Introduction

Top Summary Introduction Materials & Methods Results Discussion References

The use of tobacco is so widespread and the health hazards deriving from it are so substantial that it has been referred to as the "global tobacco epidemic" (Bartecchi, 1995). Chronic bronchitis, emphysema, and lung cancer occur frequently in tobacco smokers. They result from the direct effect of tobacco smoke, the components it contains, or both (Peto et al., 1992; Bartecchi et al., 1995; The Harvard Mental Health Letter, 1997). In the United States, the use of tobacco is responsible for one in every seven deaths. Among people 35 to 70 years old, cardiopulmonary diseases related to tobacco smoking account for one in every three deaths (Peto et al., 1992; The Harvard Mental Health Letter, 1997).

Nic, although addictive and likely responsible for the substance dependence resulting from tobacco use (Bock and Marsh, 1990), is considered to be one of the less dangerous components of tobacco smoke (Bartecchi et al., 1995; The Harvard Mental Health Letter, 1997). It has been suggested that the amount of Nic to which one is exposed as a result of tobacco smoking may not pose a serious health risk (The Harvard Mental Health Letter, 1997). Low tar cigarettes have been considered an acceptable solution to satisfy the smoker's craving for Nic. Devices for aerosol delivery of Nic without tar-related carcinogens are actively developed and tested as safe alternatives to tobacco smoking. Nic is highly soluble in water, and its concentration in the saliva of tobacco smokers can be very high (an average of 8 µM during "smoking days") (Lindell et al., 1993). Comparable concentrations are likely present on the bronchial and lung surface.

Nic binds to and activates the nicotinic receptors for ACh. These are a family of proteins formed by five homologous or identical subunits, arranged symmetrically around a central ion channel (Conti-Fine et al., 1994; Galzi and Changeux, 1995). Different AChR isotypes exist in muscle and neurons. Muscle AChRs are composed of four different types of subunit (, ,  or , and ) (Conti-Fine et al., 1994; Galzi and Changeux, 1995). Neuronal AChRs may include only two types of subunits ( and ) or five copies of the same  subunit (Conti-Fine et al., 1994; Galzi and Changeux, 1995). Neurons express at least eight different  subunits (2-9) and three  subunits (2-4) (Conti-Fine et al., 1994; Galzi and Changeux, 1995). A large variety of neuronal AChRs results from the combinatorial association of different  and  subunits (Conti-Fine et al., 1994; Galzi and Changeux, 1995).

Human skin keratinocytes express AChRs sensitive to ACh and Nic, similar to those expressed by ganglionic neurons, that compose the 3 subunit (Grando et al., 1995). Skin keratinocyte AChRs are likely activated in an autocrine or a paracrine fashion by ACh synthesized and secreted by the keratinocytes (Grando et al., 1993). They seem to regulate cell adhesion and motility because their block by ganglionic AChR-specific antagonists, such as Mec and -BTX, causes cell paralysis and cell/cell detachment (Grando et al., 1995). Because of those findings and because of reports that ACh is synthesized and released by the bronchial epithelium (Wessler et al., 1995; Klapproth et al. 1997), we investigated whether the epithelial cells that line the surface of the bronchial tree express functional AChRs sensitive to Nic.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Cell cultures. Primary cultures of human BECs (Clonetics/BioWhittaker, San Diego, CA) were seeded in T-25 culture flasks (Corning Glassworks, New York, NY) [10,000 cells/ml in 10 ml of culture medium (Clonetics/BioWhittaker)]. When the cells reached 80-90% confluence, they were detached from the plastic by mild trypsinization using 0.25% trypsin EDTA (Clonetics/BioWhittaker), according to the manufacturer's instruction. They were used for the [³H]epibatidine binding assay, or plated onto glass coverslips in 24-well plates (Corning Glass works) for other experimental uses, or reseeded in T-25 culture flasks for further expansion. For plating the cells onto glass, coverslips (Fisherbrand Microscope Cover Glass, 12-mm circle; Fisher Scientific, Pittsburg, PA) were wiped with 70% ethanol and set in the well until dry. A small drop of medium containing 500-800 cells was placed in the center of the coverslip, and 1 ml of medium was slowly delivered to the well. The cultures were grown until they reached confluence.

