Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Liquid present in the fetal lung is secreted by the epithelia lining the developing airway and acinar structures. The distending pressure resulting from liquid secretion is thought to be important for lung growth (1). Early studies in fetal sheep determined that liquid secretion results from active movement of chloride (Cl) ions across the fetal lung epithelia (2, 3).

The mechanisms for transepithelial Cl transport have been probed by studies with selective agonists and inhibitors of Cl secretion and, more recently, with molecular studies of genetically altered mouse models. Cl secretion by fetal lung is regulated by both intracellular cyclic adenosine monophosphate (cAMP) (4) and purinoceptor- activated intracellular Ca²+ (7). Studies of cystic fibrosis transmembrane conductance regulator (CFTR) knockout mice show that CFTR is not required for lung fluid homeostasis and normal lung growth in the fetal mouse. A number of additional candidate channels are proposed to transport Cl through the apical membrane (7, 8).

This study focuses on paths for Cl entry through the basolateral membrane of the pulmonary epithelial cells. The sodium-potassium-2-Cl cotransporter (NKCC) is a logical candidate for this function. It is an important conduit for Cl entry in other liquid-transporting epithelia (reviewed in Ref. 9). Two isoforms of the NKCC have been identified. NKCC1 is expressed widely in a number of liquid-secreting epithelia (10), including the murine lung (11), whereas NKCC2 expression is confined to the epithelium lining the thick ascending loop of Henle (12).

Previous animal studies have shown the importance of NKCC1 for Cl entry through the basolateral membrane of fetal respiratory epithelia. In fetal sheep and guinea-pig lung, addition of the loop diuretic bumetanide or furosemide (specific NKCC inhibitors) to fetal lung liquid caused slowing of lung-liquid secretion or liquid absorption (13, 14). Likewise, loop diuretics inhibited liquid secretion by distal lung explants from fetal rat lung (3, 15) and decreased basal short-circuit current by excised fetal canine or rabbit trachea (16, 17). In studies of bioelectric properties of fetal mouse trachea and distal lung explants, bumetanide also inhibited cAMP-activated Cl secretion (3). Similar inhibition of cAMP-activated Cl secretion is reported for human fetal lung (18). The existence of Cl entry mechanisms other than NKCC1 in the basolateral membrane of fetal lung epithelia is suggested by the minimal effect of bumetanide on basal Cl secretion by human fetal distal lung epithelium (18) and incomplete inhibition of cAMP-induced Cl secretion by this inhibitor (18, 19).

Mouse lines deficient in NKCC1 exhibit phenotypes that indicate the importance of this protein in salivary secretion (20), spermatogenesis (21), and intestinal secretion (22). We studied the role of NKCC1 in fetal lung liquid and Cl secretion in mice carrying a null allele for the gene encoding NKCC1. Affected mice are born with apparently normal lung function, and survive and live into adulthood with no respiratory problems. The apparent good respiratory health of these mice raises several questions about the dependence of Cl ion transport on NKCC. In this study we show the role of NKCC1 in fetal lung Cl and liquid secretion. We conclude that NKCC1 plays an important role in fetal lung Cl secretion, but that alternative Cl mechanisms exist that can compensate for the absence of NKCC1 in NKCC1-null mice.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of NKCC1-Null Mice

The generation of NKCC1-null mice has been previously described (21). Briefly, a genomic clone containing a 3' region of the Slc12a2 gene was isolated from a FixII SV129  phage library using a complementary DNA (cDNA) probe corresponding to base pairs (bp) 1239 to 2449 of the published mouse Slc12a2 sequence (20) and subcloned into pBluescript (Stratagene, La Jolla CA), pBS/NKCC3a. DNA fragments from this clone were used to construct the Slc12a2^506-621 targeting plasmid, which is designed to replace a 344-bp region containing exons 9, 10, and 11 with the neomycin gene. A 5.1-kb Xba I fragment located 5' to exon 9 and a second fragment of 7.1 kb beginning just 3' to exon 11 and extending to the Not I site 3' of exon 15 of the Slc12a2 gene were cloned in the targeting plasmid on either side of a neomycin gene. The plasmid was linearized and introduced into the ES cell line E14TG2a and transformants were isolated as previously described (23). Cells were screened for targeted integration of the plasmid by Southern blot analysis, and targeted cell lines were microinjected into 3.5-d C57BL/6J blastocysts to generate chimeric animals. These were subsequently mated to B6D2 (C57BL/6J × DBA/2J F1) animals to generate animals heterozygous for the mutant allele.

