The epithelial sodium channel γ-subunit is processed proteolytically in human kidney

Abstract

The epithelial sodium channel (ENaC) of the kidney is necessary for extracellular volume homeostasis and normal arterial BP. Activity of ENaC is enhanced by proteolytic cleavage of the γ-subunit and putative release of a 43-amino acid inhibitory tract from the γ-subunit ectodomain. We hypothesized that proteolytic processing of γENaC occurs in the human kidney under physiologic conditions and that proteinuria contributes to aberrant proteolytic activation. Here, we used monoclonal antibodies (mAbs) with specificity to the human 43-mer inhibitory tract (N and C termini, mAbinhibit, and mAb4C11) and the neoepitope generated after proteolytic cleavage at the prostasin/kallikrein cleavage site (K181-V182 and mAbprostasin) to examine human nephrectomy specimens. By immunoblotting, kidney cortex homogenate from patients treated with angiotensin II type 1 receptor antagonists (n=6) or angiotensin-converting enzyme inhibitors (n=6) exhibited no significant difference in the amount of full-length or furin-cleaved γENaC or the furin-cleaved-to-full-length ratio of γENaC compared with homogenate from patients on no medication (n=5). Patients treated with diuretics (n=4) displayed higher abundance of full-length and furin-cleaved γENaC, with no significant change in the furin-cleaved-to-full-length γENaC ratio. In patients with proteinuria (n=6), the inhibitory tract was detected only in full-length γENaC by mAbinhibit. Prostasin/kallikrein-cleaved γENaC was detected consistently only in tissue from patients with proteinuria and observed in collecting ducts. In conclusion, human kidney γENaC is subject to proteolytic cleavage, yielding fragments compatible with furin cleavage, and proteinuria is associated with cleavage at the putative prostasin/kallikrein site and removal of the inhibitory tract within γENaC.

Keywords: cell and transport physiology; epithelial sodium channel; ion channel; kidney; proteinuria.

Copyright © 2015 by the American Society of Nephrology.

Figures

Figure 1.

mAbinhibit recognizes intact, furin-cleaved and prostasin/kallikrein-cleaved γENaC by Western immunoblotting. (A) mAbinhibit was generated against an epitope within the N-terminal one half of the putative inhibitory tract in the human γENaC subunit. The proteolytic processing of γENaC theoretically leads to four fragments: intact, furin-cleaved, prostasin/kallikrein-cleaved, and furin- and prostasin/kallikrein-cleaved γENaC. The experimentally measured sizes are indicated as well as the formula molecular mass (in parenthesis). The furin- and prostasin/kallikrein-released fragment has a molecular mass of 5 kD, which is not resolvable by Western blotting in our setup. (B) Dot blot of the N- and C-terminal one half of the inhibitory peptide as well as the full-length inhibitory peptide (Intact) shows that mAbinhibit only displayed immunoreactivity to the N-terminal one half of the full-length inhibitory peptide. PBS was used as a negative control. (C) In HEK-293 cells transfected with human γENaC expression vector, a band at approximately 90 kD and a less abundant product at 75 kD were detected using mAbinhibit, whereas no bands were detected in mock-transfected HEK293 cells. In contrast to the HCP from five patients, no bands were detected in VSMCs. GADPH was used as loading control. (D) Western blotting of a serially diluted pool of five human kidney cortex samples displayed three bands with molecular mass of approximately 90, 75, and 37 kD. (E) The immunoreactivity of the three γENaC fragments was abolished by peptide preabsorption with the immunogenic peptide. (F) Fractioning of the cellular compartments of HCP samples showed that mAbinhibit reacted with an epitope present in the IC and MF. Note the presence of 75- and 37-kD fragments in the MF (open arrows). GAPDH was used as a loading control for the intracellular fraction, whereas E-cadherin was used a loading control for the membrane fraction. (G) Using urine exosomes from a healthy control person for immunoblotting, mAbinhibit yielded molecular sizes at 90, 75, and 37 kD. The intensity of the bands showed an opposite pattern of relative abundances compared with the HCP (i.e., the 90-kD band was low in abundance compared with the fragment at 37 kD that was predominant).

