Research advances on the impact of gestational diabetes mellitus on offspring neurodevelopment Authors: Jiale Liu, Yanpo Gao, Yan Yi, Tao Wang, Yahui Chen, Jianying Ma Published: 2026 ABSTRACT Gestational diabetes mellitus (GDM) is one of the most common complications during pregnancy and can severely impair fetal neurodevelopment, leading to abnormalities in emotional, cognitive, and memory functions. However, the precise mechanisms by which GDM affects fetal neural development remain incompletely understood, which is closely associated with the lack of animal models that accurately replicate the key clinical and pathological features of human GDM. Therefore, this review summarizes and evaluates current methodologies for establishing GDM animal models, elucidates the impact of GDM on offspring neurodevelopment and its potential underlying mechanisms, and provides a theoretical foundation for developing and assessing appropriate GDM animal models, clarifying the pathophysiological mechanisms of GDM, and exploring innovative therapeutic strategies. Introduction Gestational diabetes mellitus (GDM) represents one of the most common pregnancy-related complications, with its global prevalence demonstrating a consistent upward trend. Recent epidemiological data from the International Diabetes Federation (IDF) estimate the worldwide prevalence of GDM at approximately 14.0%, with China reporting a slightly higher rate of 14.8% [1, 2]. This metabolic disorder imposes substantial burdens on both public health systems and socioeconomic resources. Recent evidence suggests that environmental disruptors, particularly endocrine-disrupting chemicals (EDCs), may contribute to the rising prevalence of GDM [3]. Humans have long been exposed to these compounds through diet and other daily consumables, and their widespread use has raised increasing concerns regarding public health and chronic metabolic disorders [4]. Chemicals such as bisphenols, phthalates, and per- and polyfluoroalkyl substances can interfere with hormonal regulation of glucose metabolism, leading to insulin resistance, β-cell dysfunction, and placental alterations [5]. Epidemiological and mechanistic studies further indicate that maternal exposure to these environmental toxicants is associated with an elevated risk of GDM [6, 7], highlighting environmental exposure as an emerging and potentially modifiable risk factor alongside traditional metabolic determinants. The maternal hyperglycemic intrauterine environment induced by GDM increases not only the risk of adverse pregnancy outcomes (including preeclampsia, preterm birth, macrosomia, neonatal hypoglycemia, and dystocia) [8], but also the offspring’s susceptibility to obesity, diabetes mellitus, cardiovascular diseases, and neurodevelopmental disorders [9, 10]. A growing body of evidence has demonstrated that GDM adversely affects fetal neurodevelopment, leading to postnatal abnormalities in emotional regulation, cognitive function, and memory processing in offspring [11, 12]. Epidemiological studies indicate that GDM is associated with an increased risk of neurodevelopmental disorders in offspring. Although precise prevalence rates vary across studies due to differences in diagnostic criteria and follow-up duration. Compared with offspring not exposed to GDM, offspring exposed to GDM exhibit a median prevalence of 16.3% for Autism Spectrum Disorder(ASD) and 14.4% for Attention-Deficit/Hyperactivity Disorder(ADHD). However, the underlying mechanisms remain incompletely elucidated. Animal models have significantly advanced our understanding of disease pathogenesis and served as cornerstones for discovering novel therapeutic targets and developing pharmacological interventions [13]. In GDM research, establishing animal models that accurately recapitulate the key clinical and pathological features of human GDM carries substantial clinical significance. Therefore, this review systematically synthesizes current methodologies for establishing various GDM animal models, elucidates the impact of GDM on offspring neurodevelopment and its underlying mechanisms, and provides a theoretical foundation for developing and validating appropriate GDM animal models, deciphering the pathophysiological mechanisms of GDM, and facilitating innovative therapeutic strategies. Establishment of GDM animal models and their applications in studying offspring neurodevelopment Establishment methodologies for GDM animal models Due to ethical constraints and sample heterogeneity in human studies, developing animal models that closely recapitulate the pathological features of human GDM has become a critical research pathway. The establishment of standardized GDM animal models is essential for investigating disease mechanisms and developing clinical interventions. Common methodologies for creating GDM animal models are summarized in Table 1. Streptozotocin (STZ), the most widely employed pharmacologic agent for GDM modeling, induces sustained hyperglycemia through selective destruction of pancreatic β-cells via DNA alkylation [14, 15]. Precise dose titration is critical for model success, as high-dose STZ triggers acute pancreatic necrosis and fatal hypoglycemia, whereas insufficient dosing fails to achieve adequate β-cell impairment [16]. Alloxan (ALX) selectively destroys pancreatic β-cells via superoxide radical generation, thereby inhibiting insulin synthesis. Its modeling efficacy in rodents is comparable to STZ [17, 18]. Although both compounds are applicable to large animal models such as porcine and ovine species [19–23], ALX is typically restricted to short-term experiments due to its more pronounced toxicity profile [24]. In contrast, the standardized single-dose STZ paradigm offers both methodological robustness and cost-effectiveness, enabling reliable induction of hyperglycemia with reduced inter-experimental variability. This approach facilitates temporal control over diabetes progression, permitting dose-dependent modeling of graded metabolic dysfunction from subclinical to overt diabetic states. Based on the available literature, low to middle-dose STZ was administered during pregnancy to induce a GDM-like model characterized by partial β-cell dysfunction [25–27], whereas higher STZ doses have been employed in some studies to induce more severe gestational hyperglycemia, primarily to model the metabolic consequences of pronounced maternal hyperglycemia rather than the precise pathogenesis of clinical GDM [28, 29]. Notwithstanding the irreversible β-cell cytotoxicity of this model [16], single-dose STZ induction remains a widely used and practical approach for studying gestational hyperglycemia and β-cell dysfunction. However, it should be acknowledged that this model does not fully recapitulate the multifactorial pathophysiology of human GDM, particularly obesity-associated insulin resistance. Therefore, its continued utility lies primarily in mechanistic investigations of hyperglycemia-driven metabolic and developmental alterations during pregnancy. Recognizing obesity as a well-established risk factor for GDM, high-fat diet (HFD) and high-sugar diet (HSD) induction models elicit insulin resistance after 6–10 weeks of