Abstract
ACh receptors sensitive to nicotine (nAChR) are present in human skin keratinocytes and in bronchial epithelial cells. They are stimulated by ACh secreted by the same cells that express them, and they modulate cell motility and shape. A variety of non-neuronal tissues, including endothelial cells, synthesize ACh, which raises the possibility that they are sensitive to nicotine. We demonstrate here that endothelial cells that line blood vessels express functional nAChRs. Their structure and ion-gating properties are similar to those of the nAChRs expressed by ganglionic neurons and by skin keratinocytes and bronchial epithelial cells. In situ hybridization experiments using primary cultures of endothelial cells from human aorta demonstrated the presence in these cells of the subunits believed to contribute to ganglionic ACh receptors (AChRs) of the α3 subtype: α3, α5, β2 and β4. Binding of radiolabeled epibatidine—a high-affinity specific ligand of certain neuronal AChRs, including the α3 subtypes—revealed the presence of approximately 900 specific binding sites per cell. We assessed the presence of functional AChRs by patch-clamp experiments. Cultured human endothelial cells express ion channels that are opened by (+)-anatoxin-a and are blocked by dihydro-β-erythroidine. These are specific agonist and antagonist, respectively, of neuronal AChRs of the α3 subtype. The ion-gating properties of the endothelial AChRs were similar to those of neuronal ganglionic AChRs. The presence of AChRs sensitive to nicotine in endothelial cells may be related to the toxic effects of nicotine on the vascular system.
ACh and its receptors are among the best-characterized neurotransmitter/receptor systems (Conti-Fine et al., 1994;Changeux, 1995; Lindstrom, 1995; Albuquerque et al., 1997). Cholinergic neurotransmission is used by a variety of neuronal systems and in a broad range of animals, from invertebrates to mammals (Conti-Fine et al., 1994; Changeux, 1995; Lindstrom, 1995). Many non-neuronal cells synthesize and secrete ACh (Sastry and Sadavongvivad, 1979; Wessler et al., 1995). This raises the possibility that non-neuronal tissues also use ACh as a chemical message, mediating cell signaling in an autocrine or paracrine manner.
ACh binds and activates two types of receptors, the muscarinic and the nicotinic receptors. The muscarinic ACh receptors are members of the superfamily of single-subunit, G protein-coupled metabotropic receptors (Eglen et al., 1996). The nAChRs are so designed because they are activated by nicotine. They are the best-known members of the superfamily of the ionotropic neurotransmitter receptors (Conti-Fineet al., 1994; Changeux, 1995; Lindstrom, 1995; Albuquerqueet al., 1997). The nAChRs are a family of pentameric proteins formed either by a single type of subunit (homo-oligomeric nAChRs) or by different, homologous subunits, which are symmetrically arranged around a central ion channel (Conti-Fine et al., 1994; Changeux, 1995; Lindstrom, 1995; Albuquerque et al., 1997). Different nAChR isotypes exist in muscle and neurons. Mammalian neurons express at least eight different α subunits (α2to α9) and three β subunits (β2 to β4). The combinatorial association of different α and β subunits results in a large variety of neuronal nAChR subtypes (Conti-Fine et al., 1994; Changeux, 1995; Lindstrom, 1995;Albuquerque et al., 1997).
We recently demonstrated that nAChRs sensitive to nicotine are present in human skin keratinocytes (Grando et al., 1995) and in the epithelial cells that line the bronchi of humans and rodents (Mauset al., 1998). Keratinocytes and bronchial nAChR are similar in their structure and ion-gating properties to nAChRs of the α3 subtype (that is, containing the α3subunit) expressed by certain neurons, particularly by the neurons of sympathetic ganglia. They are probably activated by ACh synthesized and secreted by the same cells that express the nAChR (Grando et al., 1993, 1995; Maus et al., 1998). The keratinocytes and bronchial epithelial nAChRs appear to modulate cell motility and adhesion. Their block by antagonists specific for ganglionic nAChRs, such as κ-bungarotoxin and mecamylamine, causes cell paralysis and cell-cell detachment (Grando et al., 1993; Maus et al., 1998).
