Nadph oxidase regulates alveolar epithelial sodium channel activity and lung fluid balance in vivo via O₂⁻ signaling

Abstract

To define roles for reactive oxygen species (ROS) and epithelial sodium channel (ENaC) in maintaining lung fluid balance in vivo, we used two novel whole animal imaging approaches. Live X-ray fluoroscopy enabled quantification of air space fluid content of C57BL/6J mouse lungs challenged by intratracheal (IT) instillation of saline; results were confirmed by using conventional lung wet-to-dry weight ratios and Evans blue as measures of pulmonary edema. Visualization and quantification of ROS produced in lungs was performed in mice that had been administered a redox-sensitive dye, hydro-Cy7, by IT instillation. We found that inhibition of NADPH oxidase with a Rac-1 inhibitor, NSC23766, resulted in alveolar flooding, which correlated with a decrease in lung ROS production in vivo. Consistent with a role for Nox2 in alveolar fluid balance, Nox2^−/− mice showed increased retention of air space fluid compared with wild-type controls. Interestingly, fluoroscopic analysis of C57BL/6J lungs IT instilled with LPS showed an acute stimulation of lung fluid clearance and ROS production in vivo that was abrogated by the ROS scavenger tetramethylpiperidine-N-oxyl (TEMPO). Acute application of LPS increased the activity of 20 pS nonselective ENaC channels in rat type 1 cells; the average number of channel and single-channel open probability (NPo) increased from 0.14 ± 0.04 to 0.62 ± 0.23. Application of TEMPO to the same cell-attached recording caused an immediate significant decrease in ENaC NPo to 0.04 ± 0.03. These data demonstrate that, in vivo, ROS has the capacity to stimulate lung fluid clearance by increasing ENaC activity.

alveolar type 1 (AT1) and type 2 (AT2) cells comprise the alveolar epithelium, which together serve to separate the luminal air space from the vasculature. The role of AT1 and AT2 cells in gas exchange and surfactant production is well established (5, 54). We have recently shown that both cell types express functional epithelial sodium channels (ENaC) using single-channel patch-clamp analysis (15, 17). Sodium uptake from the luminal surface generates the electroosmotic flow that pulls water molecules across the alveolar epithelium and, hence, maintains an appropriately moist luminal space for effective gas exchange. As such, regulation of lung ENaC is of utmost importance for normal respiration. Transgenic mice that lack α-ENaC expression die within days of birth, because of an inability to clear lung fluid (22). Interestingly, this lethal phenotype can be rescued by epigenetically expressing ENaC (21). Conversely, hyperactive sodium channel activity aberrantly dries the airways and leads to the excessively thick mucus present in children with cystic fibrosis disease (33, 46, 58). Although ENaC plays an important role in health and disease, the signal transduction pathways that regulate this ion channel remain largely unknown in both AT1 and AT2 cells.

Fetal lungs transition from low to high oxygen tension at birth as they transition from a fluid-secreting organ to an organ that can actively reabsorb salt and water. Thus oxygen signaling has been implicated in regulating lung ENaC (3, 37, 38, 40–43, 50). Oxygen could serve as a signaling molecule following reduction to superoxide anion, O₂⁻. Superoxide anions are usually dismutated to H₂O₂, or they react with nitric oxide to form peroxynitrite; all of these molecules have been shown to be important regulators of ENaC activity (14, 19, 29, 34, 52, 56). We have recently shown in vitro that sequestering O₂⁻ with 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) inhibits ENaC activity (56), and conversely that increasing amounts of O₂⁻ by using a mix of hypoxanthine and xanthine oxidase, or superoxide dismutase inhibitors, increases ENaC activity (15, 56). The focus of this study is to determine whether these pathways play an important role in regulating lung fluid in vivo.

One of the best characterized sources of cellular reactive oxygen species (ROS) is via NADPH oxidase production of O₂⁻ (reviewed in Ref. 28). In the present study, we test the hypothesis that the Nox family of NADPH oxidases plays an integral role in regulation of ENaC. NADPH oxidases are membrane-bound enzyme complexes, with seven major isoforms identified (Nox1–5 and Duox1–2). Each isoform is structurally related to the others, and all are capable of releasing ROS. The signaling pathways that govern NADPH oxidase significantly overlap with established ENaC regulatory pathways. For example, full assembly and activation of Nox2 requires protein kinase signaling, lipid metabolizing enzymes, and conversion of Rac1 from the GDP-bound to the activated GTP-bound form. Only after full stimulation of these signaling molecules can the cytosolic regulatory proteins of Nox1–3 isoforms (p40phox, p47phox, and p67phox) associate with and activate the catalytic domains of gp91phox, for ROS production. Interestingly, the kinases and signaling molecules needed for regulation of Nox isoforms, such as phosphatidylinositol 3-kinase, phosphatidylinositols, and Rac1, have also been shown to be important components of ENaC signaling (10, 16, 32). Despite the significant overlap in signal transduction between Nox2 and ENaC activation, the intimate relationship between NOX release of O₂⁻ and ENaC activity has not been fully investigated. In the present study, we utilized multimodal animal imaging and single-channel patch-clamp analysis to study the role of NADPH oxidase in lung fluid clearance.

