Animal and cell culture models in diabetic nephropathy
The progression of renal decline in the majority of individuals with diabetic nephropathy has stimulated tremendous interest from both the academic and industrial communities in the development of novel therapies. Central to the development of new therapies is the recapitulation of pathogenetic processes in experimental models. This section reviews some of the animal and cell culture models used for the study of diabetic nephropathy and briefly considers their advantages, limitations and pitfalls.
Animal models of diabetic nephropathy
Historically, one of the principal obstacles hampering the development of new treatments for diabetic kidney disease has been the inability of rodent models to mimic the human phenotype. Indeed, most rodents are classically considered to be nephropathy-resistant, developing a moderate increase in albuminuria and mesangial expansion but failing to progress to later-stage renal decline as seen in patients. Nevertheless, recent advances in genetic manipulation strategies have led to the development of newer models that may more closely replicate some of the features of human disease.
Mouse models of diabetic nephropathy
Induction of diabetes: An insulinopaenic state, broadly considered to mimic Type 1 diabetes, can be induced in mice either pharmacologically or genetically. Commonly, diabetes is induced by treating mice with streptozotocin (STZ), which is toxic to the insulin-producing pancreatic ß-cells. Although both “low-dose” and “high-dose” protocols have been employed, the low-dose strategy is favoured as it minimises the propensity for glucose-independent toxic effects of STZ. Genetic models of insulinopaenia circumvent the potential off-target effects of STZ and include the Akita mouse, which carries a spontaneous mutation in the insulin gene that results in insulin misfolding, and the transgenic OVE26 mouse that overexpresses the calcium-binding protein, calmodulin.
Mouse models of obese Type 2 diabetes include the db/db mouse and the ob/ob mouse. The db/db mouse carries a spontaneous mutation in the gene that encodes the leptin receptor, which plays a pivotal role in the regulation of food intake and energy homeostasis. The ob/ob mouse carries a mutation in the gene that encodes for the protein leptin itself.
The importance of background strain in mouse models of diabetic nephropathy: Background genetic strain plays a major role in the susceptibility of mice to renal injury with diabetes . For instance, kidney damage induced by diabetes in the commonly employed C57BL/6 mouse strain is moderate at best, whereas other genetic backgrounds, such as the DBA/2 strain, are more susceptible to nephropathy . db/db mice are available on either a C57BL/6 or C57BKS background, with renal injury being more prominent in the latter. In comparison to some other strains, FVB mice appear to have a predisposition to develop renal fibrosis in response to diabetes. This is the background strain that OVE26 mice were initially developed on and it may contribute to the relatively advanced nature of the renal injury that has been reported in these animals .
Assessing the renal phenotype in diabetic mice: The National Institute of Health-sponsored Diabetic Complications Consortium (DCC) nephropathy committee has established a series of criteria for evaluation of mouse models of diabetic nephropathy (http://www.diacomp.org/). These criteria are: >50% decline in glomerular filtration rate (GFR) (µl/min); >50% increase in glomerular basement membrane width (µm); >5000% increase in urine albumin (µg/ml); and the presence of arteriolar hyalinosis, mesangial sclerosis and tubulointerstitial fibrosis.
A number of methods have been used to determine GFR. One technique that is commonly employed involves the administration of fluorescein isothiocyanate (FITC)-labeled inulin with interval sampling of venous blood. The Jaffé reaction which employs the alkaline picrate method is widely used for the determination of serum or plasma creatinine in patients. However, this approach overestimates creatinine in mice and HPLC is preferred .
Urine albumin excretion is expressed as 24h urine albumin excretion and/or as the ratio to urine creatinine. To enhance standardization, the DCC recommends the use of commercial ELISA kits for determination of urinary albumin. Commercial kits based on the alkaline picrate method are often employed for determination of urine creatinine.
Morphometric assessment is an essential component of renal phenotyping. Periodic acid-Schiff (PAS) and haematoxylin and eosin (H&E) are useful for the overall evaluation of glomeruli, whereas silver methenamine is more specific for extracellular matrix. Immunohistochemical approaches and electron microscopy facilitate the visualization and stereometric assessment of specific cells within the renal glomerulus. Quantitative morphometric methods are favoured over semi-quantitative scoring systems.
Newer mouse models of diabetic nephropathy: Whereas renal injury is rarely pronounced in diabetic wildtype mice, more advanced disease has been reported in some models that combine diabetes with the genetic deletion or overexpression of genes considered to play an essential role in the pathogenesis of nephropathy .
