Lipotoxic disruption of NHE1 interaction with PI(4,5)P2 expedites proximal tubule apoptosis
Shenaz Khan, Bassam G. Abu Jawdeh, Monu Goel, William P. Schilling, Mark D. Parker, Michelle A. Puchowicz, Satya P. Yadav, Raymond C. Harris, Ashraf El-Meanawy, Malcolm Hoshi, Krekwit Shinlapawittayatorn, Isabelle Deschênes, Eckhard Ficker, Jeffrey R. Schelling
Shenaz Khan, Bassam G. Abu Jawdeh, Monu Goel, William P. Schilling, Mark D. Parker, Michelle A. Puchowicz, Satya P. Yadav, Raymond C. Harris, Ashraf El-Meanawy, Malcolm Hoshi, Krekwit Shinlapawittayatorn, Isabelle Deschênes, Eckhard Ficker, Jeffrey R. Schelling
View: Text | PDF
Research Article Nephrology

Lipotoxic disruption of NHE1 interaction with PI(4,5)P2 expedites proximal tubule apoptosis

Abstract

Chronic kidney disease progression can be predicted based on the degree of tubular atrophy, which is the result of proximal tubule apoptosis. The Na+/H+ exchanger NHE1 regulates proximal tubule cell survival through interaction with phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2], but pathophysiologic triggers for NHE1 inactivation are unknown. Because glomerular injury permits proximal tubule luminal exposure and reabsorption of fatty acid/albumin complexes, we hypothesized that accumulation of amphipathic, long-chain acyl-CoA (LC-CoA) metabolites stimulates lipoapoptosis by competing with the structurally similar PI(4,5)P2 for NHE1 binding. Kidneys from mouse models of progressive, albuminuric kidney disease exhibited increased fatty acids, LC-CoAs, and caspase-2–dependent proximal tubule lipoapoptosis. LC-CoAs and the cytosolic domain of NHE1 directly interacted, with an affinity comparable to that of the PI(4,5)P2-NHE1 interaction, and competing LC-CoAs disrupted binding of the NHE1 cytosolic tail to PI(4,5)P2. Inhibition of LC-CoA catabolism reduced NHE1 activity and enhanced apoptosis, whereas inhibition of proximal tubule LC-CoA generation preserved NHE1 activity and protected against apoptosis. Our data indicate that albuminuria/lipiduria enhances lipotoxin delivery to the proximal tubule and accumulation of LC-CoAs contributes to tubular atrophy by severing the NHE1-PI(4,5)P2 interaction, thereby lowering the apoptotic threshold. Furthermore, these data suggest that NHE1 functions as a metabolic sensor for lipotoxicity.

Authors

Shenaz Khan, Bassam G. Abu Jawdeh, Monu Goel, William P. Schilling, Mark D. Parker, Michelle A. Puchowicz, Satya P. Yadav, Raymond C. Harris, Ashraf El-Meanawy, Malcolm Hoshi, Krekwit Shinlapawittayatorn, Isabelle Deschênes, Eckhard Ficker, Jeffrey R. Schelling

×

Figure 1

The eNOS–/– db/db mouse is an authentic model for DN.

Options: View larger image (or click on image) Download as PowerPoint
The eNOS–/– db/db mouse is an authentic model for DN. (A–F) Whole kidn...
(A–F) Whole kidneys from 26-week-old mice were fixed in 4% paraformaldehyde and labeled with (A and D) Masson’s trichrome stain or Oil red O and (B, C, E, and F) hematoxylin counterstain. Original magnification, ×20 (A, B, D, and E); ×40 (C and F). (G) Histomorphometric quantitation of tubular atrophy and interstitial fibrosis, as described in Methods (n = 5 kidneys per group). IS, interstitium. *P < 0.01 compared to eNOS+/– db/m group by t test. (H) Proximal tubule apoptosis was determined by counting TUNEL-positive cells from Tapinauchenius purpureus–counterstained kidney frozen sections (n = 3 per group). *P < 0.05 compared to eNOS+/– db/m group. **P < 0.05 compared to all other groups. (I) Kidney frozen sections were labeled with rat anti-active (cleaved) caspase-2 IgG and T. purpureus proximal tubule counterstain and quantitated (n = 3 per group). *P < 0.05 compared to eNOS+/– db/m group. (J) Frozen sections were labeled with Ki-67 and T. purpureus for proliferating proximal tubule cells and quantitated (n = 5 per group). (K) Cortical tubule suspensions from wild-type, db/db, eNOS–/–, and eNOS–/– db/db mice were assayed for NHE1 activity by determining maximum rate of change in cytosolic pH following NH4Cl washout protocol and ratiometric BCECF fluorescence by spectrofluorimetry (n = 3 per group). *P < 0.05 compared to wild-type group by ANOVA.