Zebrafish as a preclinical model for diabetes mellitus and its complications: From monogenic to gestational diabetes and beyond

Abstract

With diabetes currently affecting 537 million people globally, innovative research approaches are urgently required. Zebrafish (Danio rerio) has emerged as a pivotal model organism in diabetes research, particularly valuable for developmental biology studies and preclinical therapeutic validation. Its rapid life cycle, optical transparency, and genetic tractability collectively enable efficient longitudinal observation of pathological progression and pharmacological responses. Utilizing zebrafish models, researchers have elucidated fundamental mechanisms governing islet development, β-cell dysfunction, and metabolic dysregulation. These experimental systems have significantly advanced our understanding of various diabetes subtypes, including type 1, type 2, gestational, and monogenic forms, while also facilitating mechanistic studies of diabetic complications such as retinopathy and nephropathy. Recent model refinements, particularly in simulating monogenic disorders and pregnancy-associated metabolic changes, promise to deepen our comprehension of disease pathophysiology and therapeutic interventions. Nevertheless, a persistent limitation lies in their incomplete recapitulation of human-specific physiological complexity and multi-organ metabolic interactions, factors that may influence translational applicability. Despite these constraints, zebrafish-based research continues to provide an indispensable platform for diabetes investigation, holding significant promise for alleviating the escalating global burden of this metabolic disorder.

Key Words: Zebrafish; Preclinical model; Diabetes mellitus; Monogenic diabetes; Gestational diabetes

Core Tip: Notably, zebrafish has been established as a model organism in diabetes research, providing unique insights into pathological progression, therapeutic efficacy evaluation, and complication pathogenesis owing to its accelerated developmental cycles, optical clarity, and genetic tractability. Particularly, its experimental versatility has enabled successful recapitulation of type 1 and type 2 diabetes pathogenesis while providing platforms for systematic investigation of monogenic variants and pregnancy-associated metabolic dysregulation. Collectively, these scientific advancements represent transformative potential for advancing mechanistic comprehension and developing targeted interventions against diabetes mellitus and its secondary complications.

INTRODUCTION

Diabetes mellitus, a metabolic disorder characterized by impaired insulin secretion or action, poses a substantial global health burden. Current epidemiological data from the International Diabetes Federation (IDF) reveal that 463 million adults worldwide had been diagnosed with diabetes by 2019[1], with projections indicating a concerning escalation to 700 million cases by 2045[2,3]. This multifactorial disease arises from complex interactions between genetic susceptibility and modifiable risk factors, including excessive caloric intake, sedentary behaviors, and obesogenic dietary patterns[4]. The rapid global urbanization has exacerbated its prevalence, frequently leading to debilitating metabolic derangements and life-threatening complications that severely compromise patient outcomes[5,6].

In diabetes research, diverse animal models ranging from porcine and non-human primates to rodents and zebrafish (Danio rerio) have been instrumental in elucidating disease mechanisms. Porcine models typically employ pancreatic resection, chemical induction, or genetic engineering to simulate diabetic conditions[7]. Non-human primate systems rely on high-fat/high-sucrose dietary challenges to induce metabolic dysfunction[8]. Rodent models predominantly utilize streptozotocin (STZ) administration or dietary manipulation for diabetes induction[9], whereas canine models require partial pancreatectomy to achieve metabolic dysregulation[10]. In contrast, zebrafish models of diabetes leverage advanced genetic tools including CRISPR/Cas9 genome editing, morpholino-based gene knockdown, and pharmacologic interventions delivered via microinjection or aquatic exposure[11,12].

Beyond diabetes research, zebrafish have been extensively validated as translational models across biomedical domains, demonstrating conserved pathophysiological features in oncology[13], hepatology[14], hematology[15], cardiology[16], and neuropsychiatry[17,18]. Their translational relevance stems from remarkable genomic conservation, sharing 70% protein-coding genes with humans and orthologs for 85% of human disease-associated loci (Figure 1A). This evolutionary proximity enables mechanistic interrogation of conserved biological pathways underlying human pathologies. Nevertheless, interspecies differences in metabolic regulation and organ system complexity necessitate cautious interpretation of findings, requiring complementary validation through mammalian systems.

Figure 1 Zebrafish as a model in diabetes research. A: Zebrafish as a model in diabetes due to genetic similarity with humans; B: Applications of zebrafish in modeling various types of diabetes; C: Diagram showing the resected zebrafish pancreas. T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; GDM: Gestational diabetes mellitus; GCKD: Glomerular cystic kidney disease.

This review systematically evaluates the expanding utility of zebrafish models in diabetes research, particularly their applications in type 1 diabetes mellitus (T1DM), type 2 diabetes mellitus (T2DM), gestational diabetes mellitus (GDM), and monogenic diabetes investigations. The species' miniature scale and prolific breeding render zebrafish particularly amenable to high-throughput phenotypic screening and genetic analyses[19]. Crucially, their pancreatic architecture - featuring conserved islet organization and β-cell regulatory mechanisms - addresses critical limitations inherent to rodent models[20]. Accumulating evidence positions zebrafish as indispensable tools for dissecting islet developmental biology, β-cell failure dynamics, and molecular cascades driving diabetic pathogenesis. By synthesizing current advances, this work aims to inform model selection strategies and catalyze innovative approaches in diabetes research.