MATERIAL AND METHODS

Animal care.

Wild-type C57BL/6J mice (8 and 12 wk old) and 8-wk-old C57BL/6J mice with Nox2 gene expression disrupted (B6.129S-Cybb^tm1Din/J) were purchased from The Jackson Laboratory (Bar Harbor, ME) and housed with access to standard diet and water ad libitum. All animal protocols and handling conform to National Institutes of Health animal care and use guidelines, as approved by Emory University Institutional Animal Care and Use Committee.

IT instillations.

Intratracheal (IT) instillation of solute into the alveolar air space has been described in Ref. 18. Mice were anesthetized with an intraperitoneal injection of 15 mg/kg xylazine (Anased 100 mg/ml, Lloyd Laboratories; Shenandoah, IA) and 125 mg/kg ketamine (ketamine hydrochloride injection USP 50 mg/ml, Bioniche Pharma USA; Lake Forest, IL) prepared in phosphate-buffered saline (PBS, pH 7.4; Invitrogen, Grand Island, NY) for a final intraperitoneal injection volume of 250 μl per 20-g mouse. The instillate vehicle for all experiments consisted of 140 mM NaCl, 5 mM KCl, 1 mM CaCl₂, and 10 mM HEPES with pH = 7.4; hereafter referred to as control saline. The osmolality of this saline solution is 319 mosmol/kgH₂O_, which is near the reported plasma osmolality of mice (1, 12). Tracheal instillation volumes of 5 μl per gram of body weight (100 μl final volume in 20 g C57BL/6J mice) were delivered by use of a 26-gauge, 5/8-in. needle and a 1-ml syringe. Experimental groups included instillations of 100 μl saline, LPS (1 mg/ml), 1 mM amiloride, 250 μM 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), or 100 μM NSC23766. All compounds are available commercially. Following instillation, the incision site was closed and animals were immediately placed into the imaging station described below under fluoroscopy.

Fluoroscopy.

Animals were X-rayed or imaged for fluorescence signal by using an In-Vivo MS FX Pro small animal imager (Carestream Health; Rochester, NY) with an E-Z Anesthesia unit (Palmer, PA) attached to deliver oxygen and 2% isoflurane. Ambient temperature in the imager was elevated ∼4° above RT to keep the animals warm. Animals were X-ray imaged at 5-min intervals up to 240 min with an acquisition period of 120 s and kept on oxygen mixed with gas anesthesia for the duration of imaging. X-ray settings were 2×2 binning, 180-mm field of view (FOV), 149 μA X-ray current, 35 kVp, and 0.4-mm aluminum filter.

X-ray density was quantified by using Carestream Health MI software with a 5-mm² region of interest (ROI). In all whole animal studies, the ROI was the upper left lobe of the lung (see Fig. 1). In fluoroscopic studies, all 5-mm² ROI placements were caudal to the third rib on the left side of the chest X-ray. All fluoroscopic ROIs were background corrected and normalized to the initial X-ray intensity (I_o) to make comparisons between time points and experimental groups. Specifically, changes in lung fluid volume are expressed as I−I_o, where I is the X-ray density at a respective time point and I_o is the initial X-ray density following saline challenge.

Fig. 1.

Distribution and detection of intratracheally (IT) instilled saline in C57BL/6J mouse lung. A: schematic of mouse lung, indicating IT administration of a saline challenge specifically into the left lobe. B: mouse lung instilled IT with 25 μl bromophenol blue dye, excised, and photographed to show region of dye accumulation in the upper left lung lobe, indicated here as region of interest (ROI). C: 10-μm tissue section prepared from the lung shown in B, demonstrating that bromophenol blue dye-labeled instillate reached the distal alveolar air space (*). D: confocal images of uninstilled (left) and 100 μl saline-instilled (right) left lung lobes fluorescently labeled with anti-zonula occludens-1 antibody to label tight junctions [488 nm (green) excitation], rhodamine phalloidin to label actin [538 nm (red) excitation]; and 4,6-diaminido-2-phenylindole (DAPI) to label nuclei [360 nm (blue) excitation]. IT saline administration did not induce gross lung damage.

Fluorescent imaging.

