Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The fetal pulmonary epithelium is primarily secretory (1- 3). The liquid produced may provide important distending forces that stimulate lung development. Disruption of these forces in fetal lambs by tracheal ligation or chronic drainage of lung fluid results in abnormal lung development (4, 5). The production of lung liquid is inhibited by bumetanide, suggesting that chloride (Cl) secretion underlies its production (6, 7). Presumably, the process involves a basolateral sodium-potassium-chloride cotransporter coupled to an apical Cl exit pathway. The lack of pathologic lung development in human newborns with cystic fibrosis (CF) suggests that the fetal CF lung is inflated with a normal amount of liquid. Further, lung development, distal lung water, and basal transepithelial potential differences across the epithelium of cultured cystic tracheal explants are also normal in CF transmembrane conductance regulator (CFTR) (/) mice (8). Fetal Cl secretion is therefore likely to occur through pathways other than the CFTR.

A number of other apical epithelial Cl conductances have been identified, including: (1) an outwardly rectifying Cl channel (ORCC) (9), (2) a calcium-activated Cl conductance (ICl_Ca) (8) (12, 13), and (3) a swelling-activated Cl channel (ICl_vol) (14). The ORCC appears to be regulated by CFTR in airway epithelia (9, 15). Both CFTR and extracellular adenosine triphosphate (ATP) are required for activation of ORCC by protein kinase (PK) A and ATP (11). To explain this relationship, a model has been proposed in which ATP is transported extracellularly through CFTR channels and subsequently binds to purinergic receptors, thereby activating ORCCs (16). The molecular identity of the ORCC remains unknown. Cl transport across fetal mouse tracheal epithelium is stimulated by calcium ionophore (8), presumably via activation of ICl_Ca. The functional importance of this channel is highlighted by the fact that when CFTR is absent, ICl_Ca may partially compensate for the deficiency in CFTR (12). The level of expression of ICl_Ca in various epithelia of CFTR-deficient mice has been correlated with the degree of phenotypic derangement (13). A new family of genes has been identified which may code for epithelial calcium-activated Cl channels (17). However, the functional significance of these gene products remains to be determined. Virtually every cell possesses an ICl_vol which can be activated by cell swelling (18, 19) or by a reduction in intracellular ionic strength (20). Protein tyrosine phosphorylation and G proteins also regulate this current (21).

The ClC family of human voltage-dependent Cl ion channel genes (CLCN1 through CLCN7 and two kidney-specific genes) code for a number of physiologically important proteins (22, 23). The proteins encoded by these genes are named: ClC-1 through -7, ClCK_a and ClCK_b. All ClC channels have 10 (less likely 12) transmembrane domains with intracellular N- and C-termini (24) and functional ClC channels probably exist as multimers of the gene product (25). Mutations in the CLCN1 gene account for both dominant and recessive forms of human myotonia (26), whereas mutations in CLCN5 cause Dent's disease and result in nephrolithiasis (27). ClCK_b mutations produce a form of Bartter's syndrome type III (28).

Overexpression of ClC-3 in vitro produced a volume-dependent current having many biophysical similarities to native ICl_vol (29). A current similar to that described for ClC-3 is present in pulmonary epithelia (30). In vitro expression of ClC-2 also produces chloride currents that are swelling-induced (31), however these currents are hyperpolarization-activated and bear little resemblance to the native ICl_vol of most cells. Rat ClC-2 is expressed in pulmonary epithelial cells and this expression peaks in fetal life and declines after birth (32, 33).

The goal of the present study was to determine which members of the CLCN family of voltage-dependent Cl channel genes are expressed in the developing human at a time when the pulmonary epithelium is primarily secretory. Northern analysis was used as an initial screen for gene expression and suggested that the most abundant CLCN message in human fetal lung was CLCN3. These results were supported by ribonuclease (RNase) protection assays (RPAs) and in situ hybridization that defined the pulmonary cell types expressing CLCN3. Identification of the genes that code for non-CFTR human pulmonary epithelial Cl currents will improve our ability to determine their functional significance during both fetal and postnatal life and may also suggest strategies for the activation of alternative Cl transport pathways for the treatment of CF.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Northern Analysis

