Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The aquaporins (AQPs) are a family of water-selective channels that function to increase plasma membrane water permeability and thus provide a route for rapid fluid movement (1). AQPs are expressed in many tissues such as the kidney, eye, and lung, where the rapid regulated transport of water is necessary (2). Several known AQPs are expressed in the lung (3): AQP1 in the peribronchiolar and alveolar endothelia and in the visceral pleura, AQP3 in the trachea, AQP4 in airway epithelia and in the trachea, and AQP5 at the apical membrane of type I alveolar epithelial cells. A number of pathologic conditions that affect the lung are characterized by disrupted fluid transport, including congestive heart failure, respiratory distress syndrome, and pulmonary edema resulting from injury or infection (9). Pulmonary edema formation and resolution involve fluid movement between the air space, interstitial, and capillary compartments. The specific localization of AQPs to both endothelial (AQP1) and epithelial (AQP5) cells supports a role for AQPs in this process. In addition, recent physiologic studies have demonstrated that both the alveolar (3) and microvascular (4) barriers have high water permeabilities, with biophysical properties indicative of water movement through AQPs. Osmotic water permeability has been determined for isolated alveolar type I cells, which express AQP5, and was found to be the highest reported for any mammalian cell membrane to date (3). The specific involvement of AQP1 in air space- capillary fluid movement was recently demonstrated in AQP1 knockout mice, which were found to have significantly reduced alveolar-capillary fluid movement as well as a more than 10-fold reduction in microvascular endothelial water permeability (10, 11).

Recent evidence has suggested a role for AQPs in the maintenance of fluid homeostasis, and that alterations in AQP expression through mutations, gene regulation, or tissue damage lead to pathophysiologic conditions (10- 16). For example, mutations in AQP2 in the kidney have been demonstrated to result in a human disease, namely nephrogenic diabetes insipidus (NDI) (12). Reduced levels of AQP2 protein are associated with several models of acquired NDI, and increased AQP2 messenger RNA (mRNA), protein, and membrane targeting have been demonstrated in many pathologic conditions in which water is retained (for reviews, see References 1 and 12-15). In addition, rats with chronic renal failure have reduced levels of AQP1, AQP2, and AQP3 (13, 16). Thus, there is substantial experimental evidence that supports the importance of a role for AQPs in fluid regulation in normal and pathophysiologic states. However, although many studies have been conducted on AQP regulation in the context of pathophysiologic states in the kidney, very few studies have addressed the regulation of AQPs under pathologic conditions in the lung.

Fluid movement through AQPs is governed by an osmotic gradient that is generated by transcellular ion movement through ion channels and pumps. The general paradigm for fluid movement in the lung is that active sodium transport drives osmotic water transport (17). Therefore, fluid movement in the lung is likely dependent upon both sodium and water transporters in the alveolar epithelium and capillary endothelium in the distal lung. Mice deficient in the  subunit of the epithelial sodium channel (ENaC), which is expressed in alveolar type II cells, die within 40 h after birth due to an inability to clear fluid in the lung (18). In addition, treatment of lungs with amiloride (inhibits ENaC) or ouabain (inhibits Na,K-adenosine triphosphatase [NKA]) severely decreases the rate of fluid clearance from the lungs (17, 19). Thus, under conditions of altered fluid movement in the lung, ENaC and the NKA may undergo altered regulation.

Pulmonary inflammation generally results in excess fluid accumulation manifested as increased-permeability pulmonary edema. Viral infection with adenoviruses and adenoviral vectors is a well-characterized model of acute lung inflammation (20). Experimentally, the intratracheal administration of adenovirus to mice is associated with a readily discernible inflammatory response in the lung, including both humoral and cell-mediated immune responses (23). The use of recombinant adenoviral vectors provides a useful in vivo model for studying the pathophysiology of viral infection of the lung (20, 21).

The cellular colocalization of AQP1 and AQP5 in the distal lung, the observed decrease in permeability in AQP1 knockout mice, and the permeability measurements of alveolar type I cells support a role for AQP1 and AQP5 in the maintenance of water homeostasis in the lung. Several AQPs in the kidney have been demonstrated to undergo altered gene regulation under pathologic conditions. For these reasons, we hypothesized that the gene expression of AQPs 1 and 5 in the lung is subject to inflammation/ edema-associated regulation. In the present study we examined the expression of AQP1 and AQP5 in the context of pulmonary inflammation and edema resulting from intratracheal infection with adenovirus. Expression of ENaC ( subunit) and the NKA (1 subunit) were also determined as indicators of genes involved in transcellular ion movement after infection with adenovirus. To the best of our knowledge, this study is the first report of the regulation of AQP1 and AQP5 gene expression in an in vivo model of pulmonary edema and provides the first demonstration of the decreased expression of AQP1 and AQP5 in the lung.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Adenovirus Vector Administration to the Mouse Respiratory Epithelium

