Renal Physiology

Fulvene-5 inhibition of Nadph oxidases attenuates activation of epithelial sodium channels in A6 distal nephron cells

David Trac, Bingchen Liu, Alan C. Pao, Sheela V. Thomas, Michael Park, Charles A. Downs, He-Ping Ma, My N. Helms


Nadph oxidase 4 is an important cellular source of reactive oxygen species (ROS) generation in the kidney. Novel antioxidant drugs, such as Nox4 inhibitor compounds, are being developed. There is, however, very little experimental evidence for the biological role and regulation of Nadph oxidase isoforms in the kidney. Herein, we show that Fulvene-5 is an effective inhibitor of Nox-generated ROS and report the role of Nox isoforms in activating epithelial sodium channels (ENaC) in A6 distal nephron cells via oxidant signaling and cell stretch activation. Using single-channel patch-clamp analysis, we report that Fulvene-5 blocked the increase in ENaC activity that is typically observed with H2O2 treatment of A6 cells: average ENaC NPo values decreased from a baseline level of 1.04 ± 0.18 (means ± SE) to 0.25 ± 0.08 following Fulvene-5 treatment. H2O2 treatment failed to increase ENaC activity in the presence of Fulvene-5. Moreover, Fulvene-5 treatment of A6 cells blocked the osmotic cell stretch response of A6 cells, indicating that stretch activation of Nox-derived ROS plays an important role in ENaC regulation. Together, these findings indicate that Fulvene-5, and perhaps other classes of antioxidant inhibitors, may represent a novel class of compounds useful for the treatment of pathological disorders stemming from inappropriate ion channel activity, such as hypertension.

  • oxidative stress
  • (H2)DCF-DA
  • cell-attached patch clamp
  • kidney slices
  • aldosterone

reversible oxidation of proteins, lipids, and DNA is an important mechanism for signal transduction and cell control. Because cellular responses to oxidants and antioxidants remain unclear, redox signaling and regulation require further investigation. It is apparent that when the level of oxidants exceeds the buffering capacity of antioxidants, cell damage and tissue injury ensue, and antioxidant therapy does not completely protect or reverse oxidative damage. As such, it is important to study cellular sources of reactive oxygen species (ROS) release.

Typically, cellular sources of ROS come from functional uncoupling of the respiratory chain caused by inactivation of complex 1, or from stimulation of enzymatic sources such as xanthine oxidase, cyclooxygenase, nitric oxide synthase, and nonphagocytic Nadph oxidases (Noxes) (57, 75). Because enzymatic production of ROS via Nadph oxidases can be a highly regulated process, with regulatory factors that remain to be identified, we investigated the mode of Nadph oxidase activation in distal nephron cells of the kidney.

There are seven isoforms of Nadph oxidases, which all produce ROS by transferring electrons across the membrane from Nadph to molecular oxygen. Nadph oxidase 4 was originally cloned and is robustly expressed in the kidney (22). Nox4 is unique in that p22phox appears to be the only subunit that is essential for enzyme activity (5), whereas Nox1–3 isoforms require small G protein activation of p47phox, p67phox, and p40phox assembly. Functionally, Nox4 has been linked to senescence (22), proinflammatory responses (60, 66), migration (30, 48, 62), proliferation (49, 71), differentiation (14), cell death/apoptosis (51), and oxygen sensing (41). However, the physiological role of Nox4 in the kidney and the regulatory factors that regulate Nox4 in a normal healthy kidney remain unknown (47, 68).

The causal connection between oxidative stress and hypertensive disorders has been established in various experimental models (1, 3, 29, 40, 46, 78). Indeed, we (25, 44, 73) and others (37, 64, 69) showed that ROS activates epithelial sodium channels (ENaC). Hence, investigating the cellular source of ROS generation in the kidney may be key to understanding (and preventing) hypertensive diseases linked to ENaC dysregulation. Herein, we report that specific and pan inhibitors of Nox enzymes attenuate ROS production and decrease transepithelial sodium current across A6 distal nephron cells. Moreover, single-channel analysis indicates that Nox enzyme activity plays an important role in cell stretch and ROS-mediated activation of ENaC in A6 cells.


Chemicals and reagents.

Unless otherwise stated, all chemicals and reagents were purchased from Sigma.

Nadph oxidase inhibitors.

