Establishment and Evaluation Methods of Diabetic Animal Models

Diabetic animal modeling should establish the corresponding pathological state according to the research objective. Type 1 diabetes models focus on simulating pancreatic β-cell destruction and insulin deficiency, while type 2 diabetes models focus on simulating obesity, insulin resistance, and compensatory β-cell failure. Model evaluation should not rely only on a single increase in blood glucose; glucose tolerance, insulin sensitivity, islet function, and target organ injury should also be assessed to determine whether the model is stable.

Keywords: diabetic animal modeling; STZ-induced model; alloxan model; high-fat diet model; HFD+STZ model; NOD mice; db/db mice; ZDF rats

1 Design Principles Before Diabetic Animal Modeling

1.1 Clarifying the Model Type

Diabetic animal models should first be selected according to the research objective, rather than directly applying a fixed model. If the study focuses on autoimmune type 1 diabetes, spontaneous models such as NOD mice, BB rats, or LEW-IDDM rats should be selected. If the study focuses on rapid hyperglycemia and β-cell injury, STZ- or alloxan-induced models can be selected. If the study focuses on obesity, insulin resistance, and type 2 diabetes progression, high-fat diet, HFD+STZ, db/db mice, ob/ob mice, ZDF rats, or KK-Ay mice can be selected.

1.2 Clarifying Modeling Endpoints

The endpoint of diabetic modeling should not be defined only as “increased blood glucose.” Type 1 diabetes models should focus on decreased insulin, reduced β-cell area, insulitis, or β-cell necrosis. Type 2 diabetes models should focus on weight gain, impaired glucose tolerance, insulin resistance, hyperinsulinemia, or insufficient insulin secretion at later stages. If diabetic nephropathy, neuropathy, retinopathy, or cardiovascular complications are being studied, the modeling period should be extended, and pathological changes in target organs should be detected.

1.3 Standardizing Animals and Operating Conditions

Before modeling, animal strain, sex, age in weeks, body weight range, housing environment, feed batch, fasting duration, administration route, and blood sampling time should be standardized. STZ, alloxan, and high-fat diet models are highly sensitive to animal condition. Excessively young animals, large differences in body weight, inconsistent fasting duration, or unstable administration procedures can all cause fluctuations in model success rate and mortality.

Table 1 Selection of Diabetic Animal Modeling Routes

Research Objective	Recommended Model	Modeling Method	Core Evaluation Indicators
Rapid establishment of type 1 diabetes-like hyperglycemia	STZ model	Intraperitoneal or intravenous STZ administration to destroy β cells	Sustained hyperglycemia, decreased insulin, reduced β cells
β-cell oxidative injury model	Alloxan model	ALX administration to induce free radical-mediated β-cell injury	Hyperglycemia, islet injury, increased oxidative stress
Autoimmune type 1 diabetes	NOD mice, BB rats, LEW-IDDM rats	No exogenous induction; spontaneous onset monitored by age	Insulitis, hyperglycemia, β-cell destruction
β-cell endoplasmic reticulum stress	Akita mice	Diabetes develops spontaneously due to INS2 mutation	Hyperglycemia, low insulin, β-cell stress
Prediabetes/T2D early stage	HFD model	Long-term high-fat diet induces obesity and insulin resistance	Weight gain, abnormal OGTT, abnormal ITT
Stable type 2 diabetes	HFD+low-dose STZ model	High-fat diet induces insulin resistance, followed by low-dose STZ	Insulin resistance, sustained hyperglycemia, decreased β-cell function
Genetic obesity-related T2D	ob/ob, db/db, ZDF, KK-Ay	Select animals with corresponding genetic backgrounds and observe by age	Obesity, hyperglycemia, insulin resistance
Chronic diabetes and complications	OLETF, db/db, Akita, long-term STZ model	Maintain long-term hyperglycemia and observe target organ injury	Urinary albumin, renal pathology, nerve/retinal injury

2 Animal Modeling Methods for Type 1 Diabetes

2.1 STZ-Induced Diabetic Model

(1) Modeling principle

Streptozotocin (STZ) can enter pancreatic β cells through the GLUT2 transporter and induce DNA alkylation, PARP activation, NAD⁺ depletion, ATP reduction, and β-cell death. Because β-cell injury reduces insulin secretion, animals develop sustained hyperglycemia. Therefore, STZ is commonly used to establish type 1 diabetes-like hyperglycemic models and can also be combined with high-fat diet to establish composite type 2 diabetes models.