These findings and the frequent presence of ACh in non-neuronal tissues raise the possibility that different non-neuronal cell types may express nAChRs sensitive to nicotine. We previously found, in cells that line external and internal surfaces (“tegumental” cells), nAChR that appeared to be involved in maintaining the integrity of the lining of those surfaces. The finding that many surface cells synthesize ACh (Grando et al., 1993; Klapproth et al., 1997) supports the hypothesis that tegumental cells may use a nicotinic cholinergic signaling system to modulate their own motility and shape.
We undertook this study to determine whether autocrine activation of nAChRs might be a common mechanism by which tegumental cells modulate their shape. Endothelial cells, although they are of nonepithelial type, line large internal surfaces of the body of vertebrates. Therefore, we searched for nAChRs in human endothelial cells. The presence of nAChRs on endothelial cells would have implications for human pathology, because in tobacco users such nAChRs would be exposed to high concentrations of nicotine and therefore might be overactivated, and/or overdesensitized. We demonstrated here that the endothelial cells that line the human aorta express functional AChRs of the nicotinic type.
Materials and Methods
Cell cultures.
Primary cultures of human aortic endothelial cells (Clonetics; BioWhittaker, San Diego, CA) were seeded in T-25 culture flasks (25 cm2, Corning, New York, NY) at 4 × 103 cells/cm2 in endothelial growth medium (Clonetics; BioWhittaker, San Diego, CA). We maintained the cells at 37°C in an atmosphere containing 5% CO2 and changed the medium every 48 to 72 h. When the cells reached 80% to 90% confluence, we detached them from the plastic by mild trypsinization, for further expansion in culture or for experimentation. For3H-epibatidine binding assays, subsequent passages (2–5) in T-75 culture flasks (75 cm2) were necessary to obtain the necessary numbers of cells. For in situ hybridization, we plated the cells on 12-mm glass coverslip circles (5 × 102 cells per slip) in 24-well plates (Corning, New York, NY), and they were grown until they reached confluence.
We obtained primary cultures of bovine aortic endothelial cells from freshly dissected bovine aorta, as follows. The endothelial layer was isolated by overnight digestion with collagenase, followed by scraping of the lumenal surface with sterile swabs. We propagated bovine aortic endothelial cells in culture, using the conditions described above for the human cells. Positive staining with DiI-LDL (Biomedical Technologies, Stoughton, MA), which is specific for endothelial cells (Voyta et al., 1984), confirmed that the cells were of the endothelial lineage.
Assay of AChR subunit transcripts by in situhybridization.
We carried out in situ hybridization experiments (Cox et al., 1984) by using confluent cultures of human endothelial cells and probes specific for each of the nAChR subunits that we found to be expressed by skin keratinocytes (Grandoet al., 1995) and bronchial epithelial cells (Maus et al., 1998)—the α3, α5, β2 and β4 subunits. The probes were transcribed in vitro from DNA clones (a generous gift of Dr. C. Lobron, University of Mainz) and labeled with Digoxygenin-UTP (Boehringer Mannheim, Ingelheim, Germany). The labeled single-stranded probes were hybridized to mRNA of the cell under conditions of high-stringency hybridization. The conditions we used allowed the probes to bind only to their corresponding mRNA (Cox et al., 1984). To detect the bound probe, we added anti-Digoxygenin antibody coupled to alkaline phosphatase (Boehringer Mannheim, Ingelheim, Germany). We used the NBT/BCIP mixture (Boehringer Mannheim, Ingelheim, Germany) as a substrate for alkaline phosphatase. Absence of the signal when we used the corresponding “sense” probe demonstrated the specificity of the binding of the probes.
3H-epibatidine binding assay.