Figure 2.

mAbinhibit Yields specific collecting duct labeling of human kidney sections. (A and C) The fluorescence labeling signal obtained by mAbinhibit was confined to (B) AQP2-expressing cells as seen in (C) the overlay and associated primarily with cytoplasm in human kidney collecting ducts. Scale bar, 50 µm. By immunohistochemistry, mAbinhibit labeling yielded signals from collecting duct in (D) cortical and (E) medullary regions of human kidney tissue sections. Scale bar, 50 µm. (D) Note that mAbinhibit reacts with the majority of cells in the cortical collecting duct (likely principal cells) in human kidneys, whereas it spares single cells, likely intercalated cells (arrows in D). Omission of mAbinhibit and (F) preabsorption with (G) the immunogenic peptide abolished labeling. Scale bar, 200 µm.

Figure 3.

mAbprostasin Recognizes the neoepitope generated after proteolytic cleavage at the prostasin/kallikrein-site in γENaC. (A) mAbprostasin was directed against the C-terminal neoepitope exposed after proteolytic processing at the putative prostasin/kallikrein cleavage site in the human γENaC subunit. (B) The reactivity of mAbprostasin was tested by dot blot using full-length γENaC inhibitory peptide (γENaC) and peptides spanning the proposed prostasin/kallikrein cleavage site (Intact) or exposing the distal part of the γENaC subunit after proteolytic processing by prostasin/kallikrein (Cleaved). mAbprostasin only reacted with the exposed cleavage site. PBS was used as a negative control. (C) Western blot of M-1 cells showing that the approximately 45-kD band is detected with or without plasmin stimulation. (D) Western blot of a pool of five HCP samples displayed a single band with immunoreactivity to mAbprostasin with a molecular mass of approximately 45 kD. (E) In contrast to the HCP, VSMCs displayed no detectable prostasin-cleaved γENaC. (F) The immunoreactivity was neutralized by peptide preabsorption with the immunogenic peptide. (G) Western blotting after cellular fractioning of HCP showed that mAbprostasin displayed immunoreactivity to an epitope present predominantly in the MF compared with low abundance in the cytoplasmic fraction. (H) Immunohistochemical labeling using mAbprostasin was associated with the collecting duct. (I) Omission of mAbprostasin and (J) preabsorption with the immunogenic peptide abolished labeling. Scale bar, 50 μm. Immunofluorescence double labeling for the prostasin/kallikrein cleavage site with (K and M) mAbprostasin (red) and (L and M) an AQP2-specific antibody (green) showed immunofluorescence signal in the same cell and mAbprostasin-labeled apical membranes and subapical vesicular structures. Scale bar, 30 µm.

Figure 4.

Treatment with angiotensin II AT1 receptor blocker (AT1) and ACE inhibition in human patients does not significantly alter the proteolytic processing of the γENaC subunit. (A) Western blot of kidney cortex from patients receiving AT1 receptor antagonists (n=6) and control patients (n=5) displayed no significant difference between the GAPDH-normalized densities of the full-length (90-kD band) and cleaved (75-kD band) γENaC. The uncropped Western blot is shown in Supplemental Figure 2A. (B) Western blot of kidney cortex from patients receiving ACE inhibitors (n=6) compared with control patients (n=5). No significant differences were detected between the densities of the full-length (90-kD band), cleaved (75-kD band), and cleaved-to-noncleaved ratio of γENaC. The uncropped Western blot is shown in Supplemental Figure 2B. The HCP from five patients used for validation was included in each blot. (C, upper panel) Immunohistochemical staining of kidney sections from control patients yielded readily detectable labeling in the collecting ducts. (C, lower panel) Magnification of the area covered by the square in C, upper panel. The kidney section corresponds to lane 9 in A. A similar labeling was seen in kidney sections from patients treated with (D) AT1 receptor antagonist and (E) ACE inhibitors. The kidney section corresponds to (D) lane 1 in A and (E) lane 4 in B. HCP, human kidney cortex pool. Scale bar, 50 µm.