Thoracic fur was removed prior to imaging. The redox-sensitive dye hydrocyanine (Cy 7; 2 mM) was reconstituted in saline solution and IT instilled into C57BL/6J lung after 4 h following LPS, LPS+ NSC23766, NSC23766, or saline instillation. Animals were placed into the small rodent imager configured for epi-illumination and fluorescence detection. Specifically, imager settings were 140-mm FOV, 4×4 binning, 700 nm excitation, and 830 nm emission. Data were collected by a charge-coupled device camera during a 5-min acquisition period and ROI was analyzed by use of Carestream Health Multispectral (MI) Software.

The fluorescence intensity of ROS-activated (Cy7) was also quantified from the upper left lobe of the lung by use of Carestream Health MI software and expressed as photons per second per square centimeter.

Statistical analysis.

All data are summarized as means ± SE. Single comparisons were performed by Student's t-test. Multiple comparisons were performed by one-way ANOVA followed by the Holms test for pairwise comparisons. P values of <0.05 were regarded as statistically significant.

Detection of 0.5% bromophenol dye in the alveoli.

After IT instillation of bromophenol blue dye under anesthesia, mice were exsanguinated by cutting the renal artery, and lungs were perfused via the right ventricle with PBS. Subsequently, lungs were excised from the thoracic cavity and covered with a thin layer of optimal cutting temperature (OCT) compound. Lung tissue embedded in OCT was snap frozen in liquid N₂-cooled isopentane and sectioned in a cryostat. Bromophenol blue was detected in the alveoli by use of an inverted IX-70 Olympus microscope.

Confocal microscopy.

Lung tissue slices were prepared as described in Ref. 15. Cells were fixed in 2% paraformaldehyde before immunolabeling with anti-zona occludin antibody used at a 1:1,000-fold dilution. Anti-rabbit secondary antibodies conjugated to Alexa 488 (Molecular Probes) were used at a 1:20,000-fold dilution. Lung tissues were also counterstained with 1:1,000 phalloidin (Upstate) prior to being mounted on glass coverslips with 4,6-diaminido-2-phenylindole Vectashield (Vector Laboratories). Labeled cells were imaged by use of an Olympus BX61WI confocal microscope and Fluoview FV10-ASW 1.7 software.

Electrophysiology.

AT1 cells were accessed from a live lung tissue preparation and patch clamped as described in Ref. 17. Briefly, 250-μm lung slices were prepared from agarose-instilled rat lungs by use of a vibrating microtome and transferred to 50:50 ice-cold DMEM/F-12 media containing FBS (10%), l-glutamine (2 mM), dexamethasone (1 μM), gentamicin (84 μM), and penicillin-streptomycin (20 U/ml) prior to patch-clamp analysis. Following GΩ seal formation between the apical membrane and patch-clamp electrode, a control recording period (5 min) was obtained prior to adding LPS (1 μg/ml) and then TEMPO (250 μM) to the same cell-attached patch-clamp recording to determine the role of ROS signaling on lung ENaC activity. The product of the number of channels (N) and the single-channel open probability (P_o) was calculated by use of Clampfit 10.1 software (Molecular Devices).

Wet-to-dry lung weight ratios.

Under ketamine and xylazine anesthesia (described above) C57BL/6J lungs were IT instilled with 100 μl saline. The lungs were excised and weighed 30 min following instillation. Lungs were dried at 60°C for 48 h, then weighed again to obtain the wet-to-dry weight ratios.

Evans blue-labeled albumin measurements.

Twelve-week-old mice were tracheally instilled (as described above) with solution containing 5% bovine serum albumin (BSA) and Evans blue dye (0.1 mg/ml). Alveolar fluid clearance (AFC) was assessed 60 min after instillation by measuring the change in concentrations of alveolar Evans blue-labeled albumin. As described in Refs. 44, 53, AFC was calculated as follows:

$$\text{AFC}=[(\text{Vi}-\text{Vf})/\text{Vi}]\times 100,$$

where Vi is the volume of instilled Evans blue-labeled albumin and Vf is the final alveolar fluid calculated as follows:

$$\text{Vf}=(\text{Vi}\times \text{EBi})/\text{EBf},$$

where EB is the concentration of Evans blue-labeled albumin in the instilled solution (i) and the final alveolar fluid (f). Protein concentration was determined with the use of spectrophotometry.

RESULTS

Distribution and detection of intratracheally instilled saline in C57BL/6J mouse lung.

