Oxidative Stress-Induced Alterations of Cellular Localization and Expression of Aquaporin 1 Lead to Defected Water Transport upon Peritoneal Fibrosis

Abstract

Being one of the renal replacement therapies, peritoneal dialysis (PD) maintains around 15% of end-stage kidney disease patients' lives; however, complications such as peritoneal fibrosis and ultrafiltration failure during long-term PD compromise its application. Previously, we established a sodium hypochlorite (NaClO)-induced peritoneal fibrosis porcine model, which helped to bridge the rodent model toward pre-clinical human peritoneal fibrosis research. In this study, the peritoneal equilibration test (PET) was established to evaluate instant functional changes in the peritoneum in the pig model. Similar to observations from long-term PD patients, increasing small solutes transport and loss of sodium sieving were observed. Mechanistic investigation from both in vivo and in vitro data suggested that disruption of cytoskeleton induced by excessive reactive oxygen species defected intracellular transport of aquaporin 1, this likely resulted in the disappearance of sodium sieving upon PET. Functional interference of aquaporin 1 on free water transport would result in PD failure in patients.

Keywords: aquaporin; cytoskeleton; oxidative stress; peritoneal dialysis; porcine.

Figure 1

Histological and functional evaluations on peritoneum in NaClO-injured pig models. (A) Experimental schedule of single NaClO injection-induced peritoneal fibrosis. (B) PET results of a single 0.1% NaClO injury model showed minor differences in D/P sodium but no difference in D/P glucose, D/P urea, and D/P creatinine after a single 0.1% NaClO injection when compared to baseline PET (in grey). (C) Experimental schedule of twice NaClO injections-induced peritoneal fibrosis. (D) PET results of twice 0.1% NaClO injury model showed significant differences in all parameters measured, D/P glucose, D/P urea, D/P creatinine, and D/P sodium values measured in NaClO injury animals were significantly different from baseline values. (E) Significant thickening (yellow double arrow) of parietal peritoneum was observed on Masson’s trichrome staining after two injections of 0.1% NaClO. Fragmentation of mesothelium (labeled with CK in green) and infiltration of myofibroblasts (labeled with α-SMA in red) was also observed. (F) Quantitative analyses showed significant thickening of the peritoneum, loss of mesothelial cells, and accumulation of myofibroblasts under both injury models by 0.1% NaClO and 0.1%*2 NaClO. Representative images from 3 individual pigs of each group were presented. Asterisks indicated significant differences between groups (* p < 0.05, ** p < 0.01, N.S. not statistical different).

Figure 2

Detection of oxidative damages on both mesothelium and vessel endothelium in NaClO-injured pigs. (A) Primary antibody of 8-OHdG was not added for negative control. (B,C) After 0.1%*2 NaClO injury, signal of 8-OHdG, a marker for DNA oxidative damage became significantly stronger and distributed across all kinds of cells on the parietal peritoneum. (B1) Enlarged image depicted a minimal expression of 8-OHdG on mesothelial cells (yellow arrowheads) in control pigs; a significant increase in 8-OHdG signals were detected on both mesothelium ((C1), yellow arrowhead) and endothelial cells ((C2), white arrowhead) that surrounded vessels. For each experimental condition, 10 images from 3 different pigs were evaluated and representative images were presented.

Figure 3

Cellular localization and protein expression of peritoneum AQP1 was altered in NaClO-injured pigs. (A) Immunofluorescent staining showed that AQP1 appeared on both mesothelial cells (A1) and endothelial cells (A2) that surrounded vessels. AQP1 was double stained with vWF to confirm the identity of an endothelial cell. In the vehicle control group, AQP1 mainly appeared on the apical and adluminal region of cells (white arrowheads); however, after 0.1%*2 NaClO injury, signals of AQP1 on both types of cells became weakened (A3) and more scattered into the cytosol (A4). (B) Quantitative analyses on whole peritoneal tissue homogenates showed that total protein expression of AQP1 on parietal peritoneum was significantly decreased (80% less) in 0.1%*2 NaClO-injured pigs. Eukaryotic elongation factor 2 (EEF2) was used as protein loading control, AQP1 signal was first corrected with loading control EEF2 before fold change comparisons to control animals. L lumen of vessels. Representative images from 3–4 individual pigs of each group were presented. For tissue sampling, 2 samples from different regions of the peritoneum were taken from every individual for further analysis. Whole membrane images of western blotting could be found in Figure S4. Asterisks indicate significant differences between groups (** p < 0.01).

