Glutamatergic signaling maintains the epithelial phenotype of proximal tubular cells

Abstract

Epithelial-mesenchymal transition (EMT) contributes to the progression of renal tubulointerstitial fibrosis. The N-methyl-d-aspartate receptor (NMDAR), which is present in proximal tubular epithelium, is a glutamate receptor that acts as a calcium channel. Activation of NMDAR induces actin rearrangement in cells of the central nervous system, but whether it helps maintain the epithelial phenotype of the proximal tubule is unknown. Here, knockdown of NMDAR1 in a proximal tubule cell line (HK-2) induced changes in cell morphology, reduced E-cadherin expression, and increased α-SMA expression. Induction of EMT with TGF-β1 led to downregulation of both E-cadherin and membrane-associated β-catenin, reorganization of F-actin, expression of mesenchymal markers de novo, upregulation of Snail1, and increased cell migration; co-treatment with NMDA attenuated all of these changes. Furthermore, NMDA reduced TGF-β1-induced phosphorylation of Erk1/2 and Akt and the activation of Ras, suggesting that NMDA antagonizes TGF-β1-induced EMT by inhibiting the Ras-MEK pathway. In the unilateral ureteral obstruction model, treatment with NMDA blunted obstruction-induced upregulation of α-SMA, FSP1, and collagen I and downregulation of E-cadherin. Taken together, these results suggest that NMDAR plays a critical role in preserving the normal epithelial phenotype and modulating tubular EMT.

Figure 1.

HK-2 cells express both NR1 and N2B subunit of NMDAR. Total RNA was submitted to RT with an oligo dT reverse primer followed by PCR with different set of primers for NMDAR subunits and GAPDH as an internal control. Representative image after agarose gel electrophoresis shows differential expression of investigated receptor subunits (NMDA R1∼370 bp; R2A∼300 bp; R2B∼376 bp; R2C∼397 bp; R2D∼280 bp).

Figure 2.

NMDAR1 gene knockdown originates loss of epithelial phenotype in PTECs. Representative Western blot (A) and quantitative analysis (B) show a decrease in NMDAR1 expression in HK-2 infected with FSVsi-NMDAR1 (shNR1) compared with FSVsi (control); *P < 0.05 versus control. Downregulation of NMDAR1 in HK-2 induced changes in epithelial phenotype, evident as a decrease of E-cadherin (C and D) and an increase of α-SMA (E and F). *P < 0.05 versus control. (G and H) Representative photomicrographs show morphologic changes in the shNR1 group of cells (H) compared with control (G). Magnification: ×20.

Figure 3.

NMDAR activation restored expression of E-cadherin, α-SMA, Snail1, and pSmad2/3 altered by TGF-β1. HK-2 cells were incubated in serum-free medium (control), TGF-β1, or TGF-β1 + NMDA for 72 hours (E-cadherin, α-SMA), 6 hours (Snail1), or 120 minutes (pSmad2/3). Whole cell lysates were immunoblotted with antibodies against E-cadherin, α-SMA, and Snail1. The same samples were reprobed with tubulin to ensure equal loading. Representative Western blots and quantitative analysis show a decrease in E-cadherin (A and B) and an increase in α-SMA (A and B) and Snail 1 (C and D) expression induced by TGF-β1. Co-treatment with NMDA restored expression of these molecules close to control levels. (E and F) NMDA managed to reduce TGF-β1–induced translocation of pSmad2/3 into the nucleus after 120 minutes (nuclear extracts). Histone 1 was used as a loading control for pSmad2/3 expression. (B, D, and F) *P < 0.05 versus control; #P < 0.05 versus TGF-β1.

Figure 4.

Calcium influx through NMDAR is responsible for the attenuation of TGF-b1-induced tubular EMT. (A–F) Thapsigargin does not have the same inhibitory effect as NMDA treatment on TGF-β1–induced alterations in HK-2 cells. HK-2 cells were incubated in serum-free media (control), TGF-β1, or TGF-β1 + Thapsigargin for 72 hours (E-cadherin and α-SMA), 30 minutes (pErk), and 10 minutes (Ras-GTP). Representative Western blots (A, C, and E) and quantitative analysis (B, D, and F) show that the incubation of HK-2 with TGF-β1 and TG for indicated periods of time did not have the same effect as NMDA treatment on expression of E-cadherin and α-SMA (A and B) nor induced decreases in phosporylation of Erk (C and D). (E) Total cell extracts were prepared and incubated with GST-RBD to measure the amount of Ras-GTP (top panel). Aliquots of total cell lysates (10 μg) were run in parallel for detection of total Ras protein (bottom panel). Co-incubation of TG and TGF-β1 did not induce deactivation of Ras, pointing to a specific effect of Ca²⁺ entry through NMDAR. (B, D, and F) *P < 0.05 versus control. (G) Co-incubation of cells with MK-801 abolished the inhibitory effect of NMDA on TGF-β1–induced overexpression of Snail1. HK-2 cells were incubated for 6 hours in serum-free medium (control), TGF-β1, TGF-β1 + NMDA, TGF-β1 + NMDA + MK-801, TGF-β1 + MK-801, and MK-801 alone. Whole cell lysates were immunoblotted with antibody against Snail1. The samples were reprobed with antibody against tubulin to ensure equal loading. (H–L) NMDAR activation failed to ameliorate TGF-β1–induced alterations in HK-2 cells in the absence of Ca²⁺ in the culture media. HK-2 cells were incubated in serum free medium (control), TGF-β1, or TGF-β1 + NMDA (in serum free EpiLife [Ca²⁺-free] medium) for 72 hours (E-cadherin), 6 hours (Snail), and 30 minutes (pErk, pAkt). Representative Western blots (H and J) and quantitative analysis (I and K) show decreases in E-cadherin, increases in Snail 1 expression, and phosphorylation of Erk and Akt induced by TGF-β1. Co-treatment with NMDA in Ca²⁺-free medium does not modify TGF-β1 effects. (I and K) *P < 0.05 versus control. (L) Light microscopy shows morphologic changes caused by TGF-β1 and NMDA treatment in Ca²⁺-free medium. Magnification: ×20.