intervention, more closely mimicking type 2 diabetes mellitus characteristics [30–32]. This approach has been successfully adapted to large animal models including canines and ovine species [33, 34]. The primary advantage of dietary induction lies in its preservation of intact gestational metabolic regulation systems. However, this method presents limitations such as prolonged modeling duration and relatively modest glycemic fluctuations. Currently, no standardized protocols exist for dietary composition or intervention duration. The combined use of HFD and low-dose STZ is designed to model the coexistence of peripheral insulin resistance and impaired β-cell compensatory capacity. HFD feeding primarily induces insulin resistance through lipid accumulation and chronic low-grade inflammation, resulting in impaired insulin signaling [35, 36]. In contrast, low-dose STZ selectively targets pancreatic β-cells via GLUT2-mediated uptake and induces partial β-cell dysfunction through DNA damage without causing complete cell death [25, 37, 38]. The combination of increased insulin demand driven by insulin resistance and limited β-cell compensatory capacity creates a metabolic state characterized by inadequate insulin compensation. Studies demonstrate that 4–8 weeks of pre-gestational HFD preconditioning, combined with first-trimester low-dose STZ administration, enhances the phenotypic penetrance of glucose intolerance and dyslipidemia by inducing partial β-cell dysfunction in the context of pregnancy-associated metabolic stress [39]. It should be noted that the term “first trimester” in rodents is used as a temporal analogy to early gestational stages in humans, rather than implying a strictly defined anatomical or functional trimester equivalence. This diet-pharmacology co-induction approach effectively integrates gestational metabolic dysregulation with tissue-specific functional alterations, though it requires precise temporal control of the STZ treatment window. Aberrant androgen levels demonstrate a significant association with dysregulated glucose and lipid metabolism, primarily mediated through androgen receptor binding and subsequent signaling modulation [40]. Experimental studies in rat models reveal that elevated androgen concentrations substantially increase the risk of insulin resistance and diabetes mellitus, which is closely linked to their disruptive effects on metabolic homeostasis [41]. Regarding pregnancy-specific hormonal regulation, placental growth hormone (PGH) has been identified as a pivotal modulator. Animal studies demonstrate that PGH overexpression primarily induces systemic insulin resistance accompanied by compensatory hyperinsulinemia, while significant alterations in glycemic levels may occur depending on the extent of metabolic compensation and disease progression [42]. Significant progress has been made in developing genetically engineered animal models, among which leptin receptor mutant (db/+) mice have been most extensively utilized. This model disrupts the leptin signaling pathway, resulting in pathological obesity and insulin resistance that recapitulate core GDM phenotypes. Notably, female db/+ mice can develop gestational glucose intolerance accompanied by pregnancy-associated weight gain, demonstrating partial pathophysiological concordance with human GDM and providing a valuable platform for investigating genetic susceptibility to metabolic dysregulation [43, 44]. In addition, prolactin receptor heterozygous mutants (PrlR+/−) and serotonin receptor 2B knockout (Htr2B−/−) mouse models have shown considerable research value by exhibiting reduced pancreatic β-cell mass and impaired glucose tolerance [45, 46]. Compared with pharmacologically induced models, genetically engineered approaches offer advantages in dissecting gene-specific mechanisms underlying metabolic regulation. However, the db/+ model is characterized by notable phenotypic variability and a strong dependence on genetic background, which may result in variable metabolic outcomes across studies. Therefore, rigorous genotyping and careful control of experimental strains are essential to ensure reproducibility and accurate interpretation of experimental results [47]. In the field of large animal model development, partial pancreatectomy has been established as a classical surgical intervention for inducing persistent hyperglycemia through reduction of functional islet cell mass and consequent insulin deficiency. This technique has been successfully implemented in species with prolonged gestational periods, including porcine, canine, and ovine models [48–50]. However, it should be noted that this invasive surgical approach presents significant limitations. Primarily, the prolonged period required for post-operative hyperglycemia development fundamentally conflicts with the relatively short gestational cycle of rodent models. Secondly, surgical trauma may induce reproductive system complications including ovarian dysfunction and embryo implantation failure, resulting in substantially elevated abortion rates [16, 51]. Consequently, pancreatectomy is only applicable to large mammalian species capable of sustaining long-term pregnancy, necessitating rigorous species-specific suitability assessments during model selection. The establishment of GDM animal models encompasses diverse methodologies, each exhibiting distinct advantages and limitations in experimental costs and applicability, necessitating judicious selection based on research objectives and experimental conditions. Pharmacologically induced models, primarily utilizing STZ and ALX, rapidly induce hyperglycemia through selective β-cell destruction. Among these, the single-dose STZ protocol has become the preferred choice for rodent models due to its cost-effectiveness and standardized procedures. Diet-induced models employing long-term HFD/HSD feeding recapitulate insulin resistance while preserving gestational metabolic regulation integrity, closely mimicking type 2 diabetes characteristics. However, these models are constrained by prolonged induction periods, modest glycemic fluctuations, and lack of standardized protocols. Combined induction models integrate dual pathological manifestations of insulin resistance and islet dysfunction, allowing precise temporal control of drug administration to enhance modeling success rates, yet require careful optimization of the synergistic effects between pharmacological and dietary interventions. Hormone modulation models elucidate hormone-metabolism interactions but depend on specific animal species and involve complex experimental designs. Genetically engineered models precisely replicate genetically predisposed GDM phenotypes, though genetic instability and variable induction rates may compromise experimental reproducibility. Surgical models, establishing persistent hyperglycemia through functional islet mass reduction, are suitable for large animals with extended gestation periods but carry risks of reproductive complications and demonstrate limited clinical translatability due to temporal mismatches with rodent pregnancy durations. In summary, rodent models with STZ induction remain the preferred choice for basic