We and others have shown that IT instillation of solutes serves as an effective and minimally invasive technique for targeted delivery of compounds into lungs (9, 18, 31, 48, 55). In Fig. 1A, we show a schematic of the five lobes of mouse lung and the site of IT instillation used to localize the site of fluid accumulation. Accordingly, 25 μl tracheally instilled bromophenol blue dye accumulates in the upper left lobe of the excised lung (Fig. 1B). From the same lung shown in 1B, 10-μm cryoslices were prepared to verify that the small volume of instilled dye uniformly extends into the distal regions (alveoli) of the left lobe (Fig. 1C). This treatment did not have a significant effect on lung morphology, since the localization of a tight junction protein, zonula occludens-1 (ZO-1), and F-actin did not differ significantly between uninstilled (left) and 100 μl saline-instilled portions of the lung (Fig. 1D). That ZO-1 and actin labeling remained largely unchanged indicates that the barrier property of the alveolar epithelium remained intact and that a single saline challenge does not cause gross lung damage. The upper left lobe of the lung is the consistent site of fluid accumulation following IT instillation (35, 48). As such, we designate this area of the lung as the ROI in all subsequent studies.

Air space fluid is opaque in X-ray imaging.

Fluid attenuates X-ray photons. As such, soft tissue and fluid filled compartments appear more opaque (white) than air-filled compartments, which, in contrast, appear black. This is demonstrated in Fig. 2A, which shows an X-ray image taken from a live C57BL/6J mouse following IT saline instillation into the upper left lobe of the lung (ROI). Between the ribs, the ROI has a cloudy, white appearance compared with the uninstilled right lobes of the lung. To confirm that this method gives us a quantitative method to measure fluid volume, we X-ray imaged plastic 35-mm culture dishes containing fluid volumes ranging from 0 to 7 ml (Fig. 2, B and C), which showed a linear increase in X-ray opacity (lower density of X-ray transmission) with increasing fluid volume. On the basis of this relationship between fluid volume and X-ray absorption, we are able to track changes in lung fluid volume.

Fig. 2.

Air space fluid is opaque in X-ray imaging. A: high-resolution X-ray imaging of C57BL/6J mice IT instilled. Arrow points to a fluid-filled ROI in the upper left lung lobe, which produces a cloudy white image due to X-ray attenuation. By contrast, air-filled lobes appear darker and are penetrated by X-rays. B and C: X-ray imaging of 0–7 ml saline in a 35-mm plastic petri dish demonstrated a linear increase in X-ray opacity with increasing fluid volume.

Air space fluid volumes correspond to altered lung ion transport properties.

Activation of amiloride-sensitive sodium channels serves as the major pathway for salt and water reabsorption across the alveolar epithelium and out of the airways (4, 11, 12, 57). Conversely, outward flow of anions via CFTR influences the rate of pulmonary secretion driving in a net increase in lung airway fluid (7). To assess the utility of X-ray fluoroscopic imaging to assess lung liquid volume, we examined 12-wk-old C57BL6 mice that were IT instilled with either 140 mM saline alone, saline containing 1 mM amiloride (to inhibit ENaC), or saline containing 1 mM glibencamide (to block CFTR). The mice were challenged with 5 μl per gram body weight (∼100 μl total). Immediately following IT instillation, animals were imaged by using a 2-min exposure every 5 min for a total of up to 240 min. X-ray intensities were quantified by I−I_o, where I_o is the X-ray intensity at an ROI immediately after instillation and I is the intensity at a given time following instillation. Thus I−I_o decreases with increasing air space fluid volume (Fig. 3).

Fig. 3.

X-ray fluoroscopy enabled in vivo imaging of lung fluid. A: 3-mo-old C57BL/6J mice were challenged by IT instillation with 100 μl of either control saline (n = 10), 1 mM glibenclamide (n = 10), or 1 mM amiloride (n = 10). The mice were imaged by X-ray fluoroscopy and quantified by I−I_o, where I_o is the X-ray intensity at an ROI immediately after instillation and I is the intensity at a given time following instillation. Each point represents a 2-min rolling average of X-ray intensity measurements corresponding to lung air space fluid volume. Amiloride instilled into C57BL mice significantly impaired the clearance of IT saline challenge compared with control animals instilled with an equal volume of control saline solution at all time points following I_o; *P < 0.01. By contrast, glibenclamide inhibition of chloride channels effectively blocked secretion and enhanced fluid clearance at all time points following I_o. (*P < 0.05) B: conventional wet-to-dry lung weight ratio measurements confirmed that amiloride inhibited lung fluid clearance 30 min postinstillation and correlate with in vivo X-ray fluoroscopic measurements (*P < 0.05). C and D: representative X-ray images of amiloride and glibenclamide instilled lung 5 and 240 min postinstillation. ROI is circled in the left lobe of the lung. X-ray opacity and area of alveolar flooding (indicated by gray circle) increases 240 min following amiloride instillation. The ROI is more X-ray dense 240 min following glibenclamide instillation. Mean X-ray intensities of ROI reported under each figure.