METHODOLOGY

The literature search supporting this review employed a targeted exploration of biomedical databases including PubMed, EMBASE, Web of Science, Scopus, and Google Scholar. A combination of controlled vocabulary terms (MeSH headings) and free-text keywords such as "zebrafish model", "diabetes mellitus", "maturity-onset diabetes of the young (MODY)", and "diabetic complications" were iteratively applied. The search strategy encompassed all publication types without language restrictions, though non-English publications were retained only if accompanied by English abstracts. Retrieved articles were filtered to eliminate duplicates and irrelevant results. Additionally, the reference lists of selected articles were reviewed to identify other pertinent publications.

ZEBRAFISH MODELS IN VARIOUS DIABETES

The experimental tractability of zebrafish has enabled the development of mechanistically distinct models encompassing monogenic diabetes, T1DM, T2DM, and GDM, thereby establishing a multifunctional experimental paradigm for comparative pathophysiological investigations (Figure 1B).

T1DM

Characteristics: T1DM is a chronic autoimmune disorder characterized by immune-mediated destruction of pancreatic β-cells, leading to absolute insulin deficiency. With a global prevalence affecting millions of individuals, this condition represents a significant public health challenge[20]. The etiology involves complex interactions between genetic predisposition and environmental factors, though the precise mechanisms underlying disease initiation remain incompletely understood[21]. Current clinical management mandates lifelong exogenous insulin replacement therapy, while patients remain vulnerable to both acute metabolic derangements and long-term microvascular/macrovascular complications. Substantial progress has been made in T1DM research through translational studies employing rodent models and human longitudinal cohorts. Recent advances have elucidated critical immunoregulatory mechanisms, particularly the dynamic equilibrium between autoreactive effector T cells and immunosuppressive regulatory T cells (Tregs). This immunological balance not only influences disease susceptibility but also modulates the tempo of β-cell destruction and clinical progression[22]. Contemporary therapeutic strategies emphasize precision glycemic control through advanced insulin delivery systems and continuous glucose monitoring technologies, aiming to mitigate complication risks while improving quality of life.

Model construction methods: Three principal approaches dominate zebrafish T1DM modeling: Surgical pancreatectomy, transgenic induction, and pharmacological β-cell modulation. Selection criteria encompass experimental objectives, resource availability, and phenotypic specificity requirements.

Surgical resection: This pioneering technique utilizes transgenic zebrafish with GFP-labeled islets (Figure 1C), enabling real-time pancreatic visualization under fluorescence microscopy for partial resection[23-26]. However, technical constraints arise from the pancreas' miniature dimensions (typically < 50 μm) and intricate anatomical adjacency to hepatic/intestinal structures. Consequently, incomplete β-cell ablation and variable postoperative regeneration limit its utility in sustained hyperglycemia modeling.

Transgenic induction: The Tg (ins:Nfsb-mCherry) reporter line exemplifies genetic ablation models. This system co-expresses nitroreductase (NTR) and mCherry fluorescence under insulin promoter control[27]. Upon metronidazole (10 mM) exposure, NTR-mediated prodrug conversion generates cytotoxic derivatives, inducing β-cell apoptosis via DNA interstrand crosslinking. Successful ablation is confirmed by quantitative mCherry signal attenuation in pancreatic regions.

β-cell proliferation induction: (1) Drug modification: Chemical agents including STZ (350 mg/kg), difluoromethylornithine (DFMO, 5 mM), and imatinib (10 μM) demonstrate dose-dependent β-cell destruction in zebrafish[28,29] (Figure 2). Standardized protocols maintain animals at 28.5 °C with 14:10 light cycles, employing double-transgenic [Tg (elastase:GFP/ins:Nfsb)] models for simultaneous exocrine/endocrine tracking. Post-ablation assessment combines morphometric analysis (pancreatic length) with β-cell quantification via fluorescent insulin reporters; and (2) Small molecules: High-throughput screening in T1DM models identified (E)-3-[3-phenylbenzo(c)isoxazol-5-yl] acrylic acid (PIAA) as a potent β-cell proliferator (Figure 3)[30]. PIAA activates the cAMP/PKA-mTOR axis, specifically enhancing β-cell replication (3.8-fold vs controls). Complementary approaches exploit pancreatic plasticity: Α-to-β transdifferentiation via Arx suppression (artemisinin 50 μM induces GABA receptor-mediated nuclear export of Arx); and 12-LOX inhibition (baicalein 20 μM blocks arachidonic acid signaling, reducing islet macrophage infiltration by 67%)[31,32].

Figure 2 Drug modification approaches for inducing β-cell regeneration. DMFO: Difluoromethylornithine; STZ: Streptozotocin.

Figure 3 Small molecules promoting β-cell regeneration. Arx: Aristaless related homeobox; 12-LOX: 12-lipoxygenase.

Model applications: Zebrafish models of T1DM serve as versatile platforms for investigating diabetes pathophysiology and its systemic complications, with demonstrated utility in neuroendocrine regulation, reproductive dysfunction, and metabolic memory studies.

Mechanistic exploration: Through transcriptomic profiling of gonadal tissues, these models reveal impaired spermatogenesis correlating with downregulation of insulin receptor substrate 2 (IRS2, 0.45-fold) and glucose transporter GLUT4 (0.32-fold), establishing molecular links between chronic hyperglycemia and male infertility[20].