Six- to 12-wk-old FVB/N wild-type mice (n = 5 mice per group) were used. Intratracheal administration of the adenoviral vector Av1Luc1 was conducted as described previously (22). Briefly, mice were anesthetized with methoxyflurane vapor and an anterior midline incision was made to expose the trachea. Intratracheal inoculation of 1 × 10⁹ plaque-forming units (PFU) of Av1Luc1, an E1-E3-deleted adenoviral vector expressing firefly luciferase from the Rous sarcoma virus promoter, in 100 µl of delivery vehicle (10 mM Tris, 1 mM MgCl₂, and 10% glycerol, pH 7.4) was performed using bent, 27-gauge tuberculin syringes (Monoject; Sherwood Medical, St. Louis, MO). The incision was closed with one drop of Nexaband liquid and the mice were allowed to recover. Mice recovered rapidly and remained active after the procedure. At 7 and 14 d after infection, mice were killed by lethal injection of sodium pentobarbital. A midline incision was made in the abdomen and mice were exsanguinated by transection of the inferior vena cava to reduce hemorrhage in the lung. Right upper, middle, and lower lobes were clamped with hemostats and removed for analysis of lung protein and RNA. The left lung was inflated with 4% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, PA) and fixed overnight for histologic examination.

Reverse Transcriptase/Polymerase Chain Reaction of Inflammatory Mediators Tumor Necrosis Factor- and Interferon-

Total RNA was isolated using TriReagent (Molecular Research Center, Inc., Cincinnati, OH) as per manufacturer's instructions from lung homogenates of uninfected mice and mice 7 and 14 d after infection with Av1Luc1. RNA (0.5 µg per sample) was converted to complementary DNA (cDNA) by the oligo dT-primed reverse transcriptase (RT) reaction according to the manufacturer's protocol (Superscript II Reverse Transcriptase; GIBCO BRL, Gaithersburg, MD). Reverse-transcribed cDNA products were amplified by the polymerase chain reaction (PCR) with primers specific for tumor necrosis factor (TNF)-, interferon (IFN)-, or -actin (see Table 1).

PCR reactions consisted of 10 pmol of each primer, 0.25 mM of each deoxynucleotide triphosphate, either 2.0 mM MgCl₂ (TNF-) or 2.5 mM MgCl₂ (IFN- and -actin), either 60 mM Tris-HCl (pH 9.0) (TNF-) or pH 8.5 (IFN- and -actin), 12.5 mM (NH₄)₂SO₄, and 0.1 U Taq polymerase (GIBCO BRL) in a total reaction volume of 20 µl. PCR was performed in an MJ PTC-100 Thermocycler under the following conditions: 94°C for 2 min, then 30 cycles of 92°C for 30 s, 58°C for 1 min and 72°C for 1 min. The amplification products were separated by agarose gel electrophoresis and visualized by staining with ethidium bromide.

Wet-to-Dry Weight Ratios

Lung tissue was excised from 10-wk-old uninfected mice and mice 7 and 14 d after Av1Luc1 infection (n = 5-8 mice per group), and weighed to obtain lung "wet" weights. Lungs were placed in a 60°C oven with desiccant and weighed after 3 to 5 d. Dry weights were obtained after the weights no longer changed on successive days.

Generation of AQP1, AQP5, ENaC, NKA1, and Glyceraldehyde-3-Phosphate Dehydrogenase Screening Probes

The AQP1 probe was obtained by a PCR amplification of mouse genomic DNA using heterologous primers designed to the rat CHIP28 mRNA sequence (GenBank X71069). AQP5, ENaC, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probes were obtained from mouse lung cDNA. Oligo dT-primed RT-PCR was performed on mouse lung RNA according to the manufacturer's protocol (Superscript II Reverse Transcriptase; GIBCO BRL). RT products were PCR-amplified with primers specific for AQP5, ENaC, and GAPDH sequences (see Table 2). The NKA1 probe was kindly provided by Dr. J. Lingrel, University of Cincinnati (25).

PCR reactions were conducted as described earlier with an annealing temperature of 58°C for all primer sets except AQP5, for which 60°C was determined to be optimal. Buffer conditions were as stated above for the TNF- primers except for the AQP1 primers, for which 3.5 mM MgCl₂ and pH 8.5 were determined to be optimal. All other components were held constant for all PCR reactions. The GAPDH probe was used as a control for equal loading of RNA samples in Northern hybridization.