(2) Modeling procedure

STZ modeling usually includes animal acclimation, short-term fasting, fresh STZ preparation, intraperitoneal or intravenous administration, acute-phase observation, blood glucose screening, and repeated monitoring. Experimental animals are generally acclimated for about 1 week before modeling, and fasting before modeling should be standardized without water deprivation. STZ is commonly prepared immediately before use in 0.05 mol/L citrate buffer (pH 4.5), protected from light, kept at low temperature, and injected as soon as possible.

When establishing a type 1 diabetes-like model in rats, a commonly used single dose of STZ is 40–65 mg/kg, administered by intraperitoneal injection or tail vein injection. In mice, a commonly used single dose is 100–200 mg/kg, also administered by intraperitoneal or intravenous injection. Multiple low-dose protocols can reduce acute toxicity and, to some extent, simulate a slower β-cell injury process. The specific dose should be optimized through preliminary experiments according to animal strain, age, sex, body weight, and experimental objective.

(3) Model validation

Blood glucose is usually measured 3–7 days after STZ administration, followed by repeated testing to confirm whether hyperglycemia is stable. Common indicators of successful modeling include sustained elevation of random or fasting blood glucose, increased water intake, increased urine output, weight loss, decreased insulin levels, and reduced pancreatic β-cell area. A single increase in blood glucose should not be directly judged as successful modeling; blood glucose should be monitored at least 2–3 consecutive times and interpreted together with insulin levels and islet pathology.

(4) Applicable scenarios

A single high-dose STZ model is suitable for rapidly establishing insulin-deficient hyperglycemia and can be used for preliminary screening of hypoglycemic drugs, evaluation of islet-protective agents, oxidative stress intervention, and diabetic complication studies. Multiple low-dose STZ can be used for relatively mild β-cell injury models. HFD+low-dose STZ is more suitable for establishing type 2 diabetes-like models.

(5) Precautions

STZ has poor stability, and prolonged standing after preparation can reduce its induction effect. STZ may also cause liver and kidney toxicity, especially at high doses or when animals are in poor condition. After modeling, animals may develop dehydration, weight loss, and short-term mortality risk. Humane endpoints should be set, and monitoring should comply with animal ethics requirements. In drug intervention experiments, hyperglycemic injury, direct STZ toxicity, and drug-protective effects should be distinguished.

2.2 Alloxan-Induced Diabetic Model

(1) Modeling principle

Alloxan (ALX) can be rapidly taken up by pancreatic β cells and undergo intracellular redox cycling, generating superoxide, hydrogen peroxide, and hydroxyl radicals, which cause β-cell DNA breakage and cell death. Because β cells have relatively weak antioxidant defenses and are sensitive to ALX-induced free radical injury, ALX can be used to establish diabetic models characterized by β-cell oxidative injury.

(2) Modeling procedure

ALX is generally prepared immediately before use as a 1%–3% solution and should be administered as soon as possible after preparation. Before modeling, animals usually require standardized fasting without water deprivation to reduce baseline blood glucose differences. After administration, animal condition should be closely observed, and blood glucose should be measured within several days. Because ALX can induce acute β-cell injury, some animals may show marked early blood glucose fluctuations, requiring strengthened monitoring.

Common intraperitoneal doses in rats are 150–200 mg/kg, and common intravenous doses are 40–60 mg/kg. In mice, the commonly used intraperitoneal dose is about 200 mg/kg, and the commonly used intravenous dose is about 85–100 mg/kg. Different animal strains vary greatly in ALX sensitivity, so the dose window should be determined by preliminary experiments before formal experiments.

(3) Model validation

Blood glucose can be measured within 3–7 days after ALX modeling, followed by continuous monitoring to confirm the stability of hyperglycemia. Successfully modeled animals usually show sustained hyperglycemia, increased water intake and urine output, weight loss, pancreatic β-cell injury, and decreased insulin secretion. If mortality is too high or blood glucose fluctuates excessively, the dose, animal condition, or operating conditions should be adjusted.

(4) Applicable scenarios

The ALX model is suitable for studying β-cell oxidative injury, hyperglycemia induction, and antioxidant intervention. Because ALX has poor stability and relatively large model variability, many experiments now more commonly select STZ as the chemical induction model for diabetes. If the research focus is free radical-mediated β-cell injury, ALX still has certain value.