Metabolic memory paradigm: The models uniquely capture persistent diabetic complications despite glycemic normalization - a phenomenon mediated through epigenetic reprogramming. Key molecular drivers include: (1) Histone modifications: Sustained H3K4me1 deposition at pro-inflammatory loci (e.g., NF-κB promoter); (2) Non-coding RNA dynamics: Upregulation of miR-192-5p (4.7-fold) targeting sirtuin-1 in pancreatic remnants; and (3) Cellular plasticity: Compromised β-cell regeneration capacity (< 15% recovery vs wild-type) through DNA methyltransferase 3a hyperactivation[33].

Therapeutic development: Pharmacological interventions targeting these mechanisms demonstrate disease-modifying potential: (1) HDAC inhibitors (trichostatin A 100 nM) restore histone acetylation levels (H3K9ac + 82%); (2) Antagomir-192-5p treatment rescues β-cell mass (38% increase) via sirtuin-1 reactivation; and (3) 5-azacytidine (10 μM) reverses DNA methylation at Pdx1 enhancer regions, enhancing neogenesis[34].

Significance: Zebrafish has emerged as a pivotal model organism for T1DM research through diverse experimental approaches including pharmacological induction, genetic modification, and targeted gene ablation. These established models provide robust platforms for systematically evaluating therapeutic candidates capable of stimulating β-cell regeneration or promoting pancreatic tissue repair, thereby advancing fundamental discoveries in diabetes pathophysiology. Notably, recent breakthroughs utilizing zebrafish models have revealed gephyrin as a promising therapeutic target that facilitates the transdifferentiation of pancreatic α-cells into functional β-like cells - a groundbreaking discovery in cellular reprogramming for diabetes treatment. Furthermore, the unique advantages of zebrafish in high-throughput screening have enabled comprehensive pharmacological evaluations, as exemplified by the successful identification of anti-diabetic properties in Pseudomonas malayi aqueous extracts through integrated analyses combining blood glucose monitoring with liquid chromatography-mass spectrometry metabolic fingerprinting[31,35]. These methodological innovations underscore the zebrafish model's dual capability in both mechanistic exploration of β-cell regeneration pathways and rapid screening of potential therapeutic compounds.

T2DM

Characteristics: T2DM is defined by a dual pathophysiology comprising impaired insulin secretion from pancreatic β-cells and diminished insulin sensitivity in peripheral target tissues, including skeletal muscle, hepatic tissue, and adipose depots[36]. Epidemiologically, the global prevalence of diabetes among adults has surged dramatically, escalating from 4.7% in 1980 to 8.5% by 2014, with T2DM accounting for 90%-95% of all diagnosed adult diabetes cases[37]. The pathogenesis of T2DM manifests as a complex interplay between polygenic susceptibility and environmental determinants. Large-scale genomic investigations, particularly genome-wide association studies (GWAS), have mapped over 150 susceptibility loci associated with T2DM risk, highlighting the polygenic architecture of this metabolic disorder. Concurrently, emerging evidence implicates epigenetic dysregulation - such as aberrant DNA methylation patterns and histone modifications - as critical mediators of gene-environment interactions in T2DM progression. At the functional level, T2DM arises from either insufficient insulin production or systemic insulin resistance, culminating in chronic hyperglycemia. This persistent elevation of blood glucose drives progressive pathological alterations across multiple organ systems, including macrovascular complications (e.g., atherosclerosis), microvascular damage (e.g., retinopathy and nephropathy), and neuropathic degeneration. Left unmanaged, these complications pose severe risks to cardiovascular integrity, renal function, and neurological health, underscoring the systemic nature of T2DM-related morbidity[38].

Model construction methods: Current therapeutic interventions for T2DM focus on dietary regulation, physical activity enhancement, and pharmacological agents targeting glycemic control and lipid metabolism. Mirroring these clinical approaches, four distinct methodologies have been developed for constructing zebrafish T2DM models: Glucose solution immersion method[39], high-fat food induction method[36], CRISPR/Cas9 gene knockout method[40], and targeted genetic ablation[41] (Figure 4).

Figure 4 Construction methods for zebrafish models of type 2 diabetes mellitus models. T2DM: Type 2 diabetes mellitus.

(1) Glucose solution immersion: This metabolic stress model induces T2DM phenotypes through chronic hyperglycemic exposure. Adult zebrafish, with baseline fasting blood glucose levels of 74 ± 8.5 mg/dL (hyperglycemia threshold: > 200 mg/dL), were subjected to alternating 2% and 0% glucose solutions over 28 days. By day 14, this regimen elicited sustained hyperglycemia (> 200 mg/dL), insulin receptor mRNA downregulation in muscle tissue, and diminished responsiveness to exogenous insulin. Partial glucose normalization following glimepiride and metformin treatment validated the model's pharmacological relevance, demonstrating its utility for anti-diabetic drug screening[39].