We verified the presence of neuronal-type AChR on confluent cultured human and bovine endothelial cells by determining the binding of 3H-labeled epibatidine. Epibatidine is a specific, high-affinity ligand of several neuronal AChRs, including the α3 AChR subtypes (Gerzanichet al., 1995; Wang et al., 1996). We used the neuronal PC12 cell line, which expresses α3 neuronal AChR (Conti-Fine et al., 1994), to identify the saturating concentration of epibatidine in our experimental system. Samples containing 0.5 to 1 × 106 PC12 cells were incubated with increasing concentrations of 3H-epibatidine (NEN; 0.1–10 nM; specific activity 48 Ci/nMol) for 4 h at 4°C and harvested by vacuum filtration over Whatman GF/C filters. We washed the filters three times by vacuum filtration with 4 ml of ice-cold phosphate-buffered saline, pH 7.4, and counted them by liquid scintillation. We determined nonspecific binding by preincubating the cells with 10 μM nonradiolabeled epibatidine for 30 min at 4°C before addition of 3H-epibatidine.
We determined the binding of 3H-epibatidine to human and bovine endothelial cells by using single-dose3H-epibatidine binding assays because of the small number of cells that we could grow. We used suspensions of trypsinized endothelial cells (0.3–2.6 × 106 cells/sample in a final volume 100 μl). For each condition, we set up samples at least in triplicate, and up to nine replicates. It is advisable to use a large number of replicates for this assay, because the small number of nAChRs expressed by the endothelial cells makes the specific signal small compared with the nonspecific binding. We incubated the cells with 10 nM 3H-epibatidine for 4 h at 4°C and harvested them by vacuum filtration over Whatman GF/C filters, as described above, or by centrifugation. We counted the filters or the solubilized pellets (in 1% sodium dodecyl sulfate in water) by liquid scintillation. We determined nonspecific binding by incubating the cells with 10 mM nonradiolabeled epibatidine for 30 min at 4°C before adding 3H-epibatidine.
Patch-clamp recording of single-channel currents.
We recorded currents from confluent cultured human and bovine endothelial cells using standard patch-clamp techniques (Hamill et al., 1981) and an LM-EPC-7 patch-clamp system (List Electronic, Darmstadt, FRG). We recorded single-channel currents from outside-out patches excised from endothelial cells (Hamill et al., 1981). The resistance of the recording pipettes was 6 to 8 megohms. We delivered the solutions containing the cholinergic ligands to be tested by using a multibarrel perfusion system that consisted of an array of glass capillary tubes. The patches were normally perfused for 1 to 2 min with agonist-free external solution. Then they were perfused for 3 to 4 min with an external solution containing 1 μM of the nicotinic cholinergic agonist, AnTX. Finally, they were perfused for 1 to 3 min with drug-free external solution. In one experiment, we tested the ability of the nicotinic antagonist dihydro-β-erythroidine (10 nM) to affect the frequency of the nicotinic ion channel activity.
The 10-kHz signal output from the EPC-7 apparatus was transferred to a video cassette recorder via a pulse-code modulation device (Neurodata Neurorecorder DR-384, Neuro Data Inst. Corp.) for off-line analysis. The electrical signal was filtered at 3 kHz in a Bessel filter (8-pole, −3 dB, Frequency Devices 902, Frequency Devices, Inc., Haverville, MA), digitized at 12.5 kHz, and was analyzed with the IPROC-2 program (Axon Instruments). 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 with the NFITS program (Axon Instruments).
The composition of the external solution used to bath the cells was (in mM) NaCl 165, KCl 5, CaCl2 2, HEPES 5 and dextrose 10 (pH = 7.3; osmolarity = 340 mOsM). The composition of the internal solution used for outside-out patches was (in mM): CsCl 80, CsF 80, EGTA 10, HEPES 10 (pH = 7.3; osmolarity = 330 mOsM).
Results
Presence of neuronal nAChR subunit transcripts in cultured human endothelial cells as demonstrated by in situhybridization.
We used in situ hybridization to determine the presence, in human endothelial cells, of mRNA transcripts for each of the subunits forming the previously described tegumental nAChRs (Grando et al., 1995, Maus et al., 1998). We used probes specific for each of the nAChR subunits previously detected in human bronchial epithelial cells and keratinocytes—the α3, α5, β2 and β4 subunits (Grando et al., 1995; Mauset al., 1998)—and confluent cultures of human endothelial cells. All probes yielded clear and specific signals, which were absent when we used the corresponding “sense” probes (fig.1).