Consistent with previously published reports, real-time X-ray imaging of the lungs showed that blocking ENaC activity with amiloride (n = 10) significantly slowed the rate of fluid reabsorption compared with control animals instilled with equal volume saline (n = 10) (Fig. 3A). In Fig. 3, C and D, we show representative chest X-rays obtained 5 and 240 min after instillation of amiloride and glibenclamide. We also validated our X-ray fluoroscopic approach by measuring lung wet-to-dry weight ratios following IT instillation of saline and amiloride by comparison with mock-instilled lungs. As shown in Fig. 3B, mice with IT instilled amiloride had the highest levels of edema, consistent with our findings by live X-ray fluoroscopy.

We also examined the acute effect of glibenclamide, a blocker of CFTR transport (45), on lung fluid content. Blocking chloride channel activity significantly increased air space X-ray density (n = 10, P < 0.05), consistent with a decrease in lung fluid accumulation due to continued absorption combined with inhibited fluid secretion. Thus the nearly consistent I−I_o values measured for saline-instilled animals are likely to reflect a steady-state equilibration between rates of fluid absorption and secretion.

Rac1 inhibition attenuated lung fluid clearance in vivo.

Using single-channel patch-clamp analysis and dihydroethidium labeling, we have recently shown that the specific Rac1 inhibitor NSC23766 significantly decreases ENaC activity in both AT1 and AT2 cells and reduces cellular production of O₂⁻ (49). To determine whether these measurements at the individual cell level translate into net changes in lung fluid balance, we assessed the clearance of 100 μl 1 μM NSC23766 IT instilled into C57BL/6J mice (n = 10). As shown in Fig. 4A, mice IT instilled with NSC23766 failed to clear the saline challenge compared with control-instilled animals at several time points between I_o and 240 min. We verified the X-ray fluoroscopic measurements of fluid clearance using wet-to-dry weight ratios, which gave comparable results (Fig. 4B).

Fig. 4.

Rac1 inhibition attenuated lung fluid clearance in vivo. A: C57BL/6J mice were challenged by IT instillation with 100 μl of either control saline (n = 10) or a Rac inhibitor compound 100 μM NSC23766 (n = 10). Mice treated with NSC23766 had significantly increased lung fluid volume, compared with control animals. *P < 0.05. B: wet-to-dry lung weight ratio measurements confirm that NSC23766 inhibited fluid clearance 30 min following instillation. *P < 0.05.

NADPH oxidase 2-deficient mice have impaired lung fluid clearance.

Although Rac1 is implicated in several physiological responses, we hypothesized that Rac1 stimulates production of ROS derived from NADPH oxidase to stimulate lung ion transport. We tested this hypothesis by assessing the ability of Nox2-deficient mice (Nox2^−/−) to clear a saline challenge. As shown in Fig. 5, 8-wk-old Nox2^−/− mice challenged with 5 μl saline per gram showed an impaired ability to clear IT instilled saline (n = 8) compared with 8-wk-old wild-type mice (n = 10).

Fig. 5.

NADPH oxidase 2-deficient mice have impaired lung fluid clearance. Eight-week-old Nox2^−/− mice (n = 7) showed diminished clearance of a 100-μl saline IT challenge compared with age-matched control mice (n = 10), over a 4-h period. *P < 0.05.

LPS-stimulated O₂⁻ production promotes ENaC activity and regulates lung fluid volume in vivo.

LPS is a known activator of NADPH oxidase activity that works through activation of Rac1 (23, 25). Thus we examined the ability of 12-wk-old C57BL/6J mice to clear IT instilled 100 μl of LPS (1 μg/μl). In a separate group of animals, we coinstilled TEMPO to sequester any ROS generated by instilled LPS. As shown in Fig. 6A, LPS immediately increased alveolar flooding. However, roughly 30 min after challenge, the LPS-instilled animals showed increased fluid clearance (n = 10). By contrast, mice cotreated with TEMPO did not (n = 10). Thus, by scavenging ROS, we abolished the ability of IT instilled LPS to acutely stimulate lung fluid clearance.

Fig. 6.

Lipopolysaccharide (LPS)-stimulated O₂⁻ production promotes epithelial sodium channel (ENaC) activity and regulates lung fluid volume in vivo. A: C57BL/6J mice were challenged by IT instillation with 100 μl 1 mg/ml LPS (n = 10) or LPS + 250 μM 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO; n = 10). In both cases, there was an initial increase in air space fluid volume; however, fluid clearance by LPS instilled animals was abrogated by TEMPO treatment (*P < 0.05). B: representative cell-attached patch-clamp recording of an alveolar type 1 cell at −Vp (−20 mV) holding potential before LPS treatment (top trace) and acutely following 1 mg/ml LPS treatment (middle trace). In the same cell-attached patch-clamp recording, TEMPO was added, which sequestered O₂⁻ and significantly decreased Na current (bottom trace). Downward deflections from arrows (indicating closed state) are representative of inward current. Breaks in continuous cell recording are denoted by ##. C: slope conductance of a representative nonselective cation channel shown (C) is 20 pS. D: single-channel data analysis of ENaC activity in continuous patch-clamp recording demonstrates a significant increase in the number of open ENaC (NPo) abrogated by TEMPO. At baseline, ENaC had NPo = 0.14 ± 0.04, which significantly increased to 0.62 ± 0.22 following LPS treatment with LPS (P < 0.05). TEMPO immediately decreased ENaC NPo to 0.04 ± 0.03, which statistically differs from control (CTR) and LPS average values, n = 6; P < 0.05; Vp, pipette potential.