Table 2 Comparison of Chemically Induced Diabetic Models

Item	STZ Model	ALX Model
Main mechanism	DNA alkylation, NAD⁺/ATP depletion, β-cell death	Free radical generation and β-cell oxidative injury
Common animals	Mice and rats	Mice and rats
Modeling speed	Fast	Fast
Common administration routes	Intraperitoneal injection, tail vein injection	Intraperitoneal injection, intravenous injection
Main advantages	High modeling success rate and broad application	Suitable for β-cell oxidative injury research
Main limitations	Possible liver and kidney toxicity; strong dose dependence	Poor stability and large variability
Model validation	Sustained hyperglycemia, decreased insulin, reduced β cells	Sustained hyperglycemia and oxidative injury of pancreatic β cells

2.3 Spontaneous Type 1 Diabetes Modeling in NOD Mice

(1) Modeling method

NOD mice do not require exogenous drug induction. The essential modeling approach is to select animals of appropriate age and continuously monitor the spontaneous disease process. NOD mice usually develop insulitis at 3–4 weeks of age, and obvious diabetes often appears after 10–14 weeks and may continue to progress until about 30 weeks. The incidence is usually higher in females than in males, so female NOD mice are often preferentially selected, or male and female animals are analyzed separately according to the research objective.

(2) Operating procedure

When establishing the NOD model, body weight, urine glucose, and blood glucose should be measured regularly from the defined age. In the early stage, pancreatic tissue can be collected during the insulitis phase to observe immune cell infiltration. During the disease-onset stage, diabetic animals are screened according to continuous blood glucose elevation. Intervention studies can set different treatment windows at the early insulitis stage, before blood glucose elevation, or after overt diabetes is confirmed.

(3) Model validation

Successful modeling in NOD mice should not be evaluated only by blood glucose. Insulitis score, β-cell area, insulin staining, T-cell infiltration, and autoimmunity-related indicators should also be assessed. If immune prevention is being studied, diabetes incidence, delayed disease onset, and insulitis severity should be emphasized. If therapeutic intervention is being studied, blood glucose stability, residual β-cell function, and changes in immune cell phenotype should be observed.

(4) Precautions

The incidence of diabetes in NOD mice is strongly affected by housing environment, microbial status, barrier conditions, diet, sex, and animal source. The same experiment should use animals from the same source, under the same barrier environment, and preferably from the same batch, to avoid incidence bias caused by environmental differences.

2.4 Spontaneous Models in BB Rats and LEW-IDDM Rats

BB rats and LEW-IDDM rats are both spontaneous insulin-dependent diabetes models and do not require chemical induction. BB rats often develop diabetes at 8–16 weeks of age, with a severe phenotype, and usually require insulin support for long-term survival. LEW-IDDM rats often develop diabetes at 8–9 weeks of age, with an incidence of about 60%, similar incidence in males and females, and relatively better survival after diabetes develops.

For these models, the key modeling approach is continuous monitoring of blood glucose and body weight according to age, with islet histological and immunological analyses performed before and after disease onset. BB rats are suitable for studying insulin-dependent diabetes and immune mechanisms, but long-term studies should consider the effect of insulin maintenance therapy on experimental endpoints. LEW-IDDM rats are more suitable for T1D mechanisms and observation of some complications.

2.5 Akita Mouse Genetic Diabetic Model

(1) Modeling method

Akita mice develop diabetes due to an INS2 gene mutation that causes proinsulin misfolding, β-cell endoplasmic reticulum stress, insufficient insulin secretion, and hyperglycemia. This model does not require exogenous induction. Target genotype animals are usually obtained by purchase or breeding, followed by blood glucose and phenotype screening according to age.

(2) Monitoring procedure

Akita heterozygous mice usually begin to develop hyperglycemia at 3–4 weeks of age. Body weight, water intake, urine output, fasting blood glucose, random blood glucose, and insulin levels can be measured regularly. If diabetic complications are being studied, observation can be extended for several months, and renal, retinal, neural, or cardiovascular histopathology can be examined.

(3) Applicable scenarios

Akita mice are suitable for studying β-cell endoplasmic reticulum stress, abnormal proinsulin processing, pancreatic β-cell functional failure, and long-term hyperglycemic injury. However, it is not an autoimmune type 1 diabetes model and cannot replace NOD mice for studies of T cell-mediated insulitis.

3 Animal Modeling Methods for Type 2 Diabetes

3.1 High-Fat Diet-Induced Model

(1) Modeling principle

High-fat diet (HFD) induces obesity, adipose tissue expansion, chronic low-grade inflammation, insulin resistance, impaired glucose tolerance, and compensatory pancreatic β-cell responses through long-term nutritional excess. This model more closely resembles the early course of lifestyle-related type 2 diabetes and is suitable for studying obesity, insulin resistance, adipose tissue inflammation, gut microbiota, and metabolic syndrome.

(2) Modeling procedure

HFD modeling usually selects metabolically sensitive mice or rats. After acclimation, animals are assigned to a normal diet group and a high-fat diet group. Common high-fat diets provide 45%–60% of energy from fat, and 60% fat energy is more likely to induce obesity and insulin resistance. During modeling, body weight, food intake, and general condition should be recorded weekly, and fasting blood glucose, random blood glucose, OGTT, ITT, and insulin levels should be measured at defined time points.

