy averages just two years, underscoring the severe clinical and socioeconomic burden [74].

DPN most frequently presents as distal symmetric polyneuropathy, with sensory symptoms such as paresthesia, burning pain, and hypersensitivity, typically beginning in the feet and progressing proximally [75]. With disease progression, motor deficits emerge, leading to muscle weakness, atrophy, gait disturbances, and impaired coordination that interfere with daily activities [71]. The pathogenesis of DPN involves multiple interconnected mechanisms, prominently including metabolic dysregulation and immune system dysfunction.

A key metabolic driver of DPN is the polyol pathway. Under hyperglycemic conditions, aldose reductase (AR)—the rate-limiting enzyme—converts glucose to sorbitol in a Nicotinamide adenine dinucleotide phosphate (NADPH)-dependent reaction [76]. This process leads to sorbitol accumulation, NADPH depletion, and increased ROS production, promoting oxidative stress [77]. Mouse studies support this mechanism; AR overexpression in Schwann cells exacerbates motor nerve conduction velocity (MNCV) deficits in diabetic and galactosemic mice, despite similar polyol accumulation, suggesting oxidative stress—not sorbitol per se—as the primary pathogenic factor [78]. Transgenic mice expressing human AR exhibit worsened DPN phenotypes, including increased sorbitol, decreased MNCV, and nerve fiber atrophy. These effects are reversed by AR inhibition [79].

Beyond metabolic perturbations, emerging evidence reveals that immune cell dysfunction plays a crucial role in DPN pathogenesis. Recent single-cell transcriptomic analyses have identified mast cells as critical mediators in DPN development. In STZ-induced diabetic mice, hyperglycemic conditions enhance glucose transporter type 3 (GLUT3)-mediated glucose uptake in mast cells. This triggers extracellular signal-regulated kinases 1 and 2 (ERK1/2) activation and mechanistic target of rapamycin (mTOR) hyperactivity, leading to metabolic dysregulation that drives mast cell degranulation and pro-inflammatory mediator release. Ultimately, this disrupts the neural microenvironment [80].

These findings demonstrate that DPN results from interconnected metabolic and immunological disturbances. AR-induced oxidative stress and mast cell–mediated inflammation converge to impair peripheral nerve integrity under hyperglycemic conditions. This dual-pathway mechanism highlights the need for therapeutic strategies that target both metabolic dysfunction and immune activation.

DR constitutes a leading cause of blindness in working-age adults. It features progressive retinal microvascular damage and neurodegeneration. This microvascular complication affects both the retinal vasculature and neural retina, ultimately leading to vision loss [81, 82]. DR progresses through distinct stages, from an early non-proliferative stage to an advanced proliferative stage. The early stage features pericyte loss, microaneurysms, retinal capillary leakage, and acellular capillaries. As the condition advances, the proliferative stage becomes marked by retinal neovascularization and fibrous tissue proliferation [83]. The pathogenesis of DR involves multiple interconnected pathways, including protein kinase C (PKC) activation, VEGF-mediated angiogenesis, and oxidative stress-induced cellular dysfunction.

PKC activation plays a central role in the pathogenesis of diabetic retinopathy through distinct isoform-specific mechanisms [84]. Pericytes are essential for maintaining blood vessel integrity and regulating endothelial cell growth; however, in the diabetic retina, pericyte loss leads to the formation of microaneurysms and acellular capillaries [85]. A clinical study evaluated the therapeutic efficacy of an oral PKC$\beta$ inhibitor, ruboxistaurin (RBX), in 252 patients with moderately severe to very severe non-proliferative DR [86]. The randomized, placebo-controlled trial was conducted over 36 to 46 months. Although RBX (32 mg/day) was well tolerated, it did not prevent the progression of DR. However, it significantly reduced the risk of moderate visual loss compared to placebo (hazard ratio: 0.37, p = 0.012), particularly in patients with baseline diabetic macular edema. Separately, hyperglycemia induces pericyte apoptosis via activation of PKC$\delta$, a mechanism particularly prominent in the retinal microvasculature. In diabetic mice, overactivation of PKC$\delta$ in the retina stimulates the p38$\alpha$ MAPK/Src homology region 2 domain-containing phosphatase-1 (SHP-1) signaling pathway. This cascade leads to platelet-derived growth factor receptor beta (PDGFR-$\beta$) dephosphorylation, resulting in pericyte dysfunction and apoptosis [87]. In STZ-induced rat models, PKC activity was increased in retinal tissues. Treatment with the PKC$\beta$2 inhibitor GF109203X partially alleviated diabetic retinopathy pathology by blocking this signaling cascade [88].