Fulvene-5, a gift from Dr. Arbison (Emory University), was used to inhibit Nox4 and Nox2 with specificity (6), whereas diphenyleneiodonium (DPI) was used as a general inhibitor of Nox and other flavin-containing enzymes (42). The synthesis of fulvene compound is referenced in Ref. 12.

Tissue culture.

A6–2F3 cells, a distal nephron cell line abundantly expressing ENaC (originally obtained from Dr. Bernard C. Rossier, University of Lausanne, Switzerland), were maintained in plastic tissue culture flasks, as described previously (83). For single-channel patch-clamp experiments, A6 cells were seeded onto permeable polyester inserts coated with crude rat tail collagen at a density that reached maximum confluency within 7–10 days. A6 cell complete culture medium was replaced 3 times/wk and consisted of 50/50 DMEM/Ham's F-12 base media (GIBCO, Grand Island, NY), 5% fetal bovine serum (GIBCO), 1.5 μM aldosterone, 1.0% streptomycin, and 0.6% penicillin (Irvine Scientific, Santa Ana, CA) at pH 7.4. Before experimentation, cells were washed 3× with 1× PBS and fed deprived media devoid of steroid hormones and fetal bovine serum for 72 h. All experiments were carried out using A6 cells between passages 98 and 106.

Western blotting.

Polarized A6 cells were lysed in RIPA buffer containing (in mM) 150 NaCl, 10 NaPO4, 0.15 SDS, 1% Nonidet P-40, and 0.25% Na deoxycholate. Approximately 30 μg protein were electrophoresed on a 10% polyacrylamide gel under reducing conditions. Protein was then transferred to Protran nitrocellulose membrane (Schleicher and Schuell BioScience) for Western blot analysis. Nitrocellulose was blocked in TBST buffer (10 mM Tris, pH 7.5, 70 mM NaCl, and 0.1% Tween) with 5% dry milk before incubation with primary antibodies, as indicated. Antibodies for Nox2 catalytic and regulatory subunits as well as antibodies for p40phox, p47phox, and p67phox were purchased from Millipore Upstate; anti-Nox4 antibody was purchased from Novus Biologicals; and anti-Rac1 antibody and anti-p22phox antibody were obtained from Santa Cruz Biotechnology. All membranes were incubated with respective antibody for 1 h at room temperature (RT) and washed extensively in TBST buffer before incubation with IgG alkaline phosphatase (AP)-labeled secondary antibody (KPL) at a 1:10,000-fold dilution in TBST for 1 h at RT. AP luminescent signal was detected using Nitro-Block chemiluminescence enhancer and CDP-Star substrate (Tropix) and Carestream Image Station Gel Logic 4000 Pro (Carestream Health) and compatible imaging software.


For coimmunoprecipitation (co-IP) studies, 20 μg of rabbit anti-Rac1 antibody (Santa Cruz Biotechnology) or rabbit anti-p67phox antibody (Milllipore) were incubated in A6 cell lysate for 1 h at 4°C. Protein A-Sepharose beads (Sigma) were used to pull down the primary antibody-bound protein complexes. Following IP, Western blot analysis was performed as indicated above.

Immunohistochemistry and confocal imaging.

Two-hundred-micrometer mouse kidney slices were prepared using a Vibratome Series 1000 tissue slicer. Unfixed tissue slices, and A6 cells seeded on coverslip slides, were incubated in anti-Nox4 antibody diluted 1:1,000, followed by incubation in secondary IgG conjugated to Alexa 568 (Invitrogen) diluted 1:50,000 in PBS containing 1% BSA and 1× sodium azide. Dolichos biflorus agglutinin (DBA) lectin (Vector Labs) and anti-aquaporin 2 (AQP2) antibody were used to label collecting duct cells in kidney slices. After being labeled, kidney slices and cell monolayers were fixed in 4% paraformaldehyde and mounted in VECTASHIELD HardSet mounting medium. AQP2 and Nox 4 dual labeling in tissue slices were performed in paraffin-embedded tissue slices using standard immunohistochemical protocol.

Confocal imaging was conducted using an Olympus FV1000 confocal laser-scanning microscope and associated software. Colocalization analysis was performed using the intensity correlation analysis plugin available in the Wright Cell Imaging Facility ImageJ collection.

ROS measurements.

