Contribution of animal models to diabetes research: Its history, significance, and translation to humans

ABSTRACT

Diabetes mellitus is still expanding globally and is epidemic in developing countries. The combat of this plague has caused enormous economic and social burdens related to a lowered quality of life in people with diabetes. Despite recent significant improvements of life expectancy in patients with diabetes, there is still a need for efforts to elucidate the complexities and mechanisms of the disease processes to overcome this difficult disorder. To this end, the use of appropriate animal models in diabetes studies is invaluable for translation to humans and for the development of effective treatment. In this review, a variety of animal models of diabetes with spontaneous onset in particular will be introduced and discussed for their implication in diabetes research.

Animal models of diabetes and application to human studies

INTRODUCTION

For a better understanding of the disease mechanisms and for the development of effective treatments, animal models have contributed enormously to the knowledge gained in diabetes research ¹ , ² . The results obtained have translated well to humans in daily practice. On September 3, 2020, Professor Yoshio Goto, a world renowned diabetologist and a lifelong researcher in experimental diabetes and its complications (Figure 1), passed away at the age of 94. He was a founder of the Japan Society of Experimental Diabetes and Obesity the meeting of which is held annually, now in its 36th year. He also established the global study group of animal diabetes, holding the Lessons from Animal Diabetes (LAD) workshop together with Professors Arnold Renold in Geneva, Dan Porte in Seattle, USA and Eleazar Shafrir in Jerusalem, Israel. His work is well remembered for the development of the Goto Kakizaki rat (GK rat), an animal model of spontaneous onset type 2 diabetes ³ , ⁴ . Currently, GK rats are distributed worldwide for studies on diabetes. The advent of GK rats promoted discoveries of many models of spontaneous onset diabetes, in particular with the selective breeding of animals. It is fortunate that we are now able to choose suitable animal models for our studies depending on the study design ⁵ . This review, with a tribute to the establishment of the GK rat by Goto and Kakizaki, introduces mainly rodent models of diabetes and their characteristic features. Some larger mammalian models as well as non‐mammalians will also be described briefly.

Figure 1.

Late Professor Yoshio Goto, Founder of the Japan Society of Experimental Diabetes and Obesity, Emeritus Professor of Tohoku University.

RATIONALE FOR THE USE OF ANIMAL MODELS

Pandy and Dvorakova ⁶ explained the rationale of the use of animal models for medicine. It is mainly based on the intention to understand biological function in humans, being designated according to their use such as: (1) Exploratory models, aimed at understanding the mechanism of action in biological systems and could be related to fundamental or basic research on biological systems or a mechanism associated with disease or abnormal biological function. (2) Explanatory models, aimed at understanding the complex biological problem. They should not necessarily be reliant only on animal usage, but could be a physical, bioinformatics, or mathematical model system developed to unravel complex mechanisms. They are used to develop a scientific hypothesis and to discover fundamental laws. (3) Predictive models are the most important animal models and are generally used for preclinical or applied research. They are unique animal models, which are aimed at assessing a possible effect in humans ⁷ . They are also used to discover and quantify the impact of the treatment, in addition to the evaluation of therapy such as pharmacokinetic/pharmacodynamics and toxicity of drugs ⁸ . Translation of the results to humans may be an ultimate goal for benefits to human health care and improvement of quality of life.

DIAGNOSTIC CRITERIA OF DIABETES IN ANIMAL MODELS

There is no consensus for the diagnosis of diabetes applicable to animal models. This situation is different to the status of human diabetes. The Study Group of the Japan Society of Experimental Diabetes and Obesity (JSEDO) provided the guideline for the use of animal diabetes and the definition of diabetes in animals ⁹ . The criterion of human diabetes as an elevated fasting blood glucose (>126 mg/dL; 7.0 mmoL/L) is based on significant correlation with the probabilities of the presence of retinal complications, either in the American Diabetes Association or in the Japan Diabetic Society. There are no data, however, on the relationship between blood glucose levels and microvascular complications in animal models. In addition, blood glucose dynamics should be variable among animal species. To obtain some guidance, if any, for markers of diabetes, the Japan Society of Animal Diabetes and Obesity sent out a questionnaire on the diagnosis of animal diabetes to investigators who were involved in studies on diabetic animals. Most commonly the investigators used levels of casual blood glucose for the designation of diabetes. For convenience, animals with blood glucose levels more than 300 mg/dL (16.7 mmoL/L) are regarded as animals with overt diabetes. When fasting blood glucose levels were used as indices of elevation, values above 150 (8.3 mmoL/L) or 200 mg/dL (11 mmoL/L) were arbitrarily adopted as a sign of hyperglycemia depending on the purpose of their studies. On the other hand, the continuous presence of sugar in the urine is regarded as a sign of diabetes in NOD mice when repeated blood withdrawal is not feasible. In contrast to models of type 1 diabetes, glucose tolerance tests with an oral or intraperitoneal load of glucose are commonly applied for models of type 2 diabetes. In either case, comparison with a non‐diabetic healthy control group is essential. The study group advocated the use of animal models for diabetes research under a sufficient rationale following ethical guidelines (Table 1).

Table 1.

Rationale and recommendation for the use of animal models in diabetes research

The decision of which model will be used should be determined by the purpose of the study, knowing characteristic pathophysiology of the model. The obtained results are reproducible with clear documentation of age (week, or month), body weight, blood glucose, blood lipids, as well as other critical data. Conditions of animal care, data on food and water intake should also be recorded. Under these circumstances, meticulous comparison of the data with those in suitable control animals should be conducted. Confirm the validity of control animals. Number of animals used for the experiment and unnecessary pain or suffering should be minimized, but be sufficient for scientific evaluation including ample statistical power.

ANIMAL MODELS FOR TYPE 1 DIABETES

Bio Breeding (BB) rats and NOD mice are the most popular models for type 1 diabetes. The clinical features of human type 1 diabetes are characterized by the acute onset of polyuria, polydipsia, hyperglycemia, and insulin deficiency. Autoimmune processes to the islet β‐cells serve as the pathologic basis in these animals. The exact pathogenesis of type 1 diabetes in humans, however, remains unclear. Consequently, the data obtained from these models should be interpreted carefully. In contrast to islet damage, the presence of long‐term hyperglycemia (‘glucotoxicity’) can be used to explore the pathogenesis of diabetic complications of the target organs, but the animals should survive long enough to show tissue damage to this end (Table 2).

Table 2.

List of major rodent animal models for type 1 diabetes mellitus

Strain	Founder(s)/place or institution	Gender (M vs F)	DM onset	Immune basis	Major islet pathology	Application and complication	Others
Type 1
NOD mice (non‐obese diabetes)	Makino, 1980 17  (Shiga, Japan) Jcl: ICR origin (cataract)	F > M DMpop; 90% of F 60% of M	Insulitis 4–5 weeks β‐cell loss 10–14 weeks DM onset ~40 weeks Autoab+, GAD+	Autoimmune basis Analogous to human type 1 diabetes Islet contains CD4,8+ & cytokines	Insulitis β‐cell loss	Useful for intervention study DM onset promoted by cyclophosphamide Nephropathy onset Neuropathy (mainly autonomic)	NON as control
BB rat (Bio Breeding)	Nakhooda, 1974 10 Ottawa, Canada, and Worcestor, MA, USA	F > M 90% of F 60% of M	Insulitis 4–5 weeks DM ~12 weeks of age after puberty 8–16 weeks Autoab+, GAD+	Autoimmune basis Analogous to human type 1 diabetes, weight loss T‐lymphopenia (ART2+)	Insulitis β‐cell loss MHC‐II CD4↓Mφ	Useful for intervention study Polyendocrinopathy (thyroid, salivary & lacrimal gland) Nephropathy Neuropathy	BBDR (diabetes resistant strain)
LEW/AR1/‐IDDM	Lenzen, 2001 35 Hannover, Germany (congenic)	M ≒ F 20–60%	20–60% DM onset; around puberty (8–9 weeks)	Autoimmune basis T, Mφ, NK infiltration MHC‐II upregulation No lymphopenia	Insulitis β‐cell loss	Suitable for complication study No other organ abnormality Polyendocrinopathy
LETL	Kawano, 1980 42 Tokushima, Japan	M ≒ F	Low DM onset ca 20% Autoab	No lymphopenia Autoimmune basis	Insulitis β‐cell loss	Low incidence of DM	LETO as a control
KDP	Komeda, 1988 43 Tokyo, Japan	M ≒ F	80%	Autoimmune basis	Insulitis+ β‐cell loss		KND as a control
Chemically & virally induced STZ rat and mice	Rakieten, STZ, 1963 4 Indiana, USA Lukens, Alloxan, 1948 46 Like & Rossini 1986 53 Worcestor, MA, USA Virus (EMC) Craighead, 1968 55	M ≒ F	Any time but susceptible around puberty	β‐cell toxicity	Massive β cell loss	Over 90% success for STZ but care for survival (10% show recovery) STZ adverse action of tumor induction (liver, kidney, soft tissues) Direct cytotoxicity of virus (EMC‐D variant)

In contrast to spontaneous diabetes, animals with diabetes induced by chemicals are much more commonly used as a model of insulin‐deficient diabetes animals because of the convenience and lower expenses. Streptozotocin and alloxan are frequently used as diabetogenic agents. Such chemically induced diabetic animals are not equivalent to those in human diabetes, but are well studied for the evaluation of the effects of long‐term hyperglycemia on target tissues with diabetic complications.