In general, high-fat diet feeding for 1–2 weeks can show trends of altered glucose metabolism, 2–4 weeks can produce more obvious weight gain, 8–12 weeks can show relatively stable impaired glucose tolerance and insulin resistance, and 16–20 weeks can produce more obvious obesity and metabolic abnormalities. The specific duration depends on animal strain, dietary fat ratio, and research endpoint.

(3) Model validation

Successful HFD modeling should not be judged only by hyperglycemia. A successful HFD model usually shows weight gain, increased fat mass, impaired glucose tolerance, abnormal insulin tolerance, hyperinsulinemia, dyslipidemia, and adipose tissue inflammation. Some animals may not show obvious hyperglycemia but still develop a typical insulin resistance model.

(4) Precautions

The HFD model requires a relatively long modeling period and is strongly affected by diet formulation, fat source, animal strain, sex, age, and housing environment. If the research objective requires stable hyperglycemia, HFD alone may be insufficient, and extension of the modeling period or combination with low-dose STZ may be considered.

3.2 HFD+Low-Dose STZ Composite Model

(1) Modeling principle

The HFD+low-dose STZ model first uses high-fat diet to induce obesity and insulin resistance, followed by low-dose STZ to cause partial β-cell injury, allowing animals to shift from compensatory hyperinsulinemia to sustained hyperglycemia. This model can simultaneously simulate insulin resistance and insufficient β-cell function in type 2 diabetes and is a commonly used composite induction model for T2D.

(2) Modeling procedure

A common procedure is to provide HFD feeding for 4–8 weeks after animal acclimation, allowing the animals to develop weight gain, impaired glucose tolerance, or insulin resistance. Low-dose STZ is then administered by intraperitoneal or intravenous injection to cause partial β-cell injury. Blood glucose is measured 3–7 days after administration, followed by repeated testing to confirm whether hyperglycemia is stable. HFD feeding is then continued, and animals are observed for 4–12 weeks or longer according to research needs.

The principle of low-dose STZ is to “partially damage β cells,” not to completely destroy β cells. In mice, common protocols may use repeated low-dose administration or a single low-dose induction. In rats, low-dose STZ can also be combined with high-fat diet. The specific dose should be determined through preliminary experiments to achieve both a high modeling success rate and low mortality.

(3) Model validation

The HFD+STZ model should meet two conditions simultaneously: insulin resistance and sustained hyperglycemia. Body weight, fasting blood glucose, random blood glucose, OGTT, ITT, fasting insulin, blood lipids, islet histology, and β-cell area can be evaluated. If animals show severe wasting, extremely low insulin, and massive β-cell destruction, the STZ dose is too high and the model is more biased toward T1D-like injury.

(4) Applicable scenarios

The HFD+STZ model is suitable for evaluating hypoglycemic drugs for T2D, studying insulin sensitizers, β-cell protection, metabolic inflammation, and diabetic complications. This model more easily produces stable hyperglycemia than HFD alone, but it is more dependent on operating parameters than monogenic models.

Table 3 Comparison Between HFD and HFD+STZ Modeling

Item	HFD Model	HFD+STZ Model
Modeling core	Diet-induced obesity and insulin resistance	Insulin resistance combined with partial β-cell injury
Degree of hyperglycemia	May be absent or mild	Usually more stable
Insulin status	Mostly hyperinsulinemia in the early stage	May shift from compensation to secretion insufficiency
Modeling period	Relatively long	Moderate
Applicable directions	Prediabetes/T2D early stage, obesity, metabolic syndrome	T2D pharmacodynamics, stable hyperglycemia, complications
Main risk	Unstable hyperglycemia	Excess STZ causing T1D-like conversion

3.3 ob/ob Mouse Modeling

ob/ob mice develop leptin deficiency due to a leptin gene defect, resulting in increased food intake, obesity, hyperinsulinemia, and insulin resistance. This model does not require artificial induction. The modeling method mainly involves selecting animals with the target genotype and monitoring body weight, blood glucose, insulin, blood lipids, and glucose tolerance starting at an appropriate age.

ob/ob mice are usually suitable for studies of obesity, leptin deficiency, insulin resistance, adipose tissue inflammation, and metabolic syndrome. The experimental focus is to control differences in animal age and body weight and to select either the early obesity stage or the stage with obvious metabolic abnormalities according to the research endpoint. Because its etiology is monogenic leptin deficiency, which differs from most human T2D, the mechanistic boundaries of the model should be emphasized in result interpretation.

