3.1.2. Models of diabetic nephropathy and research outputs
For many years, mouse models have been the primary and most extensively studied species to investigate kidney pathogenesis and for the development of novel therapies to limit diabetic nephropathy. As previously mentioned, the diabetogenic ß-islet cell toxin, STZ, is also by far the most widely used drug to induce diabetic nephropathy. However, development of overt diabetic nephropathy is strain dependent. C57BL/6J and BALB/c strains of mice are highly responsive to developing STZ-mediated diabetes although C57BL/6J mice are relatively resistant to developing diabetic nephropathy[107]. Hence, C57Bl/6J mice are commonly crossbred with genetic mouse models known to have impaired renal function. One approach is to breed eNOS-/- mice onto the C57BL/6 background. Deletion of eNOS diminishes NO availability, thereby phenocopying the well-documented decreased NO levels and its effects on the vasculature in human diabetic nephropathy. STZ-induced diabetic eNOS-/- mice on the C57BL/6 background develop significant impairment of renal function in which clear pathophysiological changes are visible after 24 weeks of diabetes[108]. On its own, a reduction in eNOS functionality is sufficient to develop vascular dysfunction, hypertension and albuminuria, all of which accelerate renal injury, however, when superimposed on a diabetic background the severity of the albuminuria appears to be enhanced, and more advanced glomerular lesions and glomerular endothelial injury are exhibited in the STZ-induced diabetic eNOS-/- mouse model. However, high dose administration of STZ should be avoided in this model since this can be highly lethal, although the extent of albuminuria and renal injury observed is milder than that of the eNOS-/- db/db (Type2 diabetes) mouse model[108].
Genetic mouse models, such as the ob/ob and db/db mice, have been widely studied as models of diabetic nephropathy, although the degree of severity differs between these two models. In the db/db mice, larger glomeruli with increased mesangial matrix are visible by 5-6month of age, becoming more pronounced over time, along with thickening of the glomerular basement membrane. The extent of renal disease is far less in ob/ob mice when compared to db/db mice. When on the C57Bl/6 background, ob/ob mice are insulin resistant but do not develop severe diabetes. The major renal pathology in ob/ob mice primarily includes diffuse and nodular lipo-hyaline changes in the glomerulus with a relative paucity in mesangial expansion[109]. Both ob/ob and db/db mice have been crossbreed with numerous genetically modified mice. Initially established by Zhao et al. the eNOS-/-;db/db mouse model has been identified as the most robust model of Type2 diabetes[110]. These mice display many human pathophysiological features of diabetic nephropathy including significant albuminuria (more than 30-fold that of controls), reduction in glomerular filtration rate and glomerular basement membrane thickening[110].
The importance of performing animal studies in both male and female mice, as opposed to studies in one gender only (mostly male), is becoming increasingly apparent to improve the translation of novel therapies to both males and females. Recent findings showed that there were no significant differences in renal structural and functional damage between male and female diabetic eNOS-/-db/db mice [111]. Both male and female eNOS-/-db/db mice showed hyperglycaemia and obesity as early as 8 weeks of age. Although male diabetic mice had a tendency towards greater body weights than female counterparts, the difference was not significant. Both male and female diabetic mice showed progressive albuminuria, renal hypertrophy, and significantly elevated serum creatinine, although this was slightly greater in female diabetic mice[111]. Notably, both male and female mice displayed significantly greater glomerular injury compared to their wild-type counterparts, but no differences were noted between genders. Hence, the eNOS-/-db/db mouse model represents a robust model of diabetic nephropathy enabling the study of drug treatment in both male and female mice without gender impacting experimental outcomes. However, it should not be assumed that drug treatments will always necessitate the same outcomes, and for this reason both genders should be evaluated. However, historically male mice strains have been used to demonstrate more robust hyperglycaemia and renal injury.