Spontaneous onset type 1 diabetes model

BB (Bio Breeding) rat

The BB rats were discovered in 1974 in a colony of outbred Wistar rats at Bio Breeding Laboratories in Ottawa, Canada. They manifested a sudden onset of polyuria, hyperglycemia, insulin deficiency, and weight loss after puberty ¹⁰ . Without insulin treatment, BB rats develop severe ketoacidosis sufficient to be fatal. The clinical features appear to be consistent with human type 1 diabetes. Shortly after the discovery, the incidence of diabetes onset was only 10%. By transfer to a suitable environment, nearly 90% of rats developed diabetes between 8 and 16 weeks of age, but 6–8 months later in some substrains. The frequency of the overt diabetic phenotype is comparable in males and females. When diabetes starts early, maturation is impaired ¹¹ . All substrains have been derived from two founder colonies of inbred [dp (diabetes‐prone) BB/Wor (Worcestor)] and outbred (BBdp) ¹¹ . BB rats resistant to diabetes (BBDR) have also been bred to serve as controls.

The pathologic basis of diabetes is characterized by inflamed islets with infiltration of T cells, B cells, NK cells, and macrophages, consistent with the feature of so‐called insulitis encountered in human type 1 diabetes. The progressive decline of islet β cells is in parallel with insulin deficiency. Distinct from human type 1 diabetes, blood analysis of BB rats disclosed lymphopenia with a severe reduction in CD4+ T cells and a near absence of CD8+ T cells ¹¹ . This model has contributed substantially to elucidating the role of genetics in type 1 diabetes ¹² , ¹³ . BB rats have been used in detailed studies of diabetic neuropathy ¹⁴ . Systemic observation in BB rats disclosed robust lymphocytic inflammation in the thyroid gland and gastric mucosa, in which polyendocrine autoimmune processes are exerted ¹⁵ .

Recently, a new substrain named BB/O(ttawa)K(arlsburg) (BB/OK) rats was further cross‐bred with SHR (spontaneously hypertensive rats) to obtain diabetic rats with vascular complications named BB.SHR ¹⁶ . BB.SHR rats are not lymphopenic nor diabetic, but obese and hypertensive, typical of metabolic disorder. This congenic BB.SHR strain is now used in studies on the effects of hypertension and diabetes on cardiovascular systems ¹⁶ .

NOD mouse

The NOD mouse was found at the Shionogi Research Laboratories in Osaka, Japan in 1974 ¹⁷ , ¹⁸ . This model develops marked inflammatory changes of pancreatic islets at 3–4 weeks of age. In the prediabetic stage, the islets are infiltrated mainly by CD4+ and CD8+ lymphocytes with a lesser population of B cells and NK cells ¹⁹ , ²⁰ . The insulitis is accompanied by β cell destruction, but the onset of overt diabetes is not apparent until approximately 90% of pancreatic insulin is lost at around 10–14 weeks of age. Diabetes is more prevalent in females ranging from 60% to 90%, whereas the incidence in males ranges from 10% to 30% ²⁰ . When the mice become overtly diabetic, they rapidly lose weight and require insulin treatment for survival. The MHC class 2 in NOD mice shares structural similarities to those in humans, which may confer resistance or susceptibility to the disease in both NOD mice and humans ²¹ . Genes that influence the onset of diabetes are mapped on chromosome 3, 11, and outside MHC ²² .

The NOD mice are suitable for exploring therapies in which modulation of the autoimmune response is being targeted ²³ , ²⁴ . Caution should be taken, however, that they are kept in pathogen‐free (SPF) conditions to maintain the incidence of diabetes because microbial exposure suppresses the development of diabetes ²⁵ . The NOD mouse has helped to identify many of the genetic and signaling pathways that can lead to type 1 diabetes, as it does represent many aspects of human disease. In fact, the onset of diabetes was significantly inhibited by cyclosporin treatment ²⁶ , whereas cyclophosphamide enhanced the onset and severity of diabetes ²⁷ , ²⁸ . Despite the fact that there are a number of drugs that are effective in NOD mice, they were found to be ineffective in humans. One of the major reasons for this discrepancy may be ascribed to the time point of intervention. The lack of suitable biomarkers and translating dosing from the NOD mouse to humans add problems to effective treatment. NOD mice can also be used in a recurrent autoimmunity model, where syngeneic islets from young non‐diabetic NOD mice are transplanted into diabetic NOD mice ²⁹ . The graft is rapidly destroyed by autoimmune mechanisms.

Extrapancreatic tissue damage is found in NOD mice. Destruction of islet sympathetic nerves concomitant with marked degenerative, dystrophic changes of neurites are salient features of the peripheral nerves, reminiscent of the structure encountered in experimental and human diabetic autonomic neuropathy ³⁰ . Genetic control is implicated in the manifestation of autoimmune neuropathy in this model ³¹ . There are substrains of NOD mice with features of autoimmune neuropathy comparable to chronic inflammatory demyelinating polyneuropathy (CIDP) ³² . There have been some investigations on kidney changes, such as renal hypertrophy and glomerular sclerosis. These alterations are accounted for by activated immune and inflammatory reactions, but the precise mechanisms are largely unknown. A variety of compounds have been used in an experimental attempt to prevent or restore the renal changes ³³ . Complete recovery was not obtained because of the lack of complete inhibition of the autoimmune processes ³⁴ .

LEW.1AR1/‐IDDM rat

LEW rats are derived from congenic Lewis rats and represent a model of spontaneous autoimmune type 1 diabetes with a defined MHC haplotype. This model was established at the Institute of Laboratory Animal Science of Hannover Medical School in Germany ³⁵ . Original LEW rats manifested diabetes at a rate of ~20% at puberty (between around 8 and 9 weeks of age without any gender differences between male and female). With further inbreeding of LEW diabetic rats, the incidence of diabetes increased to ~60% with a comparable incidence in both genders. Pancreatic islets contained apoptotic β cells infiltrated with macrophages and lymphocytes about a week before the animals became hyperglycemic ³⁶ . The diabetic phenotype of LEW rats is characterized by hyperglycemia, glycosuria, ketonuria, and polyuria, but without lymphopenia in which T cells are normally populated. Proinflammatory cytokines are upregulated to exert damage to the islets ³⁷ . In contrast to robust inflamed islets, LEW rats usually survive well without insulin supplementation ³⁶ , and thus are suitable for studies on the progression of diabetic complications. The slow progression of autoimmune diabetes resembles latent autoimmune diabetes in adults (LADA) ³⁸ , alternatively slowly progressive insulin dependent diabetes mellitus (SPIDDM) reported from Japan ³⁹ .

The treatment approach for type 1 diabetes may start with initial monotherapy followed by combination therapy to target proinflammatory cytokines ⁴⁰ . So far in this model, most studies have been investigating the mechanisms involved in the development of diabetes. Among these, Fingolimod (FTY720) treatment protected islet infiltration of immune cells in the prediabetic and early diabetic phase in this model, but not in humans ⁴¹ . The rats with diabetic syndrome exhibit an autosomal recessive mode of inheritance with an incomplete penetrance of the mutant phenotype. Other endocrine and exocrine organs such as adrenals, thyroid, and salivary glands as well as lacrimal glands are unaffected in this model.

LETL rat and KDP rat

The Long‐Evans Tokushima Lean (LETL) rat is an inbred strain with type 1 diabetes based on autoimmune destruction of islet β cells without lymphopenia ⁴² . This strain was established in 1989 from an outbred colony of Long‐Evans rats in Otsuka Pharm Co., Tokushima, Japan. These rats develop sudden onset of diabetes with polyuria, polyphagia, hyperglycemia, and weight loss. Islets are marked with lymphocytic infiltration which is also detected in the salivary and lacrimal glands. Two recessive genes were suggested to be involved in the pathogenesis of insulitis, which is closely linked with RT1^u. There is no gender difference in the incidence of diabetes. The incidence of diabetes is low, ranging at about ~20%. This drawback of a low incidence was overcome by the efforts of Komeda and his group who established substrains of diabetes‐prone and non‐diabetic rats (Komeda non‐diabetes rats, KND) and diabetes prone diabetic rats (Komeda Diabetes Prone rats, KDP) ⁴³ . The KDP rats show a high incidence of diabetes (overall ~70%) and 100% development of mild to severe insulitis at 120–220 days of age. Different from other type 1 diabetes models of NOD, BB, and LEW rats in which islets are infiltrated with IL‐1β and TNFα, the major cytokines in KDP are IFNγ and TNFα ⁴⁴ . There is a complete absence of diabetes incidence in the KND substrain. Among 165 SSLP marker loci throughout all the rat chromosomes, there was no loci of variation in either KDP, KND, or their parenteral LETL rats, indicating that the genetic background of these substrains appears to be uniform except for the major gene(s), responsible for diabetes ⁴⁵ .