Assay of neuronal-type AChRs by binding of 3H-labeled epibatidine.
We determined the dose dependence of 3H-epibatidine binding to nAChR in our system by using PC12 cells, which express nAChRs of the α3subtype (Boulter et al., 1986). 3H-epibatidine bound in a specific manner, and the binding was saturable at concentrations of 5 nM or higher. Scatchard analysis of3H-epibatidine binding to PC12 cells detected two populations of binding sites. Their Kd values were 70 pM (Bmax = ∼1400 sites/cell) and 720 pM (Bmax = ∼6600 sites/cell), respectively, which are in the range of those described for3H-epibatidine binding to neurons (Wang et al., 1996).
We carried out four independent 3H-epibatidine binding experiments with different batches of 3H-epibatidine and different batches of cells. In three of them, we used confluent cultures of human endothelial cells. In one we used cultures of confluent bovine endothelial cells. We used a concentration of3H-epibatidine (10 nM) that, on the basis of experiments with PC12 cells and cultured human endothelial cells, we expected to be saturating. The experiment yielded similar results regardless of the source of endothelial cells. The average specific3H-epibatidine binding obtained in the four independent experiments was 1448 ± 450 binding sites/cell. Figure2 shows the actual data that we obtained in one of the experiments that used human endothelial cells.
Patch-clamp recording of single-channel currents induced by nicotinic agonists in cultured human and bovine endothelial cells.
We excised stable patches from 5 out of 11 human endothelial cells and from 6 out of 11 bovine endothelial cells. The nicotinic agonist AnTX, applied to outside-out patches from both human and bovine endothelial cells, evoked nicotinic currents. This indicates that endothelial cells express functional nAChRs. The characteristics of the single channel that we detected were the same in both species.
Application of 1 μM AnTX to patches held at −80 mV elicited single-channel currents (fig. 3) whose mean amplitude was ∼2.8 pA and whose lifetime was 0.74 ms. Considering the reversal potential of the currents to be close to zero, the conductance of these channels would be close to 35 pS. The frequency of AnTX-induced channels was low (0.1–0.9 events/s). The nicotinic antagonist dihydro-β-erythroidine (10 nM) reduced the frequency of nicotinic channel activity by approximately 75% (fig. 3).
Discussion
This study offers several lines of evidence suggesting that the endothelial cells that line blood vessels express functional nAChRs of the nicotinic type, similar to the nAChRs expressed by ganglionic neurons. The results of both structural and functional studies demonstrated the presence of nAChRs in endothelial cells. First, the results of in situ hybridization experiments indicated that human endothelial cells express mRNA that encodes each and all the subunits that contribute to nAChRs of ganglionic type—α3, α5, β2 and β4 subunits (fig. 1). Second, the binding of3H-epibatidine demonstrated the presence in both human and bovine endothelial cells of a nicotinic cholinergic binding site (fig.2). Third, the patch-clamp experiments demonstrated that human and bovine endothelial cells express functional nAChRs that are activated by the specific nicotinic agonist AnTX, are blocked by the specific antagonist dihydro-β-erythroidine (fig. 3) and have ion-gating properties similar to those of ganglionic nAChRs, formed by α3, α5 and β2 or β4 subunits (Papke, 1993). The structural and functional properties of the endothelial cell nAChRs are similar to those of the nAChRs expressed by other tegumental cells, the skin keratinocytes and the bronchial epithelial cells (Grando et al., 1995; Mauset al., 1998).
