Abstract
Diabetes is a chronic metabolic disorder associated with the accelerated development of macrovascular (atherosclerosis, coronary artery disease) and microvascular complications (nephropathy, retinopathy and neuropathy), which remain the principal cause of mortality and morbidity in this population. Current understanding of cellular and molecular pathways of diabetes-driven vascular complications as well as therapeutic interventions have arisen from studying disease pathogenesis in animal models. Diabetes-associated vascular complications are multi-faceted, involving the interaction between various cellular and molecular pathways. Thus, the choice of an appropriate animal model to study vascular pathogenesis is important in our quest to identify innovative and mechanism-based targeted therapies to reduce the burden of diabetic complications. Herein, we provide up-to-date information on available mouse models of both Type 1 and Type 2 diabetic vascular complications as well as experimental analysis and research outputs.
Introduction
Diabetes Mellitus (DM), either Type1 (insulin-deficient) or Type2 (adult onset, insulin resistance), is a chronic metabolic disorder that has serious implications on cardiovascular health. Indeed, macro- and microvascular complications associated with diabetes account for approximately 65% of deaths in the diabetic population[1, 2]. Macrovascular complications associated with diabetes are mainly represented by atherosclerosis and/or hypertension and their sequelae, which include heart attacks and stroke, whilst microvascular complications include chronic kidney disease, retinopathy and neuropathy. The increased risk of cardiovascular complications in diabetes is attributed to the elevated blood glucose levels and the interplay between different cellular mechanisms in the disease setting as a consequence of hyperglycemia[3]. These include, but are not limited to, genetic, epigenetic, and cellular signalling pathways, particularly the interplay between oxidative stress and inflammation, which then result in imbalances in cellular metabolism contributing to the development of cardiovascular complications.
Current standard of care for diabetic patients focusses on glucose-lowering, lipid lowering and blood pressure control, however these treatment options have yielded suboptimal outcomes on cardiovascular end points, with the development of tolerance and resistance to these therapies, particularly with respect to blood pressure lowing medications limiting their long-term use[4, 5]. Newer, more innovative CV-specific drug treatments are urgently needed to prevent diabetes-mediated vascular diseases. In this regard, animal models have proven to be an invaluable tool in exploring and characterising novel pathophysiological pathways, identifying therapeutic targets and evaluating their in vivo potential. However, it is critical to choose appropriate animal models that most closely mimic the human pathophysiology of diabetes and its sequalae such as the cardiovascular complications, to investigate potential treatments and therapeutic strategies. To date, there is no perfect animal model that encompasses all the pathophysiology and characteristics of human disease but with advances in genetic engineering and chemical manipulation, rodents have proven to be the most reproducible, effective and preferred animal model for studies of diabetic vascular diseases, particularly with respect to emulating specific conditions that occur in humans. In this review, we provide a comprehensive and up-to-date analysis on proven animal models of diabetes-associated vascular diseases, with a focus on the analysis and experimental outputs that can be achieved with these models.
Animal models of diabetes-associated macrovascular diseases
Vascular pathophysiology: key features that need to be represented in animal models
Atherosclerosis is the principal cause of death and disability in the diabetic population, with the incidence of an atherosclerotic event occurring approximately 14.6 years earlier in this patient cohort. The pathophysiological events that lead to disease progression in the presence of diabetes have been extensively reviewed[1, 6, 7], however, key factors that are important to consider in choosing a relevant animal model of diabetes–associated atherosclerosis will be discussed. A main clinical feature that presents in uncontrolled or poorly controlled Type1 and Type2 diabetic patients is hyperglycemia, the key driver of vascular injury. Elevated blood glucose, primarily via advanced glycation end products (AGEs) and their interaction with receptor for advanced glycated end products (RAGE) drives pro-inflammatory/pro-oxidant pathways to elicit molecular, cellular and vascular injury[8, 9]. A further key feature of diabetes is dyslipidemia manifested through elevated triglycerides, decreased high density lipoproteins (HDL) and changes in low density lipoprotein (LDL) structure to a more atherogenic profile[10], which results from the ineffective clearance of post-prandial lipoprotein rather than increased circulating levels of fasting lipoproteins[11]. In addition, the heightened state of oxidative stress and inflammation perturbs cellular metabolism particularly within endothelial cells which leads to endothelial dysfunction. Together with a build-up of oxidised LDL particles, endothelial dysfunction promotes increased adhesion and infiltration of leukocytes into the vessel wall and the formation of foam cells, thereby leading to the acceleration of atherosclerotic processes[1].