(2) High-fat diet: Nutrient overload via high-fat feeding recapitulates human metabolic syndrome progression in zebrafish. Prolonged administration induces obesity (elevated body mass index), hepatic steatosis (visceral lipid accumulation), and systemic insulin resistance, evidenced by elevated fasting glucose and tissue-specific insulin mRNA upregulation[42]. Notably, zebrafish larvae exhibit lipid dysregulation detectable at 5 days post-fertilization (dpf) through circulating triacylglycerol quantification and lipid droplet visualization, possessing a significant temporal advantage over rodent models restricted to post-weaning interventions (≥ 3 weeks)[43-45]. This approach combines physiological relevance with experimental tractability for large-scale metabolic studies.

(3) CRISPR/Cas9 technology: Precision genome engineering enables targeted modeling of T2DM-associated genetic variants. A standardized protocol involves microinjecting 1 nL of CRISPR cocktail (1000 ng Cas9 mRNA, 200 ng Mitfa sgRNA, 500 ng donor DNA, and 1% phenol red in 5 μL total volume) into single-cell embryos, followed by incubation in E3 medium at 28 °C. Successful integration of Cas9 expression cassettes permits systematic investigation of gene-specific contributions to β-cell dysfunction and insulin signaling pathways[40,46].

(4) Inducible genetic ablation: This optogenetic-chemogenetic hybrid system enables spatiotemporal control of β-cell dynamics. Transgenic zebrafish expressing Nfsb-mCherry fusion proteins in β-cells permit nitroreductase-mediated ablation upon metronidazole exposure. Pharmacological withdrawal triggers β-cell regeneration within 36 hours, with lineage tracing (ptf1a reporter) and proliferation analysis (BrdU incorporation) confirming de novo β-cell formation independent of exocrine pancreatic pathways[27].

Model applications: Zebrafish has emerged as a premier in vivo platform for investigating metabolic pathophysiology, owing to its evolutionarily conserved mechanisms governing lipid metabolism, adipocyte biology, pancreatic organogenesis, and glucose regulatory systems. These cross-species parallels extend to T2DM modeling, where zebrafish demonstrate remarkable translational value in: (1) Deciphering conserved molecular pathways underlying insulin resistance and β-cell dysfunction; (2) High-throughput screening of therapeutic compounds targeting obesity-related metabolic syndrome; and (3) Functional validation of human GWAS-identified risk alleles through CRISPR-based genetic manipulation. Particularly noteworthy is their capacity to model diet-induced metabolic perturbations - from hepatic steatosis to systemic insulin resistance - within compressed developmental timelines, enabling rapid identification of druggable targets for obesity management and diabetes prevention[47,48].

Significance: T2DM, a complex metabolic syndrome manifesting insulin resistance and β-cell dysfunction[49,50], has been mechanistically investigated through zebrafish models demonstrating three key advances: (1) Elucidation of metformin-silybin combinatorial therapy mechanisms via tumor necrosis factor-α/interleukin-6 inflammatory axis modulation[51]; (2) Establishment of insulin resistance-neurodegeneration causality through real-time imaging of blood-brain barrier permeability alterations[52]; and (3) Functional genomics validation of therapeutic targets including Pdx1-mediated pancreatic regeneration and Kcnj11-encoded Kir6.2 potassium channel regulation using CRISPR/Cas9 ablation models[53]. These breakthroughs underscore zebrafish models' capacity for simultaneous interrogation of metabolic dysregulation and neural pathophysiology, accelerating antidiabetic drug discovery pipelines.

Monogenic diabetes

Characteristics: Monogenic diabetes, a genetically heterogeneous disorder, encompasses distinct clinical entities including neonatal diabetes mellitus (NDM), MODY, and syndromic diabetes variants. Among these, MODY predominates, representing 0.5%-5% of non-autoimmune diabetes cases[54]. Current classification recognizes 14 molecular subtypes (MODY1-MODY14) based on causative gene mutations (Table 1)[55], with MODY1, MODY2, MODY3, and MODY5 constituting > 85% of diagnoses. Rarer subtypes (MODY4, MODY6-MODY14) demonstrate unique genotype-phenotype correlations, often involving transcription factors critical to pancreatic development (e.g., mutation of pancreatic and duodenal homeobox, and neurogenic differentiation factor 1)[56].

Table 1 Characteristics of maturity-onset diabetes of the young subtypes 1 to 14.

Type of MODY	Mutant gene	Clinical characteristics	Therapeutic method	Ref.
MODY 1	HNF4α mutations	Macrosomia; transient neonatal hypoglycemia; insulin secretory defect in adolescence or early adulthood	Low dose sulfonylurea; diet; insulin	[81-83]
MODY2	GCK gene mutation	Small rise in 2-hour PG level on OGTT; higher glucose threshold; nonprogressive elevated fasting blood glucose	Diet and exercise modification (usually no treatment is needed)	[82,84]
MODY3	HNF1A mutation	Lowered renal threshold for glucosuria; insulin secretory defect in adolescence or early adulthood; rise in 2-hour PG level on OGTT	Sulfonylureas	[83,85,86]
MODY4	Mutation of PDX1	Postpuberty; mild form of diabetes	Insulin; OHAs	[87]
MODY5	HNF1B mutation	Renal disease; genitourinary abnormalities; atrophy of the pancreas; hyperuricemia; gout	Early insulin therapy; OHAs (sulfonylurea or repaglinide)	[13]
MODY6	NEUROD1 mutation	Variable age of onset; different degrees of hyperglycemia	Insulin; OHAs	[88]
MODY7	Mutation of KLF11 gene	Variable age of onset; decreased sensitivity to insulin; mild hyperglycemia	Insulin	[89]
MODY8	Mutation of CEL gene	> 25 years; Impaired pancreatic function	Insulin; OHAs	[90]
MODY9	PAX4 gene mutation	Post-puberty; progressive hyperglycemia; occurrences of ketoacidosis	Diet; insulin; OHAs	[91]
MODY10	INS gene mutation	> 10 years; hyperglycemia; diabetes	Insulin; diet	[92]
MODY11	Mutation of BLK gene	Variable age of onset; hyperglycemia; diabetes	Diet; insulin; OHAs	[93]
MODY12	Mutation of ABCC8 gene	Variable age of onset: Diabetes; rarely developmental delay	Sulfonylureas	[94]
MODY13	Gene mutation in member 11 of the inward rectifying potassium ion channel subfamily J (KCNJ11)	After second decade of life; diabetes; possible developmental delay and seizures	Low dose sulfonylures	[95-97]
MODY14	Heterozygous mutation in the APPL1 gene on chromosome 3p14	10-50 years; hyperglycemia; diabetes	Diet; insulin; OHAs	[98]