Two separate yet complementary approaches were used to quantify ROS levels. 2′,7′-Dichlorodihydrofluorescein (DCFH2) was used as a fluorescent indicator of cellular ROS production in A6 cells grown to confluence on Transwell-permeable supports (Corning, Acton, MA) (9). Following 1-h treatment with 5 μM Fulvene-5, 5 μM DPI, 50 μg/ml catalase, or 50 μg/ml SOD treatment, confluent A6 cells were incubated with 5 μM DCFH2 for 5 min. Images were captured with a single 485-nm excitation event using an Olympus FV1000 upright confocal microscope; the emission filter was 520 nm. Dihydroethidium (DHE; Invitrogen) was used as a fluorescent indicator of cellular ROS production in A6 cells grown to confluence on poly-l-lysine-coated glass coverslips. DHE oxidizes in the presence of ROS. We (25, 31) and others (19) previously reported that fluorescence detection of oxidized DHE is a reliable method for quantification of intracellular ROS levels. Confluent A6 cell monolayers were incubated with 10 μM DHE in saline solution containing (in mM) 96 NaCl, 3.4 KCl, 0.8 MgCl2, 0.8 CaCl2, and 10 HEPES (∼300 mosmol/l) at 37°C for 30 min protected from light. Osmotic stretch was induced by concomitant incubation with 10 μM DHE in dilute saline solution (DSS) containing (in mM) 48 NaCl, 1.7 KCl, 0.4 MgCl2, 0.4 CaCl2, and 5 HEPES (∼150 mosmol/l). Images were captured using an Olympus FV1000 inverted confocal microscope (518/605-nm excitation/emission wavelengths). To verify osmotic cell stretch, subconfluent A6 cells were incubated for 15 min in DSS and plasma membranes were stained with CellMask Deep Red Plasma Membrane Stain (649/666-nm excitation/emission wavelengths; Invitrogen). Changes in cell size were visualized with confocal microscopy.

Transepithelial current measurements.

A6 cells were grown to confluence on Transwell-permeable supports in complete A6 cell culture medium. At confluency, cells were washed 3× using 1× PBS and then cultured in media lacking fetal bovine serum and steroid hormone for 72 h before experimentation. Transepithelial resistance values were measured before experimentation to verify cell viability and integrity of monolayer. One hour following treatment (2 μl DMSO vehicle control, 1 mM H2O2, 5 μM Fulvene-5, or 5 μM DPI in the presence or absence of 1 μM aldosterone) the potential difference (PD) and transepithelial resistance (RTE) across cell monolayers were measured using an epithelial voltohmeter equipped with stick electrodes (World Precision Instruments, Sarasota, FL). The equivalent short-circuit current (Isc) was calculated according to Ohm's law (Isc = PD/RTE) and then corrected for the surface area of the Transwell insert.

Single-channel recordings.

The cell-attached configuration was used in all patch-clamp studies. Micropipettes were pulled from filamented borosilicate glass capillaries (TW-150F, World Precision Instruments) with a two-stage vertical puller (Narishige, Tokyo, Japan). The resistances of the pipettes were between 7 and 10 MΩ when filled with and immersed in patch solution containing (in mM) 96 NaCl, 3.4 KCl, 0.8 MgCl2, 0.8 CaCl2, and 10 HEPES, with pH adjusted to 7.4 by NaOH.

Channel currents were recorded at 1 kHz with an Axopatch 1-D amplifier (Molecular Devices) with a low-pass, 100-Hz, eight-pole Bessel filter. To measure the number of channels and the open probability (NPo), we used pCLAMP 9 software (Molecular Devices). The channel NPo can be calculated from the single-channel record without any assumptions about the total number of channels in a patch or the Po of a single channel using the following relationship: NPo = ∑ i·ti/T, where T is the total recording time, i is the number of channels open, and ti is the time during the recording when there were i channels open.

Statistical evaluation.

Data are reported as means ± SE. Statistical analysis was performed with SigmaPlot and SigmaStat software (Jandel Scientific) and SAS 9.3 (SAS Institute). Differences between groups were evaluated with one-way ANOVA with Scheffe's post hoc procedure. Repeated paired t-tests were performed on patch-clamp data. Results were considered significant if P < 0.05.


Noxes are robustly expressed in A6 distal nephron cells.