Diabetes-associated atherosclerosis is most commonly studied in mice despite the fact that humans and mice differ in certain characteristics that affect the development of disease progression. Firstly, mice are generally protected from atherogenesis due to differences in their lipoprotein profile, with mice exhibiting elevated HDL levels due to the absence of cholesteryl ester transfer protein (CETP)[12, 13]. Therefore, atherosclerotic plaque generation requires genetic manipulation such as the deletion of the apolipoprotein E (ApoE) gene or the LDL receptor (LDLR) gene together with dietary interventions. Additionally, humans and mice have differences in cardiovascular anatomy and haemodynamic physiological forces that influence the sites of atherosclerosis[14]. Lesions predominantly occur in the coronary arteries, carotid and peripheral vessels in humans whilst in mice, lesions are commonly studied in the aortic sinus and the aortic arch[14]. A summary of the models of diabetes-associated atherosclerosis is shown in Figure 1.
Rodent models of diabetes-associated atherosclerosis2.2.1. Models of Type1 diabetes-associated atherosclerosis
To replicate human disease, most mouse models of Type1 diabetes are insulin deficient, either by chemical induction or genetic manipulation. The most extensively studied Type1 model of diabetes-associated atherosclerosis is the streptozotocin (STZ) induction model in the ApoE knockout (-/-) mouse, initially characterised by Park et al[15]. A major advantage of the STZ-induced ApoE-/- mouse model is the vast amount of published literature available utilising this model allowing researchers to compare and contrast data, cellular mechanisms and therapeutic interventions. STZ, a pancreatic β-islet cell toxin, enters the cell via Glut-2 transporters to mediate its cytotoxic effect, thereby limiting insulin production, leading to hyperglycaemia[16]. Generally, STZ is administered as five consecutive low doses (55mg/kg/day)[17-19] but certain groups have demonstrated more consistent elevations in blood glucose by using two consecutive high doses (100mg/kg/day)[20-24]. This becomes particularly important for longer term studies where blood glucose levels may decline due to recovery of β-islet cells. Despite the benefits of a robust and consistent increase in hyperglycemia of around 25-30mmol/L, there are limitations that may preclude the use of STZ under certain circumstances. For example, STZ at high doses has been shown to affect the function of other cells expressing the Glut-2 transporter, such as hepatocytes and epithelial cells of the renal tubules, which may be exposed to the toxic action of STZ leading to unwanted side effects in the liver and kidneys[25]. In addition, a further limitation of using STZ as the chemical inducer of diabetes is that females, particularly mice, have been shown to be resistant to STZ, often requiring higher doses of STZ to obtain equivalent levels of hyperglycemia compared to male counterparts[26]. Thus, female mice are commonly omitting in pre-clinical research for diabetic complications when STZ is the preferred method of Type1 diabetes induction. STZ also exhibits broad spectrum antibacterial properties which may alter the gut microbiota[27].