Northern Blot Analysis

Total RNA was isolated from mouse lungs using TriReagent solution (Molecular Research Center, Inc.) as per manufacturer's instructions. RNA concentrations were determined via spectrophotometry and confirmed with agarose gel electrophoresis. The amount of 15 µg of total RNA per sample was heat-denatured at 55°C for 20 min in formaldehyde and formamide and size-fractionated on a 1% formaldehyde agarose gel in 1× 3-(N-morpholino)propanesulfonic acid buffer for 4.5 h at 85 V. The gel was transferred overnight in 10× saline sodium citrate (SSC) to Hybond N+ nylon membrane and ultraviolet-crosslinked. The blot was prehybridized in ExpressHybe buffer (Clontech, Palo Alto, CA) for 30 min at 68°C and hybridized at 68°C for 1 h with 2 × 10⁶ cpm/ml of ³²P random-labeled probe (Random primer; GIBCO BRL). The membrane was washed once with 2× SSC and 0.05% sodium dodecyl sulfate (SDS) for 30 min at room temperature, washed twice at 60°C for 20 min with 0.1× SSC and 0.1% SDS, and exposed to Kodak-MS film (Eastman Kodak, Inc., Rochester, NY) overnight at 70°C with an intensity screen. The membrane was stripped in boiling 0.5% SDS for 30 min and re-probed. As controls for loading of total RNA, the gel was examined by staining with ethidium bromide and the membrane was probed with an ³²P-labeled cDNA probe of mouse GAPDH (described previously).

Protein Preparation

Total membrane proteins were isolated from lungs of uninfected mice and mice 7 and 14 d after adenoviral infection. Right-lung tissue was isolated and placed in chilled membrane isolation solution containing 250 mM sucrose, 10 mM triethanolamine (Sigma, St. Louis, MO), 1 µg/ml leupeptin (Sigma), and 0.1 mg/ml phenylmethylsulfonyl fluoride (Sigma) adjusted to pH 7.6. Whole tissues were homogenized in 3 ml of membrane isolation solution.

Homogenates were centrifuged at 4,000 × g for 10 min at 4°C. A total of 2 ml of isolation solution was added to the pellet and the pellet was rehomogenized and centrifuged to increase yields. The supernatants were pooled and centrifuged at 200,000 × g for 1 h. The resulting pellets were resuspended in 250 µl of isolation solution. The total protein concentration in each sample was measured using the Pierce BCA protein assay reagent kit (Pierce, Rockford, IL) according to manufacturer's protocol.

Antibodies

The peptide-derived, affinity-purified rabbit polyclonal antibody specific to the COOH-terminal domain of AQP1 (LL266) was a generous gift of Dr. M. Knepper, NIH, Bethesda, MD (28). A rabbit polyclonal antibody (LL639; Lofstrand Labs, Gaithersburg, MD) was generated against a synthetic peptide corresponding to the mouse AQP5 COOH-terminal residues N-CEPEEDWEDHREERKKTIELTAH-COOH and affinity-purified on a SulfoLink column (Pierce) conjugated with the immunizing peptide (29). The antihuman surfactant protein C propeptide (proSPC) antibody is a rabbit polyclonal antibody directed against human proSPC (30) and was a generous gift of Dr. J. Whitsett, Children's Hospital Medical Center, Cincinnati, OH.

Western Blot Analysis

Total membrane proteins (5-10 µg/sample) were solubilized in Laemmli sample buffer and boiled for 5 min. SDS-polyacrylamide gel electrophoresis (MiniProtean II apparatus; Bio-Rad, Hercules, CA) was performed using 12% pH 8.8 separating gel and 4% pH 6.8 stacking gel. Electrotransfer was carried out for 1.5 h on ice onto polyvinylidene difluoride (PVDF) membranes (SequiBlot; Bio-Rad). Membranes were blocked for 1 h at room temperature in 1% blocking solution (Boehringer Mannheim, Indianapolis, IN), followed by incubation overnight at 4°C either with anti-AQP1 antibody at a dilution of 0.25 µg/ml or with anti-AQP5 antibody at a dilution of 0.5 µg/ml in 0.5% blocking solution (Boehringer Mannheim). After washing in TBST (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, and 0.1% Tween 20) and 0.5% blocking solution, the membranes were incubated with 50 mU/ml horseradish peroxidase (POD)-labeled antimouse antirabbit secondary antibody (Boehringer Mannheim) for 1 h at room temperature, washed four times in TBST, and visualized via enhanced chemiluminescence (SuperSignal; Pierce) with variable film exposures (Kodak-MS film). To confirm equivalent loading of samples, PVDF membranes were stripped of the antibodies after probing in Tris-buffered saline containing 100 mM -mercaptoethanol and 2% SDS at 50°C for 30 min, washed in TBS, and stained with 0.3% india ink in TBS for ~ 2 h.