Chemically induced type 1 diabetes and virus‐induced type 1 diabetes

The most potent and popular diabetogenic chemicals are alloxan and streptozotocin (STZ) [2‐deoxy‐2‐(3‐(methyl‐3‐nitrosoureido)‐D‐glucopyranose)], the latter of which is synthesized by Streptomycetes chromogenes ⁴⁶ , ⁴⁷ . The diabetogenic effects are different among the species and also in individuals ⁴⁸ , ⁴⁹ . These chemicals are cytotoxic glucose analogs, which accumulate in β‐cells via glucose transporter 2 (GLUT2) ⁵⁰ , ⁵¹ . A higher dosage such as 50 mg/kg STZ and above can produce irreversible diabetes in rats, whereas a dose of 150 mg/kg is required to develop diabetes in other species (e.g. Cynomologus monkey, pigs). Partial correction of hyperglycemia may occur in pigs 4 weeks after STZ injection ⁵⁰ . The mechanism of STZ action depends on poly(ADP)‐ribose activation (the DNA alkylating activity) by its methyl‐nitrosourea moiety ⁵² . The methyl group from STZ is transferred to the DNA molecule and causes damage along with other defined events, which leads to DNA fragmentation. In chemically induced diabetes, the majority of the endogenous β‐cells are destroyed, resulting in little endogenous insulin production. Diabetes is usually induced 5–7 days after treatment with streptozotocin or alloxan. Due to their similarity in structure to glucose, glucose can compete with alloxan and streptozotocin, and thus, fasting animals tend to be more susceptible. Both alloxan and streptozotocin are relatively unstable, and the solutions should ideally be made immediately prior to injection. Repeated injection of a low dose of streptozotocin to neonates of rat and mice was used for the production of a non‐insulin‐dependent diabetes model in which β‐cell destruction is not marked enough to end with fatal hyperglycemia ⁵³ , ⁵⁴ . It is problematic, however, that streptozotocin is a potent carcinogenic compound yielding epithelial and mesenchymal tumors in the kidney, liver, and soft tissues over long‐term observations.

Similar to streptozotocin, viruses have a tropism for the islet to elicit β‐cell damage resulting in type 1 diabetes. Among these, Coxackie B virus, Encephalomyocarditis (EMC) virus, Kilham rat virus, and LCMV under an insulin promoter have been found to be diabetogenic inducing insulitis, hypoinsulinemia, and hyperglycemia ⁵⁵ , ⁵⁶ . Classic studies on virus‐induced diabetes disclosed a clear difference from human type 1 diabetes ⁵⁷ . However, the exact role of viruses in human type 1 diabetes such as fulminant type 1 diabetes, LADA (latent autoimmune diabetes in adult), or SPIDDM (slowly progressive insulin‐dependent diabetes mellitus) is yet to be elucidated, and the relationship of viruses with islet inflammation and exocrine pancreatic changes are largely unknown, still worthy of critical studies.

ANIMAL MODELS FOR TYPE 2 DIABETES

Type 2 diabetes is the most common type of diabetes in humans (Table 3). The clinical features are variable but usually silent for symptoms until various vascular and neurological complications become manifest. The underlying mechanisms of this disorder are ascribed to impaired insulin secretion (insulin deficit) and ineffective insulin action (insulin resistance), resulting in glucose intolerance and overt hyperglycemia ⁵⁸ . Obesity causes insulin resistance with resultant hyperglycemia in animals and contributes to the onset of diabetes. The C57BL6/J mouse develops hyperphagia‐induced obesity subserving as a model for human obesity ⁵⁹ . The induction of obesity by a high fat diet (HFD) appears to depend on epigenetic differences, reminiscent of the diverse range of susceptibility to obesity among humans ⁶⁰ . Not all animals with diabetes are obese and different mechanisms may operate in obesity and diabetes. In this section, spontaneously occurring diabetic animals with polygenic and with monogenic involvement will be discussed.

Table 3.

List of major rodent animal models for type 2 diabetes mellitus

	Founder(s)/place or institution	Gender (M vs F)	DM onset	Metabolic basis	Major islet pathology	Complication	Others
Type 2 diabetes polygenic
GK rat	Goto, Sendai & Hirosaki, 1975 3 , 4 , Japan Portha, 2009 62 Ostensson, 1993 61	M > F	Around 8–12 weeks Age‐related decline of β‐cell function and structure Type 2 diabetes mellitus Selectively bred	Impaired insulin secretion Insulin resistance Inherent decrease in β‐cell mass	Progressive β‐cell loss Fibrosis with irregular contour	Abnormal OGTT Vasculopathy; thickened basal lamina Neuropathy; distal axonal changes Polycystic ovary syndrome model 71	Control Wistar
SDT rat	Shinohara, 2000 72 , Chiba, Japan Sasase, 2013 74 Osaka	M > F Male	20–24 weeks Non‐obese type 2 diabetes mellitus	Ins deficiency Islet pathology Hyperglycemia Hypoinsulinemia	β‐cell loss Islet hemorrhage	Proliferative retinopathy, cataract, glaucoma Nephropathy, mesangial growth Nodular lesion Neuropathy	Non‐obese Wistar
Sand rat Desert gerbil	Marquie, 1984 78 Shafrir, 1992 79 Israel	M ≒ F	80–90% on high energy diet ~8 month age	Obesity Ins resist	Islet hyperplasia Hyperinsulinemia	Hyperglycemia Diet‐induced obesity No hyperphagia
UCD‐T2DM rat	Cummings, 2008 163 , UCSF, USA Cummings, 2011 166 , Davis, CA, USA	M > F M 92% F 43%	Polygenic HFD‐obesity Hybrid of SD&ZDF	Insulin resistance Type 2 diabetes mellitus Dyslipidemia	Islet atrophy hypoplasia β‐cell loss	GLP‐1 and leptin responsive Impaired insulin secretion	LSD rat
NZO mice	Bielschowsky, 1953 85 Bielschowsky, 1970 86	M ≫ F	Type 2 diabetes, polygenic HFD‐induced obesity	Insulin resistance Type 2 diabetes, lipidemia	Islet hypertrophy	Metabolic syndrome model Hyperinsulinemia, leptin resist Hyperglycemia	F: infertile Kept in Dusseldorf
KK & KK‐Ay mice	Nakamura, 1967 90 Michaud, 1994 93	Agouti mutated	Type 2 diabetes, polygenic Obesity Yellow fur	Insulin resistance Obesity	Islet hypertrophy	Hyperinsulinemia Hyperglycemia, hyperleptinemia
NSY mice	Shibata, 1980 96 Nagoya, Japan STZ origin, JCl:ICR	M > F M 98% F 31%	HNFa‐mut Selectively bred Polygenic	Type 2 diabetes	No distinct Islet pathology	Hyperglycemia, fatty liver Hyper → hypoinsulinemia Amyloid in kidney & other organs
ON mice	Nagao‐Oikawa, 2012 104 , 106 Tokyo, Japan	M > F	Selectively bred Polygenic HFD‐induced	Mildly obese Modest type 2 diabetes	Mild islet changes Insulin resistance	Atherosclerosis Hyperglycemia, hypoinsulinemia
TallyHO mice	Kim, 2001 109 , 2006 111 Theiler Original colony, W Virginia, USA	Type 2 diabetes mellitus Male only	polygenic	Obesity both M&F Dyslipidemia Insulin resistance	Islet changes Islet hyperplasia Degranulated β‐cell	Hyperinsulinemia Hyperglycemia lipidemia
TSOD mice	Suzuki, 1992 118 (Tsumura Co, Ibaraki, Japan) Slc:ddY origin	Type 2 diabetes mellitus M ≫ F 100%	Type 2 diabetes mellitus, polygenic Metabolic syndrome	Obesity with marked insulin resistance	Lipidemia Hyperglycemia	Fatty liver, hepatoma Hyper → hypoinsulinemia
Monogenic
OLETF	Kawano, 1992 130 Otsuka Long Evans Tokushima Fatty rat Tokushima, Japan	CCK‐R Type A mutated	Obesity Type 2 diabetes	Obesity Lipidemia Type 2 diabetes	Islet atrophy β‐cell loss Lipid deposition	Fatty liver Type 2 diabetes; chronic hyperglycemia Insulin resistance Lipotoxicity	LETO
ZF	Zucker fatty 1972 141 Zucker and Zucker Cross Merck M‐strain and Sherman	Male Leptin−/− (missense mut)	Fa/fa	Obesity at 4 weeks Hyperphagia Mets, lipidemia Hyperinsulinemia Not DM but IGT	Islet hyperplasia	Marked insulin resistance Hyperinsulinemia Lipotoxicity Infertility
ZDF	Peterson, 1990 147 Indianapolis, IN, USA Genetic Models Inc.	Male Leptin‐R(−/−) (mutated)	ZDF diabetes‐prone (dp) 7–10 weeks DM Selectively bred Hyperglycemic rats	Type 2 diabetes with obesity Insulin resistance Lipidemia Female nonDM, but become DM by HFD	Islet degeneration Progressive β‐cell loss with apoptosis Lipid droplets in β‐cell	Infertility Lipotoxicity Initial hypo‐ to hyperinsulinemia Premature death Mild retinal microvessel changes Neuropathy changes
Koletskey (SHROB) rat	Kolestky (SHROB), 1973 160 , Cleveland, OH, USA Obese SHR	M > F	Mets 5w obese Hybrid SHR&ZF	Model of Mets Insulin resistance Hypertension Lipidemia	Lipidemia (type 4)	Vasculopathy Membranous glomerulonephritis Inflamed large vessels (mesenteric arteries)
ob/ob mice	Jackson lab (USA) Ingalls, 1950 171	> F Leptin (−)	Obese, nonDM Defect of leptin	Insulin resistance Hyperleptinemia Hyperinsulinemia	Lipidemia, hyperphagia Islet hypertrophy	Transient hyperglycemia	B6‐ +/+ mice
db/db mice	Jackson lab (USA) Coleman, 1969 183 LeptinR(−/−)	M > F Mutated Lepr	Type 2 diabetes mellitus 6–8 weeks Mutated leptin receptor (defect of lepr)	Initial hyperphagia Non‐obese	Hyper‐to hypoinsulinemia, islet atrophy and β‐cell loss	Premature death Microangiopathy (retina, kidney) Cardiomyopathy Neuropathy	BKS‐ +/+ mice
Akita mice (Mody2)	Yoshioka et al., 1997 194 Kayo et al., 1998 195	Ins2‐mutated ER tress	C57Bl/6NSlc origin Male only Mimic type 1	Onset 3–4 weeks Insulin deficiency Hyperglycemia Polyuria	Hypoinsulinemia Islet atrophy β‐cell apoptosis	Premature death ~12 weeks of age Neuropathy Nephropathy, albuminuria Microangiopathy
Unique	Pancreatic diabetes
WBN/Kob rat	Wistar‐Bonn Kobori Tsuchitani, 1985 203	Insulin def M > F Pancreatic DM	Exocrine inflame Age‐related Hyperglycemia, hemorrhage	Endocrine Exocrine insufficiency Non‐obese Lean	Islet fibrosis Hemorrhage Chronic pancreatitis Hemosiderin deposit	Motor neuropathy Exocrine insufficiency Vascular complication Sclerotic changes CIDP model