PG: Plasma glucose; OGTT: Oral glucose tolerance test; OHAs: Oral hypoglycemic agents; MODY: Maturity-onset diabetes of the young; GCK: Glucokinase; HNF1A: Hepatocyte nuclear factor 1A; PDX1: Mutation of pancreatic and duodenal homeobox; NEUROD1: Neurogenic differentiation factor 1; KLF11: Krüppel like factor 11; CEL: Carboxyl ester lipase; PAX4: Paired box gene 4; BLK: B lymphocyte kinase; ABCC8: ATP-binding cassette subfamily C member 8.

Mutations in the glucokinase (GCK) gene (MODY2) and hepatocyte nuclear factor 1A/4A (HNF1A/4A) genes (MODY3 and MODY1, respectively) are the most common causes of MODY. GCK mutations typically cause fasting hyperglycemia, which is mild and usually did not require special treatment. In contrast, HNF1A and HNF4A mutations result in β-cell dysfunction, and persistent hyperglycemia could lead to microvascular complications, while sulfonylurea drugs are effective for managing these conditions, though some patients might eventually require insulin treatment. Mutations in HNF1B (MODY5) were associated with a broader spectrum of abnormalities, including pancreatic hypoplasia, renal dysfunction, reproductive tract anomalies, and liver dysfunction.

Model construction methods: Zebrafish models of monogenic diabetes are broadly categorized into two pathophysiological paradigms: NDM models and MODY models, each employing distinct genetic engineering strategies to mirror human disease mechanisms. NDM pathogenesis is replicated through targeted CRISPR/Cas9 editing, exemplified by introducing the dominant-negative ATP-binding cassette subfamily C member 8 (SUR1) truncating mutation (p.Lys499Ter) into zebrafish homologs[57]. Complementary strategies involve translational cloning of human transcription factors (notably, HNF3B (FOXA2) isolated from hepatic cDNA libraries), followed by systematic mutagenesis screening to identify pathogenic variants affecting pancreatic β-cell development[58]. Additionally, CRISPR/Cas9 had been utilized to construct a MODY3 model to study the functional loss of the HNF1a-Q125ter variant[59]. Furthermore, Hnf1ba hypomorphic mutants exhibit conserved multiorgan pathology including pancreatic hypoplasia (75% β-cell mass reduction) and renal dysplasia, while maintaining intact endodermal patterning, precisely mirroring human HNF1B syndrome phenotypes[20].

Model applications: Zebrafish models have become pivotal in monogenic diabetes research, primarily serving to investigate unidentified gene mutations or deletions. These models have proven invaluable for conducting familial DNA analyses, detecting pregnancy-associated MODY, and developing personalized care strategies. For instance, they enable clinicians to formulate tailored nursing protocols for MODY patients during pregnancy and optimize neonatal care plans for postpartum management, ensuring that both maternal and infant health outcomes are systematically addressed[60,61].

Significance: Although some forms of MODY, such as GCK-MODY, do not require treatment, others like HNF1A-MODY show significant responsiveness to oral medications. The risks of microvascular and macrovascular complications vary substantially across MODY subtypes. However, research indicates that 50%-90% of MODY cases are misdiagnosed as T1DM or T2DM. While clinical features often suggest MODY, no unified diagnostic standard exists. Thus, these models will enhance our understanding of monogenic diabetes[62]. Although monogenic diabetes is frequently misclassified as T1DM or T2DM, targeted research can address diagnostic gaps and ensure accurate patient management.

GDM

Characteristics: Diabetes in pregnancy is typically classified into two distinct categories: Pre-existing diabetes diagnosed prior to gestation and GDM first identified during pregnancy[60]. European epidemiological studies have reported prevalence rates ranging from 0.15% to 4% for GDM, compared to 0.2%-0.4% for pre-pregnancy diabetes mellitus[63]. The diagnostic criteria for GDM generally involve the detection of glucose intolerance characterized by maternal glucose metabolism abnormalities that are initially recognized during the gestational period. This review primarily focuses on pregnancy-associated diabetes, specifically addressing cases where glucose metabolism disorders are first diagnosed during pregnancy.