Figure 4

NaClO-induced NOX4-associated oxidative damages on both mesothelial and endothelial cells. (A) Significant cellular ROS production was detected when MeT5A was treated with 0.005% and 0.01% NaClO after 30 and 60 min. (B) Immunofluorescence results showed MeT5A with increased signals of 8-OHdG and NOX4 after being treated with 0.01% NaClO. (C) Time-dependent incr in NOX4 protein expression of ease was observed after 1 h. Subsequent activation of the ERK pathway as evidenced by increased expression of phosphorylated ERK after three hours. Arrowheads indicated peak protein expression. (D) Significant cellular ROS production was detected when EA.hy926 was treated by 0.05% NaClO after 15, 30 and 60 min. (E) Immunofluorescence results showed increased 8-OHdG and NOX4 signal intensity as well as round morphology of EA.hy926 after being treated with 0.05% NaClO. (F) Similar to MeT5A, time-dependent increases in ROS-related protein expressions on EA.hy926 under 0.05% NaClO were quantified. A peak of NOX4 protein expression was observed after 1 h. Then, activation of the ERK pathway was demonstrated by increased expression of phosphorylated ERK after 3 h; however, the protein expression of phosphorylated ERK went back to normal after 24 h. Peak protein expression was labeled in arrowheads. Triplet experiments were performed for cellular ROS production assay. For each experimental condition, 10 images were evaluated, and representative images were presented. Whole membrane images of western blotting could be found in Figure S4. Asterisks indicate significant differences between groups (* p < 0.05, ** p < 0.01, N.S. not statistical different).

Figure 5

Changes of the cytoskeleton on mesothelial and endothelial cells under NaClO-induced oxidative damages. Double staining of F-actin (labeled in red) and tubulin (labeled in green) was performed on MeT5A and EA.hy926 to observe the changes in cytoskeletal under NaClO stimulation. (A) MeT5A showed obvious shrinkage and disorganized patterns of F-actin and tubulin under 0.01%NaClO after 30 and 60 min. Also, punctate accumulation of cytoskeleton (red and green arrowheads) was observed at the edges of the cells. (B) Under 0.05% NaClO, EA.hy926 showed a dispersed signal of F-actin and blurred tubulin signal after 30 and 60 min. Besides, ring-shaped membrane ruffling (green arrowheads) was observed after co-incubation of 0.05% NaClO. For each experimental condition, 10 image frames were evaluated, and representative images were presented.

Figure 6

Cellular localization and protein expression of AQP1 in MeT5A and EA.hy926 under NaClO-induced oxidative damages. (A) Total AQP1 protein expression of MeT5A showed no differences after being treated with 0.01% NaClO within 60 min. (B) Immunofluorescence staining of AQP1 in MeT5A showed an aggregated signal (yellow arrowheads) around the nucleus region under 0.01% NaClO after 30 and 60 min when compared to a dispersed membrane pattern in control cells. (C) Self-defined cellular AQP1 distribution analyses showed significant differences in AQP1 distribution after MeT5A was treated with 0.01% NaClO after 30 and 60 min. (D) Total AQP1 protein expression of EA.hy926 showed no differences after being treated with 0.05% NaClO within 60 min. (E) Immunofluorescence staining of AQP1 in EA.hy926 showed aggregated signals (yellow arrowheads) around the nucleus region under 0.05% NaClO after 30 and 60 min when compared to control cells. (F) Quantification analyses showed the more dispersed cellular distribution of AQP1 in control EA.hy926 cells, and a concentrated (aggregated) AQP1 distribution after treatment with 0.05% NaClO. For each experimental condition, except for EA.hy926 at 60 min, 50 cells were evaluated, and representative images were presented. (* p < 0.05, ** p < 0.01, **** p < 0.0001, N.S. not statistical different).

Figure 7

Summarized model for oxidative stress-induced disruption on the cellular transport of AQP1. Upon NaClO stimulations, direct and indirect production of excessive ROS activated NOX4 associated signaling pathways and promoted phosphorylation of ERK, which subsequently enforced a positive feedback loop in producing ROS. Accumulation of ROS led to oxidative stress and enhanced depolymerization of the cytoskeleton. Disruption of both actin and tubulin interfered with intracellular transport of AQP1 from the endoplasmic reticulum to the cell membrane. Failure for AQP1 to reach the cell membrane resulted in the impairment of water transport, which led to the loss of sodium sieving on PET.