Patch-clamp recording of single-channel and whole-cell currents. Currents were recorded from confluent cultured human BECs using standard patch-clamp techniques (Hamill et al., 1981) and an LM-EPC-7 patch-clamp system (List Electronic, Darmstadt, Germany). Single-channel currents were recorded from either outside-out or cell-attached patches (Hamill et al., 1981). The resistance of the recording pipettes was 6-8 M. The test solutions to be applied to the cell-attached patches were placed in the recording pipette. For outside-out patches, we applied the test solutions with a perfusion system consisting of an array of glass capillary tubes. The 10-kHz signal output from the EPC-7 apparatus was transferred to a video cassette recorder using a pulse-code modulation device (PCM: Neurodata Neurorecorder DR-384, Neuro Data Institute, Pasadena, CA) for off-line analysis. The electrical signal was filtered at 3 kHz in a Bessel filter (eight-pole, 3 dB, Frequency Devices 902; Frequency Devices, Haverville, MA), digitized at 12.5 kHz, and analyzed with the IPROC-2 program (Axon Instruments, Foster City, CA). Open events were considered finished when the amplitude decreased to below 50% of the estimated mean single-channel amplitude. We obtained the time constants by fitting an exponential equation to histograms of the channel dwell times using the NFITS program (Axon Instruments).

Assay of neuronal-type AChRs by binding of ³H-labeled epibatidine. We verified the presence of neuronal-type AChR on confluent cultured human BECs by studying the binding of ³H-labeled epibatidine. Epibatidine is a specific, high affinity ligand of several neuronal AChRs, including the 3 AChR subtypes (Gerzanich et al., 1995; Wang et al., 1996). We used the neuronal PC12 cell line, which expresses 3 neuronal AChR (Conti-Fine et al., 1994), to identify the saturating concentration of epibatidine in our experimental system. For Scatchard analysis, samples containing 0.5-1 × 10⁶ PC12 cells were incubated with increasing concentrations of [³H]epibatidine (NEN, Boston, MA; 0.1-10 nM; specific activity 48 Ci/nmol). For BECs, because of the small number of cells that we could grow, we used single-dose [³H]epibatidine binding assays. In this assay, we used suspensions of trypsinized human BECs (3 × 10⁵ to 2 × 10⁶ cells/sample in a final volume of 100 µl). For all and each condition, we set up samples at least in triplicate. The cells were incubated with 5-10 nM [³H]epibatidine for 4 hr at 4° and harvested by vacuum filtration over Whatman GF/C filters or by centrifugation at 300 × g for 3 min. The filters were counted by liquid scintillation counting. In each experiment, we determined the nonspecific binding by preincubating the cells with 10 µM nonradiolabeled epibatidine for 30 min at 4° before the addition of [³H]epibatidine.

Detection of neuronal AChR subunits by RT-PCR assay. The results of the patch-clamp and [³H]epibatidine binding experiments suggested the presence on BECs of AChRs of isotype or isotypes similar to those expressed by the neurons in sympathetic ganglia, formed by 3, 5, 2, and 4 subunits (Vernallis et al., 1993; Conti-Fine et al., 1994; Galzi and Changeux, 1995; Wang et al., 1996). We investigated the presence of mRNA transcripts for AChR subunits by RT-PCR experiments using mRNA isolated from human BEC cultures and primers specific for the 2, 3, 4, 5, 6, 2, 3, and 4 AChR subunits and for actin as a positive control. As a positive control tissue, we used mRNA isolated from human brain (a generous gift of Dr. James Howard, University of North Carolina at Chapel Hill) or commercially available human brain cDNA (Quick Clone cDNA, Clontech Laboratories, Palo Alto, CA). As a negative control tissue, we used mRNA isolated from adult human muscle (a generous gift of Dr. James Howard). Adult muscle does not expresses neuronal AChRs but instead expresses the muscle AChR isotype, which includes the 1 subunit subtype and the homologous , , and  subunits. In the experiments that used muscle mRNA, we used primers specific for the 3, 5, 2, and 4 AChR subunits and for the 1 subunit as a positive control.