3.4 db/db Mouse Modeling

db/db mice develop leptin signaling impairment due to leptin receptor gene mutation and show overeating, obesity, insulin resistance, hyperinsulinemia, and hyperglycemia. This model does not require exogenous induction. Animals of a specific background strain and age are usually selected directly for experiments.

During modeling, body weight, blood glucose, urine glucose, insulin, OGTT, ITT, and renal indicators should be dynamically monitored according to age. db/db mice on the commonly used C57BLKS/J background show more obvious hyperglycemia and may develop β-cell functional decline and diabetic nephropathy-like changes. db/db mice are suitable for T2D drug efficacy evaluation, diabetic nephropathy, obesity-related inflammation, and insulin resistance studies.

3.5 ZDF Rat Modeling

ZDF rats are an obesity-related type 2 diabetes model formed on a leptin receptor mutation background. Male ZDF rats usually develop obvious diabetes at about 8–10 weeks of age, characterized by insulin resistance, early hyperinsulinemia, and later β-cell functional decline. Female ZDF rats are less likely to develop stable diabetes, so male animals are usually preferred when establishing diabetic models.

The key to ZDF rat modeling is selecting male animals at an appropriate age and dynamically measuring body weight, blood glucose, blood lipids, insulin, OGTT, and ITT. This model is suitable for studying T2D progression, β-cell compensatory failure, insulin resistance, and diabetic complications.

3.6 KK-Ay Mouse Modeling

KK-Ay mice are an obesity-related diabetic model generated by introducing the Ay mutation into the KK mouse background. These animals can show increased food intake, obesity, hyperinsulinemia, hyperglycemia, and insulin resistance, with the phenotype usually more obvious in males.

KK-Ay mice do not require chemical induction. The modeling method is to select the target strain and animals of appropriate age, followed by monitoring of body weight, blood glucose, blood lipids, OGTT, ITT, and insulin levels at defined time points. This model is suitable for obesity-related type 2 diabetes, metabolic syndrome, insulin resistance, and evaluation of hypoglycemic drugs.

3.7 OLETF Rat Modeling

OLETF rats are a spontaneous type 2 diabetic rat model characterized by mild obesity, late-onset hyperglycemia, and progressive changes in islet structure. Diabetes usually becomes gradually evident after 18 weeks of age, and the disease course is relatively long, making this model suitable for simulating chronic T2D progression.

The OLETF model does not require artificial induction. The key is long-term follow-up according to age. In experiments, blood glucose, body weight, blood lipids, OGTT, ITT, insulin, and renal function indicators should be measured regularly, and islet pathology should be observed at different disease stages. This model is suitable for studying the natural course of T2D, β-cell compensation and failure, diabetic nephropathy, and chronic metabolic injury.

3.8 NZO Mouse Modeling

NZO mice are a polygenic obesity model that may develop hyperleptinemia, leptin resistance, obesity, impaired glucose tolerance, and diabetes in some males. This model does not require exogenous induction, but individual variability is large. Therefore, clear inclusion criteria should be set during modeling, such as body weight, fasting blood glucose, OGTT results, or insulin levels.

NZO mice are suitable for studying polygenic obesity, leptin resistance, insulin resistance, and β-cell compensatory failure. Because the diabetes incidence is not 100%, a sufficient number of animals should be reserved before experiments, and metabolic phenotype screening should be performed before formal grouping.

Table 4 Modeling Methods for Common Type 2 Diabetes Models

Model	Induction Required	Specific Modeling Method	Key Modeling Focus
HFD model	Yes	Long-term high-fat diet feeding	Obesity, insulin resistance, impaired glucose tolerance
HFD+STZ model	Yes	High-fat diet first, followed by low-dose STZ	Insulin resistance plus partial β-cell injury
ob/ob mice	No	Select leptin-deficient animals and observe by age	Obesity, hyperinsulinemia, insulin resistance
db/db mice	No	Select leptin receptor-deficient animals and observe by age	Obesity, hyperglycemia, nephropathy-like changes
ZDF rats	No	Select male ZDF rats and observe by age	Stable T2D phenotype in males
KK-Ay mice	No	Select target strain and appropriate age	Obesity, hyperglycemia, hyperinsulinemia
OLETF rats	No	Long-term follow-up of spontaneous T2D progression	Late-onset hyperglycemia and complications
NZO mice	No	Screen polygenic obese animals for inclusion	Large individual variability; phenotype screening required

4 Evaluation Indicators After Successful Modeling

4.1 Blood Glucose Detection

Fasting blood glucose and random blood glucose are the most basic indicators for diabetic modeling. STZ and ALX models usually begin blood glucose testing 3–7 days after administration. HFD models should be tested regularly by week. Genetic models should be dynamically tested according to age. A common practice is to judge successful modeling after blood glucose exceeds the experimental threshold multiple consecutive times, rather than relying on a single blood glucose result.