PKC activation, in addition to directly affecting pericytes, also contributes to retinal ischemia through vascular dysfunction. In advanced diabetic retinopathy, this ischemia triggers compensatory but pathological angiogenesis mediated by VEGF. VEGF becomes upregulated in response to retinal ischemia from capillary dropout, ultimately driving proliferative diabetic retinopathy. The Ins2^{A⁢k⁢i⁢t⁢a} mouse model reveals retinal neovascularization by 9 months, developing new blood vessels and preretinal neovascular tufts accompanied by increased retinal VEGF levels [89]. Therapeutic interventions targeting the VEGF pathway have shown promise in experimental models. Inhibition of the VEGF pathway reduced the severity of retinal damage in STZ-induced rat models [88]. Additionally, the glucagon-like peptide-1 (GLP-1) analogue exendin-4 exhibited protective effects by downregulating placental growth factor (PLGF) and VEGF expression via the ERK and protein kinase B (AKT/PKB) signaling pathways, while preserving blood–retinal barrier integrity [90].

These findings support a mechanistic model in which hyperglycemia-induced PKC activation impairs pericyte survival, leading to retinal ischemia and triggering VEGF-driven neovascularization—key events in the progression of DR.

STZ-treated C57BL/6 mice represent a commonly used model of type 1 DM. However, they exhibit relatively mild DN and typically require additional insults to manifest robust pathology. These additional insults include uninephrectomy or endothelial nitric oxide synthase (eNOS) deletion [93, 94]. The nephropathy phenotype can be enhanced through combinations with advanced oxidation protein products or high-protein diets, exacerbating albuminuria and tubulointerstitial injury [95].

Functionally, STZ-induced diabetic mice exhibit moderate albuminuria and glomerular hyperfiltration in early stages. The Animal Models of Diabetic Complications Consortium defined criteria for validating murine models require greater than 50% decline in glomerular filtration rate (GFR) over the lifetime of the animal and greater than 10-fold increase in albuminuria compared with controls for that strain at the same age and gender [96].

Histologically, STZ-treated mouse strains show moderate mesangial matrix expansion, increased glomerular surface area, and tubular cell damage but lack nodular lesions [97].

Therapeutic target validation in STZ models emphasizes glycemic control and antioxidant interventions. Moreover, the involvement of both RAGE-dependent and RAGE-independent mechanisms highlights the complexity of AGE-mediated renal injury. Targeting AGEs directly, as shown by the efficacy of alagebrium even in RAGE-deficient mice, suggests that AGE-lowering therapies may offer additive benefits beyond RAGE blockade alone [98]. Recombinant human bone morphogenetic protein-7 significantly inhibits glomerular hypertrophy and tubulointerstitial fibrosis by restoring expression of Ski-related novel protein N via activation of Smad1/5. This suppresses partial epithelial-to-mesenchymal transition and ECM accumulation in diabetic kidneys [99].

Akita mice carry a spontaneous Ins2 gene mutation causing pancreatic $\beta$-cell failure and sustained hyperglycemia. This model exhibits early-onset hyperglycemia, hypoinsulinemia, and impaired insulin secretion. It offers insight into spontaneous, progressive type 1 diabetes [100]. They serve as a valuable genetic model for studying DN [100, 101].