Cloning and sequencing of RT-PCR products. First-strand cDNA was synthesized from human BEC RNA (Grando et al., 1995). Primers (5'-CCAGTGGCCAGGGCCTCCAGA-3' and 5'-TATGCATCTTCCCTGGCCATCA-3') were designed to amplify the full-sequence region encoding the mature 3 subunit protein. Hot-start PCR (Horton et al., 1994) was used in 10 separate 100-µl reactions containing 10 mM Tris·HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 200 µM concentration each dNTP, 12% sucrose, 250 nM concentration each primer, 1 µl of cDNA template, and ~2 units of Taq DNA polymerase. The reaction yielded a prominent band that was excised and reamplified using histidine tag-encoding T-linkers (Horton et al., 1996b). The reamplified product was cloned directionally into the plasmid pT7-7 using NdeI and HindIII restriction sites in the linkers and sequenced using the fmol sequencing kit (Promega, Madison, WI).

Assay of AChR subunit transcripts by in situ hybridization. We carried out in situ hybridization experiments (Cox et al., 1984) using cultured human BECs and sections of rat trachea and probes specific for each of the AChR subunits detected by the RT-PCR experiments. The probes were transcribed in vitro from DNA clones (a generous gift of Dr. C. Lobron, University of Mainz, Mainz, Germany) and labeled with digoxygenin-UTP (Boehringer-Mannheim, Mannheim, Germany). The labeled single-stranded probes were hybridized to mRNA of the cell under conditions of high stringency of the hybridization. The conditions we used allowed the probes to bind only to their corresponding mRNA (Cox et al., 1984). To detect the bound probe, anti-digoxygenin antibody coupled to alkaline phosphatase (Boehringer-Mannheim) was added. The NBT/BCIP mixture (Boehringer-Mannheim) was added as a substrate for alkaline phosphatase. The specificity of the binding of the probes is demonstrated by absence of the signal when the corresponding "sense" probe is used.

Detection of AChR subunit proteins in tissue sections by immunofluorescence. AChR subunit-specific antisera were obtained by immunizing individual rabbits with a mixture of peptides corresponding to unique sequence regions of a given AChR subunit (Bellone et al., 1993). The peptides corresponded to nonconserved sequence regions of the 1, 3, 4, and 5 subunits: 1: residues 23-41, 154-172, 322-341, 343-362, and 423-437; 3: residues 23-41, 153-172, 319-338, 379-398, and 459-472; 4: residues 338-357, 391-410, 411-430, 431-45, 451-470, 471-490, and 491-510; and 5: residues 31-50 and 151-170. The peptides were coupled to keyhole limpet hemocyanin (0.5 mg/mg peptide) by reaction with a 5-molar excess of 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide for 24 hr at room temperature, in a mixture of 1.5 volumes of 50 mM Na phosphate, 50 mM NaCl, 1 volume of water, and as much ethylene glycol as needed to keep the peptide in solution. After coupling, the mixture was dialyzed extensively against PBS (10 mM Na phosphate buffer, pH 7.4 containing 137 mM NaCl, and 2.7 mM KCl). Rabbits were injected several times with 0.5 mg of conjugate emulsified in complete Freund's adjuvant for the first injection and incomplete Freund's adjuvant for the subsequent injections. The presence of anti-peptide antibodies in the sera was verified by enzyme-linked immunosorbent assay (Lei et al., 1993).

Assay of the effects of cholinergic antagonists on the cell shape and cell/cell adhesion. We tested the effects of Mec (from 100 µM to 5 mM) and -BTX (~10 nM and ~1 µM) on cell shape and cell/cell adhesion in confluent human BEC cultures. In each experiment, we measured the area covered by three to five randomly chosen individual cells just before administering the drugs and at different intervals up to 20 min. We also measured the area covered by three to five randomly chosen individual control cells that received drug-free medium, at different intervals up to 20 min. Phase-contrast microscopic images were captured on a Macintosh Quadra 650 computer equipped with a LG3 NuBus frame grabber card (Scion Corporation, Fredrick, MD). Frame capture at fixed intervals was automated using the "Make movie to disk" macro of NIH Image (a public domain software available on the Internet at http://rsb.info. nih.gov/nih-image/; the macro is accessed under "Special, Load Macros, Movie Making"). Cell areas (in pixels) were determined be selecting "Area" under "Analyze, Options," tracing the outlines of the cells with the Freehand Selection Tool, and then selecting "Analyze, Measure." The results were analyzed with Microsoft Excel (Redmond, WA).