Model construction methods: A zebrafish model of GDM was established through dietary induction, wherein gravid females were administered a high-glucose dietary regimen throughout the gestational phase. This intervention successfully replicated key pathological features of GDM, manifested as sustained maternal hyperglycemia exceeding physiological thresholds, accompanied by observable maternal-fetal complications[64]. Experimental protocols involved controlled exposure of wild-type embryos to 4% and 5% D-glucose solutions, administered from 3 hpf through 5 dpf. Comparative analysis revealed significant developmental toxicity: Control groups exhibited 4% mortality, whereas glucose-exposed groups demonstrated a dose-dependent mortality pattern with 28.4% (4% D-glucose) and 37.5% (5% D-glucose) lethality rates[64].

Model applications: Zebrafish GDM models have served as valuable experimental platforms for investigating the teratogenic effects and pathophysiological sequelae of maternal hyperglycemia. These models have been successfully employed to: (1) Characterize aberrant retinal vascular patterning during embryogenesis, yielding mechanistic insights into hyperglycemia-induced ocular developmental anomalies[61]; and (2) Delineate the regulatory role of pancreatic β-cell mass in cardiovascular development and neural crest cell migration[65,66]. Notably, the translucent zebrafish embryos enable real-time observation of neuro-cardiac axis formation under hyperglycemic conditions, providing unprecedented spatial-temporal resolution for studying developmental crosstalk between organ systems.

Significance: Zebrafish GDM models have emerged as powerful experimental systems for elucidating transgenerational metabolic programming effects induced by maternal hyperglycemia. These models have been successfully utilized to: (1) Decode hyperglycemia-mediated disruptions in embryonic retinal angiogenesis through longitudinal analysis of vascular branching patterns[64], establishing an evaluative framework for predicting developmental trajectories in offspring of diabetic pregnancies; and (2) Map the molecular cascades linking maternal glucose dysregulation to fetal epigenetic modifications. Particularly valuable is the models' capacity for real-time monitoring of maternal-fetal metabolic crosstalk via advanced live-imaging techniques. Through systematic interrogation of these pathophysiological interactions, researchers are pioneering novel biomarkers for early complication detection and developing targeted therapeutic strategies to optimize perinatal outcomes.

DIABETIC COMPLICATIONS

Zebrafish models for T1DM complications

T1DM remains a significant public health concern, affecting over 160000 Americans annually and manifesting as chronic hyperglycemia with microvascular complications including retinopathy, neuropathy, and nephropathy. Zebrafish models of T1DM have yielded critical insights into its systemic complications through three principal research avenues: (1) Reproductive pathophysiology studies demonstrating impaired spermatogenesis characterized by diminished sperm motility and compromised DNA integrity[20]; (2) Immunometabolic investigations using pancreatitis models that elucidated macrophage-mediated β-cell degradation mechanisms, where β-cell loss was identified as a critical driver of metabolic decompensation[32]; and (3) Ocular pathology models employing cyclical hyperglycemic induction protocols (alternating glucose/water immersion) that revealed intraretinal stratification anomalies, particularly altered spatial relationships between inner plexiform and nuclear layers[67]. These findings collectively validate zebrafish models of T1DM as versatile platforms for investigating the multidimensional pathophysiology spanning reproductive, immunological, and neurosensory systems.

Zebrafish models for T2DM complications

T2DM, characterized by insulin deficiency stemming from pancreatic β-cell dysfunction[68], manifests through microvascular and macrovascular complications, with diabetic nephropathy representing one of the most devastating microvascular complications that frequently progresses to end-stage renal disease[19]. Zebrafish models have been used to elucidate key pathological mechanisms through three principal approaches: (1) Genetic manipulation studies targeting Pdx1 (a master regulator of pancreatic development), where morpholino-mediated knockdown in embryos recapitulated critical diabetic phenotypes including sustained hyperglycemia, anterior glomerular hypertrophy, renal filtration barrier defects, and podocyte developmental abnormalities[69]; (2) Metabolic pathway analyses identifying mitochondrial thiosulfate sulfurtransferase (TST) as a novel biomarker for obesity-associated T2DM, where pharmacological activation of the TST axis in zebrafish forekidney models demonstrated therapeutic potential for hyperglycemia mitigation and organ protection[70]; and (3) Systemic complication modeling revealing chronic hyperglycemia-induced bone matrix disorganization with concomitant osteocyte dysfunction and metabolic impairment, alongside retinal vascular anomalies mirroring human diabetic retinopathy patterns[42]. These findings collectively position zebrafish T2DM models as indispensable tools for deciphering the multiorgan pathophysiology spanning renal failure progression, skeletal deterioration, and metabolic syndrome components.

Zebrafish models for monogenic diabetes complications

Monogenic diabetes manifests through two principal subtypes: MODY and NDM. Mechanistic studies have identified three key genetic determinants: (1) Pathogenic PCBD1 variants causing concomitant hypomagnesemia and MODY5-like diabetes phenotypes through impaired magnesium homeostasis[71]; (2) vHnf1 homeobox gene mutations associated with MODY5 progression and familial glomerulocystic kidney disease (GCKD) via disrupted renal tubular morphogenesis[72]; and (3) HNF1A heterozygous mutations driving MODY3 pathogenesis characterized by altered renal glucose reabsorption thresholds and enhanced susceptibility to diabetic nephropathy[59]. Crucially, zebrafish models of monogenic diabetes have proven instrumental in elucidating the molecular pathogenesis of these pleiotropic complications - particularly magnesium dysregulation, cystic kidney degeneration, and glucose transporter dysfunction - through real-time visualization of developmental metabolic cascades.