Polygenic models

GK (Goto‐Kakizaki) rat

This model was developed in Hirosaki and Sendai, Japan by Professor Yoshio Goto and his group ³ , ⁴ , ⁶¹ , ⁶² . They first selected normal Wistar rats with mild glucose intolerance and repeated selective breeding of rats with hyperglycemia. After the 6th–7th generation, all offspring developed overt hyperglycemia fulfilling the criteria of diabetes. They do not present obesity. Both insulin resistance and impaired insulin secretion are present. Inheritance is polygenic. The pancreatic islets became mildly inflamed with macrophage infiltration together with distortion and fibrosis with robust reduction of β‐cells ⁶³ , ⁶⁴ . Amyloid deposition is not apparent, possibly due to genetic differences in the amyloid peptide of rodents ⁶⁵ . This model is characterized by lean animals, impaired insulin secretion, moderate hyperglycemia, and insulin resistance. At the age of 6–8 weeks, there is a gradual increase in postprandial hyperglycemia, and impaired glucose tolerance. Studies on this model disclosed an early emergence of microangiopathic changes in the kidney, peripheral nerves, validating this model for type 2 diabetes ⁶⁶ , ⁶⁷ , ⁶⁸ . The GK rat is used widely in USA, Europe, UK, and Japan. Genetic analysis identified the genetic foci related to diabetes onset ⁶⁹ , ⁷⁰ . More recently, this model has exhibited a resemblance to the ovarian structure and hormonal perturbations of polycystic ovary syndrome ⁷¹ .

SDT (spontaneously diabetic Torii) rat

The spontaneously diabetic Torii rat is an inbred strain of spontaneously onset non‐obese diabetic rat of the type 2 diabetes phenotype established in 1992 at Torii Pharmaceutical Laboratory in Japan ⁷² . Diabetes develops in male rats after 20 weeks of age to nearly 100% at 40 weeks of age. Diabetes manifests with severe hyperglycemia, polydipsia, and polyuria, impaired insulin secretion and hypoinsulinemia. Pancreatic islets undergo atrophy with a depletion of β‐cells, mixed with changes of hemorrhage and fibrosis. In this model, it is noteworthy that ocular complications of diabetes characterized by cataract, proliferative retinopathy, and retinal detachment develop with the longer duration of diabetes ⁷³ . Vascular complications also accompany peripheral neuropathy ⁷⁴ . They develop diabetes with a polygenic basis, regulated by at least three quantitative trait loci (QTLs) each located on rat chromosome 1, 2, and X ⁷⁵ . These investigators further developed a new, obese type animal model with diabetes, incorporating Lepr (fatty) gene by breeding with Zucker fatty rat, to explore whether the specific complications of diabetes become worse ⁷⁶ . They found aggravated signs of somatic neuropathy in this model, but still await further confirmation of the neuropathy phenotype ⁷⁷ .

Sand rat (Psammomys obesus)

The sand rat, a diurnal gerbil (Psammomys obesus), is an outbred polygenic model of nutrition‐dependent obesity and diabetes ⁷⁸ . In its native desert habitat, with access to a low‐calorie plant‐based diet, the sand rat is lean and normoglycemic. In the laboratory environment, however, with free access to an energy rich laboratory diet, the sand rat becomes obese and rapidly develops hyperglycemia. Diabetes is characterized by severe hypoinsulinemia, hyperlipidemia, and ketosis, often observed within 3–4 weeks of birth ⁷⁹ . Similar to the progression of type 2 diabetes in humans, the diabetic sand rat suffers from β‐cell degradation, retinopathy, nephropathy, body weight loss, and premature death. Induction of glucotoxicity elicited by excessive diet accelerates the aggravation of β‐cell loss ⁸⁰ , ⁸¹ . Suppressed expression of PDX1 is associated with progressive β‐cell reduction and diabetes onset ⁸¹ . On a cholesterol‐rich diet, the sand rat also develops non‐alcoholic steatohepatitis with morphological and functional consequences that are similar to the human pathology ⁸² . Through selective breeding, researchers in Israel have created diabetes‐prone and diabetes‐resistant sand rat groups in order to study the interaction between diet and genes in the development of diabetes and obesity ⁸³ . Careful investigations in the chronically observed sand rat disclosed the presence of retinal changes comparable to those in humans with diabetes ⁸⁴ .

NZO (New Zealand Obese) mouse

The NZO mouse is a model of type 2 diabetes with a polygenic nature exhibiting classic characteristics such as insulin resistance and impaired glucose‐mediated insulin secretion. This model was developed from a mixed colony at the University of Otago Medical School in Dunedin, New Zealand ⁸⁵ , ⁸⁶ . Initially, three different coat colors were found in this colony, namely agouti, tan, and chocolate. Inbred animals with agouti color commonly manifested obesity and were established as a new strain of obese and agouti‐colored mice termed NZO.B1. The other line of mice developed from agouti was the New Zealand Black (NZB/B1). Inbreeding of the tan colored mice gave rise to the NZY/B1 line and inbreeding of chocolate‐colored mice the NZC line, respectively ⁸⁶ . The lack of a suitable control strain has been a problem, because the NZO mouse is an inbred model. A number of lean mice were used as controls including the C57BL/6, Balb C, and the albino ICR mouse. Some groups attempted to use the lean NZC mouse as a control since it was derived from the same population ⁸⁷ . To raise the inbreeding efficiency, treatment with a β3 adrenergic receptor agonist was adopted. Diet restriction to reduce body weight was also effective. This model typically exhibits hyperphagia, inactivity, and obesity ⁸⁸ . Blood analysis demonstrates lipidemia, hyperglycemia, and hyperinsulinemia in particular in the fed state and elevated levels of leptin which do not respond to its receptor ⁸⁸ . Islet pathology depicts enlarged islets composed mostly of β cells and lack of peripheral mixed composition of α‐ and β‐cells ⁸⁹ .