Figure 1: Periodic acid-Schiff (PAS)-stained kidney sections from a wildtype mouse (A), a streptozotocin (STZ)-diabetic eNOS-/- mouse on a C57BL/6 background (B) and a db/db eNOS-/- mouse on a C57BKS background. Original magnification x400. The thin arrows point to the modest mesangial expansion seen in the STZ-eNOS-/- mouse and the thick arrow points to nodular glomerulosclerosis in the db/db-eNOS-/- mouse kidney. [click on figure to enlarge]Heavy albuminuria occurs in mice genetically deficient in the gene encoding endothelial nitric oxide synthase (eNOS) when diabetes is induced either with STZ or by crossing onto a db/db background. However, structural indices of nephropathy are more marked in the latter, which may be a consequence of the different background strains (C57BL/6 and C57BKS respectively)  (Figure 1).
Deletion of Bdkrb2, the gene encoding the bradykinin B2 receptor (B2R), in the Akita mouse on a C57BL/6 background results in nephropathy that appears to mimic human diabetic glomerulosclerosis . Renal injury is augmented further by the simultaneous deletion of both the bradykinin B1 and B2 receptors .
The OVE26 mouse overexpresses the calmodulin gene. By nine months of age, OVE26 mice have been reported to develop marked albuminuria, GFR decline, diffuse and nodular glomerulosclerosis and tubulointerstitial fibrosis .
Introduction of the ob/ob mutation into black and tan brachyuric (BTBR) mice that are naturally hyperinsulinaemic has resulted in accelerated nephropathy development. Proteinuria is evident in BTBR ob/ob mice as early as four weeks of age and glomerular lesions resembling human nephropathy may occur by approximately 18 weeks of age . Reversibility of the structural and functional changes of nephropathy in BTBR ob/ob mice has been observed with leptin replacement .
The table summarizes some of the more commonly studied mouse models of diabetic nephropathy together with some of their advantages and disadvantages.
|STZ-wildtype||Inexpensive, Easy to breed, Suitable for rapid throughput studies||Magnitude of renal injury is strain dependent, Mild/modest renal injury under most circumstances, Off-target effects of STZ, STZ is a biohazard|
|db/db||Widely used, Predictable elevation in albuminuria and mesangial expansion||Autosomal recessive trait, Homozygotes are infertile, Albuminuria may not be progressive|
|Akita||Autosomal dominant trait||Avoids off-target effects of STZ|Modest renal injury on a C57BL/6 background|
|STZ-eNOS-/-||Rapid onset heavy albuminuria, eNOS-/- mice are commercially available||Progressive histopathological changes may not be robust and may be dependent on the STZ protocol|
|db/db-eNOS-/-||Heavy albuminuria, GFR decline, nodular glomerulosclerosis and sensitivity to renin angiotensin system blockade||Double knockout mice are difficult to breed, Small litter size, Difficulty in generation may impede their utility for the evaluation of novel therapies|
|OVE26||Heavy albuminuria, hypertension, GFR decline, nodular glomerulosclerosis and tubulointerstitial fibrosis have been described, Available on the more sensitive FVB background||May be difficult to breed and maintain, Requirement for FVB background for phenotype impedes breeding of OVE26 mice with knockout mice on other backgrounds|
|BTBR ob/ob||Progressive nephropathy development over approximately 18 weeks, Reversible renal injury with leptin replacement||Infertile, Difficulty in breeding impedes use for interventional studies|
Rat models of diabetic nephropathy
Rats are generally more sensitive than mice to the diabetogenic effects of STZ. Several rat models of Type 2 diabetes that develop renal injury have also been described.
Zucker Diabetic Fatty (ZDF) rats carry a recessive mutation in the leptin receptor (fa/fa). Hyperglycaemia is more marked in males than females. These animals have been reported to develop progressive albuminuria together with focal and segmental glomerulosclerosis and a decline in creatinine clearance by 22 weeks of age .
Otsuka Long-Evans Tokushima Fatty (OLETF) rats and Goto Kakizaki (GK) rats are polygenic models of Type 2 diabetes. Albuminuria, glomerulosclerosis and tubulointerstitial fibrosis have been described in both OLETF and GK rats. However, the long natural history (>1 year) over which renal disease develops may practically impede the applicability of these models to the study of new therapies.
Cell culture models
Recapitulating diabetes in the in vitro setting
The most widely used method for recapitulating the pathogenetic mechanisms of diabetes in renal cells is exposure to high glucose concentrations, with control of osmolar effects achieved by comparison with cells exposed to mannitol or L-glucose. Additional pathogenetic mediators (eg angiotensin II, transforming growth factor-ß, inducers of apoptosis and hypoxia) implicated in the development of nephropathy may be employed in the culture system either alone or in combination with high glucose. To mimic the hemodynamic stresses of diabetes, specialized culture equipment is available that subjects cells to mechanical or shear forces. Finally, since cells of the renal glomerulus do not exist in isolation, co-culture systems utilizing Transwell inserts have been developed enabling the study of paracrine communication systems between glomerular cell-types, increasingly recognized for their importance in the maintenance of filtration barrier integrity in vivo .