research due to their cost-effectiveness and high efficiency, while diet–drug combination models are better suited for investigating metabolic regulation mechanisms. Genetically engineered models offer unique advantages in dissecting genetic susceptibility, and large animal models provide an essential experimental foundation for preclinical translation. It should be noted that, unlike human pregnancy, the short gestational period and distinct placental physiology of rodents impose substantial limitations on modeling trimester-specific metabolic changes [52]. Precise prediction or replication of human trimester-dependent pathophysiology during gestation is therefore challenging in rodent models. Consequently, these models are more appropriate for elucidating general mechanistic of pregnancy-associated metabolic dysregulation rather than exact temporal parallels to human trimesters. Accordingly, study design should comprehensively consider the model’s pathological relevance, experimental timeline, species-specific gestational characteristics, and technical feasibility to achieve optimal alignment between clinical observations and basic mechanistic research. Research on the neurodevelopmental impacts in offspring based on GDM animal models The establishment of animal models for studying the neurodevelopmental effects of GDM on offspring primarily involves rodents, utilizing two main strategies: pharmacological induction and dietary intervention (Table 2). Rodent models offer advantages such as low cost, short reproductive cycles, and high reproducibility of results. Compared to dietary induction, pharmacological induction is more widely adopted due to its operational simplicity, shorter modeling duration, and higher success rate. STZ serves as the core inducing agent. Although STZ can cross the placental barrier and enter the fetal circulation, its short half-life and limited placental transfer efficiency reduce the likelihood of direct teratogenic effects during early embryogenesis [53–55]. Notably, fetal growth retardation reported in some STZ-induced GDM models is generally considered to result from sustained maternal hyperglycemia and impaired metabolic regulation during pregnancy rather than direct fetal toxicity of STZ itself [56–58]. Further optimization of the dosage and timing of STZ administration can markedly reduce its direct toxicity to the embryo. Therefore, when administered at appropriate doses and gestational stages, STZ-based models remain a valuable and widely accepted approach for investigating maternal–fetal metabolic interactions, and the developmental consequences of maternal hyperglycemia [26, 27]. The use of STZ for pharmacological induction of GDM animal models has been traditionally predominant in studies investigating offspring neurodevelopment, although debates persist regarding the optimal timing and dosage of administration. To mimic the clinical features of mid- to late-stage GDM in humans, researchers often establish the model during mid-gestation in rodents to assess the impact of gestational hyperglycemia on offspring neurodevelopment. Studies employing a single intraperitoneal injection of 45 mg/kg STZ in Sprague-Dawley (SD) rats on gestational day 13 (GD13) have reported that offspring exposed to STZ-induced GDM exhibited reduced white matter myelination and increased cellular apoptosis in the brain, suggesting potential neurodevelopmental alterations [59]. Studies using a Wistar rat model with a single injection of 38 mg/kg STZ on GD9 have reported a significant association between increased oxidative stress in the offspring’s cerebral cortex and hippocampus and impaired cognitive function [12]. However, some researchers advocate for earlier induction timing, such as administering STZ on the day of conception, to minimize potential embryotoxicity. In an ICR mouse model receiving a single 150 mg/kg STZ injection on the day of conception, offspring exhibited reduced expression of hippocampal neuron differentiation–related genes, which was associated with alterations in synaptic plasticity [29]. Similarly, a C57BL/6J mouse model injected with 120 mg/kg STZ on the day of conception revealed that disrupted offspring brain transcriptomes may contribute to autism-like repetitive behaviors [11]. Thus, whether different gestational time points for GDM induction exert distinct effects on offspring neurodevelopment and whether the underlying mechanisms are consistent remain incompletely understood, warranting further investigation. Dietary induction, employing HFD or HSD to simulate metabolic dysregulation, offers the advantage of avoiding drug toxicity while better recapitulating the clinical phenotype of obesity-associated GDM. However, this approach is limited by prolonged modeling duration and insignificant glucose fluctuations [16]. Studies using a 10-week HFD-induced C57BL/6J mouse model demonstrated that upregulated expression of placental fibroblast growth factor 4 (FGF4), pro-inflammatory factors (CD68 and TNF-α) was strongly associated with an increased incidence of fetal neural tube defects (NTDs) [60]. Studies using combined HFD/HSD–induced GDM animal models have reported impaired cognitive performance in adult offspring, accompanied by alterations in hippocampal synaptic integrity and microglial function [61]. In contrast, HFD-alone models indicated that maternal immune activation (MIA) exacerbates offspring neuroinflammation and dysregulates the expression of neurodevelopment-associated genes [62]. Notably, non-specific effects secondary to dietary induction, such as aberrant lipid metabolism, may confound the independent analysis of GDM-specific mechanisms underlying offspring neurodevelopmental impairments [62]. The application of the novel insulin receptor antagonist S961 has introduced a new approach for establishing GDM models. In C57BL/6J mice, continuous administration from GD7 until delivery successfully replicated the phenotypic characteristics of neurodevelopmental abnormalities in GDM offspring. However, its placental permeability and developmental toxicity require systematic validation [63]. Currently, the STZ-induced GDM model remains the preferred methodology for investigating the neurodevelopmental impacts on GDM offspring, owing to its well-established modeling system and clearly defined mechanisms of action. Impact of GDM on offspring neurodevelopment and underlying mechanisms The influence of GDM on offspring neurodevelopment and its pathogenesis has attracted extensive attention. Studies using population experiments, animal experiments, and cell experiments have confirmed that GDM has profound effects on offspring neurodevelopment (Tables 3, 4 and 5). Effects of GDM on offspring nervous system development Embryonic nervous system The nervous system is characterized by highly complex and tightly regulated developmental processes. During early embryogenesis, particularly under conditions of maternal hyperglycemia present before or during early pregnancy, excessive transplacental glucose transfer can expose the developing embryo to a hyperglycemic intrauterine environment at a stage when fetal pancreatic β-cell function and. Table 1Commonly used