Despite the noted limitations, the STZ-induced diabetic ApoE-/- mouse model is the preferred model for studying atherosclerosis, since these mice exhibit hyperglycemia as well as dyslipidemia, as indicated by the elevated blood glucose levels, total cholesterol, LDL and triglyceride levels. Plaque formation is increased 3-6 fold in this model over a 10-20 week period, with early discrete lesions observed within the aortic sinus and the aortic arch; however, over a longer course of diabetes, lesions additionally develop over the thoracic and abdominal aortic regions. This model has been associated with endothelial dysfunction, increased inflammation and oxidative stress as well as macrophage foam cell formation, all of which are hallmarks of atherosclerotic pathogenesis. More importantly, the STZ-induced ApoE-/- mouse model has been particularly useful in investigating potential therapies, such as but not limited to soluble RAGE (sRAGE), peroxisome proliferator-activator gamma agonists, anti-oxidants (ebselen, dimethyl furate), SGLT2 inhibitors and ApoA1 mimetics, in their ability to attenuate diabetes-associated atherosclerosis, independent of blood glucose lowering effects [21, 22, 28-32]. As the ApoE-/- mouse model is well established and easy to breed, cross-breeding ApoE-/-mice with mice harbouring genetically altered proteins such as transcription factors, receptors and/or enzymes enables the elucidation of their role in diabetic atherosclerotic mechanisms. On the contrary, STZ-induced diabetic LDLR-/- mice do not exhibit extensive atherosclerotic lesions despite having hyperglycemia and elevated LDL/vLDL levels and often require additional diet supplementation[33] making them a more ideal model of Type2 diabetes.
A less well-known model that is used in Type1 diabetic atherosclerotic research, particularly if the side-effects of STZ are of concern (as detailed above), is the spontaneous genetic Akita mouse, cross bred with either the ApoE-/- or LDLR-/- mouse. The Akita mouse, first selected in Japan, carries an autosomal dominant missense mutation in the insulin2 gene, resulting in impaired proinsulin processing and aggregation of misfolded insulin protein, rendering it insulin deficient. Akita mice develop typical Type1 diabetes characteristics, including hyperglycemia, polyuria and polydipsia, as well as impaired vascular tone as early as 3-4 weeks of age, making it an ideal model to study the effects of diabetes on endothelial dysfunction and macrovascular complications[34]. When cross-bred to the ApoE-/- or LDLR-/- mouse, this model exhibits dyslipidemia, mainly manisfested as increases in triglycleride and non-HDL cholesterol levels, and an approximately 2 to 3-fold increase in atherosclerotic lesions[35-37]. This is accompanied by increased infiltration of immune cells into the plaque, in particular macrophages and T-cells, as well as augmentation of vascular inflammation as demonstrated by increased levels of TNF-α, IL-1β and MCP-1, all of which are key mediators involved in plaque progression[36, 37]. Interestingly, the differences in plaque size are not as pronounced in female Akita/LDLR-/- mice as they are in male mice, suggesting gender differences could play a part in diabetic atherosclerosis in this model.
The models described thus far are promising to investigate the mechanisms of endothelial dysfunction, vascular inflammation and plaque formation in a diabetic setting, however, it is difficult to discern whether these arise as a direct effect of the hyperglycemia or as a consequence of other confounding atherogenic factors such as the dyslipidemia under these conditions. To circumvent this problem, Vikramadithyan et al have developed a mouse model to study the effects of hyperglycemia in accelerating atherogenesis by manipulating Aldose reductase, an enzyme involved in catalysing the reduction of glucose to sorbitol, in combination with STZ[38]. Aldose reductase has been shown to accelerate diabetic vascular complications through the increased production of reactive oxygen species (ROS). Indeed, overexpression of human aldose reductase in the diabetic LDLR-/- mouse resulted in a 2-fold increase in lesion size as compared to diabetic LDLR-/- mice alone, driven by defects in glucose-driven ROS signalling[38]. Another model developed to closely mimic the autoimmune destruction of pancreatic β cells as observed in human Type1 diabetes is the induction of lymphocytic choriomeningitis virus (LCMV) infection [39, 40]. The LCMV glycoprotein (GP) transgene is controlled by the insulin promoter in β cells, which is rapidly destroyed upon infection. These LDLR-/- mouse injected with the LCMV GP developed severe hyperglycemia that was associated with accelerated lesion formation in the absence of hypertriglyceridemia[39, 40]. Moreover, atherosclerotic lesions were morphologically similar to human lesions particularly with increased arterial macrophage accumulation and intralesional haemorrhage[39]. Nonetheless, further investigation is required to delineate the exact role of altered lipoprotein profiles and hyperglycemia in preclinical animal models of Type1 diabetes-induced atherosclerosis.