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Patch-clamp recording of single-channel and whole-cell currents induced by Nic in cultured human BECs. Nicotinic agonists applied to outside-out, cell-attached, and whole-cell patches from human BECs evoked nicotinic currents, indicating that BECs express functional AChRs. Application of the agonist AnTX (1 µM) to outside-out patches excised from three of 15 BECs elicited single-channel currents whose mean amplitude was ~1 pA at 80 mV (Fig. 1, top). Their open times were fitted by a double-exponential function. The short-lived channels had a mean open time of ~0.2 msec (_fast), and the long-lived channels had a mean open time of ~4.8 msec (_slow). The reversal potential for these nicotinic currents was close to 0 mV; thus, the conductance of the AChR channels expressed by the BECs is ~12 pS.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Functional characterization by patch-clamp of the AChRs expressed by cultured human BECs: single-channel currents. Top, samples of AnTX-induced single-channel currents recorded from excised outside-out patches. Bottom, left, samples of ACh-induced single-channel currents recorded from a cell-attached patch. Right, samples of ()-Nic-induced single-channel currents recorded from a cell-attached patch. The pipette potentials at which the currents were recorded is indicated next to the corresponding trace.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Functional characterization by patch-clamp of the AChRs expressed by cultured human BECs: whole-cell currents and effect of -BTX. A, Samples of whole-cell currents evoked by 2-sec pulse application of ACh (1 µM and 1 mM) to a BEC held at 80 mV. B, Samples of whole-cell currents evoked by 2-sec pulse application of Nic (1, 10, and 100 µM) to a BEC held at 100 mV. C, Effect of -BTX on the nicotinic responses recorded from BECs. After recording the whole-cell current evoked by Nic (100 µM), a BEC was perfused for 15 min with external solution containing 10 nM -BTX. After exposure to -BTX, the sensitivity of the cell to Nic was tested again. More than 90% of the response was blocked by -BTX. Holding potential, 100 mV.

Assay of neuronal-type AChRs by binding of ³H-labeled epibatidine. Scatchard analysis of [³H]epibatidine binding to PC12 cells detected two populations of binding sites. Their K_d values were 70 pM (B_max = ~800 sites/cell) and 720 pM (B_max = ~3700 sites/cell) respectively, which are in the range of those described for [³H]epibatidine binding to neurons (Wang et al. 1996). We carried out [³H]epibatidine binding experiments with cultured confluent human BECs using a concentration of [³H]epibatidine (10 nM) that, based on the experiments with PC12 cells and pilot experiments with human BECs using a few increasing concentrations of [³H]epibatidine, we expected to be saturating. In four independent experiments carried out with different batches of [³H]epibatidine and of BECs, we found 500, 630, 1600, and 7800 binding sites/cell, respectively. Fig. 3 reports the results of a representative experiment.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Binding of [3H]epibatidine to cultured human BECs. Total and nonspecific binding of [3H]epibatidine to confluent BECs. We used replicates of seven samples, each containing 2 × 106 cells. The nonspecific binding was obtained by incubating the cultures with an excess of nonlabeled epibatidine (10 µM) before the addition of 1 nM [3H]epibatidine. Preincubation with nonlabeled epibatidine significantly (p < 0.0003) reduced [3H]epibatidine binding. The specific binding, obtained by subtracting the nonspecific binding from the total binding, corresponded in this experiment to 630 binding sites/cell.

Detection of neuronal AChR subunit transcripts in cultured human BECs by RT-PCR: Verification of the identity of the 3 transcript. We used RT-PCR to investigate the presence of mRNA transcripts for AChR subunits in mRNA isolated from human BEC cultures. We used primers specific for the 2, 3, 4, 5, 6, 2, 3, and 4 AChR subunits. Fig. 4A reports the results of one of several consistent experiments. The primers for the 3, 5, 2, and 4 subunits yielded products of expected size. The 2 primers yielded a second product of higher molecular weight than expected, for both BEC and brain cDNA. This second band was especially prominent when we used BEC cDNA. We do not have any explanation at this point regarding the nature of the high molecular weight product obtained with the 2 primers. The 2, 4, 6, and 3 primers and the negative controls never yielded PCR products. When we used adult muscle cDNA, the 3, 5, 2, and 4 subunits did not yield detectable products even after 40 PCR cycles, whereas the 1 primers yielded two products of the expected size (Fig. 4B). The presence of two PCR products is due to the presence in adult human muscle of two isoforms of the  subunit. One of those isoforms includes a 75-bp sequence, encoded by an exon termed P3A, which is spliced out in the second isoform (Beeson et al., 1990).