methods for establishing GDM animal modelsExperimental animalsInduction methodsPharmacological inductionDietary inductionCombined dietary-pharmacological inductionHormonal inductionGenetic engineeringSurgical proceduresRat/mouseSTZ, ALXHFD, HSDHFD + Low dose STZAndrogen,PGHdb/+mouse, PrlR+/-mouse,Htr2B-/-mouse-DogSTZ, ALXHFD---Subtotal pancreatectomyPigSTZ, ALXHFD---Subtotal pancreatectomySheepSTZ, ALXHFD---Subtotal pancreatectomy Table 1Commonly used methods for establishing GDM animal modelsExperimental animalsInduction methodsPharmacological inductionDietary inductionCombined dietary-pharmacological inductionHormonal inductionGenetic engineeringSurgical proceduresRat/mouseSTZ, ALXHFD, HSDHFD + Low dose STZAndrogen,PGHdb/+mouse, PrlR+/-mouse,Htr2B-/-mouse-DogSTZ, ALXHFD---Subtotal pancreatectomyPigSTZ, ALXHFD---Subtotal pancreatectomySheepSTZ, ALXHFD---Subtotal pancreatectomy Commonly used methods for establishing GDM animal models Androgen, PGH db/+mouse, PrlR+/-mouse, Htr2B-/-mouse Table 2Establishment of animal models for studying neurodevelopmental outcomes in GDM offspringCategoryModeling MethodAnimal StrainDosageInduction TimingReferencesPharmacological inductionSTZSD rats45 mg/kgGestational day 13 [59] Pharmacological inductionSTZWistar rats38 mg/kgGestational day 9 [12] Pharmacological inductionSTZICR mice150 mg/kgDay of conception [29] Pharmacological inductionSTZC57BL/6Jmice120 mg/kgDay of conception [11] Pharmacological inductionSTZC57BL/6Jmice75 mg/kgTwo days prior to mating [64] Pharmacological inductionSTZSD rats50 mg/kgOne day prior to mating [65] Dietary inductionHFDC57BL/6J mice-High-fat diet feeding for 10 weeks, followed by mating after blood glucose normalization [60] Dietary inductionHFD + HSDSD rats-High-fat/high-sucrose diet feeding for 6 weeks, then mating with continued feeding through lactation [61] Dietary inductionHFDC57BL/6J mice-High-fat diet feeding for 6 weeks, then mating with continued feeding until pregnancy termination [62] Table 2Establishment of animal models for studying neurodevelopmental outcomes in GDM offspringCategoryModeling MethodAnimal StrainDosageInduction TimingReferencesPharmacological inductionSTZSD rats45 mg/kgGestational day 13 [59] Pharmacological inductionSTZWistar rats38 mg/kgGestational day 9 [12] Pharmacological inductionSTZICR mice150 mg/kgDay of conception [29] Pharmacological inductionSTZC57BL/6Jmice120 mg/kgDay of conception [11] Pharmacological inductionSTZC57BL/6Jmice75 mg/kgTwo days prior to mating [64] Pharmacological inductionSTZSD rats50 mg/kgOne day prior to mating [65] Dietary inductionHFDC57BL/6J mice-High-fat diet feeding for 10 weeks, followed by mating after blood glucose normalization [60] Dietary inductionHFD + HSDSD rats-High-fat/high-sucrose diet feeding for 6 weeks, then mating with continued feeding through lactation [61] Dietary inductionHFDC57BL/6J mice-High-fat diet feeding for 6 weeks, then mating with continued feeding until pregnancy termination [62] Establishment of animal models for studying neurodevelopmental outcomes in GDM offspring metabolic regulatory capacity remain immature [66, 67]. Rather than reflecting an absolute inability of the fetal pancreas to produce insulin, this state represents a limited and developmentally constrained capacity to adapt to sustained glucose overload, in conjunction with placental transport dynamics and immature fetal metabolic homeostasis [68]. Such early exposure to hyperglycemia has been shown to exert deleterious effects on early fetal neurodevelopment [69]. Previous studies have reported that GDM is associated with a broad spectrum of adverse pregnancy outcomes, including congenital anomalies such as NTDs [70]. While much of the classical evidence originates from models of pre-gestational diabetes, increasing data suggest that hyperglycemic exposure during critical windows of neural tube formation may similarly impair neural development under certain GDM-related conditions [71]. Specifically, an adverse intrauterine metabolic environment disrupts neural cell development, interferes with the transition from the neural plate to the neural tube, and thereby contributes to the occurrence of NTDs [72, 73]. Animal studies have demonstrated that NTDs are closely associated with high glucose–induced protein misfolding and Caspase-8–mediated programmed cell death [74]. Furthermore, in vitro studies have reported that hyperglycemic conditions upregulate thioredoxin-interacting protein (Txnip), enhancing its interaction with thioredoxin (Trx), an endogenous inhibitor of apoptosis signal-regulating kinase 1 (ASK1). This interaction has been associated with activation of the ASK1-dependent apoptotic pathway, which may contribute to increased neural stem cell (NSC) apoptosis in neuroepithelial cells and potentially increase susceptibility to fetal NTDs [75]. During the transition from the neural plate to the neural tube, successful neural tube closure requires adequate levels of myo-inositol (MI) [76]. Notably, studies have reported reduced MI levels in the brains of fetuses exposed to maternal hyperglycemia, including experimental models of GDM. These findings suggest that MI deficiency may represent an additional metabolic alteration associated with increased susceptibility to NTDs and potential changes in neural conduction velocity in offspring [77, 78]. Brain development In terms of brain function, studies assessing auditory evoked responses (AER) have reported prolonged postprandial AER latency in fetuses exposed to GDM. This finding has been interpreted as reflecting reduced efficiency in auditory information processing and may indicate alterations in the functional maturation of the fetal central nervous system [79]. Regarding cerebral metabolism, late-gestation imaging assessments have reported that although total brain volume remains unchanged in GDM-exposed fetuses, the choline/creatine ratio is reduced. This alteration has been interpreted as reflecting changes in neuronal membrane phospholipid metabolism and may be associated with alterations in myelination-related processes and brain microstructural organization, rather than serving as direct evidence of impaired synaptic plasticity or global developmental delay [80]. At the microstructural level, MRI texture quantitative analyses have reported neuroanatomical alterations in GDM-exposed neonates, including altered integrity of deep gray matter and white matter fiber tracts, as well as abnormal diffusion tensor parameters in specific regions such as the corpus callosum, thalamus, and right amygdala [81]. These microstructural abnormalities correlate significantly with neurobehavioral deficits, highlighting their potential clinical predictive value. Hippocampal development is closely associated with memory formation, episodic memory encoding, and spatial navigation, playing a critical role in long-term memory consolidation [82]. Animal studies have reported hippocampal structural alterations in offspring exposed to GDM, particularly in the CA1 region, including reduced neuronal density, decreased nucleolar volume, and altered synaptic morphology [29, 61]. In humans, population-based cohort studies suggest that prenatal exposure to