Genotype Determination

Experiments were done with fetal mice from females heterozygous for NKCC gene deletion (NKCC +/) mated with heterozygous males (NKCC +/). Fetuses were removed after dams were killed by CO₂ inhalation. Genomic DNA was extracted from the tail and one leg of each fetus by the method adapted from Miller and colleagues (24). The DNA was cut with a restriction enzyme and run through the Southern blot procedure.

In Situ Hybridization

The preparation of a 600-bp NKCC1 cDNA probe corresponding to bases 3127 to 3636 of the published mouse NKCC1 sequence has previously been described (20). Using this construct, 35S- labeled sense and antisense RNA was prepared (Maxiscript SP6/ T7 kit; Ambion, Austin, TX) and used to analyze tissue sections. Tissues were fixed in 4% paraformaldehyde and washed with 30% sucrose in phosphate-buffered saline to remove the fixative. Tissues were embedded in Tissue-Tek, and cryostat sections cut at 8 µm were mounted on slides and stored at 80°C. In situ hybridization was performed by standard methods, as described previously (25). Briefly, prehybridization consisted of proteinase K digestion, then acetylation. A quantity of NKCC1 sense or antisense probe producing 1 × 107 counts per min was hydridized to serial sections overnight at 54°C in 50% formamide, 1× Denhardt's solution, 0.6 M NaCl, 10 mM Tris (pH 8.0), 1 mM ethylenediaminetetraacetic acid, 0.1% sodium dodecyl sulfate, 10 mM dithiothreitol (DTT), 1 mg/ml yeast transfer RNA, and 10% dextran sulfate. After hybridization, slides were washed in 4× saline sodium citrate (SSC) at room temperature and subjected to the following sequential protocol: ribonuclease A digestion (20 mg/ml) for 30 min at 37°C, 2× SSC/1 mM DTT at room temperature, and a high-stringency wash of 0.5× SSC/1 mM DTT at 58°C (three times for 15 min each time) followed by ethanol dehydration. Dried slides were dipped in Kodak NTB2 photoemulsion and stored at 4°C until developed. Slides were developed at 2 wk, counterstained with hematoxylin and eosin (H&E), and photographed using brightfield and darkfield microscopy at ×40 magnification (Nikon Microphot-SA microscope).

Lung Water Measurements

Fetal lungs from 18.5-d fetal mice were excised and placed in a dish containing medium (Ham's F12, Invitrogen Corp., CA). To obtain an index of lung water content, the wet/dry lung weight (W/ D) was determined as described previously (26). In brief, the right lung lobes were dissected free, blotted on premoistened filter paper, weighed wet, desiccated overnight at 80°C, and reweighed.

Explant Culture

Distal lung fragments containing acinar and terminal bronchiolar epithelium were dissected from 17-d fetal lung and cultured to form liquid-filled cystic explants as described previously (27). In brief, distal lung fragments containing acinar and terminal bronchiolar structures were dissected from lungs of 17-d fetal lung and placed on a clear, permeable PET Transwell support inside a culture dish containing a 1:1 F12/Dulbecco's modified Eagle's medium solution supplemented with bovine serum albumin (1 µg/ml). These explants were cultured at 37°C in a 95% air-5%CO₂ environment for the 3-d duration of the studies.

Excised 17- and 18-d fetal tracheas were explanted into submersion culture as previously described (3, 28) and cultured at 37°C in a 95% air-5%CO₂ environment until they formed cysts (4 to 5 d). Culture medium was changed every 48 h.

Distal Lung Liquid Secretion

After 24 h, lung fragments seal and form liquid-secreting cysts. Photographs were taken of the cultures at 0, 24, and 72 h. The photographs were then analyzed with Metamorph software to determine change in a cross-sectional area over time. Because of the heterogeneity of explant shape, an estimate of volume was not attempted. Values are expressed as change in cross-sectional area of the explant. The 0-to-24-h photos were used to determine basal growth rates. Inhibitor or vehicle was added at 24 h, after cysts had sealed and cyst formation was evident. The effect of the drugs on liquid secretion was determined from the change in cross-sectional area between 24 and 72 h. Drugs used in this study were: bumetanide (NKCC1 inhibitor), 4,4'-diisothiocyanto-stilbene-2,2'-disulfonic acid (DIDS) (anion exchange inhibitor), acetazolamide (carbonic anhydrase inhibitor), and amiloride (Na⁺ channel inhibitor); all were obtained from Sigma Chemical Co. (St. Louis, MO). Terbutaline sulfate (selective ₂-adrenergic agonist) was obtained from Geigy (Ardsley, NY). Final concentration for all drugs was 10⁴ M, except for terbutaline (3 × 10⁵ M). Bumetanide and acetazolamide were dissolved in dimethyl sulfoxide (DMSO) (10¹ M stock); amiloride and DIDS were dissolved in water (10² M stock). Terbutaline was purchased in a buffered saline solution at 3 × 10³ M strength.