KK (Kondo Kasukabe) and KKAy (KK‐Agouti‐yellow) mice

Nakamura and Yamada ⁹⁰ introduced a new mouse model of diabetes named KK mice after the founder's name (Kondo) and the location (Kasukabe in Saitama, Japan). The features are modest obesity (30–35 g), polyphagia, polyuria, glucosuria, glucose intolerance, moderate hyperglycemia, lipidemia, insulin resistance, hyperinsulinemia, and histological changes in the pancreas ⁹¹ , ⁹² . Because of the modest diabetes and obesity, Nishimura, a co‐worker of Kondo, transferred the yellow gene (Ay) into KK mice by repeated crossing of yellow obese mice and KK mice. The Ay allele (dominant allele at the mouse agouti locus) is associated phenotypically with yellow fur, hyperphagia, and obesity ⁹³ . This congenic strain of KK mice was named yellow KK or KKAy mice. Yellow KKAy mice are easily distinguished by the black color of the fur from KK mice at weaning. The diabetic phenotype manifests in KKAy mice at young ages (6–8 weeks) but reverts to normal after 40 weeks of age. This may be due to a reduction in food intake. They have a latent prediabetic state of glucose intolerance and insulin resistance before the development of hyperglycemia and glucosuria. Treatment with an α‐glucosidase inhibitor effectively suppressed hyperglycemia more potently in KK‐Ay mice than in normal control mice, mediated by intestinal sucrase activities ⁹⁴ . The KK mice develop diabetes based on polygenic inheritance. A retinoid X receptor antagonist successfully improved leptin resistance without affecting plasma leptin levels in KK‐Ay mice ⁹⁵ .

NSY (Nagoya Shibata Yasuda) mice

As with the KK mice, this model was developed by selective inbreeding using a laboratory strain of mouse termed Jcl:ICR, originally chosen from hyperglycemic mice susceptible to STZ ⁹⁶ . NSY mice spontaneously develop diabetes in an age‐dependent manner. The key features include the type 2 diabetes phenotype exemplified by impaired insulin secretion and mild insulin resistance. Obesity is not a major feature and there is a marked gender difference with almost all males developing hyperglycemia, but less than a third of females, being affected.

The NSY mouse is particularly useful when considering age‐related phenotypes (e.g. decline of β‐cell function). A genome‐wide linkage scan identified a sequence variation in the hepatic nuclear factor 1β gene, a gene also implicated in human MODY syndrome ⁹⁷ , ⁹⁸ , ⁹⁹ . Thus, it showed the polygenic transmission of diabetes genes ¹⁰⁰ showing impaired insulin secretion in response to glucose as well as to arginine. Ikegami and his group uncovered the genetic loci of NSY mice responsible for the diabetes phenotype ⁹⁹ , ¹⁰¹ . It is interesting to note that environmental factors interrelate with this genetic background for the onset of diabetes in which enhanced autonomic nerve activity is characteristic around the endocrine pancreas ¹⁰² . A substrain of obese type of NSY mice with yellow fur was obtained by cross breeding with agouti Ay mice ¹⁰² . This strain has a good fertility and does not display inter‐male aggression, making them useful as a new model for type 2 diabetes with early onset and persistent hyperglycemia ¹⁰³ .

ON (Oikawa‐Nagao) mice

A new mouse model of type 2 diabetes exhibiting the clinical phenotype of diabetes was established by repeated selective inbreeding of mouse lines of hyperglycemia fed with a high fat diet (HFD) ¹⁰⁴ . Inbred breeding started using hybrid mice of C57BL/6J, C3H/HeJ, and AKR/N backgrounds. After 5 weeks of HFD feeding, mice with both high and low 2 h blood glucose levels in an oral glucose tolerance test were separately and repeatedly selected over 14 generations. Two lines of mice prone and resistant to diet‐induced glucose intolerance designated ON‐DP (Oikawa‐Nagao diabetes prone) and ON‐DR (Oikawa‐Nagao diabetes resistant) mice were obtained ¹⁰⁵ , ¹⁰⁶ . The ON‐DP mice exhibit mild diabetes and obesity. These mice likely carry the polygenic and heritable nature of type 2 diabetes, showing impaired insulin secretion and reduced insulin sensitivity (insulin resistance). They are mildly obese with impaired glucose tolerance, hyperinsulinemia and hyperglycemia. Pancreatic islets were slightly hyperplastic with an increased β‐cell area ¹⁰⁵ . The clinical phenotype of glucose intolerance and hyperphagia was likely associated with impaired glucose uptake and methylation of leptin gene promoter in adipose tissue in this model ¹⁰⁷ . This model may also be applicable to the exploration of the role of lipid in the genesis of macroangiopathy and therapeutic effects on atherosclerotic lesions ¹⁰⁸ .

TH (TALLYHO/Jng) mice

This model is a newly developed model of type 2 diabetes with mild obesity named the TALLYHO/Jng (TH) mouse of an inbred polygenic nature ¹⁰⁹ . The TH mice develop overt diabetes rather late (~40 weeks of age) and exhibit reduced insulin‐stimulated glucose uptake in adipose tissue and skeletal muscle and abnormal pancreatic morphology and function. The clinical and pathologic features, in many aspects, resemble polygenic human type 2 diabetes ¹¹⁰ in which leptin has a role in impaired insulin secretion in response to glucose. Despite hyperinsulinemia and obesity, the female TH mice remain euglycemic. The molecular basis of the diabetic phenotype appeared to relate to lower levels of insulin receptor substrate 1 (IRS‐1) and impaired GLUT4 mobilization in the adipose tissue ¹¹¹ . This model also manifests cerebrovascular dysfunction ¹¹² , and male TH mice present drastically increased plasma triglyceride levels. The obesity is related to leptin resistance ¹¹³ . Finally, in addition to the obese diabetic phenotype ¹¹⁴ , the TH mouse displays reduced bone mineral density, making it an interesting model for studying the interplay between type 2 diabetes and the skeleton ¹¹⁵ and cognitive decline in diabetes ¹¹⁶ . Dysregulated mitochondria are invoked in the complicated diabetic phenotype ¹¹⁷ .

TSOD (Tsumura, Suzuki, Obese diabetic) mice

The TSOD mouse is a polygenic type 2 diabetes model with obesity, hyperglycemia, and hyperlipidemia. This model was developed 1999 in Tsumura Central Laboratories, in Ibaraki, Japan by Suzuki et al. (1999) and named after Tsumura, Suzuki, obesity, and diabetes ¹¹⁸ , ¹¹⁹ . It is derived from the selective breeding of obese male mice of the ddY strain using indices of heavy weight and the appearance of urinary glucose. The male TSOD mice constantly showed signs of obesity and urine glucose with increased intake of food and water. With a longer duration of diabetes, this model was found to develop characteristic complications of diabetes ¹²⁰ . Alterations of autonomic activity were accompanied by metabolic aberrations ¹²¹ . Concurrently, increased blood glucose, insulin, and lipid levels are found in males, with some involved with steatohepatitis ¹¹⁸ , ¹²⁰ . They are similar to ob/ob mice showing obesity and hyperinsulinemia, polydipsia, polyuria, and hyperglycemia. Pancreatic islets are hypertrophic without any signs of insulitis, or fibrosis. Although this model is alluded to accompany diabetes‐like complications, the presence of nephropathy and neuropathy is not well characterized. More recently, associated fatty liver, steatohepatitis, as well as the development of hepatic neoplasm have been well studied to constitute the current topics of this model ¹²² , ¹²³ .

Chemically induced type 2 diabetes model

Inflammatory as well as environmental insults to the islet are known to cause functional and structural damage to islet endocrine cells. If the injury is sufficient to destroy a considerable population of more than 70% of the islets, the resultant loss of β‐cells induces hyperglycemia and diabetes onset. If it is modest or not sufficient to cause diabetes, the minor injury still provides a sufficient basis for susceptibility to diabetes to cause diabetes by subsequent environmental influences on the islets. Based on this presumption, Like and Rossini succeeded in the establishment of an animal model of type 2 diabetes, first in mice and then in rats ⁵³ , ¹²⁴ . They applied repeated injections of a small dose of streptozotocin to neonatal mice or rats for the production of rodent models of type 2 diabetes ¹²⁵ , ¹²⁶ . The diabetes occurred with a male predominance. They found lymphocytic infiltration in the islets reminiscent of insulitis encountered in a virus induced model of type 1 diabetes. However, hyperglycemia is modest in these mice injected with small dose of streptozotocin, and rather mimics type 2 diabetes ¹²⁷ , ¹²⁸ . The presence of type C viral particles may also underlie the potential additional injurious effects on the islet ⁵³ , ¹²⁶ .