Genetic background significantly influences disease severity in Akita mice. DBA/2 background mice manifest severe hyperglycemia and pronounced renal dysfunction including significant albuminuria. In contrast, C57BL/6 Akita mice show mild albuminuria and limited mesangial expansion [102]. Various Akita-based strain studies have revealed marked differences in albumin-to-creatinine ratios, establishing new models for DN research [103]. Genetic modifications further modulate nephropathy severity. Bradykinin receptor, eNOS, or Angiotensin-converting enzyme (ACE) deficiency exacerbates kidney injury, while interventions like ketogenic diet can reverse diabetic nephropathy [101]. Additionally, Nrf2 knockout (Akita::Nrf2^-⁣/-) worsened diabetic kidney disease, while Nrf2 induction (Akita::Keap1FA/FA) provided protection through antioxidant mechanisms [38].

OVE26 transgenic mice exhibit severe early-onset type 1 DM within weeks of birth. They manifest comprehensive DN features including progressive albuminuria, hyperfiltration followed by GFR decline, glomerular enlargement, mesangial matrix expansion, tubulointerstitial fibrosis, and glomerular basement membrane thickening [104]. The FVB genetic background proves critical for disease severity, as these mice exhibit the highest albuminuria among diabetic models. Crossing to C57BL6 or DBA2 backgrounds reduces albuminuria 17-fold through susceptibility loci on chromosomes 9, 11, 13, and 19 that account for $>$70% of strain differences [105].

The db/db mouse constitutes one of the most extensively used models for investigating type 2 DM and its renal complication. This model harbors a G-to-T point mutation in the leptin receptor gene (Lepr), resulting in leptin resistance, hyperphagia, obesity, and persistent hyperglycemia. Consequently, db/db mice manifest hallmark features of DN, including albuminuria, mesangial matrix expansion, and glomerular basement membrane thickening [106, 107].

The severity of the diabetic phenotype in db/db mice varies markedly depending on their genetic background. On the C57BLKS/J background, db/db mice exhibit severe hyperglycemia accompanied by progressive degeneration of pancreatic islet cells. This leads to reduced insulin secretion and body weight loss by 5–6 months of age. This strain also displays increased sensitivity to $\beta$-cell toxins such as streptozotocin, indicating an intrinsic predisposition to $\beta$-cell vulnerability [107]. In contrast, C57BL/6J db/db mice show milder hyperglycemia, despite displaying similar levels of hyperphagia and weight gain. This results from islet cell hypertrophy, which helps preserve $\beta$-cell function, suggesting a relative resistance to $\beta$-cell dysfunction in this strain [107].

To model severe DN, the eNOS-deficient db/db mouse (eNOS^-⁣/-db/db) was generated on the C57BLKS/J background. These mice manifest a constellation of advanced DN features, including severe albuminuria, hypertension, arteriolar hyalinosis, mesangiolysis, nodular mesangial expansion, and tubulointerstitial fibrosis. This closely resembles the pathology seen in advanced human DN [108].

Among available models, the DBA/2J db/db mouse exhibits the most aggressive DN phenotype. It features an approximately 50-fold increase in the albumin-to-creatinine ratio by 12 weeks of age. These mice also show glomerulosclerosis, foot process effacement, and marked glomerular basement membrane thickening [109], making them particularly valuable for studying advanced DN pathogenesis.

The BTBR ob/ob mouse model represents a valuable tool in DN research. It offers a robust system that closely mirrors human diabetic kidney disease progression. This model combines the leptin-deficient ob/ob mutation with the BTBR genetic background, resulting in severe type 2 DM accompanied by hyperinsulinemia, insulin resistance, hypercholesterolemia, and elevated triglyceride levels [110]. Unlike traditional mouse models that exhibit limited renal involvement, the BTBR ob/ob mouse manifests a full spectrum of renal lesions that align with both early and advanced stages of human DN [111].