Histology and Immunohistochemistry

Histopathologic changes and AQP1 and AQP5 immunohistochemical staining were evaluated in inflation-fixed mouse lungs from uninfected mice and mice 7 and 14 d after administration of Av1Luc1. Inflation-fixed lungs were washed in phosphate-buffered saline (PBS) three times, and bisected for paraffin embedding. Paraffin-embedded lungs were sectioned at 5 µm, and stained with hematoxylin and eosin (H&E) for morphologic analysis. For immunohistochemical staining of AQP1 and AQP5, tissue sections were deparaffinized by washing in xylene three times for 10 min each, followed by rehydration through a series of ethanol washes from 100 to 70% ethanol. Slides were placed in methanol containing 0.5% hydrogen peroxide for removal of endogenous peroxidase activity. Nonspecific binding was blocked by incubation of the slides for 1 h at room temperature in 0.1 M PBS, pH 7.4, containing 0.2% Triton X and 2% normal goat serum. Lung sections were incubated with either anti-AQP1 antibody or anti-AQP5 antibody (both at dilutions of 0.5 or 0.25 µg/ml) overnight at 4°C. Titration of both antibodies was conducted from 0.1 to 1.0 µg/ml to determine optimum antibody concentrations. Sections were rinsed five times in 0.1 M PBS-0.2% Triton X and incubated for 30 min at room temperature with biotinylated goat antirabbit secondary antibody (Vector Labs, Burlingame, CA), followed by incubation with an avidin-biotin complex (Vector Labs) for 30 min at room temperature. Sections were washed with PBS, rinsed briefly in 0.1 M acetate buffer (pH 6.0), incubated with diaminobenzidine for 4 min, and counterstained with 0.1% nuclear fast red in 5% aluminum sulfate. The following labeling controls were performed under the same conditions described earlier: (1) the primary antibody was substituted with preimmune rabbit immunoglobulin G purified on a protein A column (Pierce); and (2) incubations were carried out without either the primary or secondary antibody.

Alveolar type II cells were identified by immunohistochemical staining for proSPC, following the same protocol as with the AQP antibodies except that blocking of nonspecific binding was conducted for 2 h and the antibody was utilized at dilutions of 1:1,000 and 1:2,000.

Quantitation of mRNA Levels

Northern blots probed with the specified ³²P-labeled probe were quantitated by exposure of a phosphor screen, scanned by means of a Storm 840 scanner, and analyzed using ImageQuant software (all from Molecular Dynamics, Sunnyvale, CA). Values were corrected by quantitation of the GAPDH values and were expressed as a probe/GAPDH ratio. Densitometry results are reported as volume integrated values and expressed in percentages compared with the mean value in controls (100%).

Quantitation of AQP1 and AQP5 Protein Levels

Films of Western blots were scanned using a Hewlett-Packard scanner and Adobe Photoshop software. The scanning was performed using enhanced chemiluminescence exposures that gave control bands in the lower gray scale. The labeling density was quantitated using ImageQuant software (Molecular Dynamics). Membranes stained with india ink after probing and stripping were scanned and quantitated using the average volume of three protein bands for each sample. These values were used to correct for differences in loading between samples. Densitometry results are reported as volume integrated values, and are expressed in percentages compared with the mean value in controls (100%), corrected by total protein concentration.

Statistical Analysis

Statistical analyses of wet/dry weights, AQP1/GAPDH, AQP5/GAPDH, ENaC/GAPDH, and NKA1/GAPDH density ratios for RNA expression, and AQP1 and AQP5 density ratios corrected for loading differences were performed using unpaired Student's t test with equal variance. Results are expressed as means ± standard error (SE). A P value of < 0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary Edema and Inflammation after Av1Luc1 Administration

Lung histology was used to assess pulmonary edema and inflammation in mice 7 and 14 d after intratracheal administration of Av1Luc1 (Figure 1). Pulmonary infiltrates were observed in peribronchial and perivascular regions 7 and 14 d after infection (Figures 1B and 1C). Diffuse inflammatory cell infiltration was detected in sparse areas of the lung parenchyma. However, the majority of the parenchyma of the lung was normal in appearance and the general integrity of the lung appeared to remain intact both 7 and 14 d after adenoviral infection (Figures 1B and 1C). Enlarged peribronchial and perivascular spaces were observed 7 and 14 d after infection, consistent with pulmonary edema (Figures 1B and 1C). Alveolar interstitial edema was detected as thickened alveolar septae in some regions (not shown). Evidence of pulmonary edema was generally decreased in the lungs of mice 14 d after infection as compared was the lungs of mice 7 d after infection. Thus, histologic examination of mouse lung sections indicates that mice develop moderate pulmonary inflammation as well as pulmonary edema at both 7 and 14 d after infection with adenovirus.

View larger version (74K): [in this window] [in a new window]   Figure 1.   Pulmonary histology 7 and 14 d after adenoviral infection. Lung sections from an uninfected mouse (A) and mice 7 (B) and 14 (C ) d after administration of 109 PFU of Av1Luc1 were stained with H&E. Peribronchovascular and parenchymal inflammation (arrowheads) and enlarged spaces surrounding airways and vessels (arrows) were observed in mice 7 and 14 d after infection with Av1Luc1. Presented figures are representative of three sections examined from five mice at each time point. Bars = 50 µm. A: airways; V: vessels.

