Review the main features you need to consider when selecting preclinical models for type 2 diabetes (T2D) research, including choice of species, disease progression, and disease inheritance.
Modeling Human Diabetes in Animals
Animal models of diabetes play an essential role in understanding disease mechanisms and evaluating new therapies. One major challenge in preclinical research is that human diabetes progresses from a pre-diabetic to a diabetic state, but many preclinical models miss this full disease progression.
Existing preclinical models mimic specific, limited aspects of diabetes. This means you need to use multiple models for each disease aspect to truly reflect human pathology. This is logistically cumbersome, time-consuming, and costly for drug developers. The lack of translatable models has meant that some compounds show excellent results in preclinical studies but ultimately fail in the clinic.
This blog looks at the three key factors to consider when choosing the best model for preclinical T2D research, including using more translatable larger animals or next generation models.
Species: Large Animal (including NHP) vs Rodent Models of Diabetes
The first main choice is species. Large animals are more directly translational, while small animals are much easier to handle.
Rodent models have many advantages in early preclinical studies. Their small size means easier handling and less compound used per model. Plus, there is more knowledge around the rodent genome, ease of genetic manipulation, relatively short breeding span, and access to physiological testing.
Initial type 2 diabetes studies for drug mechanism often employ “quick and dirty” experiments in knockout mice. These are followed up with studies on ob/ob or db/db mice, since sticking with the same species is key. Db/db and ob/ob mice are well-known obese models of T2D, frequently used for studying safety and efficacy alongside disease mechanism.
While ob/ob or db/db mice are useful for these early preclinical studies, they do lack translatability – without the intact leptin pathway seen in human disease and discussed later in this post. Similarly, Zucker Diabetic (ZD) and Zucker Diabetic Fatty (ZDF) rats are also popular models if this species is needed. However, these models also don’t really mimic human disease, with a mutation in the leptin molecule, no pre-diabetic phase, and development of early beta cell failure.
When you need your studies to more closely resemble human disease, larger animals are often used. This includes non-human primate diabetes models (NHP), which can develop spontaneous diabetes similar to humans.
Larger animals are especially useful for:
- Testing medical devices, such as stents, due to their larger size.
- Testing statins, as some larger animals have similar cardiovascular physiology and function as humans.
- Testing insulin biologics, because larger animals are valuable in predicting PK/PD relationships.
- Clamp studies.
NHP models are invaluable for the final phases of preclinical efficacy studies, as they most closely mimic human T2D progression. They spontaneously develop obesity and diabetes, and have distinct pre-diabetic and diabetic phases, recapitulating the full range of diabetic complications observed in humans.
This means that the preclinical data collected in NHP studies is the most predictive available for moving agents forward into clinical trials.
Disease Progression: Induced vs Spontaneous Diabetes
When choosing how models develop diabetes, the two main options are induced disease vs spontaneously-developed disease. It’s known that spontaneous models more closely resemble human disease and complications, but supplies can be limited. This means that conventional models often have diabetes induced via diet, chemical, or surgical techniques.
Diet-induced models are somewhat artificial, but can accelerate disease progression. For example, the diet induced obesity (DIO) mouse, which relies on a high fat diet to develop T2D, is a popular conventional model of diet induced diabetes and obesity. This is despite the model not developing significant hyperglycemia, reducing its translatability.
The Sprague Dawley rat is especially sensitive to high fat diet, and develops insulin resistance and diabetes more easily than other strains. This model does have a longer onset to developing diabetes and related pathogenesis; which means longer study timelines and higher costs for researchers.
Overall, DIO models are advantageous because they are relatively easy to generate compared to surgical induction, for example, and supply is rarely an issue. DIO models are largely used for evaluating treatments to improve insulin resistance, or studying beta cell function and/or diet induced obesity.
Streptozotocin (STZ) is a common chemical used to induce a type 1 diabetes phenotype in animals. In combination with a high fat diet, STZ can also induce type 2 diabetes. Alloxan is another less frequently used method of chemical diabetes induction, which has mostly been replaced by STZ induction due to lesser efficacy and side effects.
Both methods induce T2D by degrading pancreatic beta cells. This means chemical induction is mainly useful for testing anti-diabetic drugs that don’t depend on beta cell function. STZ and alloxan induced models show hyperglycemia, reduced serum insulin levels and hyperlipidemia, but typically lack insulin resistance so they aren’t the most ideal T2D model. Chemically-induced model utility is usually limited to screening anti-hyperglycemic or insulinotropic drugs.
One other main concern with this model type is toxicity, as STZ and alloxan can both affect other organs.
Surgical induction through partial pancreatectomy was developed as an alternative method to avoid liver and kidney damage caused by chemical induction. This method is occasionally used to model T2D though only moderate hyperglycemia is observed, without frank changes to blood insulin or body weight.
Pancreatectomy is used infrequently, especially on rodents, due to the high level of surgical skill required. Surgical induction also needs administration of pain medication and can occasionally lead to side effects (e.g. infection) that could affect the study.
One advantage of partial pancreatectomy is that it reduces the time to onset of hyperglycemia compared to diet and chemical induction, which can reduce costs within large scale preclinical studies.
Spontaneous models of diabetes are more predictive of the hyperglycemia and associated diabetic complications seen in humans. Obese db/db, ob/ob, and ZDF rodents spontaneously develop T2D, though as mentioned, unlike humans they have mutations in the leptin pathway and monogenic diabetes.
So, although the spontaneous disease may be more predictive than induced diabetes in these models, translatability is still an issue.
Spontaneous obese polygenic T2D models include KK, NZO, NSY mice, and OLETF rat. The GK rat is a polygenic non-obese spontaneous T2D model that is used for studying diabetic complications, though it has very early beta cell decline.
Two improved rodent models are the FATZO mouse and ZDSD rat – both are models of obesity with intact leptin pathways that spontaneously develop T2D on a regular chow diet. NHPs also develop diabetes spontaneously.
Spontaneous T2D models are the ideal preclinical model choice for maximum translatability, though researchers often choose standard induced models due to their faster time line and robust supply. The intended application is also important to consider when selecting an induced vs spontaneous model.
Disease Inheritance: Polygenic vs Monogenic
Disease inheritance is a feature which sees many conventional rodent models diverge from human disease. Human T2D is polygenic, however, conventional rodent models are often monogenic with mutations in either the leptin receptor (db/db and ZDF) or leptin molecule (ob/ob). Since leptin does not play a dominant role in human disease these models are therefore not very translatable to human-relevant research.
New rodent models of polygenic inheritance include the FATZO mouse and ZDSD rat. Spontaneously diabetic nonhuman primates also have polygenic disease. These polygenic models of obesity and dysmetabolism develop significant hyperglycemia with an intact leptin pathway, making them a more translatable model of T2D disease inheritance.
Model advantages include mimicking human disease progression more completely, with pre-diabetic states, glucose intolerance, and weight gain on a normal diet. The ZDSD rat and NHPs also develop the same diabetic complications as humans.
These are ideal options when studying the full continuum of metabolic disturbances that accompany T2D, rather than just one aspect of disease progression and related diabetic complications.
Both conventional and unique animal models are key in developing new therapeutics for type 2 diabetes. To ensure maximum success in clinical trials, specific models should be chosen at each stage for the most appropriate applications, and that best mimic human disease progression and inheritance.