GDM is associated with alterations in hippocampal volume and morphology, with limited evidence suggesting potential sex-specific effects. However, the precise laterality and subregional patterns of these changes remain inconsistent across studies and are not yet fully established [83]. Emotion and cognition GDM is associated with an increased risk of emotional and cognitive difficulties in offspring from childhood onwards, although reported effect sizes and the persistence of these outcomes into adulthood vary across studies [10]. Emotion regulation primarily relies on the prefrontal cortex-amygdala neural circuit, where the amygdala—a core brain region for emotional processing, mediates fear and anxiety responses, while the prefrontal cortex maintains emotional homeostasis by exerting top-down regulatory control to suppress excessive activation of downstream limbic systems [84, 85]. Clinical epidemiological studies suggest that prenatal exposure to GDM is associated with an increased risk of neurodevelopmental disorders, including ASD and ADHD, with most evidence based on diagnoses made during childhood and adolescence [86, 87]. ASD is characterized by deficits in social interaction and repetitive/stereotyped behaviors, whereas ADHD manifests as inattention and hyperactive/impulsive behaviors [88, 89]. Animal studies have reported that male GDM offspring exhibit reduced exploratory activity and increased repetitive behaviors in open-field tests, with sex-specific differences observed in behavioral outcomes. These findings raise the possibility that GDM-related neurodevelopmental alterations may involve sex-dependent mechanisms, although the underlying pathways remain to be fully elucidated [11]. In terms of cognitive function, population-based cohort studies suggest that prenatal exposure to GDM is associated with lower cognitive performance in offspring, including learning difficulties during childhood, with an increased risk compared to offspring from normoglycemic pregnancies [90]. The novel object recognition test further indicates that adult offspring exposed to GDM show impaired performance in discriminating between novel and familiar objects, suggesting alterations in recognition memory. Some studies have proposed that these behavioral changes may be related to disrupted hippocampal synaptic remodeling and altered microglial activity; however, the precise mechanistic links remain to be fully clarified [61]. Morris water maze testing has suggested sex- and age-dependent differences in GDM-exposed offspring. Several studies report that male offspring tend to exhibit more persistent impairments in spatial learning and memory, particularly into adulthood, whereas females often show milder or more transient deficits during early developmental stages [12]. However, these patterns are not entirely consistent across studies. Collectively, these findings indicate that structural and functional abnormalities in the prefrontal cortex and hippocampus constitute the neural basis of cognitive impairments in GDM offspring, with pathophysiological mechanisms likely involving neuroinflammation and dysregulated synaptic plasticity. Mechanisms underlying the impact of GDM on offspring neurodevelopment Although existing studies have confirmed the significant impact of GDM on offspring brain development, emotional regulation, and cognitive function, the underlying biological mechanisms remain to be systematically elucidated. The maternal metabolic disturbances induced by GDM collectively mediate neurodevelopmental abnormalities in offspring through three interconnected pathways: neuroinflammation, oxidative stress damage, and aberrant epigenetic regulation. These mechanisms interact synergistically, creating a vicious cycle that exacerbates the imbalance in the fetal neurodevelopmental microenvironment (Fig. 1). This pathological cascade ultimately disrupts neurogenesis and neuronal differentiation, leading to persistent alterations in emotional and cognitive functions in offspring. Fig. 1The mechanism of GDM’s influence on the neurodevelopment of offspring.The intrauterine hyperglycemic environment induced by GDM increases the expression of inflammatory factors and pro-apoptotic factors by triggering the expression of the advanced glycation end product receptor (RAGE) in the fetal hippocampus, thereby interfering with the proliferation and survival of fetal hippocampal neurons.GDM exacerbates intrauterine fetal oxidative stress, resulting in an increase in the DNA damage marker p-H2AX and a decrease in the synthesis of DNA repair-related proteins p-Chk1/2 and p53. Concurrently, upregulation of methylglyoxal (MG) in the fetal brain inhibits Glo1 activity—impairing the detoxification capacity of neural cells—while directly disrupting the balance of neural progenitor cell (NPC) proliferation and differentiation. These processes collectively establish a ROS–MG feed-forward interaction, which exacerbates oxidative stress and ultimately impairs fetal neurodevelopment The mechanism of GDM’s influence on the neurodevelopment of offspring.The intrauterine hyperglycemic environment induced by GDM increases the expression of inflammatory factors and pro-apoptotic factors by triggering the expression of the advanced glycation end product receptor (RAGE) in the fetal hippocampus, thereby interfering with the proliferation and survival of fetal hippocampal neurons.GDM exacerbates intrauterine fetal oxidative stress, resulting in an increase in the DNA damage marker p-H2AX and a decrease in the synthesis of DNA repair-related proteins p-Chk1/2 and p53. Concurrently, upregulation of methylglyoxal (MG) in the fetal brain inhibits Glo1 activity—impairing the detoxification capacity of neural cells—while directly disrupting the balance of neural progenitor cell (NPC) proliferation and differentiation. These processes collectively establish a ROS–MG feed-forward interaction, which exacerbates oxidative stress and ultimately impairs fetal neurodevelopment Neuroinflammation GDM is associated with a state of low-grade maternal inflammation, characterized by altered cytokines and chemokines. Increasing evidence suggests that this inflammatory milieu may interact with placental immune and metabolic functions. Such placental alterations have been proposed to influence fetal immune development and contribute to a pro-inflammatory tendency, which could potentially increase susceptibility to neuroinflammatory processes in the developing brain [62, 91]. Animal studies have reported that sustained intrauterine hyperglycemia is associated with increased expression of the receptor for advanced glycation end-products (RAGE) in the offspring hippocampus. Activation of RAGE has been linked to inflammatory and pro-apoptotic signaling pathways, which may adversely affect hippocampal neuronal proliferation and survival. These findings suggest that RAGE-mediated signaling could contribute to neurodevelopmental vulnerability, although its precise role remains to be fully clarified [92]. At the molecular level, experimental studies have reported reduced expression of the anti-apoptotic gene Bcl-2 and brain-derived neurotrophic factor (BDNF) in offspring