Figure 5.

Treatment with AT1 receptor antagonists and ACE inhibitors is not consistently associated with higher levels of prostasin/kallikrein-cleaved γENaC. (A and B) No bands are detected in Western blots made from kidney cortex from the control patients, whereas (A) two of six AT1 receptor antagonist-treated patients (lanes 2 and 6) and (B) three of six ACE inhibitor-treated patients (lanes 1, 2, and 4) displayed γENaC cleavage at the putative prostasin/kallikrein cleavage site. The uncropped Western blot is shown in Supplemental Figure 2, C and D. mAbprostasin labeling was not detectable in the cortical region of human kidneys from (C) control (corresponding to lane 9 in A) and (D) AT1 receptor antagonist-treated patients (corresponding to lane 1 in A), whereas (E) labeling was present in kidney sections from a patient receiving ACE inhibitors (corresponding to lane 4 in B). The areas covered by the squares in the upper panels of C–E are magnified in the lower panels of C–E. HCP, human kidney cortex pool. Scale bar, 50µm.

Figure 6.

Higher levels of γENaC in kidney tissue from patients treated with diuretics and patients with proteinuria. (A) Western immunoblotting for the inhibitory peptide with kidney homogenate from patients receiving diuretics (n=4) and patients who did not receive medication (control; n=5). Patients treated with diuretics displayed a significantly higher protein level of full-length (90-kD band) and cleaved (75-kD band) γENaC, whereas no significant difference was detected between the cleaved and noncleaved ratios. Exposure time: 30 seconds. *P<0.05; **P<0.01. GAPDH was used as an internal loading and housekeeping control. The uncropped Western blot is shown in Supplemental Figure 3A. (B) Western blotting on kidney cortex homogenate from patients with proteinuria (n=6) and control patients (n=5). Proteinuric patients displayed a significantly higher level of full-length (90-kD band) γENaC, whereas no significant difference was detected in the level of cleaved (75-kD band; faint signal) γENaC. The ratio between noncleaved and cleaved was significantly reduced in proteinuric patients compared with controls. Exposure time: 5 seconds. ***P<0.001. The uncropped Western blot is shown in Supplemental Figure 3B as well as a blot with a 1-minute exposure time. The HCP was a positive control. (C–E) In kidney sections from control, patient treated with diuretics, and patient with proteinuria, respectively, the mAbinhibit labeling was present in the collecting duct cells. The lower panels in C–E show magnifications of selected tubular segments. HCP, human kidney cortex pool. Scale bar, 50 µm.

Figure 7.

Proteinuria is associated with higher levels of prostasin/kallikrein-cleaved γENaC and positive immunolabeling. (A and B) Result of immunoblotting experiments with kidney cortex homogenate from (A) patients treated with diuretics, (B) patients who displayed proteinuria, and (A and B) patients who received no medication (control). (A) No bands were detected in Western blots made from control patients, whereas (B) one of four patients treated with diuretics (lane 3) and (B) five of six proteinuric patients (lanes 2–6) displayed immunoreactivity with the prostasin cleavage site-specific antibody. (B) Note that the mAbprostasin band is present as a double band in homogenate from proteinuric patients. The uncropped Western blot is shown in Supplemental Figure 3, C and D. (C) Immunohistochemical staining detected a few cells with immunoreactive protein in kidney sections from controls (corresponding to lane 7 in A). (D) In a patient displaying proteinuria, immunohistochemical staining showed labeling in the collecting duct cells (corresponding to lane 6 in B). HCP, human kidney cortex pool. Scale bar, 50 µm.