Radiolabeled probes for CLCN2 through 6 and ClCK_a or K_b were prepared using the same CLCN gene-specific probes used in previous studies (34). CLCN1 expression was not investigated due to the previously reported absence of pulmonary expression (35). Probes were designed using unique sequences with minimal similarity to other members of the CLCN family. The majority of the clones were from 3' untranslated sequence. Radiolabeling was performed by polymerase chain reaction (PCR) incorporation of ³²P-deoxycytidine triphosphate (dCTP) (36) using cloned segments from various members of the CLCN family of chloride channels as template. The cloned segments and the primers used to produce the clones were identical to those reported previously (34). The ClCK_a probe was expected to detect both ClCK_a and ClCK_b due to the high degree of homology (219 of 239 base pairs [bp] or 91.6%) of the two genes over this region. Labeling PCR reactions contained 200 pmol of complementary DNA (cDNA) as template, 250 nM concentrations of each oligonucleotide, and 2 µM concentrations of each deoxynucleotide triphosphate except for dCTP. A total of 17 pmol of ³²P-dCTP was used (50 µCi of 3,000 Ci/mmol specific activity), with 0.25 U of Taq DNA polymerase (Boehringer Mannheim, Indianapolis, IN), 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl₂, and 50 mM KCl in a total volume of 10 µl. Usual PCR parameters consisted of an initial denaturation step at 94°C for 2 min followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s. A different annealing temperature was used for CLCN2 (53°C). Labeled probe was column-purified and quantified using a Beckman L6000 (PE) liquid scintillation counter.

Total RNA was extracted from human fetal and postnatal peripheral lung tissues using TRI Reagent (Molecular Research Center, Inc., Cincinnati, OH). Fetal tissues were obtained as a result of pregnancy termination not associated with these studies and were obtained under a protocol approved by the Institutional Review Board at the University of Iowa. At early gestational ages only very small pieces of lung tissue were available and these were not trimmed in any way to select for specific areas of the lung. When larger samples were available from older donors, pieces of peripheral lung were used. Northern blots were prepared by running 20 µg of total RNA per lane plus a 0.24- to 9.5-Kb RNA ladder (GIBCO BRL, Gaithersburg, MD) out on a 1.85% formaldehyde/agarose gel and then transferring the RNA onto an S&S Nytran membrane (Schleicher & Schuell, Keene, NH). Each blot contained four lanes with whole-lung RNA obtained from: (1) a 96-d estimated gestation fetus, (2) a 22-wk gestation fetus, (3) a 5-mo-old infant, and (4) an adult. Individual blots were hybridized separately in ExpressHyb Hybridization Solution (Clontech Laboratories, Inc., Palo Alto, CA) and each probe was added at a specific activity of 1 × 10⁶ counts/min of probe per milliliter of solution. Blots were initially rinsed with 2× saline sodium citrate (SSC) solution containing 0.05% sodium dodecyl sulfate (SDS) and then washed for 40 min in this solution at room temperature. A second wash was performed in 0.1× SSC solution containing 0.1% SDS at 50°C for 40 min. Radioactive bands were detected by autoradiography. Relative levels of expression were compared visually.

RPA

CLCN2 and CLCN3 messenger RNAs (mRNAs) were identified using RPA methods previously described for quantitative assessment of HBD-1 mRNA in the lung (37). Total RNA was isolated from frozen lung-tissue samples and human airway epithelial cell cultures as described earlier. The -³²P-uridine triphosphate (UTP) (Amersham Corp., Arlington Heights, IL)-labeled CLCN2, CLCN3, and 18S ribosomal subunit (for internal standard; Ambion Co., Austin, TX) antisense riboprobes were transcribed using the same cDNA templates described earlier and containing the T7 promoter. The radiolabeled riboprobes were hybridized to tissue-specific mRNA using a Hybspeed RPA kit (Ambion). The RNA-RNA hybrids were separated by denaturing Tris-borate-ethylenediaminetetraacetic acid vertical gel electrophoresis and visualized either by autoradiographic methods as previously described (37) or by using an Ambis phosphorimager.

In Situ Hybridization

Tissue preparation. Human embryonic lung from first-trimester fetuses was obtained as described earlier and immersion-fixed in molecular biology grade HC fixative (Amresco, Solon, OH) for 12 h at 4°C, subsequently dehydrated in a series of alcohols, cleared in xylene, and paraffin-embedded by standard methods. Sections, 10 µm, were then prepared from paraffin blocks, mounted on pretreated slides (Superfrost Plus; Fisher Inc., Houston, TX), and stored desiccated at 20°C until use. Immediately before use, the tissue sections were cleared of paraffin with xylene washes and rehydrated.