Using Western blot analysis, we show that Nox4 and Nox2 are abundantly expressed in A6 distal nephron cells (Fig. 1, A and B). The Nox2 blot shows two bands of molecular weights near 55–65 and 80 kDa, which is consistent with the variability in molecular size of glycosylated protein, as previously reported in Ref. 58. Figure 1C shows that Nox4 is robustly expressed in confluent A6 distal cells, in which the single-channel patch-clamp recordings were conducted. Importantly, we show that Nox4 protein is expressed in rat cortical collecting duct cells (Fig. 1, D-F). First, in Fig. 1D, the colocalization of DBA lectin signal with AQP2 immunofluorescence signal shows that DBA lectin can be used as a marker for collecting duct cells. Colocalization is confirmed with intensity correlation analysis. Indeed, many others have used DBA lectin as a marker for collecting duct cells (70, 81). Then, in Fig. 1E we used intensity correlation analysis to show that Nox4-immunolabeled signal colocalizes with DBA lectin signal. Figure 1F shows colabeling of AQP2 and Nox 4 antibodies (left) and confirms the findings presented with DBA lectin as a marker for collecting duct cells. Together, these findings show that Nox4 expression is maintained in cultured A6 distal nephron cells and is indeed expressed in mammalian cortical collecting duct cells.

Fig. 1.

Nox4 and Nox2 expression levels in A6 distal nephron, rat kidney, and mouse collecting duct cells. Western blot analysis showing Nox4 (A) and Nox2 (B) expression in A6 distal nephron (A6) and rat kidney (RK) homogenate. Both blots were stripped and reprobed for β-actin protein expression levels. C, right: immunofluorescent detection of Nox4 protein in A6 monolayer using Novus anti-Nox4 antibody and secondary Alexa Fluor 488-nm (green) conjugated antibody. Left: corresponding white light image. D: mouse kidney cortex tissue slices were colabeled with anti-aquaporin 2 (AQP2; 568-nm red signal), DBA lectin (488-nm green signal), and DAPI (blue signal). DBA lectin signal colocalizes with AQP2 immunofluorescent signal on cortical collecting duct cell membranes (as indicated by yellow signal and intensity correlation analysis, respectively, in the bottom). E: mouse kidney tissue slices were labeled with DBA lectin (488-nm green signal) and Nox4 antibody with subsequent detection with Alexa 568 (red signal). Regions of Nox4 and DBA lectin colocalization in the kidney are indicated by yellow overlap signal and intensity correlation analysis, respectively, in the bottom. F, left: paraffin-embedded tissue slices were colabeled with AQP2 and Nox4 antibodies. Right: anti-rabbit IgG conjugated to Alexa Fluor 488 antibody labeling serves as a negative/background signal control.

Figure 2 shows that the small G protein Rac1 (a key Nox1–3 coregulator) is expressed in A6 cells (Fig. 2, A and B) and rat kidney (RK) homogenate (Fig. 2A). Moreover, the regulatory subunits p22phox, p40phox, p47phox, and p67phox are also found in A6 cells (Fig. 2C). Seemingly, the p22phox auxiliary protein appears to be the only subunit that is essential for Nox4 enzyme activity (5). To determine whether Rac1-dependent Nox isoforms (13) could play an important role in distal nephron cells, we performed additional co-IP studies in Fig. 2D. These data indicate that Rac1 and p67phox coimmunoprecipitate regardless of whether the co-IP was conducted using anti-Rac1 or anti-p67 phox antibody.

Fig. 2.

Nox subunits p22phox, p40phox, p47phox, p67phox, and small G protein Rac1 expression in A6 distal nephron cells. A: small G protein Rac1 expression in A6 cells (A6) and rat kidney (RK) homogenate. B: immunohistochemical detection of Rac1 in subconfluent monolayer of A6 distal nephron cells. C: Western blot detection of Nox regulatory subunits p22phox, p40phox, p47phox, and p67phox. The molecular weight of each protein is indicated in the figure. D, top: A6 cell protein lysate immunoprecipitated (IP) for Rac1 and p67phox and immunoblotted for Rac1. Bottom: A6 cell protein lysate IP-ed for Rac1 and p67phox and immunoblotted for p67phox. Results indicative of 3 independent observations in A–D.

Fulvene-5 attenuates ROS production in A6 cells.