Disease progression follows a predictable timeline. Podocyte loss and proteinuria emerge by 8 weeks, followed by mesangial expansion around 10 weeks. Between 18–22 weeks, advanced DN features become evident, including marked proteinuria, $>$50% increase in mesangial matrix, ~20% glomerular basement membrane thickening, and diffuse mesangial sclerosis with nodular features. Additional pathologic changes include arteriolar hyalinosis, mesangiolysis, and interstitial fibrosis. These renal changes occur alongside metabolic disturbances including elevated serum triglycerides, cholesterol, and blood urea nitrogen in both sexes [110].

A key advantage lies in the accelerated disease progression, enabling interventional studies within shorter timeframes compared to other models [112]. MicroRNA sequencing analyses have identified upregulation of 99 microRNAs linked to inflammatory and fibrotic pathways that mirror human DN. This supports the model’s translational relevance for mechanistic studies and therapeutic development [111, 113].

While traditional diabetic nephropathy models focus on hyperglycemia alone, clinical diabetic kidney disease often involves multiple metabolic abnormalities. The STZ-induced diabetic ApoE^-⁣/- mouse model addresses this limitation by combining hyperglycemic and hyperlipidemic conditions. This more closely mimics the complex metabolic environment of human diabetes [114, 115]. This dual-pathology model reveals enhanced susceptibility to renal injury compared to either STZ-induced diabetes or ApoE deficiency alone. The combined metabolic stress results in severe albuminuria, glomerular and tubulointerstitial damage, increased fibrosis, and elevated expression of profibrotic markers including TGF-$\beta$1 and collagen types I and IV [114]. A unique advantage lies in the model’s ability to simultaneously evaluate both DN progression and atherosclerotic complications, reflecting the clinical reality of diabetic patients.

Multiple therapeutic interventions have shown renoprotective efficacy in this complex model. Avosentan reduced albuminuria and glomerular pathology through suppression of profibrotic and inflammatory gene expression [115]. Omapatrilat, a dual inhibitor of ACE and neutral endopeptidase, showed superior renoprotection compared to ACE inhibition alone [116]. Additionally, $\alpha$-lipoic acid supplementation ameliorated nephropathy by reducing oxidative stress and mesangial expansion [117]. These findings highlight the model’s value for testing therapeutics targeting multiple pathogenic pathways simultaneously.

STZ-induced type 1 DM mouse models serve as widely used platforms to investigate the pathophysiology of DPN and to validate potential therapeutic targets. These models exhibit relatively rapid onset of neuropathic symptoms following STZ administration. Within 4 to 8 weeks post-injection, mice typically show reduced sensory and motor nerve conduction velocities. These changes occur alongside thermal hypoalgesia and diminished tactile sensitivity—hallmarks of early small fiber dysfunction [73].

Behaviorally, these mice initially present with hyperalgesia, which gradually transitions to hypoalgesia during the chronic phase. However, this temporal progression becomes considerably compressed compared to the course of human DPN [118]. Advanced fluorescent labeling techniques have further revealed that peptidergic and nonpeptidergic nociceptive neurons exhibit differential susceptibility to diabetic injury. Specifically, peptidergic fiber loss occurs as early as 4 weeks post-STZ. This precedes the degeneration of nonpeptidergic fibers at approximately 8 weeks [119]. The selective correlation between early behavioral deficits and peptidergic fiber loss suggests that these neurons may drive early neuropathic symptoms [119].

Given the thermal sensory deficits observed in STZ-induced models, careful behavioral assessment proves critical. Traditional thermal avoidance tests, such as the plantar assay, consistently detect thermal hypoalgesia but may fail to capture the full complexity of thermosensory dysfunction. Recent studies employing thermal gradient behavioral assays have identified altered temperature preference patterns in diabetic mice that are distinct from those observed in transient receptor potential vanilloid 1 (TRPV1)-deficient animals, despite comparable thermal avoidance behaviors [120]. These findings emphasize the utility of complementary thermal behavior paradigms in uncovering mechanisms underlying thermosensory deficits in DPN [120].