Zebrafish models for GDM complications

GDM confers significant maternal risks including preeclampsia, preterm labor, polyhydramnios, elevated cesarean delivery rates, and heightened infection susceptibility. Rigorous metabolic regulation through glycemic optimization remains critical for mitigating these adverse perinatal outcomes[73]. Women with pre-gestational diabetes mellitus face compounded risks of hypoglycemic episodes, progressive retinopathy, diabetic nephropathy exacerbation, and diabetic ketoacidosis, collectively contributing to elevated perinatal morbidity and mortality indices[74]. The zebrafish visual system, sharing evolutionary conservation with mammalian ocular architecture, has emerged as a premier model for investigating diabetic retinopathy pathogenesis. An acute hyperglycemia induction paradigm has been established in laboratory, demonstrating remarkable translational research value for: (1) High-throughput screening of novel retinoprotective agents[75]; and (2) Mechanistic analysis of glucose-mediated developmental ocular defects. Experimental data revealed that sustained maternal hyperglycemia induces embryonic retinal pathology characterized by disruption of retinal laminar architecture (particularly reduced inner nuclear layer thickness) and cellular depletion phenomena affecting both Müller glia populations and retinal ganglion cells[64]. These pathophysiological recapitulations validate zebrafish models of GDM as biologically relevant systems for decoding hyperglycemia-induced retinal degeneration mechanisms and developing targeted therapeutic interventions.

SIGNIFICANCE OF ZEBRAFISH MODELS

Zebrafish models have emerged as indispensable experimental systems for elucidating pathogenic mechanisms across diabetes subtypes, including T1DM, T2DM, GDM, and monogenic forms. Compared with conventional mammalian models (pig, monkey, mouse, and dog), zebrafish offer unparalleled advantages for diabetes research, as comprehensively analyzed in Table 2.

Table 2 Applications of zebrafish model in diabetes.

Types of diabetes	Advantages	Disadvantages	Ref.
T1DM	High degree of similarity to the human pancreas and most genes have been deciphered by the human body; it compensates for the shortcomings of rodents, such as high reproductive rate	Zebrafish are variable temperature animals and it is difficult to measure their insulin content	[77,79,99]
T2DM
Monogenic diabetes		There have been few studies on monogenic diabetes based on the zebrafish model
GDM	For assessing the health status of infants born to mothers with a history of diabetes or gestational diabetes	There have been few studies on pregnancy diabetes with zebrafish as the model	[76]

T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; GDM: Gestational diabetes mellitus.

Genetic tractability and high-throughput screening

The zebrafish model system possesses a streamlined genomic architecture exhibiting exceptional tractability for precision genome-editing technologies including CRISPR/Cas9, TALENs, and morpholino-based gene modulation. This genetic versatility enables accelerated development of pathophysiologically relevant models replicating human diabetes subtypes, particularly monogenic diabetes and GDM - conditions notoriously challenging to recapitulate in conventional mammalian systems. The optical transparency of zebrafish embryos facilitates systematic high-content screening of pharmacological candidates, allowing simultaneous interrogation of metabolic pathways, therapeutic targets, and developmental toxicities through real-time confocal imaging approaches. These combined advantages significantly expedite mechanistic deciphering of diabetic pathogenesis and validation of novel treatment modalities.

Translational relevance

Zebrafish, despite their diminutive scale and accelerated ontogeny, demonstrate profound evolutionary conservation of core physiological and metabolic pathways with humans, establishing them as premier preclinical discovery platforms for diabetes investigation. This conserved pathophysiology enables systematic interrogation of: (1) Pancreatic β-cell dynamics in diabetes pathogenesis; (2) Metabolic memory effects during disease progression; and (3) Multiorgan complication cascades spanning microvascular pathologies (retinopathy and nephropathy) to macrovascular complications involving neurovascular coupling impairments. Particularly valuable is their capacity for simultaneous analysis of hepatic gluconeogenesis regulation and endothelial dysfunction mechanisms - key processes underlying diabetic complications - through real-time metabolic imaging modalities.

Multiscale and multimodal imaging capabilities

The embryonic translucency of zebrafish enables revolutionary in vivo imaging capabilities through multiphoton microscopy and light-sheet imaging modalities. This optical accessibility permits non-invasive visualization of: (1) Pancreatic islet neogenesis with single-β-cell resolution; (2) β-cell Ca²⁺flux patterns during glucose-stimulated insulin secretion; and (3) Microvascular permeability alterations in diabetic complications. When combined with fluorescent reporter lines, longitudinal intravital microscopy achieves single-cell resolution tracking of pathophysiological cascades - from mitochondrial dynamics in β-cells to leukocyte-endothelial interactions during vascular inflammation. These integrated approaches provide unparalleled spatiotemporal resolution for mapping therapeutic response kinetics and cellular recovery mechanisms.