In situ hybridization histochemistry (ISHH). Nonradioactive ISHH was performed on selected lung sections using methods similar to that previously described (38, 39). Digoxigenin-labeled cellular RNA was generated from template DNA (see clones described earlier) linearized with the appropriate restriction enzyme and transcribed using T3 or T7 DNA-dependent RNA polymerase in the presence of a labeling mixture containing digoxigenin-UTP and UTP at a ratio of 0.54 (Boehringer-Mannheim) (38). In situ hybridization was performed at a final probe concentration of 5 µg/ml with either an antisense-specific or sense strand-specific probe. Hybridization, RNAse treatment, and washing were done as previously described with the exception of using a hybridization/wash temperature of 45°C. Immunodetection of digoxigenin was performed using alkaline phosphatase-conjugated, anti- digoxigenin-Fab fragments from sheep (Boehringer-Mannheim) at a dilution of 1:5,000. Alkaline phosphatase-conjugated antibody binding was detected by incubating the slide at 4°C for 1 to 6 h in 100 mM Tris (pH 9.5), 50 mM MgCl₂, 100 mM NaCl, 1 mM levamisole, 0.34 mg/ml nitroblue tetrazolium, and 0.175 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (Boehringer-Mannheim), which produces a blue-black precipitate in cells with positive hybridization signal. Experiments using sense and antisense strand probes were done under identical conditions.

Imaging. Images of representative sections were captured using a COHU 4900 series CCD camera (Cohu, Inc., San Diego, CA) fitted to a Nikon Optiphot microscope (Nikon, Inc., Melville, NY) equipped with Hoffman optics (Modulation Optics, Inc., Greenvale, NY).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

CLCN mRNA Expression in Developing Lung

Due to the difficulty associated with obtaining high-quality tissue samples for these studies, a limited number of time points during development were available for analysis. It was not possible to obtain any suitable samples from a third-trimester fetus. Although this limitation precluded quantitative analysis of expression levels, it was possible to determine a general pattern of gene expression.

Of the six probes used for Northern analysis of CLCN gene expression, evidence of expression was seen for only two genes, CLCN3 and CLCN6. No signal was obtained using the CLCN2, CLCN4, CLCN5, CLCN6, or ClCK_a or K_b probes. The CLCN3 blot is shown in Figure 1. Two bands, one at ~ 4 and one at ~ 7 Kb, were identified. The presence of two bands and their relative sizes agree with those previously detected by Northern analyses of CLCN3 expression (34, 40). CLCN3 expression was developmentally regulated, with the strongest signal seen at midgestation and a clear decline in expression postnatally. Interestingly, of the two CLCN3 transcripts, the longer message appeared to predominate during fetal life and the shorter postnatally. We have recently determined that the difference in the length of these transcripts is due to the presence of alternative polyadenylation sites (Schutte and Lamb, unpublished data). However, the functional significance of the different 3' untranslated regions, if any, remains unknown. CLCN6 message was identified but the signal was very faint and the magnitude of expression did not appear to differ significantly at different developmental stages (data not shown). CLCN6 message has never been functionally expressed, leading to speculation that it may encode an intracellular Cl channel (41).

View larger version (89K): [in this window] [in a new window]   Figure 1.   Northern blot of CLCN3 message expression in total RNA isolated from human lung tissues at various stages of development and postnatal life. Two transcripts are detected which are of appropriate size for CLCN3. It appears that whereas the larger band predominates in fetal life, there is a shift to the smaller message size after birth. CLCN3 expression is higher in the fetus but persists at a lower level throughout life. Lower panel shows ethidium-stained gel at the time of transfer, verifying equivalent loading of each lane.

Previous studies demonstrated that ClC-2 expression is developmentally regulated in the rat lung (32, 33). We suspected that Northern analysis was not sensitive enough to detect the CLCN2 message in total RNA from whole lung. We therefore used RPAs to simultaneously detect expression of the CLCN2, CLCN3, and 18S ribosomal RNAs (Figure 2). Two separate experiments were performed using different tissues. The first used RNA samples from 8-, 12-, 18-, 22-, and 42-wk fetuses, and 5-mo- and 12-yr-old individuals. The second used samples from 9.5-, 13-, 16-, and 19-wk fetuses, the same 12-yr-old as in the first experiment, and an adult. CLCN2 and CLCN3 messages were readily detected by this method. The sizes of the protected fragments were 275 bp for CLCN2 and 249 bp for CLCN3. Both experiments revealed a similar pattern of channel expression.

View larger version (34K): [in this window] [in a new window]   Figure 2.   RPA demonstrates the ontogeny of expression of CLCN2 and CLCN3 in human lung. Protected fragments represent CLCN2, CLCN3, and 18S ribosomal RNA. The sizes of the protected fragments were 275 and 249 bp for CLCN2 and CLCN3, respectively. (A) The first five lanes of RNA were obtained from fetal tissues with gestational age denoted in weeks. The last two lanes of RNA were postnatal tissues obtained from a 5-mo-old and a 12-yr-old individual. This blot was exposed overnight on X-ray film. Peak expression for both channels was seen at 22 wk gestation. Expression persists at lower levels through 12 yr of age. (B) A separate but similar experiment, performed using a different set of RNA samples. This blot was quantified using a phosphorimager.