To quantitatively assess pulmonary edema after infection with Av1Luc1, whole-lung wet/dry weight ratios were determined. The ratio of wet-to-dry weights of whole lungs from mice after infection with adenovirus was significantly increased both at 7 (121 ± 6%, n = 5, P < 0.02) and 14 d (112 ± 1%, n = 5, P < 0.002) after infection when compared with uninfected mice (100 ± 1%, n = 5) (Figure 2). Lung wet/dry weights in mice 7 d after infection were slightly greater than in mice 14 d after infection, but the difference did not reach statistical significance (P < 0.2).

View larger version (36K): [in this window] [in a new window]   Figure 2.   Increased lung fluid after adenoviral infection. Wet/dry weight ratios of whole lungs were determined from uninfected mice and mice 7 and 14 d after infection with Av1Luc1. Values are expressed as grams of wet weight/grams of dry weight and are means ± SE, n = 5 for each group. P values refer to comparisons of wet/dry weights for either Day 7 or Day 14 after infection with the wet/dry weights of uninfected mice.

TNF- and IFN- mRNAs were analyzed by RT-PCR to provide molecular evidence of inflammation in the lungs of mice after infection with Av1Luc1. TNF- and IFN- mRNAs were undetectable in the uninfected mouse lungs, but readily detectable in the lungs of mice 7, and to a lesser extent 14, d after infection with adenovirus (Figure 3).

View larger version (47K): [in this window] [in a new window]   Figure 3.   Lung TNF- and IFN- mRNA after adenoviral infection. Total RNA was extracted from mouse lungs (uninfected and 7 and 14 d after administration of 109 PFU of Av1Luc1; n = 5 for each group), and 0.5 µg/sample were subjected to the RT reaction followed by PCR with primers specific for TNF-, IFN-, and -actin. RT-PCR products were analyzed via ethidium-stained agarose gel electrophoresis. -actin was used as a control for the amount of RNA in each sample.

AQP1, AQP5, ENaC, and NKA1 mRNA Expression after Pulmonary Adenoviral Infection

Northern blot analysis of total RNA was used to determine the level of AQP1, AQP5, ENaC, and NKA1 mRNA in the lungs of mice 7 and 14 d after adenoviral infection. Hybridization with an AQP1-specific probe demonstrated transcripts at 3.1 and 1.35 kb, which corresponds to the previously reported size of mouse AQP1 mRNA (31) (Figure 4A). Densitometry of the predominant 3.1-kb band revealed decreased AQP1 mRNA 7 d after infection with adenovirus to 31 ± 7% of uninfected control animals (n = 5, P < 0.0004) (Figure 4B). AQP1 mRNA was also significantly decreased in the lungs of mice 14 d after infection with adenovirus to 48 ± 7% of uninfected animals (n = 5, P < 0.004) (Figure 4B). Hybridization of the Northern blots with an AQP5-specific probe demonstrated a transcript of 1.8 kb, which corresponds to the previously reported size of mouse AQP5 mRNA (29) (Figure 4A). AQP5 mRNA was significantly reduced to 62 ± 6% of uninfected mice (n = 5, P < 0.01) in the lungs of mice 7 d after infection, and decreased to 53 ± 5% of uninfected mice (n = 5, P < 0.002) 14 d after infection with adenovirus (Figures 4A and 4B). Because the movement of water through AQPs is osmotically driven, we quantitated the amount of ENaC mRNA and NKA1 mRNA. Northern blot analysis with an ENaC-specific probe revealed a transcript of approximately 3.8 kb, corresponding to the size of ENaC mRNA (19) (Figure 4A). ENaC mRNA was decreased 7 and 14 d after infection with adenovirus (Figures 4A and 4B). The decrease in ENaC mRNA in the lungs of mice 7 d after infection (39 ± 6% of uninfected mice, n = 5, P < 0.0004) was greater than the decrease 14 d after infection (47 ± 6% of uninfected mice, n = 5, P < 0.0007) (Figure 4B). NKA1 mRNA was not significantly altered 7 or 14 d after infection with adenovirus (Figures 4A and 4B). In summary, AQP1, AQP5, and ENaC mRNA were significantly decreased in the lungs of mice 7 and 14 d after infection with adenovirus, whereas NKA1 mRNA levels remained unchanged.

View larger version (48K): [in this window] [in a new window]   Figure 4.   AQP1, AQP5, ENaC, and NKA1 mRNA after adenoviral infection. Total RNA was extracted from mouse lungs and 15 µg/lane was subjected to electrophoresis. (A) Northern blot analysis of AQP1, AQP5, ENaC, NKA1, and GAPDH mRNA in the lungs of uninfected mice and mice 7 and 14 d after infection. Time (in days after infection) is indicated at the top. The same membrane was rehybridized with the cDNA probe for mouse GAPDH to correct for the amount of mRNA loaded. (B) Densitometry of Northern blot results shown in A for uninfected mice (n = 3) and mice 7 (n = 5) and 14 (n = 5) d after infection. The densitometric values were quantified as a ratio of the expression of each gene normalized for the expression level of GAPDH for each sample as an internal loading control. Values are presented as percentages of the mean uninfected values and are means ± SE. P values refer to comparisons of values for either Day 7 or Day 14 after infection with the uninfected values.