Examples of cultured renal cells
Both primary cell cultures and immortalized cell lines are employed to explore the pathogenetic mechanisms leading to cellular injury in the diabetic kidney. Being terminally differentiated, podocytes are relatively difficult to culture as primary cells and conditionally immortalized podocyte cell lines possessing the SV40 large T antigen are more commonly studied. These cells proliferate at a temperature of 33oC and can be induced to differentiate by thermoshifting to 37oC. Glomerular endothelial cells are also difficult to isolate and maintain and an immortalized glomerular endothelial cell line has been generated  . Mesangial cells are highly proliferative and relatively easier to isolate and maintain following differential sieving of glomeruli. Similarly, proximal tubular cells may be isolated following collagenase digestion, sieving and separation on a density gradient.
A number of cell lines are commercially available. For instance, the NRK-52E cell line (NRK = normal rat kidney) resembles the proximal tubule epithelium and NRK-49F cells are a renal fibroblast cell line. LLC-PK1 cells have the properties of proximal tubular epithelial cells and Madin-Darby canine kidney (MDCK) cells are considered to mimic the epithelial cells of the distal nephron. However, being of porcine and canine origin respectively, the utility of these cells is restricted by the availability of experimental tools suitable for their study.
In summary, a variety of in vivo and in vitro model systems are available to enable the study of pathogenetic mechanisms in diabetic nephropathy and to evaluate the efficacy of novel therapies. Although each has its limitations, the collective insights garnered from their study will hopefully ultimately result in the development of more effective strategies to prevent the progressive decline of kidney function in patients.
^ Breyer MD, Bottinger E, Brosius FC, 3rd, et al. (2005) Mouse models of diabetic nephropathy. J Am Soc Nephrol 16: 27-45
^ Alpers CE, Hudkins KL (2011) Mouse models of diabetic nephropathy. Current opinion in nephrology and hypertension 20: 278-284
^ Xu J, Huang Y, Li F, Zheng S, Epstein PN (2010) FVB mouse genotype confers susceptibility to OVE26 diabetic albuminuria. Am J Physiol Renal Physiol 299: F487-494
^ Meyer MH, Meyer RA, Jr., Gray RW, Irwin RL (1985) Picric acid methods greatly overestimate serum creatinine in mice: more accurate results with high-performance liquid chromatography. Anal Biochem 144: 285-290
^ Brosius FC, 3rd, Alpers CE, Bottinger EP, et al. (2009) Mouse models of diabetic nephropathy. J Am Soc Nephrol 20: 2503-2512
^ Yuen DA, Stead BE, Zhang Y, et al. (2012) eNOS deficiency predisposes podocytes to injury in diabetes. J Am Soc Nephrol 23: 1810-1823
^ Zhang MZ, Wang S, Yang S, et al. (2011) The Role of Blood Pressure and the Renin-Angiotensin System in Development of Diabetic Nephropathy (DN) in eNOS-/- db/db Mice. Am J Physiol Renal Physiol 302: F433-438
^ Kakoki M, Takahashi N, Jennette JC, Smithies O (2004) Diabetic nephropathy is markedly enhanced in mice lacking the bradykinin B2 receptor. Proc Natl Acad Sci U S A 101: 13302-13305
^ Kakoki M, Sullivan KA, Backus C, et al. (2010) Lack of both bradykinin B1 and B2 receptors enhances nephropathy, neuropathy, and bone mineral loss in Akita diabetic mice. Proc Natl Acad Sci U S A 107: 10190-10195
^ Zheng S, Noonan WT, Metreveli NS, et al. (2004) Development of late-stage diabetic nephropathy in OVE26 diabetic mice. Diabetes 53: 3248-3257
^ Hudkins KL, Pichaiwong W, Wietecha T, et al. (2010) BTBR Ob/Ob mutant mice model progressive diabetic nephropathy. J Am Soc Nephrol 21: 1533-1542
^ Pichaiwong W, Hudkins KL, Wietecha T, et al. (2013) Reversibility of structural and functional damage in a model of advanced diabetic nephropathy. J Am Soc Nephrol 24: 1088-1102
^ Chander PN, Gealekman O, Brodsky SV, et al. (2004) Nephropathy in Zucker diabetic fat rat is associated with oxidative and nitrosative stress: prevention by chronic therapy with a peroxynitrite scavenger ebselen. J Am Soc Nephrol 15: 2391-2403
^ Siddiqi FS, Advani A (2013) Endothelial-podocyte crosstalk: the missing link between endothelial dysfunction and albuminuria in diabetes. Diabetes 62: 3647-3655
^ Satchell SC, Tasman CH, Singh A, et al. (2006) Conditionally immortalized human glomerular endothelial cells expressing fenestrations in response to VEGF. Kidney Int 69: 1633-1640