ADAM10-Mediated Ectodomain Shedding Is an Essential Driver of Podocyte Damage

Abstract

Background: Podocytes embrace the glomerular capillaries with foot processes, which are interconnected by a specialized adherens junction to ultimately form the filtration barrier. Altered adhesion and loss are common features of podocyte injury, which could be mediated by shedding of cell-adhesion molecules through the regulated activity of cell surface-expressed proteases. A Disintegrin and Metalloproteinase 10 (ADAM10) is such a protease known to mediate ectodomain shedding of adhesion molecules, among others. Here we evaluate the involvement of ADAM10 in the process of antibody-induced podocyte injury.

Methods: Membrane proteomics, immunoblotting, high-resolution microscopy, and immunogold electron microscopy were used to analyze human and murine podocyte ADAM10 expression in health and kidney injury. The functionality of ADAM10 ectodomain shedding for podocyte development and injury was analyzed, in vitro and in vivo, in the anti-podocyte nephritis (APN) model in podocyte-specific, ADAM10-deficient mice.

Results: ADAM10 is selectively localized at foot processes of murine podocytes and its expression is dispensable for podocyte development. Podocyte ADAM10 expression is induced in the setting of antibody-mediated injury in humans and mice. Podocyte ADAM10 deficiency attenuates the clinical course of APN and preserves the morphologic integrity of podocytes, despite subepithelial immune-deposit formation. Functionally, ADAM10-related ectodomain shedding results in cleavage of the cell-adhesion proteins N- and P-cadherin, thus decreasing their injury-related surface levels. This favors podocyte loss and the activation of downstream signaling events through the Wnt signaling pathway in an ADAM10-dependent manner.

Conclusions: ADAM10-mediated ectodomain shedding of injury-related cadherins drives podocyte injury.

Keywords: glomerular disease; membranous nephropathy; podocyte; proteinuria.

Graphical abstract

Figure 1.

ADAM10 is expressed at podocyte FPs in mice. (A) Proteome analyses of ADAM10 and ADAM17 abundance in FACS-sorted glomerular podocytes and mesangial and endothelial cells. Plotted is the log2 of label-free quantification (LFQ) as percentage difference (n=30). (B) Representative confocal images of ADAM10 expression (green) using a rabbit polyclonal antibody in C57BL/6 glomeruli in relation to the SD protein nephrin (red). DNA is stained in blue (Hoechst). Red arrows point toward apical ADAM10 expression in tubuli; white arrow toward podocyte ADAM10 expression. Colocalization of ADAM10 and nephrin is seen as yellow overlay. (C) Immunogold EM for ADAM10 of glomerular freeze fractures, using a rat mAb to ADAM10, shows specific localization of gold particles at FPs (arrows). GBM, glomerular basement membrane; p, podocyte.

Figure 2.

Podocyte-specific deletion of ADAM10. (A) Breeding scheme used to generate podocyte-specific, ADAM10-deficient mice, termed ADAM10Δpod. (B) PCR analysis on isolated glomeruli demonstrates the successful deletion of ADAM10 exon 2 (ΔADAM10) upon Cre expression. (C) Immunoblot from isolated glomeruli reveals reduced ADAM10 abundance in glomeruli of ADAM10Δpod. Total kidney lysate was used as positive control. Graph exhibits densitometric quantification of n≥4 mice per genotype. *P≤0.05. (D) Representative confocal image proves specific reduction of ADAM10 (green) in podocytes (p) of naive ADAM10Δpod mice using a rat mAb to ADAM10. +, wild type; ctrl, control; Ex1, exon 1; F, flox; pADAM10, immature ADAM10 proform; mADAM10, mature ADAM10; tg, transgene.

Figure 3.

Podocyte development is unaffected by ADAM10 deficiency. Adult ADAM10Δpod and control littermates were compared. (A) Periodic acid–Schiff (PAS) staining of kidney sections exhibiting inconspicuous morphology. (B) Representative high-resolution confocal images of the distribution of nephrin (white) at the SD exhibits a normal, tightly meandering white line in both genotypes. DNA (blue) is visualized by Hoechst staining. (C) EM analysis demonstrates normal podocyte (p) and GFB ultrastructure. Scale bars, 1 µm. (D) Assessment of proteinuria (albumin-creatinine ratio) in urine shows no interference with podocyte function. Pooled data from two independent experiments, n≥7 per genotype. c, capillary; ec, endothelial cell; u, urinary space.

Figure 4.