Relatively low levels of CLCN2 expression were detected early in development (8, 9.5, and 12 wk of gestation) and message levels subsequently increased at 13, 16, 18, and 22 wk. The 19-wk sample had relatively low levels of CLCN2 message; this is likely simply to represent individual variability. CLCN2 expression declined by the end of pregnancy (42 wk) and remained relatively low at all postnatal time points (5 mo, 12 yr, and adult).

The developmental pattern of CLCN3 gene expression was similar to that of CLCN2, but not identical. Small quantities of CLCN3 message were detectable at 8 wk gestation. Much higher levels of CLCN3 expression were seen by 9.5 wk and message levels were consistently high until the end of intrauterine life at 42 wk gestation. An exception to this was the 18-wk sample which, again, may represent variability between individuals. An important contrast in the pattern of expression of these two genes is that only CLCN3 expression was significantly above basal levels at the time of delivery. It is worth noting that although expression was lower relative to intrauterine life, both gene products were detectable in human lung tissues obtained from a 12-yr-old and an adult patient and therefore may have functional significance throughout life.

Localization of CLCN2 and CLCN3 Expression by In Situ Hybridization

To determine whether the CLCN2 and CLCN3 message found in whole-lung total RNA samples originated from epithelial cell expression, in situ hybridization was performed. Lung tissues were obtained from first-trimester fetuses. The antisense probe for CLCN2 detected expression in both surface and submucosal gland epithelia of the trachea and in airway and acinar epithelia in this pseudoglandular stage lung (Figure 3). Hybridization signals were also detected over the media and intima of blood vessels of all sizes. There was no evidence of message in pulmonary interstitial cells or in fibroblasts of the adventitial layer of blood vessels. No significant hybridization signal was seen when a CLCN2 sense probe was used as a control.

View larger version (134K): [in this window] [in a new window]   Figure 3.   (A) CLCN2 expression within the trachea and adjacent glands as well as surrounding epithelium. (B) Similar image from adjacent section using sense-strand probe. (C) Higher-power image of CLCN2 expression in trachea and surrounding glands. (D) Sense-strand detection in trachea of section adjacent to A. Scale bar: A, 100 µm, also applies to B; C, 50 µm, also applies to D. Arrows, trachea; asterisks, cartilage; arrowheads, surrounding epithelium.

Hybridization signals for CLCN3 (Figure 4) demonstrated a pattern similar to those of CLCN2. CLCN3 mRNA hybridization was detected in surface and submucosal gland tracheal epithelia and in small airway and acinar epithelia. Signals were also detected over the media and intima of blood vessels of all sizes, as has been demonstrated previously (34). As with CLCN2, there was no evidence of CLCN3 message in pulmonary interstitial cells or in fibroblasts of the vascular adventitia and no significant hybridization signal was seen when a sense probe was used.

View larger version (140K): [in this window] [in a new window]   Figure 4.   (A) CLCN3 expression in trachea and surrounding glandular tissue. (B) CLCN3 expression in blood vessels and surrounding epithelium. (C) Higher-power image of CLCN3 expression in the surrounding lung parenchyma. (D) Higher-power image of region similar to that in B using sense-strand probe. Note lack of signal in areas similar to those seen in B and C. Scale bar: A, 100 µm, also applies to B; C, 50 µm, also applies to D. Arrows, trachea; asterisks, cartilage; arrowheads, blood vessels.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These studies demonstrate that CLCN3 expression is developmentally regulated in human fetal lung. CLCN2 is also expressed in these tissues but perhaps at slightly lower levels. The ontogeny of expression of the two genes is similar, with peak message levels found at midgestation and a substantial decline occurring around the time of birth. Both CLCN2 and CLCN3 expression localized to the pulmonary epithelium, suggesting that these gene products may contribute to transepithelial Cl transport and fetal lung liquid production.