exposed to GDM. Alterations in immune-related markers have also been described, including changes in MHC-I and IL-1β expression, as well as increased levels of the pro-inflammatory cytokine TNF-α in the brain [93]. Notably, MHC-I and BDNF, as key regulators of hippocampal neurogenesis and synaptic plasticity, demonstrate a dose-dependent relationship with neurodevelopment: physiological concentrations effectively promote neuronal differentiation and synaptic maturation, whereas insufficient expression leads to delayed neurodevelopment [94, 95]. Beyond its classical neurotrophic effects, BDNF exerts neuroprotective actions through multiple mechanisms, including suppression of apoptosis, clearance of oxidative stress products, and modulation of autophagy pathways [96]. Although IL-1β belongs to the pro-inflammatory cytokine family, it plays a crucial regulatory role in maintaining hippocampal neurogenesis at physiological levels [97]. In vitro studies have suggested that sustained hyperglycemic conditions are associated with increased microglial activation in the hippocampus, accompanied by elevated release of inflammatory mediators such as IL-1α [61]. Activated microglia have been implicated in alterations of neuronal function and synaptic plasticity through inflammatory signaling and aberrant phagocytic activity, and may contribute to disruptions in neuron–glia communication [98, 99]. However, the extent to which these processes directly drive neuronal dysfunction remains to be fully established. Importantly, these molecular alterations should not be interpreted as independent mechanisms. Research report indicate that sustained inflammatory signaling can impair neurotrophic support and influence microglia–neuron interactions [100, 101]. Accumulating evidence suggests that RAGE signaling, BDNF, MHC I–mediated immune modulation, and microglial activation form an interconnected network of inflammatory and neurotrophic pathways [102–105]. Dysregulation of this network may jointly affect hippocampal neurogenesis, synaptic function, and neuronal survival in GDM-exposed offspring. This disruption of neural microenvironment homeostasis may represent a key underlying one mechanism of neurodevelopmental abnormalities in GDM-exposed offspring. Oxidative stress GDM has been associated with oxidative stress-related alterations that may contribute to adverse neurodevelopmental outcomes in offspring. Experimental studies have reported increased levels of reactive oxygen species (ROS) and lipid peroxidation products in the cerebral cortex and hippocampus, accompanied by reduced superoxide dismutase (SOD) activity and glutathione content. These findings are consistent with a potential impairment of antioxidant defense systems [12]. First, excessive ROS disrupts mitochondrial energy metabolism, induces DNA damage, and promotes protein dysfunction. These molecular alterations have been associated with neuronal apoptosis and synaptic signaling impairment [106]. In HG-treated neural stem cell models, increased expression of DNA damage markers (p-H2A.X) and activation of DNA repair-related factors (p-Chk1/2, p53) have been reported, suggesting that hyperglycemic conditions may perturb genomic stability and transcriptional regulation in NSCs, with potential implications for neurodevelopment [107]. Second, lipid peroxidation–derived aldehydes, particularly 4-hydroxy-2-nonenal (4-HNE), have been reported in diabetic models to covalently modify neuronal structural proteins. These modifications have been associated with features such as neurofibrillary changes and altered synaptic plasticity, possibly involving effects on neuronal membrane properties [108–110]. Although direct evidence in GDM models remains limited, GDM-associated dysglycemia may similarly favor the accumulation of methylglyoxal (MG) in the developing brain. Elevated MG has been reported to suppress glyoxalase-1 (Glo1) activity under hyperglycemic conditions, which may reduce cellular detoxification capacity and potentially enhance oxidative stress via a ROS–MG feed-forward interaction [111, 112]. In parallel, experimental studies suggest that decreased Glo1 activity may disrupt the proliferation–differentiation balance of neural progenitor cells (NPCs), with possible implications for NPC maintenance and cortical neurogenesis [113]. Collectively, these findings suggest that GDM-induced oxidative stress may underlie neurodevelopmental deficits through multiple cooperating pathways, such as inducing DNA damage, causing protein dysfunction, and disrupting NPC homeostasis. These effects collectively establish a critical pathological foundation for neurological impairments in offspring. Epigenetics Epigenetics refers to heritable alterations in gene function that occur without changes to the DNA sequence, mediated through mechanisms such as DNA methylation, histone modifications, and non-coding RNA regulation, ultimately giving rise to epigenetic phenomena [114]. Normal epigenetic regulation is essential for maintaining cellular growth and metabolism, while its dysregulation is closely associated with the pathogenesis of various diseases. Accumulating evidence, primarily derived from animal and cellular models, indicates that GDM can interfere with epigenetic regulatory mechanisms relevant to neurodevelopment, including alterations in chromatin structure, disruptions of DNA methylation patterns, and modulation of miRNAs involved in NSC/NPC differentiation [115]. In humans, however, evidence for GDM-associated epigenetic alterations is largely indirect and is mainly based on analyses of placenta, cord blood, and other accessible tissues, rather than direct assessments of the developing nervous system [116, 117]. At the DNA methylation level, increased ROS levels associated with GDM have been suggested to influence DNA methylation patterns, particularly in experimental models of early embryonic development [118, 119]. DNA methylation predominantly occurs at CpG dinucleotide sites and typically exerts gene silencing effects when located in promoter regions [120]. Notably, several studies have reported alterations in DNA methylation patterns in offspring exposed to GDM, although most evidence is derived from experimental or surrogate tissue analyses. Human NPC models suggest that hyperglycemic conditions may influence methylation patterns of genes involved in neural tube development [121]. Animal studies have also described changes in methylation profiles of hippocampal genes related to synaptic plasticity [122]. These epigenetic modifications display region-specific patterns. For instance, the glycoprotein M6A (GPM6A) gene has been reported to exhibit promoter hypomethylation accompanied by increased mRNA expression in the hippocampus of GDM-exposed offspring, whereas an opposite pattern has been observed in the cerebral cortex [123]. Additionally, hypermethylation of the sterol regulatory element binding transcription factor 2 (SREBF2) gene has been described in fetal brain tissues from GDM models and has been associated with alterations in cholesterol metabolism–related pathways, which may have implications for neuronal development [124]. Recent studies have reported that