Beyond chemically induced models, spontaneous genetic models such as the Akita mouse offer an additional platform for investigating the development and progression of DPN. The Akita mice exhibit milder signs of neuropathy, potentially due to lower circulating insulin levels and the absence of obesity. This represents a pure type 1 DM phenotype [121]. However, despite their lean phenotype, Akita mice manifest marked insulin resistance—shown by an ~80% reduction in glucose infusion rate during hyperinsulinemic-euglycemic clamps. This results from decreased glucose uptake in skeletal muscle and brown adipose tissue, as well as impaired hepatic insulin action [122]. This combination of insulin deficiency and insulin resistance renders the model particularly relevant for studying diabetic complications involving both hormonal and metabolic disturbances [122].

Notably, Akita mice exhibit pronounced diabetic sympathetic autonomic neuropathy, with characteristic neuritic dystrophy in prevertebral ganglia. This includes prominently swollen axons and dendrites that progressively worsen over 2 to 8 months of diabetes [121]. Progressive neuronal loss occurs, with approximately one-third of neurons lost in the celiac ganglia by 8 months. This becomes accompanied by distinctive perikaryal abnormalities, including membranous aggregate accumulation and mitochondrial dysfunction [121]. In addition to autonomic features, Akita mice also manifest significant sensory neuropathy. This features impaired mechanical and thermal nociception and substantial intraepidermal nerve fiber loss [123]. Electrophysiological analyses reveal reduced action potential discharge in mechanonociceptors and markedly impaired heat responsiveness in dorsal root ganglion neurons, while cold-sensitive and low-threshold A-fiber functions remain largely intact [123]. By 16 weeks of age, these mice exhibit reduced sensory nerve conduction velocities, thermal and mechanical hypoalgesia, and tactile allodynia, with further deterioration in some measures by 20 weeks [124].

The extended survival and consistent pathological manifestations establish Akita mice as a valuable model for studying both autonomic and sensory components of DPN [121]. Moreover, their responsiveness to a broad range of therapeutic interventions—including complete reversal with insulin therapy, partial improvement with inhibiting poly(ADP-ribose) polymerase, and selective effects with erythropoietin-derived peptides—highlights their utility in evaluating mechanism-based treatments and dissecting distinct pathogenic pathways involved in DPN [123, 124, 125].

Alloxan-induced type 1 DM shares many pathophysiological features with STZ-induced models but represents an alternative chemical approach for inducing $\beta$-cell destruction. In experimental studies, alloxan-diabetic mice manifest significant tactile allodynia and thermal hyperalgesia, paralleling behavioral phenotypes observed in STZ models [126]. Electrophysiological assessments reveal similar reductions in both motor and sensory nerve conduction velocities [126].

Longitudinal studies in alloxan-diabetic rats have documented extensive ultrastructural abnormalities, including demyelination and remyelination, axonal degeneration and regeneration, and characteristic onion bulb formations formed by proliferating Schwann cells [127]. Unique pathological features observed in this model include crystalline deposits within vessel walls and endoneurium, glycogen accumulation in axonal mitochondria, and axoplasmic inclusions resembling Lafora bodies. These findings suggest that alloxan-induced neuropathy involves complex metabolic disruption affecting axons, Schwann cells, and microvasculature [127].

Importantly, comparative studies suggest that alloxan may offer certain advantages over STZ in isolating hyperglycemia-mediated mechanisms. While STZ induces mechanical hypersensitivity in both hyperglycemic and normoglycemic animals—likely via direct neurotoxic effects involving nociceptive neurons and TRPV1 modulation—alloxan-induced mechanical sensitization occurs only under hyperglycemic conditions [128]. This distinction highlights alloxan’s potential to more selectively model glucose-dependent neuropathic changes, reducing confounding from direct chemical neurotoxicity.