4.2 Glucose Tolerance Test

OGTT, IPGTT, or IVGTT is used to evaluate glucose clearance capacity. Type 2 diabetes models especially require glucose tolerance testing, because some animals may not yet show obvious elevation of fasting blood glucose, but delayed glucose decline after glucose loading indicates abnormal glucose metabolism regulation.

4.3 Insulin Tolerance Test

ITT is used to evaluate insulin sensitivity. HFD, db/db, ob/ob, ZDF, KK-Ay, and HFD+STZ models should all consider ITT detection. If the degree of blood glucose reduction after insulin injection is weakened, insulin resistance is indicated.

4.4 Insulin and C-Peptide Detection

Insulin and C-peptide help distinguish insulin deficiency from insulin resistance. High-dose STZ or ALX models usually show decreased insulin. Early HFD models often show hyperinsulinemia. HFD+STZ and late-stage T2D may show hyperglycemia combined with insufficient insulin secretion.

4.5 Islet Pathology

Islet histology is important for determining the modeling mechanism. T1D models should focus on pancreatic β-cell area, insulin staining intensity, insulitis, and β-cell destruction. T2D models should examine islet hyperplasia, β-cell compensation, disordered islet structure, β-cell apoptosis, and later functional failure.

4.6 Target Organ Complications

Complication studies require selection of detection indicators according to endpoints. Diabetic nephropathy can be evaluated by urinary albumin, urinary albumin/creatinine ratio, serum creatinine, glomerular hypertrophy, mesangial expansion, and renal fibrosis. Diabetic neuropathy can be evaluated by nerve conduction velocity, pain threshold, and myelin injury. Retinopathy can be evaluated by vascular permeability, inflammation, and retinal structural changes.

Table 5 Indicators for Determining Successful Diabetic Animal Modeling

Evaluation Module	Common Indicators	Main Significance
Blood glucose status	Fasting blood glucose, random blood glucose, continuous blood glucose monitoring	Determines whether hyperglycemia is stable
Glucose tolerance	OGTT, IPGTT, IVGTT	Determines glucose clearance capacity
Insulin sensitivity	ITT, HOMA-IR, insulin level	Determines insulin resistance
β-cell function	Insulin, C-peptide, insulin release test	Determines islet secretory capacity
Islet pathology	Insulin staining, β-cell area, insulitis	Determines islet injury or compensation
Obesity metabolism	Body weight, fat mass, blood lipids, adipose tissue inflammation	Determines obesity-related T2D phenotype
Oxidative stress	ROS, MDA, SOD, GSH/GSSG	Determines the degree of metabolic injury
Renal complications	Urinary albumin, serum creatinine, glomerular lesions	Determines diabetic nephropathy
Neurological complications	Nerve conduction velocity, pain threshold, myelin injury	Determines diabetic neuropathy
Retinopathy	Vascular permeability, retinal structure, inflammatory indicators	Determines diabetic retinal injury

5 Common Problems in Diabetic Animal Modeling

5.1 No Increase in Blood Glucose After STZ Induction

Common causes include STZ inactivation, prolonged preparation time, inappropriate pH, insufficient dose, administration failure, insensitive animal strain, or inconsistent fasting conditions. The STZ preparation and administration procedure should first be checked, followed by dose and administration route adjustment according to preliminary experimental results.

5.2 High Mortality After STZ Induction

High mortality is usually associated with excessive dose, poor animal condition, excessive fasting, severe dehydration, young age, or strong acute toxicity. Mortality can be reduced by lowering the dose, using a multiple low-dose protocol, optimizing fasting duration, strengthening post-modeling observation, and setting humane endpoints.

5.3 Unstable Hyperglycemia in HFD Models

HFD models mainly induce insulin resistance. In the early stage, β cells can maintain blood glucose by increasing insulin secretion, so hyperglycemia may not be obvious. If the research objective is stable hyperglycemia, the HFD period can be extended, or an HFD+low-dose STZ composite model can be used.

5.4 HFD+STZ Model Bias Toward Type 1 Diabetes

If the STZ dose is too high, β cells are severely destroyed, and animals show marked wasting, low insulin, and severe hyperglycemia. In this case, the model more closely resembles type 1 diabetes-like injury. The key to HFD+STZ modeling is that low-dose STZ causes partial β-cell injury, rather than complete destruction of islet function.