Transepithelial Potential Difference

After 5 d in culture, potential difference (PD) was measured across the wall of explants that formed liquid-filled cysts, as previously described (3, 28). Tracheal bioelectric studies were conducted in room air (as compared with volume-change studies that were conducted in 95% air-5%CO₂). The cysts were impaled first with the electrode to determine basal PD, and then with a micropipette containing amiloride (10² M; Sigma). Once the PD reached a stable value, amiloride was injected into the cyst lumen to inhibit amiloride-sensitive Na⁺ channels. The residual PD and changes in PD were assumed to reflect, predominately, Cl secretion. Terbutaline sulfate was used in the bath to stimulate Cl secretion. The NKCC inhibitor bumetanide and the anion exchanger DIDS were added to the bath to block Cl entry through the basolateral membrane. The carbonic anhydrase inhibitor acetazolamide was added to the bath to block intracellular HCO₃ production.

Lung-Liquid pH and Cl Measurements

After maternal and fetal anesthesia with maternally administered 2.5% isofluorane gas, an 18-d fetal pup head was delivered via a uterine incision. The trachea was identified after midline neck incision and a small transverse opening was made in the fetal trachea. Fetal lung liquid was aspirated via a pulled glass pipette and the sample was kept on ice until pH measurements were done (typically within 15 min of sampling). Microaliquots (0.5 to 1.0 µL) were aspirated directly from the micropipette tube into the tip of one end of a short (0.5-cm) section of C0₂-permeable silicone tubing (Helix Medical Inc., CA; .025-in inner and .047-in outer diameter). The tip of the microelectrode was placed and secured into this end of the tubing such that the tip and reference electrode were fully immersed in the sample. The other extremity of the sample was separated from the distal end of the silicone tubing by a column of air. The microelectrode and attached tubing were then placed in a water bath fully equilibrated with 5% CO₂ at 37°C. The column of air between the sample and the open end of the silicone tubing prevented sample mixing with fluid in the water bath. pH was recorded once the reading was stable for at least 1 min. Calibration with reference solutions of known pH were carried out under these circumstances before sample analysis. This was repeated at regular intervals to ensure that error due to "drift" in electrode output did not arise.

Measurements were highly accurate and reproducible to within ± 0.01 pH units. Bicarbonate concentrations were calculated from the Henderson-Hasselbach equation. For measurement of Cl concentration, fetal lung liquid was diluted in doubly distilled deionized water. Part of each diluted specimen was further diluted into a 0.1 N nitric acid matrix and analyzed for Cl by flame emission photometry as described previously (29).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

NKCC Expression in Fetal Lung Epithelia

Expression of NKCC messenger RNA (mRNA) was examined in freshly excised 17-d fetal trachea and distal lung. Intense expression of NKCC mRNA was localized to the epithelium of fetal trachea and was also expressed in the surrounding thymic tissue (Figure 1C). There was no NKCC expression in the underlying tracheal interstitial or cartilaginous tissues. In distal lung sections, NKCC expression was likewise intense in small airways and acinar epithelium and expression was confined to these cells (Figure 1F).

View larger version (71K): [in this window] [in a new window]   Figure 1.   In situ hybridization of frozen sections of 18-d-gestation fetal trachea and distal lung, showing H&E-stained sections (A and D) and sections exposed to sense probe for NKCC (B and E) and antisense probe (C and F ). A shows the tracheal epithelium (ep) and D shows the small airways (sa) and acinar (ac) structures. Bar: 70 microns.