We showed that aldosterone increases ROS production in A6 cells (82). As such, we used dihydrodiclorofluorescein diacetate [H(2)DCF-DA] to detect ROS production in aldosterone-stimulated A6 cells following treatment with Fulvene-5 (a Nox 2 and 4 inhibitor), DPI (a pan Nox inhibitor), catalase (an enzyme that decomposes H2O2), and superoxide dismutase (SOD; an enzyme that inactivates superoxide; Fig. 3). The (−) aldosterone panels serve as negative ROS signal controls. Specifically, in Fig. 3A, we show that aldosterone increases the production of ROS in A6 distal nephron cells. Importantly, Fulvene-5, DPI, catalase, and SOD significantly decreased ROS detection in A6 cells. Representative images are shown in Fig. 3B. Threshold settings for confocal analysis of DCF signal were determined based on low levels of detection in cells devoid of aldosterone. Data are representative of six independent cell culture events, with multiple data points obtained from each cultured sample.

Fig. 3.

Nox4 contributes to aldosterone-sensitive reactive oxygen species (ROS) formation. A: cell-permeant, fluorogenic DCFH2 dye (DCF) was used to measure ROS formation in intact A6 cells grown in the presence and absence of 1 μM aldosterone, 2 μl DMSO vehicle control, 5 μM Fulvene-5, 5 μM DPI, 50 μg/ml catalase, or 50 μg/ml SOD. Confocal microscopy with maximum excitation and emission spectra of 495 and 530 nm, respectively, was used to quantify DCF fluorescence, expressed as relative fluorescence units (RFU). Data are representative of n = 36 independent observations from 6 separate cell-cultured events. *P < 0.05; otherwise, mean DCF fluorescence levels are not significantly different. B: representative confocal microscopy images showing analysis of full monolayers.

Fulvene-5 attenuates transepithelial Na+ transport in A6 cells.

Figure 4 reports the equivalent Isc values obtained from A6 cell monolayers in the absence or presence of 1 μM aldosterone. Figure 4A shows that H2O2 treatment significantly increases transepithelial current in the absence of 1 μM aldosterone. As expected, Fulvene and DPI inhibition of Nox enzyme activity did not significantly alter Isc values in A6 cells cultured in the absence of aldosterone and fetal bovine serum. Interestingly, however, Fulvene-5 and DPI inhibition of Nox enzyme activity significantly attenuates Isc values in A6 cells in the presence of 1 μM aldosterone (Fig. 4B). H2O2 did not increase Isc values above and beyond the aldosterone-induced rate of sodium reabsorption in A6 distal nephron cells (Fig. 4B). Transepithelial resistance values were not negatively impacted following aldosterone deprivation before experimentation, or changes in redox potential (Fulvene-5, DPI, and H2O2 treatment), indicating that these conditions did not impair cell viability or compromise the monolayer (data not shown).

Fig. 4.

Nox-derived H2O2 regulates transepithelial current in A6 distal nephron cells. Equivalent short-circuit current (Isc) of A6 distal nephron cell monolayers obtained in the absence (A) and presence (B) of 1 μM aldosterone, and subsequent treatments of 2 μl DMSO vehicle control, 1 mM H2O2, 5 μM Fulvene-5, or 5 μM DPI. Data are expressed in units of μA/cm2. *P < 0.05 and the number of observations are as indicated on the graphs. ns, Not significant.

Osmotic cell stretch increases ROS production in A6 distal cells.

If osmotic stretch of the cell membrane increases ENaC activity, as previously shown in Ref. 45, then we would expect to observe a concomitant increase in ROS production. First, we verified osmotic cell stretch using confocal microscopy as shown in Fig. 5, A and B. Then, we examined A6 cells incubated with DHE. Representative images are shown in Fig. 5C. Osmotic cell stretch significantly increases oxidized DHE fluorescence (Fig. 5D).

Fig. 5.

Osmotic stretch increases ROS production in A6 distal cells. A: subconfluent A6 cell plasma membranes were labeled with CellMask Deep Red stain (649/666-nm excitation/emission wavelengths). Cell diameter was measured before and after cell bath osmolarity was decreased. Scale bar provides visual indication of change in cell size. B: quantification of cell size; n = 3 independent observations with multiple measurements of cell diameter. *P < 0.05. C: DHE label of A6 cell monolayer before and after osmotic stretch of cell membrane as observed using fluorescence microscopy (518/605-nm excitation/emission wavelengths). D: quantification of change in DHE fluorescence intensity using Olympus Fluoview software; n = 3 independent observations with multiple measurements of cellular DHE intensity and *P < 0.001.

Fulvene-5 inhibits H2O2-induced ENaC activation.

