Soluble HB-EGF induces epithelial-to-mesenchymal transition in inner medullary collecting duct cells by upregulating Snail-2

Abstract

Animal models of acute renal injury suggest that the epidermal growth factor receptor (EGFR) axis may have a beneficial role in the recovery from acute renal injury, but recent reports describe detrimental effects of EGFR activation in chronic renal injury. Expression of the EGFR ligand heparin-binding EGF-like growth factor (HB-EGF) increases following renal injury, but the effects of this sustained upregulation have not been well studied. Here, stable overexpression of soluble HB-EGF (sHB-EGF) in mouse inner medullary collecting duct (IMCD) cells led to marked phenotypic changes: sHB-EGF-expressing cells demonstrated a fibroblast-like morphology, did not form epithelial sheets, exhibited cytoplasmic projections, decreased expression of epithelial markers, and increased expression of fibroblast-specific protein-1. They also demonstrated anchorage-independent growth and formed tumors when injected subcutaneously into nude mice. Quantitative RT-PCR and a luciferase reporter assay suggested that sHB-EGF repressed transcription of E-cadherin, and a concomitant TGF-beta-independent upregulation of the E-cadherin repressor Snail-2 was observed. Stable downregulation of Snail-2 in sHB-EGF-overexpressing cells restored epithelial characteristics (E-cadherin and cytokeratin expression) but did not alter their anchorage-independent growth. In summary, sustained exposure to sHB-EGF induces epithelial-to-mesenchymal transition of IMCD cells, in part by upregulating the E-cadherin transcriptional repressor Snail-2.

Fig. 1.

A: immunoblot of conditioned media from wild-type inner medullary collecting duct (IMCD) cells, vector-transfected IMCD cells (vec), and soluble heparin-binding EGF-like growth factor (sHB-EGF)-overexpressing IMCD cells (stable clones 11 and 12). B: treatment of IMCD cells with HB-EGF (100 ng/ml) increases EGF receptor (EGFR) phosphorylation and decreases total EGFR expression. A similar pattern is observed in sHB-EGF-overexpressing IMCD cells. Total ERK, which is unaffected by EGFR activation, served as a loading control. C: morphology of vector-transfected and sHB-EGF-overexpressing IMCD cells. D: representative immunoblots of cell lysates; β-actin served as a loading control.

Fig. 2.

Cell proliferation (A) and cell migration (B) were increased in sHB-EGF-overexpressing IMCD cells compared with vector-transfected control cells. Treatment with the EGFR tyrosine kinase inhibitor PD153035 (500 nM) for 6 h before and during [³H]thymidine incorporation significantly inhibited cell proliferation. *P ≤ 0.008.

Fig. 3.

A: photomicrographs of colonies of sHB-EGF-overexpressing IMCD cells and vector-transfected control cells in soft agar 3 wk after plating (magnification ×100 for both). B: gross appearance of subcutaneous tumors 3 wk after injection of the designated number of sHB-EGF-overexpressing IMCD cells (1.25 × 10⁵ - 1 × 10⁶ cells). C: subcutaneous tumors formed at every injection site of sHB-EGF-overexpressing IMCD cells (16 sites in 4 mice). No tumors formed from injections of vector-transfected IMCD cells. The average tumor volume increased with dose of cells injected (P < 0.0001 by ANOVA).

Fig. 4.

E-cadherin mRNA was downregulated by sHB-EGF overexpression in IMCD cells compared with vector-transfected cells, as measured by quantitative RT-PCR (A; *P < 0.0001) and a luciferase reporter assay (B; *P = 0.006).

Fig. 5.

A: Snail-2, but not Snail-1, mRNA was significantly upregulated in sHB-EGF-overexpressing IMCD cells compared with vector-transfected control cells as determined by quantitative RT-PCR. B: representative ethidium bromide-stained agarose gels of semiquantitative RT-PCR products derived from vector-transfected and sHB-EGF-overexpressing IMCD cells. C: semiquantitative RT-PCR suggested that, similar to IMCD cells, Snail-2 is upregulated in sHB-EGF-overexpressing MDCK II cells compared with vector-transfected cells.

Fig. 6.

A: immunoblots of cell lysates for P-Smad2/3. Lysates were prepared after incubating near-confluent cells in serum-free conditions overnight. For the indicated treatments, exogenous HB-EGF (100 ng/ml) or TGF-β (10 ng/ml) was added 1 h before lysis on ice. When the ALK5 inhibitor SB431542 (SB; 10 μM) was used, it was added 15 min before either the addition of TGF-β (in the case of IMCD+SB+TGF-β cells) or lysis (in the case of IMCD^sHB+SB cells). Constitutive expression or exogenous addition of HB-EGF did not lead to phosphorylation of Smad2/3. Exogenous treatment with TGF-β served as a positive control. B: cells were grown in the presence or absence of TGF-β (10 ng/ml) and/or SB431542 (10 μM) for 5 days. RNA was collected, and quantitative RT-PCR was used to determine the relative expressions of Snail-1 and Snail-2 compared with wild-type IMCD cells. P < 0.0001 (Snail-1) and P = 0.0011 (Snail-2) by ANOVA; *P < 0.01 compared with IMCD by Dunnett's multiple comparison test.

Fig. 7.

Quantitative RT-PCR demonstrated near-complete knockdown of Snail-2 mRNA in sHB-EGF-overexpressing IMCD cells stably transfected with shRNA constructs 601 and 871. *P = 0.005 (vs. IMCD^sHB-scram, Snail-1). **P = 0.001 (vs. IMCD^sHB-scram, Snail-2).

Fig. 8.

Immunofluorescence images demonstrating significant restoration of E-cadherin and cytokeratin with knockdown of Snail-2 in sHB-EGF-overexpressing IMCD cells stably transfected with Snail-2 shRNA (IMCD^{sHB-601 and -871}). Fsp-1 expression, although reduced compared with IMCD^sHB cells, was still detectable in stable shRNA transfectants. Fsp-1 was not detected in vector-transfected IMCD cells that do not overexpress sHB-EGF (IMCD^vec).

Fig. 9.

Immunoblots of cell lysates from vector-transfected IMCD cells that do not overexpress sHB-EGF (IMCD^vec), sHB-EGF-overexpressing IMCD cells (IMCD^sHB), and sHB-EGF-overexpressing IMCD cells stably transfected with Snail-2 shRNA (IMCD^{sHB-601 and -871}). β-Actin served as a loading control.