Diabetic nephropathy (DN) contributes to nearly 50% of chronic kidney disease and end-stage renal disease cases in the United States (1). Despite clinical trials that have affirmed the importance of glycemic control and the provision of inhibitors of the renin-angiotensin-aldosterone system for slowing progression of DN, the burden of disease continues to increase (2). Therefore further mechanistic studies are needed to understand the pathogenesis of DN.
One of the earliest histologic changes in glomeruli from patients with DN is mesangial extracellular matrix (ECM) deposition (3). Ultrastructural studies also demonstrate increased glomerular basement membrane (GBM) thickness and podocyte foot process effacement (3). These lesions are associated with the development of albuminuria, but three fundamental questions remain: 1) Although transforming growth factor (TGF)-β signaling is an important mediator of these early lesions, why does treatment with an anti-TGF-β neutralizing antibody not reduce albuminuria (4)? 2) What is the contribution of each glomerular cell type to these lesions and to albuminuria? 3) What is the role of albuminuria in mediating further glomerular and tubular damage in DN? The article by Fan et al. (5) in this issue of Diabetes represents an important step forward in addressing these gaps in our knowledge of DN.
How TGF-β induces albuminuria has been debated for more than a decade. In a now classic article, Ziyadeh et al. (4) demonstrated that reducing TGF-β bioavailability by administration of a neutralizing anti-TGF-β antibody in db/db mice decreased mesangial ECM expansion and progressive renal disease but did not reduce albuminuria. TGF-β signaling can be separated into the canonical pathway, mediated through Smad2 and Smad3, and the alternative pathway through Smad1 and Smad5 (6). Genetic deletion of Smad3 in mice reduces ECM deposition and GBM thickening but does not reduce albuminuria (7). By contrast, Chen et al. (8) demonstrated that gene delivery of Smad7, an inhibitor of both canonical and alternative TGF-β signaling (6,9), significantly reduces ECM deposition, GBM thickness, and albuminuria, suggesting TGF-β could contribute to albuminuria in DN. However, none of these articles specifically investigated the activation of the TGF-β alternative pathway.
Fan et al. (5) studied the role of TGF-β alternative signaling in DN. Using knockout (KO) mice, they deleted BAMBI (BMP, activin, membrane-bound inhibitor), an endogenous antagonist of the TGF-β alternative pathway (10). When they induced diabetes in these BAMBI KO mice, activation of the alternative pathway caused podocyte foot process effacement and albuminuria but not ECM deposition or increased GBM thickness, suggesting TGF-β canonical and alternative pathways promote different components of the pathogenesis of DN (Fig. 1). A second strategy to specifically inhibit the TGF-β alternative signaling pathway is necessary to validate these results, for example, deletion of the Smad1 or Smad5 gene in diabetic mice. These results shed light on the debate of how TGF-β activation induces albuminuria and suggest several possibilities for why anti-TGF-β therapy did not reduce albuminuria. Perhaps antibody therapy preferentially inhibited the canonical pathway, or the alternative pathway is activated by decreased BAMBI expression. Moreover, since decreased BAMBI expression was observed in kidneys from both humans and mice with DN, activation of the alternative pathway may represent a modifier in the presentation of chronic kidney disease in DN with or without albuminuria (12). TGF-β also contributes to leukocyte kidney accumulation and to the epithelial-to-mesenchymal transition in DN (9,13), but which signaling pathways are responsible is still unknown. Studying these end points in Smad3 KO and BAMBI KO mice will answer these questions.
Many factors causing albuminuria in DN have been identified (14), but the contribution of each glomerular cell type (endothelial cells, mesangial cells, and podocytes) is unknown. Sison et al. (15) demonstrated that podocyte-secreted vascular endothelial growth factor (VEGF) maintains a normal glomerular filtration barrier by paracrine signaling through its receptor VEGF receptor 2 (VEGFR2) on glomerular cell types other than podocytes, but whether this mechanism contributes to DN is unknown. Guillot et al. (16) previously demonstrated endothelial injury in BAMBI KO mice. In their article, Fan et al. (5) demonstrated that glomerular VEGFR2 is expressed only in endothelial cells, and activation of the TGF-β alternative signaling pathway decreases endothelial expression of VEGFR2. These results provide an intriguing hypothesis that TGF-β alternative signaling might contribute to podocytopathy and albuminuria through primary damage of the glomerular endothelial cell. Endothelial cell injury closely correlates with albuminuria in patients with DN (17). Further mechanistic studies of the contribution of TGF-β–dependent endothelial VEGF signaling will help to elucidate the role of TGF-β in albuminuria and the deleterious role of albuminuria in patients with DN.
C57BL/6 mice, a commonly used mouse strain, develop only a very mild form of DN (18) because of unknown mechanisms. To overcome this problem, several researchers have induced diabetes on a modified C57BL/6 background, for example, endothelial nitric oxide synthase KO mice (19). Fan et al. (5) provide a valuable new option, the BAMBI KO mice, in which the TGF-β alternative signaling pathway and albuminuria are induced. Moreover, because these mice do not develop significant ECM deposition or GBM thickening, they may serve as a valuable resource to study the specific role of albuminuria in mediating kidney injury in diabetes (20).
Although many pathological end points are characterized in DN studies, the links among those end points are weak. The work by Fan et al. (5) provides a candidate link, that is, the TGF-β alternative signaling pathway, to connect endothelial cell and podocyte damage with canonical pathway–stimulated mesangial cell injury. These findings will inspire further investigation to better understand the distinct consequences of injury to different glomerular cell types and the appropriate pathways to target for novel therapies.
See accompanying article, p. 2220.
Acknowledgments. The authors thank Dr. Glenn Chertow (Stanford University) for scientific discussion and critical review of the manuscript.
Funding. X.Z. received support from the Larry L. Hillblom Foundation Postdoctoral Fellowship (2014-D-021-FEL), and V.B. received support from the National Institutes of Health (Diabetes Complications Consortium Pilot & Feasibility Award, National Institute of Diabetes and Digestive and Kidney Diseases grant U24-DK-076169-0853, subaward 25732-15).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.