AQP1 and AQP5 Protein Expression after Pulmonary Adenoviral Infection

To determine whether the decrease in AQP1 and AQP5 mRNA also results in an altered level of protein, Western blot analysis of membrane proteins was used to assess AQP1 and AQP5 protein levels in mouse lungs 7 and 14 d after infection with Av1Luc1. Using an antibody directed against an AQP1 carboxy-terminal synthetic peptide, Western blot analysis revealed bands at 28 and 35 to 40 kD, indicating native and glycosylated AQP1 protein (28) (Figure 5A). Densitometry of the 28-kD AQP1 band revealed a significant decrease in AQP1 protein levels 7 d after infection with adenovirus to 40 ± 11% of uninfected control levels (n = 5, P < 0.004) and a similar decrease 14 d after infection to 44 ± 8% of uninfected control levels (n = 5, P < 0.002) (Figure 5B). AQP5 protein expression was analyzed by Western blot analysis using an antibody prepared against an AQP5 carboxy-terminal synthetic peptide (29) (Figure 5A). Western blot analysis revealed a band at 27 kD, indicating AQP5 protein (29). AQP5 protein levels were significantly decreased to 54 ± 11% of uninfected mice (n = 5, P < 0.03) at 7 d after infection with Av1Luc1. The level of AQP5 protein was further decreased 14 d after infection to 27 ± 10% of uninfected mice (n = 5, P < 0.002) (Figure 5B).

View larger version (29K): [in this window] [in a new window]   Figure 5.   Lung AQP1 and AQP5 protein after infection with adenovirus. (A) Representative Western blot analyses of AQP1 and AQP5 protein expression in the lungs of uninfected mice and mice 7 and 14 d after infection with Av1Luc1. Immunoblots of total membranes (10 µg/lane) probed with affinity-purified anti-AQP1 identified 28- and 35- to 40-kD bands, corresponding to the native and glycosylated forms of AQP1. Use of the affinity-purified anti-AQP5 antibody detected a 27-kD band, corresponding to AQP5 protein. (B) Densitometry of all Western blot results of mouse lungs from uninfected mice (n = 4) and mice 7 (n = 5) and 14 (n = 5) d after infection with Av1Luc1. Western blot membranes were stripped and stained with india ink subsequent to hybridization for use as loading controls (not shown). AQP1 and AQP5 values are expressed as a ratio of AQP:protein loading. Densitometric values are presented as percentages of the mean values for the uninfected mouse lungs and are means ± SE. P values are given for the infected mouse lungs (either 7 or 14 d after infection) in comparison with the values for the uninfected mouse lungs.

Immunohistochemical Localization of AQP1 and AQP5 Protein in Infected and Uninfected Mouse Lungs

Immunohistochemistry was used to analyze AQP1 and AQP5 localization and expression in the lungs of mice 7 and 14 d after Av1Luc1 infection. AQP1 has previously been shown to localize to the pulmonary vascular endothelium throughout the parenchyma of the lung and surrounding vessels and airways as well as in the visceral pleura (4, 8). The localization and cell type-specific staining of AQP1 remained unchanged in infected and uninfected mice (Figures 6A, 6C, and 6E). By light microscopy, distal lung AQP1 was decreased in mice infected with Av1Luc1 for 7 d when compared with uninfected mice (n = 5) (Figures 6A and 6C). The decrease in AQP1 was observed throughout the lung and not localized to regions of intense inflammation. In mice 14 d after infection, AQP1 protein was reduced throughout the lung in comparison with uninfected mice (n = 5) (Figures 6A and 6E). Titration of the AQP1 antibody revealed decreased AQP1 staining in infected lungs at all effective antibody concentrations (data not shown).

View larger version (121K): [in this window] [in a new window]   Figure 6.   (above). Immunohistochemical localization of AQP1 (A, C, and E ) and AQP5 (B, D, and F ) protein after infection with adenovirus. Paraffin sections (5 µm) of lungs of uninfected mice (u.i.) (A and B), mice 7 d after infection with adenovirus (d.7) (C and D), and mice 14 d after infection (d.14) (E and F ). A, C, and E are stained with anti-AQP1 antibody; and B, D, and F are stained with anti-AQP5 antibody. The arrows depict AQP1 labeling of the alveolar region (A, C, and E ) and AQP5 labeling of the alveolar type I cells (B, D, and F ). Presented figures are representative of three sections examined from five mice at each time point. Bars = 25 µm.