Drug discovery and therapeutic screening

Zebrafish provide a high-throughput pharmacological discovery platform with exceptional cost-efficacy in therapeutic development. Their accelerated reproductive cycles and minimal husbandry requirements facilitate systematic assessment of compound efficacy, metabolic regulation, and toxicity profiles in diabetic models. This vertebrate system enables multiparametric screening to prioritize lead compounds for subsequent preclinical validation and translational clinical applications.

Unique insights and expansion of disease modeling

Zebrafish research reveals non-canonical disease pathways through multidimensional analysis of developmental programming, environmental determinants, and gene-environment-epigenome crosstalk. While robust models exist for T1DM/T2DM, systematic model development for monogenic diabetes and GDM with precision phenotyping remains imperative[76]. With 87% genomic conservation to humans and versatile reverse genetic approaches[20,77], this model system uniquely deciphers diabetes subtype heterogeneity through functional genomics pipelines, enabling target discovery and precision diabetology frameworks.

CHALLENGES AND FUTURE PERSPECTIVE FOR ZEBRAFISH MODELS

Challenges

Despite their translational utility in diabetes research, zebrafish models present distinct methodological constraints. The aquatic drug exposure route, while simplifying therapeutic delivery, introduces pharmacokinetic variability due to cutaneous absorption dynamics, complicating precise toxicokinetic modeling[78]. Strict hydrochemical parameter maintenance (pH 7.0-8.0; conductivity 500-1500 μS/cm) and thermal regulation (28.5 ± 0.5 °C) are critical, as environmental perturbations alter metabolic homeostasis and confound insulin quantification[79]. Pharmacokinetic analyses by Berghmans et al[80] demonstrated 40%-65% variability in embryonic uptake efficiency for nine xenobiotics, underscoring standardization imperatives. While bridging translational gaps, interspecies physiological divergences persist, particularly in replicating human-specific pathomechanisms like neuroendocrine regulation discrepancies. Current monogenic diabetes and GDM models face additional limitations from incomplete molecular characterization and partial phenotypic recapitulation (Table 3).

Table 3 Characteristics of animal models for diabetes except for zebrafish.

Animal model	Experimental cycles	Costs	Stability	Human similarity	Ref.
Pig	Long experimental cycles	Compared with other animals, increased experimental costs; larger breeding space required	High for experimental research	About 98%	[100]
Monkey	Long experimental cycles	Compared with other animals, increased experimental costs	Spontaneously develop diabetes, which interferes with experimental results	About 90%	[101]
Mouse	Short experimental cycles	Compared with other animals, lower experimental costs	Instability as there are innate immune differences between the two, adding uncertainty and risk to experimental research	About 95%	[9,102]
Dog	Long experimental cycles	High experimental costs	As they share the same environment as humans, they are naturally exposed to many risk factors	About 95%	[103]

Future perspectives

To overcome zebrafish model limitations, methodological innovations show significant promise: (1) Implementing intracardiac or intraperitoneal microinjection protocols circumvents aquatic drug delivery variability, enabling precise pharmacokinetic profiling while reducing water quality-mediated mortality; (2) Nutritional induction protocols using high-fat/high-glucose diets better recapitulate metabolic syndrome pathophysiology; and (3) Emerging precision diabetology frameworks leverage CRISPR-based somatic cell editing for monogenic diabetes modeling, particularly for HNF1A/HNF4A variants. Future advancements should integrate super-resolution live imaging with chemical genetics pipelines, enabling simultaneous visualization of β-cell regeneration dynamics and whole-organism metabolic mapping. These multimodal approaches will accelerate therapeutic target validation across diabetes subtypes through combinatorial screening of small molecule libraries and genetic modifiers.

CONCLUSION

Preclinical model selection constitutes the pivotal translational gateway between bench research and clinical implementation. Optimal model selection requires strategic alignment with study objectives, where validated preclinical models deliver indispensable pharmacokinetic/pharmacodynamic evidence while maintaining cost-efficacy. However, physiological interspecies disparities persist across all model systems. Zebrafish models, despite inherent limitations, provide unique diabetes research advantages: (1) Unparalleled β-cell regenerative capacity enabling mechanistic dissection of pancreatic islet regeneration pathways; (2) 87% conserved coding genome facilitating precision gene editing for personalized medicine prototypes; and (3) Optical transparency permitting real-time metabolic imaging. Current research disproportionately focuses on T2DM (76% of studies) versus monogenic diabetes (12%), reflecting diagnostic challenges in rare genetic disorders. While whole exome sequencing has identified > 40 monogenic diabetes genes, mechanistic understanding of complications remains fragmented. Zebrafish models address this gap through combinatorial approaches: CRISPR/Cas9-mediated knock-in of human mutations coupled with metabolic phenotyping. Compared to rodent models, zebrafish offer three key translational advantages: Accelerated generation of genetic variants (4-6 weeks vs 6-12 months), whole-organism drug response profiling, and conserved complication pathways. Beyond diabetes, this model shows expanding utility in cardiovascular pathophysiology (76% conserved cardiac genes), osteometabolic regulation, and neurodevelopmental disorders, with emerging applications in cancer metabolism and microbiome-host interactions. Future directions should prioritize establishing monogenic diabetes complication registries integrated with zebrafish phenomics databases to accelerate therapeutic discovery.

ACKNOWLEDGEMENTS

We express our gratitude to all colleagues in the nursing teaching and research department for their dedication to this study.