ADAM10 is upregulated in human and murine podocytes upon antibody-mediated injury. (A) Gene expression analysis of ADAM10 mRNA in the glomerular compartment of manually microdissected kidney biopsy specimens from patients with different kidney diseases. Values are expressed as log2 fold change compared with controls (living donors; LD). A q value <5% was considered to be statistically significant. (DN, n=14; hypertensive nephropathy [HT], n=15; MCD, n=14; IgA nephropathy [IgA], n=27; FSGS, n=23; MN [MGN], n=21; lupus nephritis [SLE], n=32; ANCA-associated GN [RPGN], n=23; LD, n=42). (B) Representative high-resolution confocal images of ADAM10 (green) expression in human biopsy specimens from patients with nephrotic syndrome due to MCD, MN (PLA₂R1 positive), and FSGS. Note the enhanced podocyte (p) expression in the MN biopsy specimen. Green arrows point toward ADAM10 localization at the podocyte side of the GFB; white arrows toward ADAM10 localization at the endothelial side of the GFB; red arrows toward tubular ADAM10 expression; nephrin shown in red; DNA shown in blue (Hoechst). (C) APN was induced in control littermates and ADAM10Δpod mice. ADAM10 (green) localization was analyzed on day 14 after disease induction. High-resolution confocal images exhibit increased signal for ADAM10 at the podocyte (green arrows) and endothelial cell side (white arrows) of the GFB in APN-treated in comparison with naive control mice. ADAM10Δpod mice show no signal for ADAM10 expression at the GFB, albeit a preserved tubular ADAM10 expression (red arrows). (D) Immunoblot for ADAM10 levels in isolated glomeruli from naive mice and APN-treated mice on day 14. Graph exhibits densitometric quantification of n≥6 mice per genotype, pooled data from two independent experiments. *P≤0.05. Abs, antibodies; mADAM10, mature ADAM10.

Figure 5.

ADAM10 deficiency protects from anti-podocyte antibody–induced podocyte injury. (A) Scheme of the APN model: podocytes are isolated from mouse kidneys (1) to immunize sheep (2). Sheep-generated, anti-podocyte antibodies are isolated and purified (3). APN antibodies intravenously injected into mice bind to podocyte target proteins and induce their injury (4). APN was induced in ADAM10Δpod and control littermates. Kidneys were analyzed on day 14 after disease induction. (B) Measurement of BUN levels (n≥15 per group), *P≤0.05. (C) Determination of proteinuria (albumin-creatinine ratio) in urine on day 14 reveals decreased development of proteinuria in ADAM10Δpod mice (n≥20 per group); pooled data from four independent experiments, ***P≤0.001. (D) Representative light micrographs of Periodic acid–Schiff (PAS) staining exhibiting the tubulointerstitial and glomerular morphology. Note the comparable occurrence of PAS-positive protein casts in the tubular lumina (*) of ADAM10Δpod and control littermate mice, contrasting the preserved glomerular morphology in ADAM10Δpod mice. The visible brown, round structures represent the perfused magnetic beads used for glomerular isolation. (E) Confocal analysis of glomerular sheep IgG (green) deposition demonstrates comparable deposition in both genotypes. (F) High-resolution confocal images of the distribution of nephrin (white) at the SD exhibits a strong broadening of the normal, tightly meandering pattern in APN-treated control littermates; whereas ADAM10Δpod showed a mostly preserved pattern. DNA (blue) is visualized by Hoechst staining. (G) Ultrastructural analysis of podocytes demonstrates FPE in control mice upon APN-antibody treatment, whereas ADAM10-deficient podocytes are mostly preserved morphologically. Scale bars, 500 nm. c, capillary lumen; u, urinary side.

Figure 6.

ADAM10-mediated N-cadherin shedding is essential for APN antibody–induced podocyte injury. (A) Quantification of podocyte number per glomerular tuft area in naive and APN-treated ADAM10Δpod and control littermates. Graph exhibits violin plot of pooled data from four independent experiments (n≥12 per group). (B) Adhesion assay demonstrates APN antibody–induced murine podocyte detachment from the culture dish, which could be suppressed by an ADAM10 inhibitor. The ADAM10 activator ionomycin served as the positive control. One representative out of three individual experiments is shown, each with six replicates per condition. (C) APN antibody (ab) reduces N-cadherin cell-surface levels from murine podocytes, which was rescued by administration of ADAM10 inhibitor GI254023X. The cell-surface protein transferrin receptor and soluble glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as control for loading and purity of cell-surface extracts. One representative out of three experiments is shown. (D) N-cadherin shedding is increased in response to APN antibody in cultured murine podocytes. ADAM10 inhibition with GI254023X suppressed generation of N-cadherin CTF (n=4). In parallel, podocytes were treated with 1 µM DAPT to stabilize the generated N-cadherin CTF. Splice variant (SV) based on the murine protein information at uniprot.org/uniprot/P15116#ptm_processing and uniprot.org/uniprot/D3YYT0#sequences. *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001.