AQP5 staining was localized to the alveolar regions of the lung in uninfected as well as infected mice (Figures 6B, 6D, and 6F). Reduced AQP5 staining was observed in mice infected with Av1Luc1 for 7 (n = 5; Figure 6D) and 14 (n = 5; Figure 6F) d when compared with uninfected mice (n = 5; Figure 6B). The reduction in AQP5 staining in the lungs of mice 7 and 14 d after infection with adenovirus was observed throughout the lung parenchyma regardless of inflammatory cell infiltration. The decrease in AQP5 staining in the lungs of mice was more prominent at 14 d than at 7 d after infection. AQP5 staining was decreased in infected mice at all concentrations of antibody tested.

Lung sections were also stained with an antibody directed against proSPC, which is localized to alveolar type II cells (30). Immunohistochemical analysis demonstrated that proSPC staining was not different in uninfected mouse lungs and mouse lungs 7 (not shown) and 14 d after infection (Figure 7), providing evidence that excessive alveolar type II cells in infected mice likely did not contribute to the observed decrease in AQP5 staining.

View larger version (61K): [in this window] [in a new window]   Figure 7.   (right). Immunohistochemical staining of alveolar type II cells with anti-proSPC antibody after infection with adenovirus. Paraffin sections (5 µm) of lungs of uninfected mice (u.i.) (A) and mice 14 d after infection with adenovirus (d.14) (B). The arrows depict proSPC labeling of alveolar type II cells. Figures are representative of two sections examined from five mice at each time point. ProSPC staining of lung sections from mice 7 d after infection was similar to 14 d after infection and uninfected mice (not shown). Bars = 25 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Several lines of evidence suggest that AQPs play a role in lung physiology: AQPs are expressed in lung epithelia and endothelia (3, 8), and this expression and function increase at birth when the lung must rapidly absorb fluid (7, 32); water permeability is high in lung epithelia, where AQP5 is expressed (3), and endothelia, where AQP1 is expressed (4); and endothelial fluid movement is significantly reduced in AQP1 knockout mice (11). In certain pathologic conditions such as adult respiratory distress syndrome, congestive heart failure, and viral infections, fluid homeostasis in the lung is severely disrupted and results in pulmonary edema (9). Multiple studies have demonstrated alterations in the gene expression of AQPs in the kidney in association with pathologic conditions and disrupted fluid homeostasis. Analysis of the gene expression of AQP1 and AQP5 in the context of pathologic conditions in the lung enables examination of the possible regulation of AQPs in the lung similar to that demonstrated for AQP2 in the kidney. The purpose of the present study was to determine whether altered gene expression of distal lung water channels and sodium transporters is associated with pulmonary inflammation and edema.

Inflammation is one of the major causes of edema formation in the lung. Intratracheal adenoviral administration results in subacute pulmonary inflammation similar to that seen in other models of viral infection and acute lung inflammation (33). Previous studies using this model of adenoviral infection have demonstrated parenchymal and peribronchovascular inflammation after adenoviral infection (21, 33). In the present study, moderate inflammation surrounding airways and vessels and sparse alveolar inflammation were observed in response to adenoviral infection. In addition, increased mRNA expression of both TNF- and IFN- was demonstrated in the lungs of mice 7 d after infection with adenovirus and to a lesser extent 14 d after infection. Evidence of expression of these proinflammatory cytokines suggests there is good agreement between molecular and cytologic markers of inflammation in this model and provides additional documentation that mice develop pulmonary inflammation after infection with adenovirus.

A second feature of adenoviral-infected mouse lungs is pulmonary edema, which represents a pathophysiologic manifestation of fluid dysregulation. Pulmonary edema was detected by histologic and wet/dry weight analyses in mouse lungs 7 and 14 d after adenoviral infection. Enlarged peribronchial and perivascular spaces were observed in the lungs of mice both 7 and 14 d after infection, consistent with the presence of pulmonary edema. Alveolar wall thickening was also evident after infection in highly inflamed areas of the lung parenchyma. Lung wet/dry weight ratios were significantly increased in mice 7 and 14 d after adenoviral infection in comparison with uninfected mice. These histologic findings and the increase in wet/dry weights are consistent with the presence of pulmonary edema in mice after infection with adenovirus.

The expression of lung AQP1 and AQP5 mRNA and protein was significantly decreased in the context of pulmonary inflammation and edema resulting from infection with adenovirus. AQP1 mRNA and protein were decreased in the lungs of mice both 7 and 14 d after adenoviral infection, but the decrease was less at 14 d, indicating that AQP1 expression was returning to baseline levels coincident with a lessening of the inflammatory response. Likewise, AQP5 mRNA and protein were decreased both 7 and 14 d after infection; however, AQP5 mRNA did not appear to be returning to normal levels 14 d after infection, indicating that the reduction in AQP5 persisted beyond the decrease in AQP1. Previously, AQP1 expression was demonstrated to increase in the lung in response to corticosteroids (7) and under cardiac-related stress (34). This study is the first report of the coordinated regulation of AQP1 and AQP5 in an in vivo model of pulmonary inflammation and associated edema, and is the first demonstration of AQP5 regulation in the lung.