We previously showed that H2O2 significantly increases ENaC activity in A6 cells using cell-attached patch-clamp analysis (44). In the current study, we examined H2O2-stimulated ENaC activity in A6 cells pretreated with Fulvene-5. Figure 6A shows a representative continuous cell-attached patch-clamp recording obtained from an A6 cell under the following conditions: control, treatment with Fulvene-5, and subsequent application with H2O2. Enlarged portions of the trace are shown in (Fig. 6B). Figure 6C shows that H2O2 fails to increase ENaC activity in cells pretreated with Fulvene-5: average ENaC NPo values decreased from 1.04 ± 0.18 (means ± SE) to 0.25 ± 0.08 and 0.18 ± 0.07, following sequential treatment with Fulvene-5 and H2O2, respectively.

Fig. 6.

Pretreatment with Fulvene-5 inhibits H2O2-induced epithelial sodium channel (ENaC) activation. A: representative cell-attached patch-clamp recording of an A6 cell at −Vp (−20 mV) holding potential before treatment (left), following 5-μM Fulvene-5 treatment (middle), and following 10-mM H2O2 treatment (right). Treatments were added directly to the luminal bath. B: channel recording enlarged to show detail. Downward deflections from the closed state are representative of channel opening and inward entry of Na+ ions. C: results are representative of n = 7 independent patch-clamp observations. The y-axis represents ENaC activity in NPo, where N represents the number of channels and Po represents the open probability. The general trend is represented as a line. Fulvene-5 treatment significantly decreases ENaC NPo (*P < 0.05). This effect is not reversed with addition of 10 mM H2O2.

Fulvene-5 inhibits cell stretch activation of ENaC in A6 cells.

We previously showed that osmotic stretch of the cell membrane significantly increases ENaC activity (45). In the current study, we osmotically induced cell membrane stretch as previously described in Ref. 45. Figure 7A shows a continual cell-attached patch-clamp recording of an A6 cell under control and osmotic stretch conditions, followed by application of Fulvene-5 to the same cell-attached recording. Figure 7B shows enlarged portions of current tracings from the cell-attached patch, which demonstrate that Fulvene-5 treatment of A6 cells reverses osmotically induced increases in ENaC activity. Average ENaC NPo values increased from 0.57 ± 0.09 (means ± SE) to 0.97 ± 0.12 following osmotic cell stretch of the membrane. In the same cell-attached recordings, Fulvene-5 significantly decreased ENaC NPo to 0.46 ± 0.10 (Fig. 7C). Interestingly, in similar continual patch-clamp recordings, A6 cells pretreated with Fulvene-5 failed to respond to osmotic cell stretch of the membrane (Fig. 8, A–C).

Fig. 7.

Cell stretch activates ENaC via Nox. A: representative cell-attached patch-clamp recording of an A6 cell at −Vp (−20 mV) holding potential before treatment (left), following addition of 200 μl distilled water (middle), and following 5-μM Fulvene-5 treatment (right). Treatments were added directly to the luminal bath. Addition of dH2O reduces the osmolarity of the solution and causes cellular swelling. B: channel recording enlarged to show detail. Downward deflections from the closed state are representative of channel opening and inward entry of Na+ ions. C: results are representative of n = 7 independent patch-clamp observations. The y-axis represents ENaC activity in NPo. General trend is represented as a line. *P < 0.05.

Fig. 8.

Pretreatment with Fulvene-5 attenuates the effect of cell stretch on ENaC activity. A: representative cell-attached patch-clamp recording of an A6 cell at −Vp (−20 mV) holding potential before treatment (left), following 5-μM Fulvene-5 treatment (middle), and following addition of 200 μl distilled water (right). Treatment was added directly to the luminal bath. B: channel recording enlarged to show detail. Downward deflections from the closed state are representative of channel opening and inward entry of Na+ ions. C: results are representative of n = 8 independent patch-clamp observations and are expressed as ENaC NPo. General trend is represented as a line. *P < 0.05.


Nadph oxidase in the kidney.