For the clearance measurements in Fig. 6A, we found that X-ray fluoroscopy was significantly more sensitive than wet-to-dry weight measurements, which were unable to detect a significant inhibition of LPS-induced fluid clearance by TEMPO (P. Goodson and M. N. Helms, unpublished data). Thus we validated our fluorometric approach by repeating each of the treatment parameters (i.e., amiloride, glibencamide, LPS, LPS+TEMPO, and NSC23766) using the Evans blue assay to measure changes in albumin concentration as a measure of lung fluid clearance (44). These results are shown in Fig. 7 and verify that 1) glibenclamide indeed increases AFC (compared with saline-instilled control) whether the results are measured by X-ray imaging or protein concentration measurements, 2) amiloride and inhibition of Rac1 significantly decreases AFC (compared to saline-instilled control) whether the results are measured by X-ray imaging or protein concentration measurements, and 3) sequestering ROS with TEMPO compound significantly attenuated the LPS-induced effect on lung fluid clearance. On the basis of these outcomes, we summarize that wet-to-dry weight determination and changes in albumin protein concentrations are useful metrics for measuring alveolar fluid clearance (at a single time point), and X-ray fluoroscopic imaging enables airway fluid to be quantified from live mice in real time (at multiple time points).

Fig. 7.

Changes in albumin protein concentration validate fluorometric data obtained from live animals. A: C57BL/6J mice challenged by IT instillation with either 100 μl glibenclamide (Glib.; 1 mM), LPS (1 mg/ml), LPS + TEMPO (250 μM), amiloride (Amil.; 1 mM), or NSC23766 (100 μM) combined with 5% BSA and Evans blue dye. Protein concentrations are reported after 60 min of treatment as mg/ml. Analysis of variance (Tukey's test) was used to determine significance between multiple groups; paired t-test was used determine statistical differences between LPS and LPS + TEMPO treatment; *P < 0.05 unless otherwise stated. Number of independent observations indicated on graph. B: alveolar fluid clearance expressed as % change in albumin instilled, calculated from AFC = [(Vi−Vf)/Vi] × 100, as described in material and methods. The percent changes greater than average control values of 22.05% indicate enhanced rates of fluid clearance (glibenclamide), whereas negative percent changes are indicative of alveolar flooding (amiloride and NSC-23766). TEMPO attenuated LPS induced changes in AFC by ∼6.5% in murine lungs.

To better understand the time course of LPS-mediated effects on transport, we performed single-channel patch-clamp analysis on AT1 cells, which makes up > 95% of the surface area of the alveolar epithelium. The single-channel analysis in Fig. 6, B–D, shows that a 20 pS nonselective channel is activated following 1 μg/ml LPS treatment. We measured NPo, which is an electrophysiological measurement of sodium channel number (N) times open probability (Po). In LPS-treated patches, NPo activity increased from 0.14 ± 0.04 to 0.62 ± 0.23 (n = 6; P < 0.05). Importantly, 250 μM TEMPO treatment, subsequent to LPS, significantly decreased channel activity in the same cell-attached recordings with average NPo values decreasing to 0.04 ± 0.03. An increase in ENaC nonselective cation channel activity could explain the latent increase in lung fluid clearance induced by LPS between 210 and 240 min following IT instillation.

In vivo detection of lung ROS.

Hydrocyanine 7 (Cy7) is a redox sensor that emits light in the near-infrared range (560–830 nm) upon interacting with ROS (26). To detect ROS production in vivo, we used a protocol that required two separate IT instillations. First, mice were IT instilled with either 1 mg/ml LPS, 1 mg/ml LPS + 100 μM NSC23766, 100 μM NSC23766 alone, or saline to stimulate or inhibit Nox2 activity. Four hours later, Cy7 was introduced by IT instillation and ROS-induced fluorescence was measured.