View larger version (65K): [in this window] [in a new window]   Fig. 4.   Detection of 3, 5, 2, and 4 subunit transcripts in cultured human BECs by RT-PCR. A and B, RT-PCR experiments using cDNA from BECs, primer specific for the different neuronal AChR subunits, as indicated, and 35 PCR cycles. All primers yielded a PCR product of the expected size when cDNA from human brain was used. When human BEC cDNA was used, the primers for actin and for the 3, 5, 2, and 4 subunits yielded products of the expected size, whereas the primers for the 2, 4, 6, and 3 subunit did not. Primers for actin were used as a positive control for RT-PCR effectiveness. Mock cDNA did not yield any product band. C, RT-PCR experiments using cDNA from adult human muscle, primer specific for the different neuronal AChR subunits, as indicated, and 40 PCR cycles. Primers specific for the muscle 1 subunit were used as a positive control for RT-PCR effectiveness. The 3, 5, 2, and 4 subunit primers did not yield detectable products even after 40 cycles, whereas the 1 primers yielded two products of the expected size. They are due to the presence in adult human muscle of two 1 isoforms, one of which includes an additional 75-bp sequence.

Detection of neuronal AChR subunit transcripts in cultured human BECs and in rat trachea by in situ hybridization. We wanted to verify that the subunit transcripts detected in the cell cultures were expressed in vivo. Toward this goal, we carried out in situ hybridization experiments using both cultured human BECs and sections of rat trachea. We used probes specific for each of the AChR subunits detected by the RT-PCR experiments. In cell cultures and in the epithelial layer of trachea sections, all probes yielded a clear and specific signal, which was absent when we used the corresponding "sense" probe (Fig. 5).

View larger version (102K): [in this window] [in a new window]   Fig. 5.   Detection of 3, 5, 2, and 4 subunit transcripts in cultured human BECs and rat BECs in intact tissue by in situ hybridization. A, In situ hybridization experiments using cultured human BECs. The 3, 5, 2, and 4 subunit probes consistently yielded a specific strong signal. Insets, negative results obtained with the corresponding "sense" control probes. B, In situ hybridization experiments using rat trachea sections. The 3, 5, 2, and 4 subunit probes consistently yielded specific signals (arrows). Insets, negative results obtained with the "sense" control probes.

Detection of AChR subunit proteins in the BECs in vivo by immunofluorescence. Because we could not use patch-clamp approaches to demonstrate the expression of AChRs in intact tissue, we studied the presence of AChR subunit proteins in sections of rat trachea by immunofluorescence, using antisera specific for unique sequence regions of the 3 and 5 subunits.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Detection of 3 and 5 subunit proteins in rat BECs in intact tissue by immunofluorescence. A, The anti-3 and anti-5 antisera yielded a clear stain of the epithelial layer in sections of trachea (anti -3 and anti 5) (magnification, 400×). The BECs were identified by double immunofluorescence staining with an anti-centrin antibody (anti centrin). The signal resulting from the anti-3 and anti-5 antisera was blocked by preincubation with the corresponding peptides (anti 3 + pep. and anti 5 + pep.). Preincubation with the 3 or 5 peptides did not affect the signal of the control anti-centrin antibody (anti-centrin + pep.). Merge, the overlay of the images obtained for the same section using the anti-5 antiserum and the anti-centrin antibody, in the absence of 5 blocking peptides (top) and in the presence of 5 blocking peptides (bottom). The overlay of the red and green signal resulting from binding of the two antibodies to the same epithelial cells results in the yellow color of the merged image. On the other hand, when the same overlay is used for antibody binding obtained in the presence of blocking peptides, only the green color of the anti-centrin antibody binding, which is not blocked by AChR peptides, is present in the merged image. B, The anti-3 and anti-5 antisera did not bind to the subunits of muscle AChR in sections of rat adult muscle (anti 3 and anti 5) (magnification, 400×). The neuromuscular junctions were identified by double immunofluorescence staining with -BTX (-BTX). Merge, the overlay of the images obtained for the same section using the anti-5 or anti-3 antisera, as indicated, and the -BTX. The green signal observed with the anti-AChR subunit antisera, which is weak and diffused, as expected by nonspecific binding, never overlapped the red signal that indicated the presence of muscle AChR at the synaptic junctions.