Immunohistochemical analyses of AQP1 and AQP5 demonstrated a reduction in AQP1 and AQP5 protein throughout the lungs of mice both 7 and 14 d after infection. AQP1 and AQP5 staining was decreased but still detectable in highly inflamed regions of the lung, such as the peribronchial and perivascular regions. AQP1 and AQP5 staining was decreased to a similar extent in lung regions in which inflammation was mild or not detectable, suggesting that lung injury per se is not the direct cause of the observed decrease in expression of both AQPs. This conclusion is supported by the previously reported demonstration that the alveolar epithelial barrier can remain intact and functional even though large numbers of inflammatory cells migrate through the tight junctions of the alveolar epithelium (17). Likewise, the widespread decrease in AQP1 and AQP5 staining also suggests that the reduction in AQP expression is not likely due to local inflammatory cell-mediated events, but due to a more widespread signaling mechanism, although a mechanism responsible for mediating this decrease is currently unknown.

In the present study, channel downregulation occurred despite the fact that the alveolar cells were morphologically intact and did not show signs of necrosis as demonstrated by light microscopy. This finding suggests that the decreased expression does not depend on disruption of normal cell architecture. However, under conditions of lung injury, alveolar type II cells proliferate and differentiate into alveolar type I cells as a means of replacing damaged type I cells (9). Therefore, it is conceivable that an increase in alveolar type II cells, which do not express AQP5, could contribute to the decrease in AQP5 expression observed in the infected mouse lungs. Immunohistochemical analysis of proSPC, which is expressed only in alveolar type II cells (30), was used to determine whether there was an increase in alveolar type II cells after adenoviral infection. No differences were detected in the amount of proSPC staining in infected and uninfected mouse lung sections, indicating that there was no detectable increase in alveolar type II cells in the mice that were analyzed in this study. Therefore, the decreased expression of AQP5 in mice 7 and 14 d after infection with adenovirus is likely not due to an increase in alveolar type II cells and is presumably due to reduced AQP5 gene expression within alveolar type I cells.

Northern blot analysis demonstrated that suppression of channel expression is not limited to water channels but also includes a sodium channel, namely, the ENaC. This channel is expressed in alveolar type II cells in the lung and has been shown to be important in the transport of sodium across the alveolar epithelium (18, 19). Sodium uptake occurs on the apical surface of type II cells primarily through ENaC, and sodium is pumped into the lung interstitium by the NKA in the basolateral membrane (17, 35). In the present study, mRNA for ENaC was found to be significantly decreased both 7 and 14 d after infection with adenovirus. Expression of the 1 subunit of the NKA mRNA was not significantly altered either 7 or 14 d after adenoviral infection, providing evidence that the suppression of transporter expression is relatively selective. Water movement through AQPs is driven by an osmotic gradient that is largely dependent upon the sodium concentration. Decreased ENaC expression could theoretically lead to a reduction in sodium movement across alveolar type II cells into the alveolar interstitium, resulting in a disruption of the sodium gradient normally found in the lung (19, 35). Disruption of the sodium gradient could therefore contribute to edema formation due to the inability of water to pass from the air space into the interstitium through water channels. Thus, suppression of ENaC expression may contribute to the edema seen after adenoviral infection.

Recently, mice deficient for AQP1 were generated and found to have a 10-fold decrease in osmotic water permeability between the air space and capillary compartments in the lung in comparison with wild-type mice (10, 11). AQP1 deletion has also been shown to result in decreased hydrostatic interstitial edema but does not affect active reabsorption of alveolar fluid (10, 11). These results suggest that the decrease in AQP1 seen in the adenovirus-infected mouse lungs may serve to decrease edema due to a reduction in the rate of passage of fluid from the blood into the interstitium. Thus, decreased AQP1 may play a compensatory role after adenoviral infection. Alternatively, a decrease in both AQP1 and AQP5 may contribute to edema by essentially reducing the transcellular rate of removal of excess water, thereby effectively trapping water in the alveolar and interstitial spaces. Further studies are needed to elucidate the mechanisms involved in the complex regulatory process of fluid movement in the distal lung.

In conclusion, adenoviral infection-induced pulmonary inflammation and edema in mice have been demonstrated to be associated with a marked reduction in the expression levels of AQP1 and AQP5 in the distal lung. In addition, ENaC mRNA was decreased in the lungs of infected mice. These decreases in expression appear to be relatively selective and are most likely associated with the lung inflammation and edema and not a general disruption of lung morphology. This study provides the first report of decreased AQP1 and AQP5 gene expression in the context of lung pathophysiology. These changes in AQP expression either may represent a response to inflammation- associated pulmonary edema or may be causal in the formation of pulmonary edema after adenoviral infection.