5.5 Do Genetic Models Require Modeling?

Genetic models such as ob/ob, db/db, ZDF, KK-Ay, OLETF, NZO, and Akita usually do not require exogenous induction. Their “modeling” mainly involves selecting the appropriate strain, genotype, sex, and age, followed by metabolic phenotype screening over time. Before formal experiments, the model status should still be confirmed using blood glucose, body weight, OGTT, ITT, and insulin levels.

5.6 Why Are Complication Models Not Obvious?

Unclear or mild complications are usually due to insufficient duration of hyperglycemia, the model itself being unlikely to develop the target organ injury, premature animal death, or inappropriate endpoint indicators. Complication studies should prioritize models with long-term stable hyperglycemia and known target organ lesions, and should set a sufficient observation period.

Table 6 Common Failure Causes and Adjustment Directions in Diabetic Animal Modeling

Problem	Possible Causes	Adjustment Direction
Blood glucose does not increase	Inducer failure, insufficient dose, administration failure	Check reagent freshness, pH, dose, and injection procedure
Large blood glucose fluctuation	Inconsistent blood sampling time, high stress, different fasting conditions	Fix detection time and standardize fasting and operating procedures
High mortality	Excessive dose, poor animal condition, severe dehydration	Reduce dose and optimize fasting and observation management
HFD model fails	Short feeding period, insensitive strain, insufficient fat ratio	Extend the period, adjust the strain, or use HFD+STZ
T2D model shifts toward T1D	Excessive STZ dose	Reduce STZ dose and detect insulin and β-cell area
Complications are not obvious	Insufficient disease duration or unsuitable model	Extend observation period or switch to a model suitable for complications
Large individual variability	Mixed sex, large body weight differences, different feed intake	Stratified enrollment and standardization of sex, age, and body weight

6 Model Selection According to Different Research Objectives

6.1 Screening of Hypoglycemic Drugs

For short-term screening of hypoglycemic drugs, STZ models, HFD+STZ models, db/db mice, or ZDF rats can be selected. If the drug mechanism is insulin sensitization, HFD, db/db, ZDF, or HFD+STZ models should be prioritized. If the drug mechanism is β-cell protection, low-dose STZ models or HFD+STZ models can be selected.

6.2 Pancreatic β-Cell Protection Research

For studying β-cell injury, STZ, ALX, Akita mice, or HFD+STZ models can be selected. STZ and ALX are suitable for chemically induced β-cell injury. Akita mice are suitable for endoplasmic reticulum stress. HFD+STZ is suitable for β-cell functional insufficiency under an insulin resistance background.

6.3 Insulin Resistance Research

For insulin resistance studies, HFD models, ob/ob mice, db/db mice, ZDF rats, or KK-Ay mice should be prioritized. Such studies should focus on ITT, insulin levels, adipose tissue inflammation, hepatic insulin signaling, and skeletal muscle glucose uptake.

6.4 Diabetic Nephropathy Research

For diabetic nephropathy, db/db mice, Akita mice, OLETF rats, or long-term STZ models can be selected. After modeling, urinary albumin, urinary albumin/creatinine ratio, serum creatinine, glomerular pathology, mesangial expansion, and renal fibrosis should be detected, rather than evaluating blood glucose alone.

6.5 Diabetic Neuropathy Research

Diabetic neuropathy usually requires a relatively long duration of hyperglycemia. Long-term STZ models, db/db models, or other chronic T2D models can be selected, and nerve conduction velocity, mechanical pain threshold, thermal pain threshold, sciatic nerve pathology, and myelin injury should be evaluated.

6.6 Diabetic Retinopathy Research

Retinopathy modeling requires a relatively long modeling period, and models capable of maintaining long-term hyperglycemia should be selected. In addition to blood glucose, retinal vascular permeability, inflammatory responses, retinal neural injury, and microvascular structural changes should be observed.