Assay of the effect of cholinergic antagonists on the cell shape and cell/cell adhesion of cultured human BECs. We investigated whether the BEC AChRs modulate cell motility by testing the effects on confluent human BEC cultures of Mec (from 100 µM to 5 mM) and -BTX (10 nM and 1 µM). Mec and -BTX block ganglionic 3 AChRs (Conti-Fine et al., 1994; Galzi and Changeux, 1995). Figs. 7 and 8 report the results of representative experiments. Within 5 min after the application of either of those compounds, the BECs retracted their flat cytoplasmic flaps. Their cytoplasm became a thin layer around the nucleus, with reduction of the area covered by the cells, and the cells detached from each other. Both Mec and -BTX had this effect at all the concentrations used, although the extent and the rapidity of the effect increased with the concentration (Fig. 7). When we used 5 mM Mec or 1 µM -BTX, all cells responded to the drug (Fig. 7). Most cell responded to Mec even at the lowest concentrations we used, although the reduction in size was not as profound as that we observed for 5 mM Mec. When we used 10 nM -BTX, several cells, did not respond (Fig. 7).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Human BECs shrink under the influence of Mec. Area covered by individual randomly selected BECs after the addition of drug-free medium (A), medium containing different concentrations of -BTX, as indicated (B), or medium containing different concentrations of Mec, as indicated (C). The area covered by the individual cells is expressed as percent of the area covered by the same cell before the addition of the antagonist or of the drug-free medium. Cells that were not exposed to any drug had small, reversible fluctuations of their size, which did not exceed 20%. The cell area decreased after -BTX or Mec addition to an extent that increased with the concentration of the drug. With 5 mM Mec or 1 µM -BTX, all cells responded to the drug. Most cells responded to Mec at all the concentrations we used. When we used 10 nM -BTX, several cells did not respond. The reduction of the cell surface induced by the lower concentrations of Mec we used reversed spontaneously over the observation period. The effect of Mec was reversible even at the highest concentration we used because the cell area increases quickly after washout of Mec (arrow in the 5 mM plot of C).

View larger version (86K): [in this window] [in a new window]   Fig. 8.   Human BECs shrink under the influence of Mec. The effect of Mec on cultured BECs in the presence of 5 mM Mec and after washing (between 6 and 7). Same cells at different time intervals are indicated inside each picture (hour, minutes, seconds) after Mec administration. Inside and below the pictures we report the outline of the largest cell in this field [note the different scale (2:1) of the outline below the pictures compared with the cell size in the actual pictures] to facilitate identification of the area occupied by the cell during the experiment.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Demonstration of functional AChRs in BECs. This study provides several lines of evidence suggesting that human and rodent BECs express functional AChRs similar to those expressed by some neurons, which can be activated by Nic. The presence of AChRs in BECs is demonstrated by the results of both structural and functional studies. First, the results of PCR and in situ hybridization experiments indicate that BECs, both in culture and in vivo, express mRNA encoding each of the subunits that contribute to AChRs of ganglionic type (3, 5, 2, and 4 subunits), whereas they do not seem to express other constituent subunits of neuronal AChRs (2, 4, 6, and 3) (Figs. 4 and 5). Second, the binding of [³H]epibatidine demonstrated the presence in cultured human BECs of a nicotinic cholinergic binding site (Fig. 3). Third, the binding of specific antibody to sections of trachea demonstrated that the 3 and 5 proteins are expressed on the BEC surface in the intact tissue (Fig. 6). Fourth, the patch-clamp experiments demonstrated that BECs express functional AChRs: they are activated by ACh and Nic, are blocked by -BTX, and have ion-gating properties similar to those of AChRs formed by 3, 5, and 2 or 4 subunits (Figs. 1 and 2).