differentially methylated genes identified in fetal rat hippocampal tissues exposed to GDM are enriched in pathways related to synaptogenesis, neural signaling transduction, and cognitive function [29]. These findings suggest a potential link between GDM-associated DNA methylation changes and neurodevelopment-related pathways, although further studies are needed to clarify their functional relevance. Non-coding RNA regulatory mechanisms have been increasingly implicated in the potential link between GDM exposure and altered neurodevelopment in offspring. In specific experimental models, GDM exposure has been associated with upregulation of miRNA-30b in offspring brain tissue, which has been reported to target the Sirtuin1 gene and may influence NSC differentiation [125]. Within this experimental context, reduced Sirtuin1 expression has been associated with increased astrocyte generation, whereas its influence on neuronal differentiation may be relatively more substantial [126]. Similarly, in vitro studies have reported upregulation of miR-5b-2p under GDM-related conditions, which was accompanied by reduced expression of its target gene methyl-CpG-binding protein 2 (MeCP2) [127]. In these experimental models, MeCP2 downregulation has been associated with alterations in synaptic protein expression and markers related to neurodevelopment [128]. Notably, there exists crosstalk between DNA methylation and non-coding RNA regulation within the epigenetic regulatory network. Evidence from selected studies suggests that GDM-induced miRNA dysregulation may exert cascade effects by influencing DNA methyltransferase activity, although this interaction likely depends on experimental context and developmental stage [115, 127]. Table 3The Effect of GDM on neurodevelopmental outcomes in offspringGestational ageMechanisms of actionResultReferences27–36 weeks-The fetal AER latency↑,longer latency impaired brain function[79]34–38 weeksCho/Cr↓ in the fetal brainBrain development↓,brain metabolism and neuronal membrane phospholipid turn-over are altered[80]After deliveryIL-1β↑,TNF-α↑,Toll-like receptor↑ in the fetal brainDeep gray and white matter development↓ ,corpus callosum, thalamus, and right amygdala structure↓[81]After delivery(childhood)-Left hippocampal CA1 subregion↓ in both boys and girls, right hippocampal CA1 subregion volume↓ in boys[83]24 weeks-Offspring scores in all intelligence and memory indices↓,frequency of ASD and ADHD↑[90]After delivery(childhood)MBH/WM↑ in the brain of fetal diagnosed with GDM before 26 weeks of gestationGliocyte proliferation↑ in the hypothalamus[129]27-31weeks-No effect of fetal AER latency and fetal brain activity[130] Table 3The Effect of GDM on neurodevelopmental outcomes in offspringGestational ageMechanisms of actionResultReferences27–36 weeks-The fetal AER latency↑,longer latency impaired brain function[79]34–38 weeksCho/Cr↓ in the fetal brainBrain development↓,brain metabolism and neuronal membrane phospholipid turn-over are altered[80]After deliveryIL-1β↑,TNF-α↑,Toll-like receptor↑ in the fetal brainDeep gray and white matter development↓ ,corpus callosum, thalamus, and right amygdala structure↓[81]After delivery(childhood)-Left hippocampal CA1 subregion↓ in both boys and girls, right hippocampal CA1 subregion volume↓ in boys[83]24 weeks-Offspring scores in all intelligence and memory indices↓,frequency of ASD and ADHD↑[90]After delivery(childhood)MBH/WM↑ in the brain of fetal diagnosed with GDM before 26 weeks of gestationGliocyte proliferation↑ in the hypothalamus[129]27-31weeks-No effect of fetal AER latency and fetal brain activity[130] The Effect of GDM on neurodevelopmental outcomes in offspring Table 4The effects of GDM on neurodevelopment in animal offspringAnimal speciesMechanism of actionResultBehavioral manifestationsReferencesC57BL/6J miceα-syn↑,Parkin↑,Htt↑,Hip1↑,Hippi↑,Caspase8↑ in the fetal brainIncidence of NTDs↑,nerve cell death↑- [74] SD rats-MI↓ in the GDM fetal brain, supplement MI: Incidence of NTDs↓- [78] SD ratsRAGE↑,NF-кB↑ in the fetal rats hippocampusNeuroinflammation↑,nerve cell apoptosis↑ in the fetal rats hippocampusAnxiety-like behavior↓,cognitive function↓ [92] Wistar ratsBcl-2↓,IL-1β↓,MHC-Ⅰ,BDNF↓,TNF-α↑ in the fetal brainNeural plasticity↓,hippocampal neurogenesis↓,Neuroinflammation↑- [93] Wistar ratsROS↑,GSH↓,SOD↓,CAT↓ in the fetal brainOxidative stress↑Anxiety behavior ↓,spatial learning ↓,working memory↓ [12] C57BL/6JmiceMyo1d↓,Nde1↓,Prox1↓,Neurod1↓,bhlhe22↓ in the fetal frontal cortex and striatumCortical development↓,forebrain development↓,neuronal circuit↓,network of neurodevelopment↓Repetitive behavior↑in MBT, displacement distance↓,exploratory activity index↓ in OFT [11] SD ratsIFNγ↑,IL-1α↑,IL-4↑,MCP-1↑,TNF-α↑ in the fetal brain, CX3CR1↓ in the fetal microgliaInflammatory response in the fetal brain↑,density of the CA1 pyramidal neuronal layer↓,hippocampal synaptic integrity↓Cognitive competence↓ [61] C57BL/6J miceMiR-27a↑,GLCM↓,GCLC↓,GSTA1↓,Nrf2↓, in the fetal brainIncidence of NTDs↑,Oxidative stress↑- [131] Wistar ratsROS↑,CAT↓ ,GSH/GSSG↓ in the cerebral cortexOxidative stress↑Anxiety-like levels↑ in F2 male offspring, Spatial memory↓ in F2 female offspring [132] Table 4The effects of GDM on neurodevelopment in animal offspringAnimal speciesMechanism of actionResultBehavioral manifestationsReferencesC57BL/6J miceα-syn↑,Parkin↑,Htt↑,Hip1↑,Hippi↑,Caspase8↑ in the fetal brainIncidence of NTDs↑,nerve cell death↑- [74] SD rats-MI↓ in the GDM fetal brain, supplement MI: Incidence of NTDs↓- [78] SD ratsRAGE↑,NF-кB↑ in the fetal rats hippocampusNeuroinflammation↑,nerve cell apoptosis↑ in the fetal rats hippocampusAnxiety-like behavior↓,cognitive function↓ [92] Wistar ratsBcl-2↓,IL-1β↓,MHC-Ⅰ,BDNF↓,TNF-α↑ in the fetal brainNeural plasticity↓,hippocampal neurogenesis↓,Neuroinflammation↑- [93] Wistar ratsROS↑,GSH↓,SOD↓,CAT↓ in the fetal brainOxidative stress↑Anxiety behavior ↓,spatial learning ↓,working memory↓ [12] C57BL/6JmiceMyo1d↓,Nde1↓,Prox1↓,Neurod1↓,bhlhe22↓ in the fetal frontal cortex and striatumCortical development↓,forebrain development↓,neuronal circuit↓,network of neurodevelopment↓Repetitive behavior↑in MBT, displacement distance↓,exploratory activity index↓ in OFT [11] SD ratsIFNγ↑,IL-1α↑,IL-4↑,MCP-1↑,TNF-α↑ in the fetal brain, CX3CR1↓ in the fetal microgliaInflammatory response in the fetal brain↑,density of the CA1 pyramidal neuronal layer↓,hippocampal synaptic integrity↓Cognitive competence↓ [61] C57BL/6J miceMiR-27a↑,GLCM↓,GCLC↓,GSTA1↓,Nrf2↓, in the fetal brainIncidence of NTDs↑,Oxidative stress↑- [131] Wistar ratsROS↑,CAT↓ ,GSH/GSSG↓ in the cerebral cortexOxidative stress↑Anxiety-like levels↑ in F2 male offspring, Spatial memory↓ in F2 female offspring [132] The effects of GDM on neurodevelopment in animal offspring Cortical development↓,forebrain development↓,neuronal circuit↓, network of neurodevelopment↓ Table 5The effect of GDM on Neural Cell Development In VitroCell linesModelMechanism of actionReferencesC17.2 mouse NSCs lines obtained from European Collection of Cell CultureNSCs were cultivated in 5 mM glucose, with 10% fetal bovine serum,100 U/ml penicillin,100 µg/ml streptomycinMiR-17↑,Txnip↑,Trx↑,Ask1↑ in the NSC[75]Mice