After birth the animals are usually fed with a high fat diet (HFD) to accelerate the incidence of diabetes ¹²⁷ , ¹²⁸ . Concurrent with the onset of hyperglycemia, polyuria and polydipsia manifest. Islets in these animals show hyperplasia and hypertrophy, but low insulin secretion. With maturation, they show a modest weight gain but not to the extent of obesity. Low insulin secretion and insulin resistance are two characteristics in this model. More recently, a distinct phenotype of neuropathic complications was found in this model ¹²⁹ .

Monogenic/multigenic models and related type 2 diabetes model

Monogenic animal models are extremely useful for understanding the function of the specific target genes in metabolic diseases and monogenic forms of diabetes and obesity. They have contributed as indispensable research tools to the development of effective drug discovery (Table 3). The obese diabetic rats named Otsuka Long Evans Tokushima Fatty (OLETF) rats, Zucker fatty (ZF) rats, Zucker Diabetic Fatty (ZDF) rats are based on the mutated cholecystokinin receptor gene, leptin gene, and leptin receptor gene, respectively ¹ , ⁵ . On the other hand, the ob/ob mouse model with a C57BL/6J background has been used frequently to assess the potency of novel anti‐obesity medications to overcome a strong hyperphagia‐driven obesity ⁵ , ⁶⁰ . Furthermore, this model was also used for studies investigating whether leptin action is required as a cofactor for the efficacy of specific therapies. The db/db mouse on the C57BLKS/J background is often used in studies investigating the efficacy of anti‐diabetic drugs. Similar to rat models, the leptin gene and leptin receptor gene are responsible for the phenotype of ob/ob mouse and db/db mouse, respectively.

OLETF (Otsuka Long Evans Tokushima Fatty) rat

The OLETF rat derived from the closed colony of Long‐Evans rat is a model of type 2 diabetes characterized by polydipsia, polyphagia, obesity, and diabetes ¹³⁰ , ¹³¹ . In this model, diabetes develops rather late around 25 to 60 weeks of age. Male OLETF rats show a high probability (nearly all) of the development of diabetes in contrast to less frequent and later onset of diabetes in female rats ¹³⁰ . The OLETF rats show higher levels of plasma insulin after 8 weeks of age, and after 60 weeks of age, the plasma insulin levels are lower than that of LETO, terminating in hypoinsulinemia ¹³² . After 12 weeks of age, this model shows impaired insulin secretion and insulin resistance in peripheral tissues. With the progression of insulin deficiency, there appears a marked hyperplasia of the islet and ongoing β cell demise. Concomitantly, the islets undergo compromised vascular architecture which is associated with marked insulin resistance in overtly diabetic animals ¹³³ . Together with robust hyperlipidemia, a marked deposition of fat droplets was noted in β‐cells and contributed to the development of diabetes ¹³⁴ . A lower frequency of diabetes onset in female rats is ascribed to ovarian hormones providing compensatory β‐cell hypertrophy in this model ¹³⁵ . Diet restriction or exercise in juvenile stages or shortly after the onset of diabetes postpones the onset of diabetes or an improvement of diabetes. It is of note that atypical hyperplasia of bile and pancreatic ducts is often encountered in this model which may provide some clues to the increased frequencies of atypical growth of ductal neoplasms ¹³⁶ .

The OLETF rats lack the cholecystokinin‐receptor (CCK‐R) type A, contributing to their phenotype ¹³⁷ . CCK is a gut derived peptide hormone that functions as a peripheral satiety signal, and the OLETF rat has been useful in studying the function of CCK in metabolism. The pancreatic growth exerted in troglitazone‐treated OLETF rats is suggested for the involvement of CCK‐R in the maturation of exocrine and endocrine pancreas ¹³⁸ . Multiple genetic components play a role in the development of diabetes in OLETF rats ¹³⁹ . Among these, odb1 is mapped on the X‐chromosome, and odb2 on chromosome 14 which is a candidate gene for CCK‐R type A. These genes cooperate mutually and result in a marked increase in plasma glucose concentrations. This model is also used to test the validity of effective compounds on vascular complications of diabetes ¹⁴⁰ .

ZF (Zucker fatty) rat and ZDF (Zucker diabetic fatty) rat

Among the healthy rat colonies bred in Massachusetts, obese hyperphagic rats were found in the Massachusetts Harriet GB Memorial laboratory by Zucker ¹⁴¹ , ¹⁴² . This strain of outbred breeding was established from a mutation named Zucker, i.e. Zucker fatty rat (mutated leptin gene). It was then transferred to the laboratory of Dr Walter Shaw at Eli Lilly Research Laboratories in Indianapolis, IN and Dr J Clark in Indiana University in 1974–75. The rats were not originally diabetic, but only showed obesity with marked insulin resistance in muscle and fatty tissues ¹⁴³ , ¹⁴⁴ . The pancreatic islets were age‐dependently hyperplastic with an increased β‐cell population ¹⁴⁵ . Degenerative changes or cell death were not documented. Among the transferred Zucker rats, several offspring of the animals were identified as hyperglycemic from which a substrain was rederived. Inbreeding of selected pairs from this rederivation yielded nonobese but a robust diabetic phenotype (type 2 diabetes) and were named Zucker diabetic rats (mutated leptin receptor) ¹⁴⁶ . Subsequently, two colonies of Zucker diabetic fatty (ZDF) rats were established in two laboratories of Dr Richard Peterson at Indiana University Medical School (IUMS) ¹⁴⁷ and Dr Roger Unger at Texas Southwestern Medical Center, Dallas, USA. Although direct comparisons were not made, the phenotype of the two lines appears to be more or less similar. In these rats, a progressive depletion of β‐cells resulting from apoptosis affected by hyperglycemia as well as lipotoxic effects are the main feature of the islet pathology in these rats ¹⁴⁸ . Accumulation of fatty acid in β‐cells with subsequent apoptotic cell death was noted ¹⁴⁹ , ¹⁵⁰ . Suppression of glucose transporters (GLUT) is also implicated in the onset of diabetic hyperglycemia ¹⁵¹ . Unger and his associates introduced the significant role of leptin in the regulation of intracellular triglycerides for the protection of lipotoxicity ¹⁵² . Genetic analysis of the ZDF rat disclosed a mutant leptin receptor (lepr ^−/− ) responsible for obesity and insulin resistance in ZDF rat (obese fa/fa) while the control ZDF (lean fa/‐) rats exhibited wild type expression ¹⁴⁷ . Forced overexpression of normal leptin receptor (OB‐Rb) in ZDF islets by perfusing ZDF pancreata with a vector containing the OB‐Rb cDNA restored the insulin secretion and inhibited lipid storage ¹⁵³ . The coat color of ZDF is black hooded with a black stripe down the length of the back. The ZDF rats carry an autosomal recessive defect in the β‐cell transcription machinery that is inherited independently from the mutation in Lepr. The ZDF rats are also used in the exploration of effective compounds for specific vascular complications of diabetes including neuropathy and micro‐ and macrovascular complications ¹⁵⁴ , ¹⁵⁵ , ¹⁵⁶ .

Efforts were extended to establish an obese type 2 diabetes substrain of the Zucker Diabetic Sprague–Dawley (ZDSD) rat. This model was established by repeated crossing of lean ZDF rats with polygenic obese Sprague–Dawley rats to eliminate the influence of the mutated leptin pathway ¹⁵⁷ . The ZDSD rat colony is maintained currently in Charles River Co. ¹⁵⁸ . Exploration of the pathogenesis of diabetes complications and treatment regimens have been conducted using this model ¹⁵⁹ .

The Koletsky rat, also known as the spontaneously hypertensive obese (SHROB) rat ¹⁶⁰ , ¹⁶¹ , lacks a functional leptin receptor because of a nonsense point mutation of Lepr gene. This model was originally based on a cross between spontaneously hypertensive rats and obese Sprague Dawley rats ¹⁶⁰ . They are hyperphagic, morbidly obese, and hypertensive with reduced energy expenditure, impaired glucose tolerance, impaired insulin sensitivity, and stunted linear growth owing to lower activity. Female and male rats that are homozygous for the mutation are infertile. In addition, and in contrast to the ZDF rat, the Koletsky rat develops hypertension starting at an age of 30 days ¹⁶² . The Koletsky rat is not prone to developing diabetes mellitus but exhibits extensive sclerotic vascular lesions.