The Cy7 fluorescence signal, coregistered with white light illumination, was measured (Fig. 8A). Note the intensity scale bar that shows relative light intensities (RLU) correlated with ROS production. In Fig. 8B, average ROS levels were determined by calculating RLU values as photons per second per centimeter squared. We found that LPS enhanced ROS production compared with all other conditions tested, including IT instillation with saline alone, which also induced ROS, albeit lower than the level stimulated by LPS. Importantly, NSC23766 significantly decreased ROS production. Together, these findings indicate that LPS stimulates lung ROS release via a Rac1-dependent pathway.

Fig. 8.

In vivo detection of lung ROS. A: C57BL/6J mice challenged by IT instillation with either 100 μl 1 mg/ml LPS, 100 μl 1 mg/ml LPS + 100 μM NSC23766, 100 μM NSC23766, or 100 μl saline (control). Four hours after initial instillation, a second IT instillation of hydrocyanine dye was added to detect lung ROS production. Left: representative images of ROS produced by mice. Right: scale of relative light intensities (RLU) correlated with ROS production. B: average ROS was determined by calculating RLU values as photons·s⁻¹·cm⁻². LPS enhanced ROS production compared with all of the other conditions tested. NSC23766 significantly decreased ROS levels regardless of whether LPS was coinstilled. *P < 0.05. Number of observations reported under the graph, with LPS + NSC23766 observations lower than all other observations because of low viability in this treatment group.

DISCUSSION

A new fluoroscopic approach to assay relative lung fluid balance.

Relative lung fluid volumes have been determined by measuring changes in albumin protein concentration (12, 36, 57) and also by comparing the lung wet-to-dry weight ratios (4, 12). Recently, Lewis et al. (30) reported the use of X-ray phase contrast in detecting lung aeration in rabbit pups delivered by caesarean section. In the present study, we expanded the use of X-ray imaging to assess lung fluid clearance in adult mice following saline challenge and were able to detect very small changes in lung fluid volume without euthanizing the animals. We are able to equate changes in lung fluid volume to ENaC activity because it has been shown that Na ions represent ∼80% of the total ion transported in alveolar epithelial cells (24) and because amiloride (sodium channel blocker) significantly decreases the lungs' ability to resolve IT saline challenges. The fluorometric data shown in Fig. 3A, where amiloride instillation leads to alveolar flooding, is similar to the effects of instilling amiloride as reported by Matthay and colleagues (12) and Matalon and colleagues (47) using alternative approaches to measure lung fluid volume.

Our approach to measuring lung fluid balance involves tracheally instilling one lobe of the lung. Practically speaking, flooding one region (of otherwise healthy lung) in our procedure precludes the need to ventilate animals, which could introduce additional lung injury. Furthermore, animals that survive the saline challenge can be used for subsequent in vivo imaging studies (such as that shown in Fig. 7), or, as a follow-up, cell isolation. For our research purposes, this model of lung injury is useful for and represents a technical advancement for studying amiloride-inhibited transport in the lung. Several research groups, namely those of Matthay (27, 39), Matalon (13, 47), and O'Brodovich (51), have also similarly evaluated the rate of fluid clearance from isolated lobes of the lung. On the basis of these studies, and our wet-to-dry lung weight measurements, we presume that changes in fluid balance observed in one lobe of the lung represent changes in total lung fluid. Although beyond the scope of the present studies, the fluoroscopic approach to measuring lung fluid volume could be extended and applied to acute lung injury, asthma, and pneumonic animal models.

Rac1-mediated NADPH oxidase regulates lung fluid balance.

We have recently reported that Nox2 is robustly expressed in the alveoli and plays an important role in regulating lung ENaC activity (49). Specifically, alveolar epithelial cells express Nox2 catalytic (gp⁹¹phox and p²²phox) and cytoplasmic (p⁶⁷phox, p⁴⁷phox, and p⁴⁰phox) subunits, as well as Rac1, which controls subunit assembly. In this prior study, single-channel recordings treated with the Rac1 inhibitor NSC23766 significantly decreased production of O₂⁻ with a concomitant decrease in lung ENaC NPo. In the present study, we examined whether redox regulation of ENaC indeed translated into physiological changes in lung fluid volume. Interestingly, we found that Rac1 inhibition of Nox2 activity led to significant increases in lung fluid volume. In fact, compared with amiloride-instilled lungs, the NSC23766-treated lungs continued to take in water during the 240-min observation period. This suggests that Rac1, and perhaps other small-molecular-weight GTPases, may be regulating additional pathways or ion channels needed for reabsorption. As an alternative to pharmacologically inhibiting NADPH oxidase activity using NSC23766, we delivered saline challenges into the distal airways of Nox2-deficient mice. Similar to the observed effect in NSC23766-treated animals, disruption of normal NADPH oxidase activity significantly impaired fluid clearance, albeit at a later time point after IT challenge (between 95 and 240 min) and to a lesser extent than NSC23766. One possible explanation for the difference in response between NSC23766 and Nox2^−/− animals is that other Nox isoforms may be compensating for the lack of normal Nox2 isoform in transgenic mice, whereas NSC23766 inhibition of Rac1 has the potential to alter all Rac-dependent Nox isoforms (Nox1–2) (20). In support of this, we have detected Nox1 and Nox4 expression in the alveolar epithelium using antibodies obtained from Dr. David Lambeth (Emory University) and Novus Biological (Littleton, CO), respectively (unpublished results). However, Nox isoform expression in the alveolar epithelium has not been systematically reported.