Effect of Genotype on Fetal Growth and Lung Liquid Content

A total of 320 fetuses from 41 litters were studied. The frequency of NKCC / fetuses at 17, 18, and 19 d (30% of all fetuses) was close to the expected Mendelian distribution, suggesting that there was no fetal loss due to effects of NKCC gene deletion. There was a trend toward lower body weights in NKCC / fetuses compared with heterozygote and wild-type littermate controls, which did not reach significance at any of the gestations studied (control versus /, respectively: 17 d, 0.68 ± 0.01 versus 0.65 ± 0.02 g; 18 d, 1.01 ± 0.03 versus 0.94 ± 0.03 g; 19 d, 1.21 ± 0.03 versus 1.11 ± 0.05 g). There was also a trend for both net lung dry weight and dry lung weight adjusted for body weight to be lower in lungs from NKCC / 19 d fetal pups compared with controls. Liquid content, expressed as WD, was significantly decreased in lungs of 19-d NKCC / fetal mice compared with that of control mice (W/D: control, 8.01 ± 0.09; NKCC /, 7.06 ± 0.14).

Basal Lung Liquid Secretion

Lung-liquid expansion was measured over 48 h (24 to 72 h after dissection). Control and NKCC / explants had similar basal rates of lateral expansion (control  area 69 ± 8% versus NKCC /  area 54 ± 6%). The effects of inhibitors of NKCC (bumetanide 10⁴ M), Cl/HCO₃ anion exchanger (AE) (DIDS, 10⁴ M), and intracellular HCO₃ production (acetazolamide 10⁴ M) were determined over a 48-h period from the time of cyst formation (approximately 24 h after dissection). Addition of bumetanide to the bath inhibited the expansion of control explants but not NKCC / explants (Figure 2), suggesting that acute pharmacologic inhibition of NKCC had a greater effect on liquid secretion than did chronic absence of NKCC in the knockout lungs. Addition of DIDS alone to the bath inhibited explant expansion in both NKCC / and control explants, more so in NKCC / explants. A combination of bumetanide and DIDS inhibited most liquid secretion in control explants but bumetanide did not inhibit secretion further in DIDS-treated NKCC / explants. DMSO (vehicle for bumetanide) did not affect cyst formation. These data are summarized in Figure 2.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Effect of inhibitors on lung liquid secretion over 72 h by distal lung explants from control (NKCC +/+ and +/) and knockout (NKCC /) 17-d fetal mouse lung. Bumet, 104 M bumetanide. * indicates values that are significantly different from basal values (P < 0.05).

Bioelectric Properties of Fetal Tracheal Explants

Basal and amiloride-sensitive ion transport. The rate of cyst formation was similar in control and NKCC / explants from excised 17- and 18-d fetal trachea. Basal PDs of control explants were higher than those of NKCC / explants (control 13.7 ± 0.5, n = 37; NKCC / 11.6 ± 0.6, n = 18; P < 0.05). Basal PD of control and NKCC / explants were inhibited to the same extent by amiloride (~ 25% inhibition).

Terbutaline-stimulated PD. Addition of 3 × 10⁵ M terbutaline to the bath of amiloride-treated explants caused hyperpolarization in both control and NKCC / explants. After addition of terbutaline, there was a rapid hyperpolarization that peaked within 4 min. Peak hyperpolarization was similar in both groups. In control explants, hyperpolarization was sustained at near 100% of peak levels; whereas in NKCC / explants, hyperpolarization was sustained at about 60% of the peak values (Figure 3). Bumetanide inhibited PD in terbutaline-treated control explants but not in NKCC / explants. DIDS added 10 min after terbutaline stimulation significantly inhibited PD in NKCC / explants. Acetazolamide had no effect on PD in control explants but profoundly inhibited PD in terbutaline-treated NKCC / explants and in control explants that were pretreated with bumetanide (Figure 4).

View larger version (13K): [in this window] [in a new window]   Figure 3.   Representative tracings of response of transepithelial PD across NKCC +/+ and NKCC / 18-d fetal mouse tracheal explants to 3 × 105 M terbutaline. Explants were pretreated by injection of 104 M amiloride into the tracheal lumen.

View larger version (14K): [in this window] [in a new window]   Figure 4.   Representative tracings of response of transepithelial PD across NKCC +/+ and NKCC / 18-d fetal mouse tracheal explants to inhibitors of Cl transport (D, 104 M DIDS; B, 104 M bumetanide) and HCO production (A, 104 M acetazolamide). Explants were pretreated by injection of 104 M amiloride into the tracheal lumen and addition of 3 × 105 M terbutaline (T) to the bath.