Table 7 Product Selection Related to Diabetic Animal Modeling

Research Module	Product Name	CAS No.	Application Positioning
Chemical induction modeling	Streptozotocin (STZ)	18883-66-4	Induces pancreatic β-cell injury and establishes a type 1 diabetes-like hyperglycemic model; low dose can be used for HFD+STZ composite models
Composite induction modeling	Nicotinamide	98-92-0	Commonly combined with STZ to partially protect β cells and construct milder T2D-like models
STZ preparation buffer system	Citric acid monohydrate	5949-29-1	Used to prepare citrate buffer for STZ and maintain acidic conditions
STZ preparation buffer system	Sodium citrate dihydrate	6132-04-3	Used with citric acid to prepare 0.05 mol/L citrate buffer
Glucose tolerance test	D-Glucose	50-99-7	Used for glucose loading in OGTT, IPGTT, and IVGTT
Insulin tolerance test	Insulin (recombinant human)	11061-68-0	Used in ITT to evaluate insulin sensitivity; can also be used for supportive treatment in severe diabetic models
Glucose uptake evaluation	2-Deoxy-D-glucose	154-17-6	Used for glucose uptake, glycolysis, and insulin response studies
Glucose uptake evaluation	2-NBDG	186689-07-6	Fluorescent glucose analog used for cellular or tissue glucose uptake detection
High-sugar/nutrient loading	D-Fructose	57-48-7	Used in high-sugar diets and fructose-induced insulin resistance and lipid metabolism abnormality studies
High-sugar/nutrient loading	Sucrose	57-50-1	Used in high-sugar diet or sugar-lipid loading model-related studies
High-fat/lipotoxicity research	Cholesterol	57-88-5	Used in high-fat diet formulation, lipid metabolism disorder, and atherosclerosis risk studies
High-fat/lipotoxicity research	Sodium cholate	361-09-1	Promotes lipid absorption in high-fat/high-cholesterol diet systems
Lipotoxicity research	Palmitic acid	57-10-3	Simulates saturated fatty acid lipotoxicity; commonly used in insulin resistance and β-cell injury studies
Lipotoxicity research	Oleic acid	112-80-1	Used in lipid loading, hepatic steatosis, and metabolic stress research
Positive control/hypoglycemic intervention	Metformin hydrochloride	1115-70-4	Common hypoglycemic positive control in T2D models; used to evaluate insulin sensitivity and glucose metabolism improvement
Positive control/insulin sensitization	Pioglitazone hydrochloride	112529-15-4	Insulin sensitizer used as an intervention control in T2D insulin resistance
Positive control/glucose absorption intervention	Acarbose	56180-94-0	α-Glucosidase inhibitor used for postprandial blood glucose and carbohydrate absorption intervention studies
Insulin resistance induction	Dexamethasone	50-02-2	Used for glucocorticoid-induced insulin resistance or metabolic abnormality models
Islet staining	Dithizone	60-10-6	Zinc-related islet staining for islet isolation and morphological observation
Histopathological staining	Hematoxylin	517-28-2	Nuclear staining in HE staining for pathological observation of pancreas, kidney, liver, and other tissues
Histopathological staining	Eosin Y	17372-87-1	Cytoplasmic and tissue structure staining in HE staining
Glycogen/basement membrane staining	Periodic acid	10450-60-9	Key oxidizing reagent in PAS staining, used for observing basement membrane and mesangial regions in diabetic nephropathy
Fibrosis evaluation	Sirius Red F3B	2610-10-8	Collagen fiber staining for evaluating diabetic nephropathy, myocardial fibrosis, and liver fibrosis
Lipid deposition evaluation	Oil Red O	1320-06-5	Lipid droplet staining for observing hepatic steatosis, adipose tissue, and arterial lipid deposition
Oxidative stress model	tert-Butyl hydroperoxide (t-BHP)	75-91-2	Stably induces oxidative stress for β-cell, renal cell, or neuronal injury studies
ROS detection	DCFH-DA	4091-99-0	Detection of total cellular ROS levels
Superoxide detection	Dihydroethidium (DHE)	104821-25-2	Detects superoxide levels in cells or tissues
Lipid peroxidation detection	Thiobarbituric acid (TBA)	504-17-6	Reagent related to the MDA/TBARS method for lipid peroxidation evaluation
Antioxidant system	Reduced glutathione (GSH)	70-18-8	Antioxidant buffering system research and redox status evaluation
Antioxidant system	Oxidized glutathione (GSSG)	27025-41-8	GSH/GSSG ratio analysis for evaluating diabetic oxidative stress
Glycation injury research	Methylglyoxal	78-98-8	Induces glycation stress and AGE-related injury; commonly used in diabetic vascular and renal injury studies

The key to diabetic animal modeling is aligning the model mechanism with the research endpoint. STZ, ALX, HFD, HFD+STZ, and genetic models each have their own applicable boundaries. Only by simultaneously controlling the modeling procedure, model validation criteria, and evaluation indicators can the experimental results be interpretable and reproducible.

For more related articles, please see below:

[1] The Technology Driving Biomedical Revolution — Animal Modeling

[2] Animal Modeling—Tumor Disease Models

[3] A Detailed Guide to the Construction of Animal Models for Metabolic Diseases

[4] Methods for Establishing Animal Models of Cardiovascular Diseases

[5] Methods for Establishing Models of Nervous System Diseases

[6] Construction of a subcutaneous hormonal tumor model of melanoma