Figure 5.

NMDAR activation modulates important key steps in tubular EMT in vitro. (A) Activation of NMDAR decreases basal actin polymerization state and TGF-β1–induced actin reorganization in HK-2 cells. Cells were grown in control conditions, TGF-β1, TGF-β1 + NMDA, and NMDA for 24 hours and examined by flow cytometry using Phalloidin fluorescence detected on FL1 (as described in Concise Methods). Values are given as percentage of the control (means ± SEM) of three independent experiments assayed in triplicate for every condition. *P < 0.05 versus control; #P < 0.05 versus TGF-β1. Immunofluorescence staining for the distribution of vimentin (B) and β-catenin (D) in HK-2 cells after incubation with TGF-β1 or TGF-β1 + NMDA. TGF-β1 induced de novo expression of vimentin (72 hours) and translocation of β-catenin (24 hours) into the nucleus. Co-treatment with NMDA reduced vimentin expression and restored β-catenin to the cell periphery. (C) Light microscopy shows morphologic changes caused by TGF-β1 and NMDA treatment. Scale bar (B) 10 μm; (C) 20 μm. (D) Original magnification: ×40.

Figure 6.

NMDAR activation inhibits basal and TGF-b1-stimulated PTEC migration in vitro. (A–H) Wound healing assay. Contrast phase micrographs of HK-2 cells migrating into the denuded area of the scratch wound at various times after monolayer wounding. One representative experiment is shown to illustrate the wound closure after 48 hours in control conditions (B), TGF-β1 (D), TGF-β1 + NMDA (F), and NMDA (H) compared with the corresponding wounds at the point 0 hours for the control (A), TGF-β1 (C), TGF-β1 + NMDA (E), and NMDA (G). (A–H) Magnification: ×4. (I) Transwell migration assay. Representative photos show parts of transwell inserts for control, TGF-β1, TGF-β1 + NMDA, and NMDA treated group of cells after 24 hours of incubation. Nuclei were stained with Hoechst. Magnification: ×20. Quantification of cell migration in wound healing assay after 24 and 48 hours (J) and transwell migration assay after 24 hours (K). Data are presented as means ± SEM (wound-healing assay) or percentage of control (means values ± SEM; transwell migration assay) of three independent experiments assayed in triplicate for each time point and condition. (J) *P < 0.05 versus control at both time points and #P < 0.05 versus TGF-β1 at both time points. (K) *P < 0.05 versus control; #P < 0.05 versus TGF-β1.

Figure 7.

NMDAR activation reduced phosphorylation of Erk and Akt and activation of Ras induced by TGF-β1 treatment. HK-2 cells were incubated in serum-free medium (control), TGF-β1, or TGF-β1 + NMDA for 30 and 60 (pErk, pAkt) and 10 minutes (Ras-GTP). Representative Western blots (A, C, and E) and quantitative analysis (B, D, and F) show alterations in protein expression induced by TGF-β1 in HK-2 cells. NMDA treatment reduced phosphorylation of Akt (A and B) and Erk (C and D) induced by TGF-β1 in HK-2 cells. After incubation with different treatments, whole cell lysates were immunoblotted with either phospho-Akt or total-Akt and phospho-Erk or total-Erk. (E and F) NMDA reduced TGF-β1–induced activation of Ras. Total cell extracts were prepared and incubated with GST-RBD to measure the amount of Ras-GTP (top panel). Aliquots of total cell lysates (10 μg) were run in parallel for detection of total Ras protein (bottom panel). (B, D, and F) *P < 0.05 versus control, #P < 0.05 versus TGF-β1.

Figure 8.

NMDA treatment reduced α-SMA and collagen I expression in the obstructed mouse kidney. (A and B) Real-time PCR analysis shows downregulation of α-SMA and collagen I mRNA expression in obstructed mouse kidney after NMDA treatment in different time points (5 and 15 days). Relative mRNA levels were calculated and expressed as fold induction over contralateral controls (value = 1.0) after normalizing with GAPDH. *P < 0.05 versus control; #P < 0.05 versus UUO. (C) Quantification of collagen content after Sirius red staining. Data are means ± SEM of seven animals per group (n = 7) *P < 0.05 versus control; #P < 0.05 versus UUO. (D and E) Western blot shows increased expression of α-SMA in the obstructed kidneys at 5 and 15 days after UUO and inhibition of α-SMA in the UUO + NMDA group of mice. Whole kidney lysates were processed for protein analysis at days 5 (D and F) and 15 (E and F) after UUO and were immunoblotted with antibodies against α-SMA and tubulin, respectively. (F) *P < 0.05 versus control; #P < 0.05 versus UUO.

Figure 9.

Activation of NMDAR in the mouse kidney attenuates renal fibrosis induced by UUO; immunohistochemical staining for E-cadherin, α-SMA, FSP1, and collagen I. Administration of NMDA reduced the loss of E-cadherin (A) and decreased α-SMA (B), FSP1 (C), and collagen I (D) expression in the obstructed mouse kidney 15 days after UUO. Kidney sections were stained with antibodies against E-cadherin (A), α-SMA (B), FSP1 (C), and with Masson-Trichrome staining (D). Representative photomicrographs of kidney sections from three investigated groups of mice are presented. (A–C) Original magnification: ×20. (D) Scale bar, 20 μm.