Effect of DIDS and acetazolamide pretreatment. Pretreatment with 10⁴ M DIDS or 10⁴ M acetazolamide had no effect on the terbutaline response in control explants. In NKCC / explants, DIDS or acetazolamide did not affect the peak response to terbutaline but inhibited the sustained hyperpolarization seen in untreated explants (Figure 5).

View larger version (21K): [in this window] [in a new window]   Figure 5.   Effect of pretreatment with inhibitors of Cl transport (D, 104 M DIDS; B, 104 M bumetanide) and HCO production (A, 104 M acetazolamide) on terbutaline-induced hyperpolarization (PD, mV) of control (NKCC +/+ and +/) and NKCC / 18-d fetal mouse tracheal explants. Peak responses were not affected in either group by pretreatment with D, B, or A. However, hyperpolarization was not sustained in NKCC / explants that were pretreated with these inhibitors.

pH, HCO₃, and Cl Measurements

Fetal lung liquid volumes ranging from 1 to 7 µl were aspirated from 19-d fetal trachea (five NKCC / and 12 control). There were no differences between NKCC / and control in the measured Cl, pH, or calculated HCO₃ concentration of lung liquid aspirated from 19-d fetuses (pH: control 7.02 ± 0.05, NKCC / 7.10 ± 0.09 mMol/liter; HCO₃: control 10.4 ± 1.2, NKCC / 12.9 ± 2.7 mMol/liter). Cl concentration of fetal lung liquid was also not different between control and NKCC / (range 123 to 160 mMol/liter).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Studies that show inhibition of fetal lung liquid and Cl secretion by loop diuretics (3, 13, 14, 28, 30) indicate that NKCC1 plays an important role in the transport of Cl across fetal lung epithelia. Our study confirms the importance of NKCC1 to basal and stimulated Cl and liquid secretion, and shows that other transporters must be present in addition to NKCC1 to sustain fetal lung liquid secretion in NKCC / fetal lung.

NKCC-Dependent Basal Cl and Liquid Secretion

Because liquid secretion is thought to be of critical importance to fetal lung growth and development (31), the apparent normal survival of NKCC / mice (21) indicates that other mechanisms must exist for Cl entry through the basolateral membrane. However, we did find differences between NKCC / and littermate controls in epithelial secretory function.

The relative importance of NKCC to basal Cl and liquid secretion was shown in studies of tracheal and distal lung explants. Basal transepithelial PD was decreased by about 15% in NKCC / tracheal explants compared with that of control explants. With studies of distal lung epithelia, we did not detect a significant difference in basal rates of liquid secretion in NKCC / and control explants, but bumetanide caused an acute inhibition of basal liquid secretion by control (but not NKCC /) explants. These findings suggest that alternative mechanisms exist for entry of Cl through the basolateral membrane, and that these alternative mechanisms may be upregulated in NKCC / lung to allow near-normal fetal lung liquid secretion. The decreased water content of NKCC / whole lung may reflect decreased rates of liquid secretion in these mice compared with those of control lung. This possibility is suggested by decreased basal PD in NKCC / trachea compared with controls. This impairment in basal transport was not sufficient to produce clinical consequences because NKCC / mice adapt normally to air-breathing at birth and have no increased mortality or respiratory morbidity after birth. Lung architecture was not different in NKCC / and control 19-d fetal lung, suggesting that non-NKCC-dependent mechanisms allowed sufficient liquid secretion for normal lung development in mice.

Role of non-NKCC Ion Transporters in Cl Secretion

Studies in which amiloride-pretreated tracheal explants were stimulated with terbutaline (to raise intracellular cAMP) tested the dependency of Cl secretion on NKCC during maximal stimulation. We found that the immediate peak of hyperpolarization after terbutaline stimulation was similar in NKCC / and control trachea, but that this hyperpolarization was sustained at a lower level in NKCC / compared with that in control explants. These findings suggest that in NKCC / airway the electrochemical gradient for Cl to leave the cell through the apical membrane (reflected by normal initial peak response to terbutaline) is normal, but that maximal sustained transepithelial Cl transport is limited in the absence of NKCC. Mechanisms for transepithelial solute transport by fetal lung epithelia other than NKCC-mediated Cl secretion include non-NKCC mechanisms of Cl entry or secretion of other ions by fetal epithelia (e.g., HCO₃ or K⁺).