The C57BL/KsJ db/db mouse constitutes a well-established model of type 2 DM that reproduces key metabolic features such as hyperglycemia, insulin resistance, and obesity. This makes it suitable for investigating DPN in the context of metabolic syndrome. Longitudinal studies in db/db mice show a characteristic progression of DPN from early functional impairments to subsequent structural degeneration, closely mirroring the human disease trajectory [129].

Functional deficits, including progressive sensory loss and electrophysiological impairments, emerge during early-to-mid disease stages. These become accompanied by reduced intraepidermal nerve fiber density (IENFD) as early as 18 weeks of age [129]. As the disease advances, evident structural alterations—such as Schwann cell apoptosis and CD3⁺ T-cell infiltration in the sciatic nerve—highlight the evolving neuropathology. This progression reflects a transition from metabolic disturbance to established neurodegeneration [129].

Notably, db/db mice exhibit a biphasic pattern of neuropathy progression [130]. An initial “metabolic” phase, which remains responsive to insulin therapy, becomes followed by a “neuronal” phase featuring progressive axonal atrophy and impaired axonal transport of acetylcholinesterase, with peak severity occurring around 180 days of age. During this later stage, neuropathy becomes refractory to insulin treatment but shows partial responsiveness to ganglioside therapy, reflecting a fundamental shift from systemic metabolic disturbance to intrinsic neuronal pathology. This biphasic progression may parallel the degenerative phases observed in human DPN, providing valuable insights into the timing and nature of therapeutic intervention windows [130].

Mechanistically, db/db mice display unique pathogenic features that distinguish them from other diabetic models. Importantly, Na⁺,K⁺-ATPase activity in sciatic and optic nerves remains unchanged across disease stages (50–280 days), suggesting that this pathway does not contribute significantly to neuropathy development in this model [131]. Additionally, the absence of AR staining in nerve tissue indicates minimal involvement of the polyol pathway, a mechanism commonly implicated in other diabetic complications [131]. These findings suggest that DPN in db/db mice may arise through alternative, model-specific biochemical pathways.

The leptin-deficient BTBR ob/ob mouse constitutes a robust model of type 2 DM that exhibits early-onset and severe DPN. These mice manifest nerve conduction deficits by 9 weeks and intraepidermal nerve fiber loss by 13 weeks, making them particularly valuable for studying accelerated DPN progression [132].

An emerging area of interest involves the role of adipose tissue innervation. BTBR ob/ob mice exhibit small fiber demyelinating neuropathy in subcutaneous white adipose tissue, along with dysregulated Schwann cell gene expression, mirroring changes seen in obese humans [133]. These findings reveal that adipose neuropathy, driven by Schwann cell dysfunction, may contribute to metabolic impairments and represents a historically underrecognized component of DPN.

Sex-specific studies have further validated the model’s utility. Female BTBR ob/ob mice manifest DPN similar to males, with comparable motor and sensory deficits but preserved IENFD, suggesting sex-dependent variations in small fiber involvement [134]. Although both sexes show similar metabolic profiles, males exhibit higher triglyceride levels, indicating subtle metabolic differences. Transcriptomic analyses of sciatic nerve and dorsal root ganglia confirm that inflammatory pathways become dysregulated in both sexes, reinforcing neuroinflammation as a central mechanism in DPN [134].