Oxidative signaling, as well as oxidative stress, has a major role in the pathophysiology of various renal diseases afflicting every portion of the nephron (18, 21, 55, 77). Although Nadph oxidases (a family of ROS-generating enzymes) have been extensively studied in phagocytes (4, 13, 23, 50, 61), recent studies show constitutive distribution of Nox components in renal cortex, vasculature, endothelium, and major segments of the nephron starting at glomeruli through the medullary collecting duct (24). Given that Nadph oxidases play a key role in producing reactive species, such as O2 and H2O2 (2, 76, 80), and that each Nox isoform is abundantly expressed in the kidney, it is important to gain a better understanding of Nox-mediated function in the kidney. To this end, we first verified that the A6 distal nephron cell line (cells derived from a region of the kidney tubule that is responsible for ENaC-mediated Na+ reabsorption) indeed expresses all components of the Nox isoform using Western blotting and confocal imaging (Fig. 1). Because the mode of Nox 1–3 activation differs significantly from Nox 4, we performed additional co-IP assays (shown in Fig. 2C) to determine whether Rac1-dependent Noxes may be functionally present in the kidney. Generally speaking, Nox1–3 activation requires complete assembly of cytoplasmic regulatory proteins p47phox, p67phox, p40phox, following Rac1 activation (5). Figure 2C shows that small G protein Rac1 indeed coimmunoprecipitates with p67phox, which represents the initial step in complete gp91phox subunit activation (and hence, O2 release). Although Nox4 is the prototypical Nox isoform (commonly referred to as “renox”) and widely investigated in kidney function (22), in Fig. 1 we [and others (24, 52)] report expression of multiple Nox isoforms in the kidney.

Nox inhibitors and clinical potential.

Although antioxidants have the potential to alleviate damage caused by excessive production of reactive species (27, 36), few have proven beneficial in improving human health (20, 27). Although apocynin and diphenyleneiodonium are frequently used to inhibit Nadph oxidase activity, several reports question both the efficacy and efficiency of attenuating ROS production (32, 42, 63, 65, 74, 79). As such, several investigative groups have worked to develop isoform-specific Nadph oxidase inhibitors (which include small molecule Nox inhibitors, peptide Nox inhibitors, and Nox siRNAs) (8, 34, 39, 53) which could significantly impact clinical diseases associated with dysregulated ROS production. In this report, we examined the effect of a novel Nox inhibitor, Fulvene-5, on ROS production and ENaC activity.

Fulvene-5, a cyclopentadiene derivative (35), has been shown to potently inhibit Nox4 oxidase activity and also limit tumor growth in mice (6). This inhibitory effect on tumor growth was similar to that of lentiviral silencing of Nox4 mRNA (6), suggesting that Fulvene-5 potently and specifically inhibits Nox4 activity. In similar experiments, Fulvene-5 was shown to decrease Nox4 (overexpressed in HEK293 cells) and Nox2 (in COSphox cells expressing inducible Nox2/p47phox/p67phox complex) production of ROS measured 1 h following treatment. Based on these results, Fulvene-5 is a promising Nox inhibitor that requires further characterization. In Fig. 3, we show that Fulvene-5 decreases H2O2 production in A6 distal nephron cells to the same extent as DPI, catalase, and aldosterone deprivation. Importantly, we show that Fulvene-5 can decrease aldosterone-sensitive transepithelial current (Fig. 4) without altering monolayer resistance and integrity (data not shown). Moreover, Fig. 6 shows that Fulvene-5 pretreatment attenuates previously reported H2O2-induced increases in ENaC activity. The 10-mM H2O2 treatment used in Fig. 6 is seemingly high; however, A6 cells reportedly have very high endogenous catalase activity, and therefore require exogenous application of high-H2O2 concentrations to elevate intracellular ROS levels (44). Dr. Ma has recently published the use of 3 mM H2O2 in similar patch-clamp studies without observing signs of toxicity in A6 cells and has shown that this concentration elevates intracellular H2O2 concentrations to ∼12 nM. The important implication is that Fulvene-5 represents a novel, potential anti-hypertensive agent that decreases net sodium reabsorption by the kidney and total body fluid through its antioxidant properties.

Nadph oxidases and clinical hypertension.

Renal ROS generation has been shown to play a critical role in severe and salt-dependent forms of hypertension in humans and animal research models (11, 53, 59). Interestingly, several studies implicate an important role for Nox isoforms in hypertension. Transgenic animals lacking normal expression of p47phox (a critical regulatory unit of Nox1–3) are protected against angiontensin II-induced elevation of systolic blood pressure (40). Likewise, chimeric peptide gp91ds-tat inhibition of p47phox association with Nox catalytic domains also attenuates the hypertensive effects of angiotensin II infusion. Similarly, Nox2−/− and Nox4−/− transgenic animals have normal basal blood pressures (28, 67). Seemingly, this could be a favorable outcome in terms of drug discovery because several novel compounds, such as Fulvene-5, have can target both Nox2 and Nox4 (39).

