There is increasing evidence that proteins in tubular fluid are “nephrotoxic.” In vivo it is difficult to study protein loading of tubular epithelial cells in isolation, i.e., without concomitant glomerular damage or changes of renal hemodynamics, etc. Recently, a unique amphibian model has been described which takes advantage of the special anatomy of the amphibian kidney in which a subset of nephrons drains the peritoneal cavity (open nephrons) so that intraperitoneal injection of protein selectively causes protein storage in and peritubular fibrosis around open but not around closed tubules. There is an ongoing debate as to what degree albumin per se is nephrotoxic and whether modification of albumin alters its nephrotoxicity. We tested the hypothesis that carbamylation and glycation render albumin more nephrotoxic compared with native albumin and alternative albumin modifications, e.g., lipid oxidation and lipid depletion. Preparations of native and modified albumin were injected into the axolotl peritoneum. The kidneys were retrieved after 10 days and studied by light microscopy as well as by immunohistochemistry [transforming growth factor (TGF)-β, PDGF, NF-κB, collagen I and IV, RAGE], nonradioactive in situ hybridization, and Western blotting. Two investigators unaware of the animal groups evaluated and scored renal histology. Compared with unmodified albumin, glycated and carbamylated albumin caused more pronounced protein storage. After no more than 10 days, selective peritubular fibrosis was seen around nephrons draining the peritoneal cavity (open nephrons), but not around closed nephrons. Additionally, more intense expression of RAGE, NF-κB, as well as PDGF, TGF-β, EGF, ET-1, and others was noted by histochemistry and confirmed by RT-PCR for fibronectin and TGF-β as well as nonradioactive in situ hybridization for TGF-β and fibronectin. The data indicate that carbamylation and glycation increase the capacity of albumin to cause tubular cell damage and peritubular fibrosis.