Northern analysis detected CLCN3 message in whole human lung tisssue. The only other CLCN message detected by this method was a faint signal for CLCN6. The lack of signal for the other members of the CLCN family was unlikely to be related purely to inadequacy of the probes because this identical set of probes successfully identified CLCN2, CLCN3, CLCN4, CLCN5, and CLCN6 in RNA isolated from cultured human vascular smooth-muscle cells (34). It seems likely that other CLCN messages either are not expressed or are less abundant. Pulmonary expression of both CLCN3 and CLCN2 was demonstrated by using RPAs. Further, expression was localized to lung epithelia and blood vessels by in situ hybridization. The presence of CLCN2 in human pulmonary epithelia agrees with previous Northern analysis of human tissue (42) and immunohistochemical findings from rat tissues (32, 33), and supports a possible role for CLCN2 in airway epithelial function. Expression of CLCN3 in airway epithelial cells represents a novel observation. The expression of both CLCN2 and CLCN3 peaked in midgestation and declined postnatally. This pattern suggests that both gene products are good candidates to mediate the apical Cl exit step underlying liquid secretion in the fetal lung and that both represent potential targets for manipulation of adult pulmonary epithelial Cl conductance.

ClC-3 was originally cloned from a rat kidney cDNA library using a PCR strategy based on conserved regions of ClC-1 and ClC-2 (43), and the human gene CLCN3 was isolated using a similar strategy (40). CLCN3 has a wide spectrum of expression when assessed by whole-tissue Northern blotting, with the highest levels seen in the brain, kidney, and skeletal muscle and the lowest levels in the liver and lung (40). However, the gene does not appear to be expressed in every cell, and the level of expression varies widely from cell type to cell type within a given organ (40, 44). These data suggest that although the distribution of ClC-3 expression is broad at the whole-tissue level (23), the channel is not ubiquitously expressed at the cellular level.

Overexpression of ClC-3 in either Xenopus oocytes or Chinese hamster ovary cells produced Cl currents that were outwardly rectified, had an ion selectivity of I > Br > Cl > Fl, and were inhibited by activation of PKC (45). It was subsequently suggested that ClC-3 represents the widely expressed ICl_vol (29). When a guinea-pig cardiac ClC-3 clone was expressed in NIH/3T3 cells, outwardly rectifying currents that inactivated at positive potentials were seen. These currents shared many properties with native ICl_vol, including a similar single-channel slope conductance (~ 40 pS), anion selectivity (I > Cl > aspartate), and sensitivity to block by tamoxifen. A specific serine residue in the channel appeared to link phosphorylation state to volume regulation of the channel (46). It should be pointed out, however, that not all groups have been able to express ClC-3 in vitro (47) and some debate remains regarding the biophysical nature of this channel (48). Downregulation of native ClC-3 expression in bovine ciliary epithelial cells resulted in only partial inhibition of swelling-induced Cl current, leading to the conclusion that ClC-3 is not the only, or even the primary, volume-activated Cl channel in these cells (49).

Does ClC-3 represent a known pulmonary epithelial Cl current that has been previously identified functionally? An ICl_vol has been characterized in epithelial cells and does not appear to be under the direct regulation of CFTR (30). No difference in this current was observed in nasal epithelial cells from control and CF individuals (50). However, at this time it may not be reasonable to equate ClC-3 and ICl_vol. Although virtually every cell tested appears to have an ICl_vol, as discussed earlier, not every cell expresses ClC-3. Expressed ClC-3 also shares many characteristics with the ORCC. In addition to outward rectification, the channels have an identical selectivity profile of I > Br > Cl and the range of single-channel conductances reported (45) overlap with that reported for ORCCs (11, 51).

It has been suggested that it may be possible to treat CF by making use of an "alternative pathway" for Cl transport. To manipulate such a conductance it is imperative to identify the genes encoding the ion currents that have been characterized biophysically. Any Cl conductance that is selectively expressed in the apical membrane and is active at physiologic membrane potentials could potentially support transepithelial Cl transport (52). Rat ClC-2 does appear to be processed to the apical membrane (33) and ClC-2 has been proposed as a viable alternative pathway for Cl conduction in CF (53). Recent data suggest that ClC-2 channels may be pH-sensitive and the relatively acidic environment of fetal lung fluid may lead to its activation (54). It remains to be determined whether ClC-3 is also expressed selectively in the apical membrane of pulmonary epithelial cells. A recent report does describe selective expression of rat ClC-3 in the canalicular membrane of hepatocytes (55).

In summary, we have shown by Northern analysis that CLCN3 is expressed in the human lung. CLCN2 is also expressed, but perhaps at lower levels that were detectable only by an RPA. Both genes are most highly expressed during development, when the pulmonary epithelium is engaged primarily in liquid secretion. In situ hybridization confirmed the presence of both CLCN2 and CLCN3 in pulmonary epithelial cells. We speculate that ClC-3 channels may contribute to the apical Cl transport pathway required for fetal pulmonary epithelial liquid production. How this gene product contributes to both lung development and adult pulmonary epithelial physiology remains to be fully defined.