Initially characterized by Clee et al., the black and tan, brachyuric (BTBR) mouse crossed onto the ob/ob background is a widely used mouse model to investigate Type2 diabetic nephropathy[112]. The BTBR mouse strain has alleles promoting insulin resistance resulting in natural hyperinsulinemia. When crossed with the ob/ob mouse, it develops a severe Type2 diabetic phenotype, maintaining sustained hyperglycaemia[112] and closely recapitulating the kidney lesions observed in human diabetic nephropathy. In particular, BTBR ob/ob mice exhibit glomerular hypertrophy, accumulation of mesangial matrix, loss of podocytes and glomerular lesions that are similar to that seen in human T2D kidney disease[113]. BTBR ob/ob mice display overactivation NF-κß and JAK/STAT pathways, thereby promoting inflammation and the infiltration of lymphocytes and macrophages compared to BTBR wild type mice. Indeed, use of the BTBR ob/ob mouse model has demonstrated the therapeutic potential of targeting the transcription factor NF-κß in the prevention of diabetic nephropathy[114].
An additional, albeit invasive approach to develop diabetic nephropathy, is through the use of unilateral nephrectomy (UNx) that can significantly shorten the time required to develop the pathophysiological signs of diabetic nephropathy. This technique involves removing the left kidney after ligation of the renal artery, vein and ureter. Initially performed on STZ-induced diabetic Sprague-Dawley rats, it was later applied to mouse models to develop enhanced renal injury. Together with STZ, it has a synergistic effect by accelerating the development of glomerular lesions and impaired renal function. It has been shown that the combination of STZ and UNx is an effective way to induce diabetic nephropathy consistent with Type1 diabetes, especially in the resistant C57BL/6J mouse strain[115]. This model includes albuminuria, thickening of the glomerular basement membrane, mesangial matrix expansion, glomerular hypertrophy and upregulation of pro-fibrotic markers[115]. However, the use of a western-diet to induce Type2 diabetes was not effective in inducing significant albuminuria and glomerular injury when combined with STZ and UNx, despite the presence of hyperglycaemia. A possible explanation for this protective effect may have been driven by metabolic changes in the renal cells of the remaining kidney after UNx, where the hyperfunctioning nephron is able to avoid hypoxia-induced injury. Moreover, 15-weeks of a western diet may not have been sufficient to induce renal injury due to the high cholesterol, potentially providing protective effectives in the shorter-term (<4 months), therefore longer HFD-feeding may be required (~9months)[115].
More recently, the incorporation of proteomics and/or metabolomics studies is providing novel insights into the molecular pathways of diabetic nephropathy, as well as identifying mouse models that more closely phenocopy human diabetic kidney disease. Metabolomics analysis predominantly involves nuclear magnetic resonance spectroscopy or chromatography-mass spectrometry, enabling either the detection of known metabolites or numerous metabolites respectively, the latter technique without any prior knowledge of the metabolites required[116]. Using a proteomics and metabolomics approach, Rossi et al showed that a tissue inhibitor of metalloproteinases (TIMP3) knockout mouse model alone or in combination with STZ-mediated diabetes, led to significant alterations in kidney lipid metabolism and peripheral acylcartinine levels, which have been implicated in Type2 diabetes[117]. Loss of TIMP3 in the kidney was associated with features of diabetic nephropathy including higher levels of medium and long chain acylcarnitine in the blood. In contrast, levels of free carnitine and short chain beta-hydroxy and dicarboxy-acylcarnitines were lower diabetic in TIMP3 knockout mice as compared to WT-STZ mice[117]. Similar metabolomics signatures were found in a study by Lim et al who showed the accumulation of a number of short acylcarnitines in the plasma of patients who developed diabetic kidney disease over time, as well as data by Gurley et al who found elevated levels of short acylcarnitines in the Akita mouse[118], indicating that this model might closely resemble the human metabolic profile. Furthermore, the incorporation of flow cytometry allows the identification of immune cell-specific population in diabetic nephropathy. Indeed, a significant increase of intrarenal CD3+ T cells that was associated with augmented production of interferon-γ and TNF-α in diabetic mice was revealed via flow cytometry[119]. In addition, transcriptomic profiling using scRNA-seq is identifying factors underlying the pathophysiology of diabetic nephropathy and is therefore a useful tool for identifying new therapeutic avenues for research. Using scRNA-seq, Fu et al demonstrated that the number of glomerular cell-specific markers that were significantly higher in diabetic mice compared with controls. These immune cells were predominantly macrophages and greater number expressed M1 phenotype than M2. Moreover, regulation of angiogenesis and migration pathways in endothelial cell from diabetic mice were altered[120]. Hence, these studies, in addition to adding valuable insights into the diseased kidney, also highlight the potential to generate a mouse model more prone to diabetic nephropathy.
Diabetic retinopathy