Nox and ENaC activity.

We (25, 73) and others (33, 72) showed that Nox plays an important role in regulating ENaC transport. The important finding from our current study is that we provide a plausible mechanism for how Nox4 activation can lead to inappropriate ENaC activation, and potentially, hypertension (illustrated in Fig. 9). To date, the mechanisms that lead to Nox4 activation remain unclear, albeit polymerase DNA-directed delta interacting protein 2 (Poldip2) association with Nox4 and p22phox complexes (43) and induction at mRNA levels (as opposed to posttranslational protein modifications and/or enzyme complex assembly) (68) have been implicated. In Figs. 6 and 7, we show that inhibition of Nox enzyme activity by Fulvene-5 treatment of A6 cells attenuates H2O2-mediated and osmotic stretch activation of ENaC. Hence, our findings implicate Nox enzyme activity as a critical mediator of ROS and osmotic stretch-induced increases in ENaC activity. The schematic presented in Fig. 9 represents our major findings in this study, which are in agreement with reports by Oeckler et al. (56) and Caldiz et al. (10) that show stretch-mediated activation of the Nadph oxidase in endothelial and myocardial cells. Tissue Nox activation typically releases low levels of ROS, as opposed to phagocytic Nox activation. As such, one would expect a mechanism for signal amplification in epithelial cells following stimulation of Nox activity. In Fig. 9, we propose that osmotic stretch activation of Nox leads to release of low levels of H2O2, which feeds back to increase Nox activity and ROS production. This mechanism is supported by data presented in Fig. 6, where cells pretreated with Fulvene-5 fail to respond to H2O2 treatment, and also by published data by Clark and colleagues (16, 17) that show both human Nox2 and Nox5 are activated by H2O2 in a c-Abl tyrosine kinase-dependent manner. In light of our observations, however, it is important to note that the mitochondria can produce a significant amount of ROS (reviewed in Ref. 54). In our studies, however, we did not delineate the subcellular source of ROS. This is an important area of future research given the recent identification of mitochondrial Nox isoforms (7, 26, 38).

Fig. 9.

Schematic showing the role of Nox-derived ROS in ENaC regulation. Fulvene-5 inhibition of renal Nadph oxidase activity attenuates H2O2 and osmotic stretch activation of ENaC. In the absence of Nox inhibition, positive feedback from inappropriate salt and water reabsorption (leading to increased blood volume, hypertension, and possible osmotic stretch of distal nephron cells) can exacerbate renal injury under oxidative stress.

In summary, our major finding is that cell stretch activation and H2O2-mediated feedback of Nadph oxidase activity play an important role in regulating ENaC activity. These findings suggest a mechanism by which inappropriate ROS production leads to an increase in ENaC activity in collecting duct cells and, potentially, hypertension. In our study, osmotically induced stretch activation of ENaC serves as a model for the established effects of cell swelling following the increase in glomerular filtration rate that accompanies elevated extracellular fluid volume and blood pressure (15). We also report the efficacy of the Nox inhibitor Fulvene-5 in decreasing both ROS production and ENaC activity. This finding serves as rationale for future testing of Fulvene-5, and perhaps other antioxidant compounds, as an anti-hypertensive agent in the clinical setting.


This work was supported by Grant R01-DK067110 awarded to H. Ma and Grant R00-HL09222601 awarded to M. N. Helms.


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: D.T., H.-P.M., and M.N.H. conception and design of research; D.T., B.-C.L., A.C.P., S.V.T., M.P., H.-P.M., and M.N.H. performed experiments; D.T., B.-C.L., A.C.P., M.P., H.-P.M., and M.N.H. analyzed data; D.T., A.C.P., C.A.D., H.-P.M., and M.N.H. interpreted results of experiments; D.T., B.-C.L., A.C.P., M.P., C.A.D., H.-P.M., and M.N.H. prepared figures; D.T., A.C.P., C.A.D., and M.N.H. drafted manuscript; D.T., A.C.P., and M.N.H. edited and revised manuscript; D.T., B.-C.L., A.C.P., S.V.T., M.P., C.A.D., H.-P.M., and M.N.H. approved final version of manuscript.


We gratefully acknowledge the technical assistance of Preston Goodson, and the 2013 APS Writing and Reviewing for Journals workshop attended by D. Trac.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
View Abstract