Figure 7.

ADAM10 deficiency results in enhanced cadherin levels on podocytes in APN antibody–induced podocyte injury. APN was induced in ADAM10Δpod and control (ctrl) littermate mice, and kidneys were analyzed on day 14. (A) Immunoblot for the cell-cell adhesion molecule N-cadherin (N-cad) levels in isolated glomeruli. Graph exhibits densitometric quantification of n=8 mice per genotype; pooled data from two independent experiments. *P≤0.05. The levels of α-actinin 4 within respective glomerular lysates suggest comparable amounts of podocytes. Splice variant (SV) based on the murine protein information at uniprot.org/uniprot/P15116#ptm_processing and uniprot.org/uniprot/D3YYT0#sequences. (B) Kidneys of mice with comparable mild proteinuria were stained for N-cadherin (green) in relation to the SD protein nephrin (red), and DNA (Hoechst, blue). Note the enhanced expression of N-cadherin in ADAM10Δpod podocytes (p) in comparison to control podocytes. White arrows point toward N-cadherin in presumably primary podocyte processes. (C) Representative confocal images demonstrate enhanced P-cadherin (green) in ADAM10Δpod podocytes in comparison to control podocytes; lectin wheat germ agglutinin (red) visualizes the GFB; DNA (Draq5) shown in blue. Care was taken to use N- and P-cadherin antibodies (Abs) raised in rabbit, so that secondary antibodies needed for visualization would not crossreact with the injected sheep IgG and the intrinsic mouse IgG. (D) Representative confocal images demonstrate podocyte N-cadherin localization in patients with MN and in anti-GBM GN in relation to nephrin (red) and DNA (Hoechst; blue). White arrows represent N-cadherin expression at the podocyte aspect of the GFB, especially in areas where nephrin is absent.

Figure 8.

β-Catenin/Wnt signaling pathway is not activated in ADAM10Δpod mice exposed to APN antibodies. APN was induced in ADAM10Δpod and control littermate mice, and kidneys were analyzed on day 14. (A) Immunoblot for β-catenin levels in isolated glomeruli. Graph exhibits densitometric quantification of n=7 mice per genotype; pooled data from two independent experiments. *P≤0.05. (B) Kidneys of mice with comparable, mild proteinuria were stained for β-catenin (green) in relation to the SD protein nephrin (red), and DNA (Hoechst; blue). Note the enhanced β-catenin signal in nuclei (white arrows) and cytoplasm of podocytes (p) and in endothelial cells (red arrow). (C) qRT-PCR analyses of Wnt response genes in isolated glomeruli. Values are depicted as relative expression to naive control (ctrl) littermates after normalization to 18S as housekeeper gene; n=3–14 mice per genotype; pooled data from two independent experiments. *P≤0.05 to control naive, ^§ P≤0.05 and ^§§ P≤0.01 to control and APN antibodies (Abs). (D) Scheme depicting proposed role of ADAM10 in podocytes in antibody-mediated podocyte injury. (1) Following a prominent embryonic expression of cell-cell adhesion molecules, such as P- and N-cadherin, in podocyte progenitors, healthy mature podocytes have low levels of cadherins at processes. (2) In the course of antibody-mediated injury, ADAM10 expression is enhanced. ADAM10 activity results in ectodomain shedding of cadherins. Cadherin cleavage results in a dissociation of β-catenin from the cytoplasmic cadherin domain and to a redistribution of β-catenin to the cytoplasm and nuclear compartment, where it initiates the transcription of Wnt response genes. Ultimately, this results in decreased cell-cell adhesion and proteinuria. (3) As a consequence of loss of ADAM10 activity, cadherin levels are stabilized at podocyte processes, after antibody-mediated injury. β-Catenin remains tethered to the cytoplasmic cadherin domain, and the Wnt/β-catenin signaling pathway is not activated. Podocyte adhesion is ameliorated, resulting in attenuated proteinuria. ADAM10Δpod, podocyte-specific ADAM10 deletion.