Using X-ray fluoroscopy, we observed a significant difference in the rate of fluid clearance in 12-wk-old animals (Fig. 3) and 8-wk-old animals (Fig. 5). We found that either younger animals were reabsorbing salt and water at a faster rate or the secretory pathways were less active in the 8-wk-old animals as opposed to 12-wk-old mice. Given the potential for age-related differences in lung fluid balance, we were careful to perform the appropriate age-matched controls for all of our experiments. Although beyond the scope of our present studies related to ROS regulation of lung ENaC and fluid balance, an additional advantage to using X-ray fluoroscopy to assess lung fluid clearance is that it offers the potential to track age-related changes in the capacity for lung fluid clearance in the same animals over time.

LPS regulates NSC and HSC channels and influences fluid clearance in the lungs.

In addition to showing the physiological role of Rac-dependent NADPH oxidase activity, our studies provide mechanistic insight into the effect of LPS on sodium channel activity. As shown in Fig. 6 by in vivo imaging, LPS immediately impedes lung fluid reabsorption (up to ∼75 min), yet lung fluid begins to steadily clear between 75 and 240 min following IT challenge. Moreover, TEMPO, a compound that sequesters ROS, antagonized this effect. These observations, coupled with single-channel analysis of the effect of LPS on ENaC expressed by AT1 cells, allowed us to correlate the effects of LPS on ion transport function with the acute effects of LPS on lung fluid clearance in vivo.

Prior to our studies, Boncoeur et al. (6) and Baines et al. (2) have shown that LPS inhibits highly selective amiloride-sensitive current in primary rat AT2 cells and human adenocarcinoma epithelial cells (H441) with characteristic type 2 phenotype. LPS inhibition of highly selective sodium channels (HSC) could indeed explain why, in our studies, tracheal instillation of LPS immediately decreased the rate of lung fluid clearance in C57BL/6J lung. Because HSC channels have Na⁺/K⁺ selectivity >40, inactivation of this family of channels may well lead to attenuation of lung fluid clearance at early times following IT instillation.

However, we also found that LPS stimulated nonselective cation channels in AT1 cells. Using single-channel recordings, Baines et al. (2) similarly reported that LPS increases the activity of NSC ENaC channels expressed by H441 airway cells (2). Likewise, Dodrill and Fedan (8) utilized Ussing chamber measurements to show that LPS increases Na⁺ transport activity in airway epithelium. Activation of nonselective cation channels, with larger conductances of 19–24 pS, could explain the latent clearance of lung fluid in LPS-instilled animals between 75 and 240 min (Fig. 6). Because TEMPO inhibited this response, we conclude that LPS-mediated activation of nonselective sodium channels is ROS dependent. By combining in vivo imaging with single-channel measurements of ENaC, we are able to uniquely assess the effect of potentially therapeutic, or harmful, compounds and to gain understanding of the underlying molecular mechanisms.

Overall conclusion and clinical implications.

We used three independent methods to disrupt superoxide release: TEMPO, Rac-1 inhibitor compound (NSC23766), and Nox2-deficient mice to show that decreases in O₂⁻ release lead to significant disruption of normal lung fluid balance. We measured these changes in vivo using a fluorometric assay, which provides a noninvasive alternative to measuring lung fluid volume in small rodents with high sensitivity. Moreover, our in vivo findings validated single-channel recordings showing that TEMPO and NSC23766 attenuated lung ENaC activity by demonstrating that this can disrupt the osmotic driving force needed for water reabsorption. Given the important role for NADPH oxidases in regulating alveolar fluid homeostasis, our data suggest that Nox (and Nox subunits) are likely to be key therapeutic targets with the capacity to reverse edematous lung injury or to promote fluid balance in hyperreabsorptive disorders, such as cystic fibrosis.

GRANTS

This research was supported from National Institutes of Health Grants K99R00 HL09222601 awarded to M. N. Helms and P50-AA-013757 and R01-HL-083120 awarded to M. Koval.

DISCLOSURES

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

CONTRIBUTIONS

P.G., L.J., M.K., and M.H. conception and design. P.G., M.K., A.K, and M.H. manuscript preparation. K.K and N.M. in vivo detection of ROS. P.G. and M.H. experiments and analysis.