Diet-induced mouse models provide a valuable platform for studying DPN, particularly in the context of metabolic syndrome and prediabetes, where neuropathy can manifest independently of marked hyperglycemia. In a comparative study involving BKS, B6, and BTBR genotypes, mice fed a 54% high-fat diet (HFD) manifested large-fiber neuropathy and impaired glucose tolerance, with B6-wt mice showing the most consistent phenotype, including obesity, hyperinsulinemia, dyslipidemia, and elevated oxidized low-density lipoproteins [135]. Notably, dietary reversal after 16 weeks of high-fat feeding completely normalized both neuropathy and metabolic parameters, suggesting that lipid and inflammatory pathways, rather than hyperglycemia alone, may serve as primary drivers of neuropathy in metabolic syndrome [135].

Mechanistic studies using these models have revealed converging pathways involving lipid metabolism, calcium signaling, and inflammation, reinforcing their value for therapeutic screening and for elucidating the multifactorial pathogenesis of obesity-associated neuropathy [136, 137]. Given the increasing global prevalence of obesity and metabolic syndrome, and the high incidence of neuropathy as a complication, these models offer critical insights into non-glucose–centric therapeutic strategies [136].

STZ-induced diabetic mice are widely used to model type 1 diabetes-related retinal pathology. After about six months of sustained hyperglycemia, they develop early signs of DR, such as pericyte loss, acellular capillary formation, mild retinal thickening, and occasional microaneurysms [85, 138, 141]. In C57BL/6J mice, transient neuronal apoptosis occurs soon after diabetes onset, accompanied by caspase-3 activation [74]. However, these neural changes normalize over time, and no significant retinal ganglion cell loss is detected even after one year, as confirmed by multiple methods.

The eNOS-deficient STZ model represents a more aggressive variant. In eNOS^-⁣/- diabetic mice, vascular leakage appears as early as 3 weeks post-induction, with more rapid and severe DR features such as acellular capillaries, persistent gliosis, and basement membrane thickening [142]. Upregulation of iNOS and total nitric oxide (NO) suggests compensatory changes in NO signaling. These findings highlight the protective role of eNOS-derived NO and make the eNOS-deficient STZ model a valuable tool for studying vascular mechanisms in DR and evaluating therapies targeting endothelial dysfunction.

The Ins2^{A⁢k⁢i⁢t⁢a} mouse develops spontaneous hyperglycemia by 2 months of age and shows progressive retinal pathology resembling human DR [89]. It is particularly useful for studying early DR features, including vascular leakage, gliosis, and neurodegeneration, typically observed over six-month periods [143]. By 6 months, vascular abnormalities become evident, such as pericyte loss, vascular leakage, and microaneurysm formation [89]. By 9 months, advanced features appear, including retinal neovascularization, new capillary bed formation, and ectopic blood vessels in the outer plexiform layer—findings rarely seen in other diabetic mouse models [89]. These structural changes are accompanied by increased apoptosis and measurable functional decline.

Early retinal pigment epithelium (RPE) dysfunction also emerges with the onset of hyperglycemia, with progressive impairment detected in both neural and RPE responses [144]. Akita mice show poor adaptation to metabolic stress; hypercapnia further impairs inner retinal function without triggering expected increases in retinal blood flow, suggesting dysfunction beyond vascular dysregulation [145]. The model’s ability to develop both early and late DR features, including neovascularization, makes it especially valuable for testing interventions across disease stages.

The db/db mouse model is particularly useful for studying retinal neurodegeneration, now recognized as an early and possibly independent contributor to DR pathogenesis [146]. Pericyte loss, a hallmark of early DR, has been compared between db/db and Akita mice. Both models show pericyte dropout and gliosis, but differ in inflammatory signaling. In db/db mice, chronic interferon-$\gamma$ exposure disrupts PDGFR-$\beta$ signaling and triggers PKC$\delta$-mediated apoptosis, contributing to microvascular damage [147].

These mechanistic insights have informed several therapeutic strategies that have shown promise in this model. Intravitreal AAV2-mediated SIRT1 overexpression improved both neuronal and vascular retinal function [148]. Additionally, natural compounds such as astragaloside IV (AR inhibitor) [149] and salidroside (oxidative stress modulator) [150] also preserved retinal integrity.