The coupling of Cl/HCO₃ (AE) and Na⁺/H⁺ exchanger (NHE) has been proposed as an alternative mechanism for Cl entry through the basolateral membrane (32, 33). Both AE and NHE isoforms (AE2, AE3, and NHE1) are expressed throughout human airways (34) and the NHE function has been localized to the basolateral membrane of airway and alveolar cells (35, 36). Our studies support the possibility that these exchangers are active in fetal lung liquid transport and that they can largely substitute for NKCC function in NKCC / mice. This dependence of liquid secretion on AE and NKCC entry pathways was strongly suggested by studies with bumetanide and DIDS. DIDS inhibits the AE on the basolateral membrane, but is also an inhibitor of other transporters (e.g., Na/HCO₃ symport [37]) and channels (e.g., Ca-mediated Cl conductance [38]). Inhibition of both basal liquid secretion in distal lung explants and terbutaline-stimulated hyperpolarization in tracheal explants by DIDS indicates the relative importance of the AE to Cl and fluid secretion by fetal lung epithelia. Both DIDS and bumetanide inhibited liquid secretion by control distal lung explants (bumetanide to a greater extent than DIDS), and most of the liquid secretion was halted when the two inhibitors were used together. There was no effect of acetazolamide alone on basal liquid secretion in either control or NKCC / explants. This contrasts with the inhibitory effect of acetazolamide on terbutaline-stimulated Cl secretion by NKCC / tracheal explants. The discrepancy can be explained by differences in experimental conditions between tracheal and distal lung studies. For distal lung studies, explants were kept in a 95% air-5% CO₂ environment except when briefly removed from the incubator for photographing. Under these conditions it is likely that the rate of HCO₃ diffusion from the bathing medium into the cell and of intracellular production of HCO₃ was sufficient to sustain the requirements of the Cl/HCO₃ exchanger for basal transcellular movement of Cl. For tracheal explant studies, it is likely that the combination of (1) deprivation of external CO₂ in the bathing solution and (2) additional demands on the AE by terbutaline-stimulated Cl secretion deprived the AE of HCO₃ substrate and limited the rate of Cl secretion by NKCC / explants (lower PD after terbutaline in NKCC / explants) (Figure 6). Further evidence of the importance of cellular production of HCO₃ for activity of the AE was shown in experiments where inhibition of carbonic anhydrase (acetazolamide) led to a prompt depolarization of terbutaline-stimulated PD in the tracheal explants.

View larger version (78K): [in this window] [in a new window]   Figure 6.   Model for NKCC-dependent and non-NKCC-dependent Cl transport across the fetal lung epithelium. In the NKCC / mouse, inhibition of liquid and Cl secretion can be explained by its effects on the AE. The effect of acetazolamide (acetazol.) is explained by the dependence of AE on a supply of cellular HCO3.

Another possible alternative mechanism for fetal liquid production is through active transepithelial movement of bicarbonate ions. An entry mechanism for HCO₃ through the basolateral membrane is the electroneutral Na⁺/HCO₃ symport shown to be present on alveolar epithelial cells (37). The apical membrane of the respiratory epithelium is known to be permeable to anions other than Cl (39) and specific paths of HCO₃ conduction have been proposed (e.g., via CFTR [40] or a Cl/HCO₃ exchanger [41]). Some of our results could be compatible with HCO₃ secretion as an alternative to Cl secretion in NKCC / lung. In such a scenario the DIDS effect on basal liquid secretion and terbutaline-sensitive hyperpolarization would be explained by inhibition of the Na⁺/HCO₃ cotransporter, and HCO₃ secretion would be inhibited by lack of substrate after acetazolamide.

The possibility that HCO₃ is secreted instead of Cl through the apical membrane in NKCC / mice was explored by measuring pH and Cl content of fetal lung liquid. In a number of species, fetal lung liquid Cl was found to be higher than that of plasma (2, 3, 42, 43), indicating that Cl is transported against the transepithelial concentration gradient for this ion. In our studies we showed no differences in measured pH, Cl, or calculated bicarbonate in lung liquid obtained from control and NKCC / fetuses. These findings imply that Cl remains the dominant ion-mediating liquid secretion in NKCC / mice and decreases the likelihood of significant HCO₃ secretion substituting for Cl secretion in these mice.

In summary, our studies with NKCC / mice indicate that NKCC-dependent Cl secretion plays an important role in fetal lung liquid secretion during development. However, alternative mechanisms for lung liquid secretion do exist in the fetal mouse lung and these mechanisms can largely compensate for the absence of NKCC. A likely candidate for alternative Cl entry through the basolateral membrane is the coupled NHE and AE transporters.