mixed glial cultures linesMixed glial cultures were prepared from corteces of newborn CD-1 mice, exposed to 16.7mM glucose levelsNO↑,superoxide↑,IL-1α↑,IL-6↓, IL-13↓, G-CSF↓in primary microglial cultures[61]C17.2 mouse NSCs lines, obtained from ECACCNSCs were cultivated in 5 mM glucose, with 10% fetal bovine serum,100 U/ml penicillin,100 µg/ml streptomycinp-H2A.X↑,p-Chk1↑,p-Chk2↑,p53↑ in the NSC[107]Mice cerebral cortex culturesCortices from E12/13 mouse embryos, were cultivated in 0.5mM glutamine and 40ng/ml FGF2Methylglyoxal↑, Glo1↓, Par3↓, aPKC↓[113]Rat hippocampus cultures induce NSCs linesNSCs were isolated from the hippocampus of the fetal rat, were cultivated in consisted of DMEM/F12 with N2 supplement and FGF-2GLP-1R↑,GLP↓,Rock1↓ in the fetal thalamus cultures[133]NPC ACS-5003 linesHuman ACS-5003 neurons were cultivated in 25 mM high glucose for 4 days, subsequent 5 mM low glucose for another 4 daysRORA↓,CYP19A1↓,SOD2↓, in the NPC[134] Table 5The effect of GDM on Neural Cell Development In VitroCell linesModelMechanism of actionReferencesC17.2 mouse NSCs lines obtained from European Collection of Cell CultureNSCs were cultivated in 5 mM glucose, with 10% fetal bovine serum,100 U/ml penicillin,100 µg/ml streptomycinMiR-17↑,Txnip↑,Trx↑,Ask1↑ in the NSC[75]Mice mixed glial cultures linesMixed glial cultures were prepared from corteces of newborn CD-1 mice, exposed to 16.7mM glucose levelsNO↑,superoxide↑,IL-1α↑,IL-6↓, IL-13↓, G-CSF↓in primary microglial cultures[61]C17.2 mouse NSCs lines, obtained from ECACCNSCs were cultivated in 5 mM glucose, with 10% fetal bovine serum,100 U/ml penicillin,100 µg/ml streptomycinp-H2A.X↑,p-Chk1↑,p-Chk2↑,p53↑ in the NSC[107]Mice cerebral cortex culturesCortices from E12/13 mouse embryos, were cultivated in 0.5mM glutamine and 40ng/ml FGF2Methylglyoxal↑, Glo1↓, Par3↓, aPKC↓[113]Rat hippocampus cultures induce NSCs linesNSCs were isolated from the hippocampus of the fetal rat, were cultivated in consisted of DMEM/F12 with N2 supplement and FGF-2GLP-1R↑,GLP↓,Rock1↓ in the fetal thalamus cultures[133]NPC ACS-5003 linesHuman ACS-5003 neurons were cultivated in 25 mM high glucose for 4 days, subsequent 5 mM low glucose for another 4 daysRORA↓,CYP19A1↓,SOD2↓, in the NPC[134] The effect of GDM on Neural Cell Development In Vitro Summary and prospect In recent years, significant progress has been made in research on the impact of GDM on offspring neurodevelopment. Researchers have established GDM animal models through pharmacological or dietary induction methods, gradually uncovering the underlying mechanisms by which GDM affects offspring neurodevelopment. Unlike classical T1DM models, the GDM model used reflects pregnancy-associated β-cell dysfunction and impaired insulin compensation. GDM exposure has been increasingly associated with alterations in offspring neurodevelopment, with accumulating evidence linking it to elevated risks of emotional and cognitive difficulties across different life stages, from childhood to adulthood. Some experimental studies have also raised the possibility of intergenerational or transgenerational effects, although further confirmation is needed [132]. The alterations in offspring emotional and cognitive functions are closely associated with GDM-induced neuroinflammation, oxidative stress, and epigenetic modifications. GDM has been associated with alterations in inflammatory signaling and increased oxidative stress in both maternal and fetal environments, which may contribute to redox imbalance and potentially affect neuronal development and synaptic function. In addition, accumulating evidence suggests that GDM exposure may be linked to long-term changes in offspring neurodevelopment, possibly mediated in part by epigenetic modifications such as DNA methylation and non-coding RNA regulation. Epigenetics, a regulatory mechanism that has garnered increasing attention in recent years, is widely studied for its potential involvement in neurodevelopmental alterations observed in GDM offspring. Given the plastic nature of epigenetic regulation, which may be modifiable through early interventions, further research in this area is warranted. The impact of GDM on offspring neurodevelopment is a complex, multi-layered process involving multiple mechanisms. Although numerous studies have investigated the effects of GDM on offspring neurodevelopment, many questions remain unresolved, particularly regarding the underlying mechanisms. Current GDM animal models remain controversial, and future research should enhance their clinical relevance to better replicate the complexity of human GDM. Additionally, the precise molecular mechanisms of GDM-induced neuroinflammation and oxidative stress, as well as their interactions, require systematic exploration. Furthermore, the role of epigenetics in GDM-associated neurodevelopmental alterations has not been fully elucidated. Future studies should employ multi-omics technologies to comprehensively dissect the regulatory networks involved. Currently, effective interventions to mitigate the adverse effects of GDM on offspring neurodevelopment remain to be established. Future research should explore various approaches, including pharmacological therapies, dietary modifications, and behavioral interventions, to alleviate GDM-induced neurodevelopmental impairments in offspring. In summary, the impact of GDM on offspring neurodevelopment has emerged as a critically important research field. A deeper understanding of its underlying mechanisms will not only facilitate the development of more effective clinical prevention and treatment strategies, but also provide novel scientific evidence for improving maternal and child health outcomes. In addition to mechanistic insights, future research should focus on the development of novel molecular biomarkers for the early diagnosis of GDM. Recent evidence suggests that circulating exosomal microRNAs (ExomiRs) may serve as promising early predictive markers due to their stability in maternal circulation and their regulatory roles in insulin signaling and β-cell function [70]. Specific ExomiRs have been shown to modulate pathways such as PI3K/AKT, Wnt, mTOR and glucose transporter regulation, highlighting their potential utility in identifying high-risk pregnancies before clinical onset [70]. The integration of molecular biomarkers with clinical risk assessment may improve early screening strategies and enable more precise prevention and intervention approaches for GDM. In short, rather than establishing a direct causal relationship, the available evidence suggests that GDM is associated with an increased vulnerability to neurodevelopmental alterations in offspring across different developmental stages. Most mechanistic evidence discussed in this review is derived from animal and cellular models. Human evidence remains relatively limited and is largely indirect. The heterogeneity in study design, diagnostic criteria for GDM, and neurodevelopmental outcome measures across studies may contribute to inconsistencies in reported findings. As a narrative review, this work summarizes existing literature but does not allow for causal inference or quantitative assessment of effect sizes.