BEC AChRs are functionally similar to AChRs of ganglionic neurons. The ion-gating properties of the ACh- and Nic-activated ion channels measured in the patch-clamp experiments and their block by -BTX (Fig. 2B) are consistent with the properties of the AChR isotypes expressed by neurons of sympathetic ganglia (Figs. 1 and 2). Other properties of BEC AChRs are consistent with the functional characteristics of certain subtypes of neuronal AChRs (Conti-Fine et al., 1994; Galzi and Changeux, 1995). First, the frequency of channel activity induced in BECs by application of the nicotinic agonists to the cell-attached patches increased with hyperpolarization (Fig. 1, bottom). Second, the open times of the BEC channels were prolonged by hyperpolarization of the membrane (Fig. 1, bottom).

Numbers of AChRs expressed by BECs. We found excellent agreement between the conclusion of the structural and the patch-clamp studies regarding the likely subunit composition of the BEC AChRs. On the other hand, our attempts at measuring the number of AChRs expressed by cultured human BECs yielded variable results. [³H]Epibatidine binding experiments yielded variable numbers of binding sites. The number of specific binding sites for [³H]epibatidine is very small as compared with the nonspecific binding (Fig. 3). Thus, the [³H]epibatidine binding assay is a qualitative verification of the presence of AChR on the BECs, rather than an accurate assessment of their number. Also, the data measuring whole-cell currents evoked by saturating concentrations of agonists in the electrophysiology experiments were variable and too scant to deduce a reliable estimate the number of functional AChR per cell. Still, the electrophysiology data suggest a lower number of receptors than those revealed by [³H]epibatidine binding.

Possible physiological function of BEC AChRs. What is the physiological ligand for the BEC AChRs? Very high levels of ACh (up to 14 µmol/g of tissue) and of choline acetyltransferase and ACh esterase (the enzymes that synthesize and degrade ACh) are present in the rabbit tracheal mucous membrane (Sastry and Sadavongvivad, 1979). Isolated human bronchi synthesize and release ACh (Wessler et al., 1995). Those findings could be explained by the cholinergic parasympathetic innervation of the airways. However, several lines of evidence indicate that BECs themselves synthesize and secrete ACh. ACh was found in extracts of surface epithelial cells isolated from human bronchi (Klapproth et al., 1997). Human BECs contain the enzyme choline acetyltransferase, which synthesizes ACh. The presence of this enzyme was demonstrated by immunohistochemistry and Western blots of sections of human bronchi and by detection of enzyme activity in isolated human BECs (Klapproth et al., 1997). In the current study, the effect of Mec and -BTX indicates that the BEC AChRs in cultures are normally being activated by endogenous agonist, as it occurs for the AChRs of skin keratinocytes.

Potential pathological consequences of BEC exposure to Nic. The characteristics of the BEC AChRs demonstrated here give clues to the pathological effects that might result from acute and chronic exposure to Nic. All AChR isotypes share the property of being desensitized after prolonged exposure to agonists (Conti-Fine et al., 1994; Galzi and Changeux, 1995). However, AChRs that contain the 3 subunit are much more resistant to desensitization than other AChR subtypes, such as the 4 and 7 subtypes (Olale et al., 1997). Human 3 AChRs can exist as hetero-oligomers of 32, 34, 325, and 345 subunits (Conti-Fine et al., 1994; Galzi and Changeux, 1995; Wang et al., 1996). In the 3 AChRs that include the 5 subunit, Nic acts as a full agonist and elicits responses as large as those obtained by full stimulation with ACh. In contrast, Nic acts only as a partial agonist on 3 AChR that do not include the 5 subunits (Wang et al., 1996). Furthermore, the presence of the 5 subunit increases the otherwise slow rate of desensitization of the 3 AChRs (Wang et al., 1996). The BEC AChRs likely include 325 and 345 AChR complexes; thus, they should be fully activated and then desensitized by Nic. Continued exposure to high Nic concentrations on the bronchial surface of cigarette smokers should desensitize the BEC AChRs, making them unable to respond to the endogenous ACh.