University of California Davis‐T2DM (UCD‐T2DM) rat

The UCD‐T2DM rat was obtained from crossing obese and insulin‐resistant Sprague–Dawley rats with lean ZDF rats which were wild‐type for the leptin receptor. The inbred breeding of the rats exhibited an overt type 2 diabetes phenotype with conspicuous β‐cell loss and marked hyperglycemia, as well as impaired glucose‐stimulated insulin secretion ¹⁶³ , ¹⁶⁴ . These animals develop type 2 diabetes on a low‐fat, low‐sugar diet, and both sexes are affected. However, the onset of diabetes is earlier in males (~4–8 months of age) than in females (~8–12 months of age) ¹⁶³ . Unlike ZDF rats, UCD‐T2DM rats are fertile and responsive to chronic leptin administration showing an improvement of hyperglycemia and a high HbA1c ¹⁶⁵ . Leptin administration results in a reduction of endoplasmic reticulum stress in the liver, muscle, adipose tissue, and pancreas ¹⁶⁶ , ¹⁶⁷ . Impairment of glucose sensing in the gut and vagal afferent nerve activity contributes to the onset of the diabetes phenotype ¹⁶⁸ . The UCD‐T2DM rats develop a number of complications of type 2 diabetes, including albuminuria, defects in vertebral disc and bone structure and function, and defects in brain and/or neuronal plasticity and metabolism ¹⁶⁹ . Overall, the UCD‐T2DM rat is a versatile rodent model that can be used in the development and evaluation of novel strategies for the therapeutic management and prevention of type 2 diabetes.

ob/ob mice (lep ^−/− mice)

Also known as B6/ob, these animals are homozygous for the obese spontaneous mutation, Lep^ob (often referred to as ob or ob/ob), show obesity, hyperphagia, transient hyperglycemia, glucose intolerance, and elevated plasma insulin ¹⁷⁰ . ob/ob mice are also hypometabolic, hypothermic, and often subfertile ¹⁷¹ , ¹⁷² . Hormone production from both pituitary and adrenal glands is increased ¹⁷³ , and wound healing is impaired ¹⁷⁴ . The obesity is characterized by a growth in the number and size of adipocytes ¹⁷⁵ . Brain neuropeptide β‐endorphin is altered and related to polyphagic behavior in this model ¹⁷⁶ . Although hyperphagia furthers the obesity, limited access to food did not influence the basal hyperglycemia nor surplus fat deposition in this model, indicating that hyperphagia alone is not responsible for the maintenance of many of the characteristic features in obesity ¹⁷⁷ . This model is sterile but treatment with the human recombinant leptin successfully corrected this defect ¹⁷⁸ . The physiology of obese‐hyperglycemic mice (ob/ob mice) is well summarized as an early phase of type 2 diabetes and obesity ¹⁷⁹ , ¹⁸⁰ .

db/db mice (lepr ^−/− mice)

db/db mice are used as a model of type 2 diabetes and obesity ¹⁸¹ , ¹⁸² . Affected mice are polyphagic, polydipsic, and polyuric. Obesity begins at 3–4 weeks of age. Mice that are homozygous for the spontaneous mutation (Lepr ^db ) demonstrate morbid obesity, chronic hyperglycemia, pancreatic β‐cell atrophy, and hypoinsulinemia at a later stage. Genetic and environmental factors are implicated in the phenotype expressions, as a parabiosis study demonstrated ¹⁸³ , ¹⁸⁴ . The connection of diabetes onset with obesity is complicated ¹⁸⁵ . Islet pathology is characterized by a robust β‐cell depletion and a modest increase in α‐cells ¹⁸⁶ . The onset of diabetes is influenced by diet behavior implicating the importance of diet composition underlying the pathologic basis of diabetes ¹⁸⁷ and autonomic nerve regulation ¹⁸⁸ . Compared with the original strain, crossbreeding of the mice with a BK background yielded the more severe diabetes phenotype black tan‐brachyury (BT‐BR) ob/ob mouse; an uncontrolled rise in blood sugar, a severe depletion of insulin‐producing β‐cells of the pancreatic islets, nephropathy, neuropathy, retinopathy, and myocardial disease and death by 10 months ¹⁸⁹ , ¹⁹⁰ , ¹⁹¹ , ¹⁹² . This strain was transferred to Charles River Co. from the Jackson laboratory in May 2002. Obviously, the leptin receptor gene is mutated, responsible for the onset of diabetes ¹⁸⁰ , ¹⁹³ .

OTHER ANIMAL MODELS WITH MONOGENIC DIABETES OR PANCREATIC DIABETES

Akita mouse (insulin 2 gene)

The Akita mouse was first discovered by Koizumi and his associates in Akita University, Japan ¹⁹⁴ , ¹⁹⁵ . The model is derived from the C57BL/6NSlc mouse with a spontaneous mutation of insulin 2 gene (Ins2+/C96Y mutation) along with incorrect proinsulin processing. The mutation causes the aggregation of misfolded proteins, resulting in the production of endoplasmic reticulum process (ER stress) ¹⁹⁶ . Excessive ER stress exerts progressive β‐cell death, and finally results in severe insulin deficiency starting from 3 to 4 weeks of age. The clinical and metabolic symptoms are characterized by hypoinsulinemia, hyperglycemia, polydispsia, and polyuria. Untreated diabetic mice rarely survive longer than 12 weeks. Although originally considered to be a model of type 2 diabetes, the rapid and total depletion of β‐cell mass in the Akita mouse renders it an alternative to a model of type 1 diabetes, comparable to STZ‐induced diabetes in transplantation studies ¹⁹⁷ . Brain orexin‐related peptide is involved in the phenotype of hyperphagia and anorexia in this model ¹⁹⁸ .

This Akita mouse is used for the study of macrovascular and neuropathic disease in type 1 diabetes. This model can also be considered as a valid animal model of diabetic sympathetic autonomic neuropathy, which closely resembles the pathology of humans as well as other diabetic rodent models ¹⁹⁹ . More recently, the mechanism of the altered pain sensation elicited by hyperglycemia was explored in this model ²⁰⁰ . High levels of albuminuria and systematic renal pathological symptoms have been observed in the Akita mice ²⁰¹ . However, the renal structural changes in this model need clear validation as to whether these changes in fact result from diabetes or other mechanisms ²⁰² . The Akita mice have advantages for developing nephropathy over the chemically induced STZ animal model of diabetes in which streptozotocin is toxic to the renal tissues.

WBN/Kob rat

This strain was originally developed for an investigation of stomach cancer at the University of Bonn in Germany. Dr Oichiro Kobori, a surgeon, and one of the Japanese investigators who worked in the laboratory brought this strain back to Japan. Among these animals, rats with the diabetic phenotype were found and named WBN/Kob (Slc Co., Hamamatsu, Japan). Only male rats that developed hyperglycemia, weight loss, cataract, and motor neuropathy were found ²⁰³ . In the beginning of the characterization of this model, this strain was represented as a model of type 2 diabetes. Further exploration disclosed that the exocrine pancreas underwent fibrosis and atrophy suggesting the presence of exocrine dysfunction, recapitulating the features of chronic pancreatitis ²⁰⁴ , ²⁰⁵ . Thus, this model was regarded as a model of pancreatic diabetes. Progressive pancreatic fibrosis invades both endocrine and exocrine pancreatic tissues, eliciting depletion of endocrine cells and acinar atrophy. The preclinical application of trypsin inhibitors successfully prevented the impaired exocrine pancreatic function ²⁰⁶ , ²⁰⁷ . In a similar manner, angiotensin converting enzyme inhibitor attenuated pancreatic inflammation and fibrosis ²⁰⁸ . Genetic analysis (QTLs) demonstrated that chromosome 7 and X; PRSS1. CFTR, and SPINK1 were increased in rats with pancreatitis in WBN rats ²⁰⁹ . The findings indicated that chronic pancreatitis in WBN/Kob rats is controlled by multiple genes and a genetic analysis in WBN/Kob rats might be useful for gene targeting for human chronic pancreatitis. Of interest this model exhibits characteristic motor neuropathy recapitulating chronic immune‐mediated demyelinating neuropathy ²¹⁰ , as suggested by the presence of the autoimmune involvement in salivary, as well as other endocrine, tissues ²¹¹ . Recently, incorporation of a modified fatty gene into WBN rat was attempted by mating with the ZDF strain carrying the gene for Lepr ^+/+ obesity ²¹² . This model exhibits enhanced pancreatic fibrosis, indicating the role of insulin resistance in the augmented pathology.

OTHER NON‐HUMAN MAMMALIAN SPECIES

Dog, cat, minipig, rabbit, and monkey

Models of diabetes and obesity are needed for an experimental treatment approach for transplantation of tissues and organs, cannulation of vessels, and access to anatomic regions not feasible in humans (e.g. the hepatic portal vein, hepatic and renal veins, or cardio‐ and cerebro‐vascular systems). Historically, dogs (canine model) were used for studies on metabolic changes, such as the discovery of insulin and its action in the brain. Originally, extirpation of the whole pancreas was applied to obtain a diabetic condition ²¹³ . In particular, a beagle dog named Marjorie is the most well known for the discovery of insulin by Banting and Best ²¹⁴ . Nowadays, mongrel hound dogs likely serve as a convenient model because of the consistency of the development of overweight and/or obesity, prediabetes or low dose STZ‐induced diabetes mellitus within a reasonable period of time ²¹⁵ . Duct ligation is also applied for the induction of diabetes in the dog, causing atrophy and marked fibrosis in the pancreas ²¹⁶ .