Beyond single-pathway interventions, the db/db model has been further used to examine interactions between hyperglycemia and dyslipidemia. Combined treatment with poloxamer 407 and high glucose aggravated retinal dysfunction more than either condition alone [55]. These observations support the use of this model in dissecting the complex metabolic contributions to diabetic retinopathy.

The ob/ob mouse carries a mutation in the leptin gene, resulting in leptin deficiency. The model has been valuable for studying leptin’s role in retinal disease. In a retinopathy of prematurity model, transgenic mice overexpressing leptin showed increased retinal neovascularization compared to wild-type littermates. In contrast, leptin-deficient ob/ob mice had significantly reduced ischemia-induced neovascularization. This effect is linked to leptin receptor signaling in retinal endothelial cells, where leptin activates VEGF mRNA expression [151].

Longitudinal studies in BTBR ob/ob mice have mapped the progression of DR. Obesity appears by 2 weeks, followed by hyperglycemia at 3 weeks. By 6 weeks, early retinal dysfunction and inner retinal thinning occur. Importantly, neuroinflammatory changes—such as glial activation, leucostasis, and microglial phenotype shifts—precede vascular degeneration and increased permeability [152].

Diet-induced models, particularly high-fat diet (HFD) feeding in C57BL/6J mice, are widely used to study the early stages and progression of type 2 DM and its complications, including diabetic retinopathy (DR). After 12 weeks of HFD feeding, mice typically develop prediabetes or early-stage diabetes with significant retinal deficits. Electroretinography reveals reduced scotopic and photopic responses, indicating impaired retinal sensitivity. At the molecular level, downregulation of key signaling proteins involved in calcium regulation and glucose transport has been observed. Notably, similar molecular changes are seen in human DR retinas, underscoring the translational relevance of this model [153].

For example, extended HFD exposure helps to characterize the temporal progression of DR. Mice fed a 60% fat diet for up to 12 months exhibit neural and vascular abnormalities, including retinal nerve infarcts, vascular leakage, and reduced vascular density, even in the absence of marked hyperglycemia [154]. These findings support the notion that retinal neurodegeneration may precede overt vascular pathology. However, standard HFD-based models often require prolonged feeding periods to induce significant pathology, posing practical limitations. To overcome these time constraints, accelerated protocols have been developed. One approach combines HFD with low-dose streptozotocin (STZ) administration, delivered via osmotic mini-pump, to more rapidly induce type 2 DM. This non-transgenic method enables the study of retinal vascular pathology, including visualization techniques such as fluorescent gelatin vascular casting [155].

Importantly, while HFD-based models are valuable, the diets used in many studies—containing 60% of total caloric intake from fat—far exceed the fat consumption typically observed in human populations, which generally ranges from 28.5% to 46.2% of total daily energy intake [156]. Additionally, growing evidence suggests that both the amount and quality of dietary carbohydrates play a crucial role in the progression of diabetes. Sucrose-rich diets have been shown to worsen hyperglycemia, while fiber-rich or resistant starch diets improve glycemic control by modulating gastric emptying and gut microbiota composition. These findings support the inclusion of carbohydrate composition as a critical factor in model design and interpretation [157]. Therefore, a more recent mouse model has employed a medium-fat diet (e.g., 34.5% of total energy from fat) combined with fructose supplementation in drinking water and low-dose STZ, which more closely mimics Western dietary patterns characterized by excess intake of both fat and sugar. This model better replicates the metabolic and vascular complications of human type 2 DM [158].

Overall, diet-induced models provide a physiologically relevant platform for studying DR progression in the context of type 2 DM. Incorporating more realistic macronutrient compositions, including both fat and carbohydrate sources, may enhance the translational fidelity of these models. Accelerated variants further improve practicality by reducing experimental timeframes, facilitating the investigation of underlying mechanisms and therapeutic interventions.