The cat is most likely to suffer from diabetes which is commonly found by veterinarians. In common with human diabetes, the pancreatic islets undergo marked amyloid deposition ²¹⁷ . The nucleic acid composition of islet amyloid polypeptide (amylin) is analogous to that of type 2 diabetes in humans ²¹⁸ . Clinically, obesity is accompanied by polydipsia, polyphagia, polyuria, and loss of hair in these animals. With progression of the disease, they lose weight ²¹⁹ .

The pig is another large animal model used for translational studies in research on obesity and diabetes ²²⁰ . In practice, the minipig (a small pig) is commonly used for convenience and ease of handling. Depending on the place of breeding, minipigs are called Ossabaw, Yucatan, or Gőttingen minipigs. Obesity in pigs is induced routinely by high‐energy high‐fat and/or high carbohydrate diet. Gőttingen minipigs are used most widely because they can be reared to adulthood at reasonable cost. Initially this model was considered to be a model of type 1 diabetes, occasionally requiring additional streptozotocin or alloxan injection to increase the reduction of the β‐cell mass ²²¹ . Subsequent exploration disclosed the correlation of the β‐cell mass and the extent of insulin deficiency, providing a model of type 2 diabetes ²²² . Inconsistency of hyperglycemia is ascribed to compensatory β‐cell activation in the remaining β‐cells ²²³ . The large size of animals allowed the sequence of perturbed insulin secretion and its association with islet pathology to be shown ²²⁴ .

There is a long history of studying non‐human primates for translational research in metabolic diseases (Figure 2). Commonly used species include rhesus macaques (Macaca mulatta), cynomologus monkeys (Macaca fascicularis), baboons (Papio species), African Green Monkeys (Chlorocebus species), and common marmocets (Callithrix jacchus) ²²⁵ , ²²⁶ . There are clear differences in the metabolic regulation of lipogenesis, thermogenesis as well as insulin‐mediated glucose utilization between monkeys and humans. Notwithstanding, studies in non‐human primates have led directly to translational studies confirming the same mechanism in humans. In particular, it is of note that the presence of islet amyloidosis with marked β‐cell loss characteristic of type 2 diabetes is common in humans and diabetic monkeys, implicating a similar etiology of islet lesions in monkeys and humans ²²⁷ . An important advantage of the use of non‐human primate models includes their close genetic relationship to humans and physiological similarity to humans. The relatively larger size of the animal is well suited to imaging studies of islets or transplantation studies ²²⁸ . The development of vascular complications of diabetes in the model is useful for studies on the treatment effects of drug compounds ²²⁹ .

Figure 2.

Characteristic features of different species of animal models for diabetes mellitus. There are a variety of animal models for diabetes mellitus research which have distinct advantages (green square) and shortcomings (red square). As shown in the figure, non‐mammalian models have the advantages of a low maintenance cost and a short life cycle. By contrast, the physiology of large animals, especially that of non‐human primates, closely resembles human physiology. However, these species have high maintenance costs and long life cycles. As such, these models are unfeasible for large scale, time efficient experiments. Small rodents represent a good combination of throughput and translational physiology (modified from Kleinert et al. ⁵ ).

NON‐MAMMALIAN SPECIES FOR MASS AND SHORT TERM STUDIES (FISH, FLIES, AND NEMATODE)

Fish (carp and zebrafish)

Classically, fish are known to suffer from high blood sugar when exposed to high osmolality. The affected fish, represented by carp, given too much food gradually become skinny and thinner. Ultrastructural investigation disclosed the degranulation of islet β‐cells in diseased fish ²³⁰ . Since such affected fish show thinned skin losing scales, and their eyes developing cataract, this condition was named “Sekoke disease (fish with falling off scales)”, comparable to the condition of human diabetes. Dr Yokote found a marked expansion of the mesangial area of glomeruli and microaneurysmal swelling of retinal microvessels in the diseased carp ²³¹ , ²³² . Recent studies suggested that the scale loss reflects skeletal fragility in the fish with the type 2 diabetes phenotype, reminiscent of impaired bone mineral composition ²³³ .

More recently, investigators have used zebrafish (Danio rerio) to uncover the molecular mechanisms of the metabolic alterations occurring in diabetes, with the advantage of their short lifespan and related genetic analogy to mammalians ²³⁴ , ²³⁵ . Further support for the convenience of the zebrafish is based on the evidence that they have mammalian‐like ghrelin and leptin activity ²³⁶ , ²³⁷ in addition to the concrete presence of insulin‐producing cells ²³⁸ , ²³⁹ . Similar to what occurs in mammals, β‐cell differentiation from common precursors in zebrafish requires both Notch and pancreas transcription factor (Ptf1) signaling ²⁴⁰ , ²⁴¹ . Ablation of β‐cells in zebrafish causes hyperglycemia ²⁴² . Other metabolic pathways will also be explored using Zebrafish to target the discovery of new mechanisms in diabetes. Zebrafish provides advantages of time‐saving analyses replacing the prolonged and transgenerational needs to explore the effects of obesity and diabetes. In addition, the availability of adequate tissues allows high‐throughput analyses such as whole‐genome RNA interference (RNAi) libraries. It may be exemplified by time‐efficient Zebrafish studies on the relationship between hyperglycemia and dementia, compared with a much longer time for humans ²⁴³ .

Nematodes (Caenorhabditis elegans) and flies (Drosophila melanogaster)

Small organisms such as the nematode Caenorhabditis elegans (C. elegans) ²⁴⁴ , ²⁴⁵ and the fruit‐fly Drosophila melanogaster ²⁴⁶ , ²⁴⁷ are also used for molecular studies on lifespan and disease. They have also served as a good model to explore the genetic and environmental factors involved in longevity ²⁴⁸ . Indeed, a high glucose diet shortened their lifespan ²⁴⁹ , whereas diet restriction of glucose successfully prolonged their life ²⁵⁰ . Although the metabolic pathways involved in the extended lifespan are not clear, involvement of autophagy ²⁵¹ or mitochondria‐driven oxidative stress ²⁵⁰ is suggested to play a role in nematode studies. Metformin exposure of C. elegans successfully delayed the process of aging which was associated with altered folate and methionine metabolism ²⁵² . In this setting, activation of AMP‐Kinase may be relevant to the extended lifespan of subjects under diet restriction ²⁵³ .

The mechanisms of diet‐induced obesity are well studied in the model of the fruitfly D. melanogaster ²⁵⁴ . Dietary composition specifies consumption, obesity, and lifespan in this species ²⁵⁵ . Interestingly, although diet restriction was confirmed to extend their lifespan, its effects depended on the amount of glucose but not of yeast which provided more potent dietary effects on the lifespan ²⁵⁶ . Extensive evidence suggests the presence of orthologous insulin and insulin‐like growth factor I (IGF‐I)‐like signaling peptides in Drosophila ²⁵⁷ . Several different cell types including neurons and hypodermic cells produce insulin and IGF‐I‐like signaling peptides ²⁵⁸ . Ablation of insulin producing cells induced the diabetes phenotype in Drosophila. This was not the case in C. elegans, however, in which it did not induce a comparable diabetes‐like state. Insulin resistance can be induced in Drosophila by a high‐calorie diet ²⁵⁹ . Collectively, non‐mammalian species possess multiple pathways of body mass control and energy balance found in mammals. They exhibit conserved features of insulin‐like signaling and secretion, develop diabetes‐like states when exposed to genetic or environmental (particularly nutritive) perturbations. After all, they serve as valuable tools to provide a conceptual basis for more detailed analyses in higher organisms.

CAUTION FOR THE USE OF ANIMAL MODELS

Ethical issues are of particular concern for animal experimentation. Minimal numbers of animals should be employed and special care may be needed for the appropriate use of animals. Unnecessary painful conditions, malnutrition, or an excessive supply of foods are all negative factors or an uncomfortable environment for the animals, which may pose stressful conditions to the animals and may influence the results. Water supply and a clean wood‐chip floor should be prepared for the animals. Investigators should pay special attention to such environmental conditions.

CONCLUDING REMARKS

In this review, I have summarized an updated list of animal models for diabetes and their characteristics with an introduction of the experimental use of non‐mammalian organisms. Investigators should be aware of the fundamental differences between humans and animals but extrapolate the translation of experimental results from animal models towards to human diabetes. There may be many other promising animal models for diabetes which are not referred to in this review, in part due to the limitation of space. I also avoided referring to the use of transgenic or knockout animal models which is an expanding area in diabetes research and should be discussed in future publications. I recommend the readers of this review to look into the related references, and only emphasize the enormous efforts of our predecessors in establishing models for human diseases.

DISCLOSURE

There is nothing to declare for the publication of this review. There is no financial support, or specific relationship to the companies for lectures or publications.

Approval of the research protocol: N/A.

Informed consent: N/A.

Registry and the registration no. of the study/trial: N/A.

Animal studies: N/A.