The ion-gating properties of the AnTX-activated ion channels measured in the patch-clamp experiments using either human or bovine endothelial cells are consistent with those of the AChR isotypes expressed by neurons of sympathetic ganglia (fig. 3). The conductance of the endothelial nAChRs in response to AnTX (35 pS) is very similar to that of the nAChRs expressed in chick ciliary ganglia (from 30 to 42 pS in different studies; reviewed in Papke, 1993), in rat cervical ganglia (25 pS; reviewed in Papke, 1993), in rat parasympathetic cardiac ganglia (32 pS; Fisber and Adams, 1991) and in primary cultures of mammalian ganglionic neurons (from 32 to 37 pS in different studies; reviewed in Papke, 1993). Also, the channel-open time of the endothelial nAChRs (0.74 ms) is consistent with those found for ganglionic nAChRs (Papke, 1993). Neuronal nAChRs formed by the α3 subunit generally include the α5 and the β2 or β4 subunits (reviewed in Conti-Fineet al., 1994; Lindstrom, 1995; Albuquerque et al., 1997). Ganglionic neurons and other neurons that express α3 subunits consistently express all those subunits (Vernallis et al., 1993; Conroy et al., 1992;Conroy and Berg, 1995; Wang et al., 1996). The genes encoding the α3, α5 and β4subunits are part of the same gene cluster in vertebrates (Boulteret al., 1990; Couturier et al., 1990), and they are expressed in highly restricted patterns (McDonough and Deneris, 1997, and references therein). In good agreement with the subunit composition found for the α3 nAChRs physiologically expressed in neurons, the endothelial nAChRs also appear to include α3, α5, β2 and β4 subunits (fig. 1).
We found about 1400 epibatidine binding sites per cell in both human and bovine cultured endothelial cells. Using the same assay, we found about 8000 epibatidine binding sites per cell in the PC12 cells, a rat pheochromocytoma cell line that expresses nAChRs of the α3 subtype. The number of sites measured in the endothelial cells, although lower than that found in the PC12 cells, is of the same order of magnitude. It also compares with the number of sites for epibatidine in human bronchial epithelial cells (500–7800 sites/cell; Maus et al., 1998) and for κ-bungarotoxin (another specific ligand of nAChRs of the α3 subtype) in skin keratinocytes (5500 sites/cells; Grando et al., 1995).
A caveat of the present study is that we used cultured cells, rather than endothelial cells in vivo. However, the following considerations reduce this concern. First, we used primary cell cultures that had been passaged for a limited number of times (six or fewer) and formed confluent monolayers with endothelial morphology (Imcke et al., 1991). Thus it is reasonable to assume that expression of the nAChRs that we detected here does not represent a consequence of lack of differentiation. Second, our previous studies on skin keratinocytes and bronchial epithelial cells (Grando et al., 1995; Maus et al., 1998) indicated that expression of nAChRs in those cultured cells reflected that of the corresponding cells in vivo. By analogy with those studies, the present results suggest that endothelial cells in vivo express nAChRs.
What is the physiological ligand for the endothelial nAChRs? Vascular endothelial cells also express muscarinic receptors for ACh, whose stimulation induces dilation of the vessel through release of nitric oxide (Furchgott and Zawadzki, 1980). They may be stimulated by the ACh present in the blood (Kawashima et al., 1993) or by ACh synthesized by the endothelial cells themselves. Endothelial cells contain choline acetyltransferase (ChAT) (Parnavelas et al., 1985, Arneric et al., 1988) and can synthesize and release ACh (Kawashima et al., 1990, Ikeda et al., 1994). Thus it appears likely that the endothelial nAChRs are physiologically activated by endogenous ACh, as probably occurs for the nAChRs of skin keratinocytes (Grando et al., 1993, 1995) and bronchial epithelial cells (Maus et al., 1998).
Further studies will be necessary to characterize the functional properties and physiological roles of these endothelial nAChRs. In skin keratinocytes and bronchial epithelial cells, nAChRs of the α3 subtype appear to be involved in the maintenance of the flat shape of the cells (which is necessary to form a continuous epithelial lining on the skin) or the bronchial surface (Grandoet al., 1995; Maus et al., 1998). The presence of the same receptor/ligand system in another tegumental cell type, the endothelial cells, and the frequent presence of ACh in surface cells (Klapproth, 1997) support the hypothesis that self-stimulation of α3 nAChRs such as those described here represents a general cellular mechanism for maintenance of the integrity of both external and internal surfaces.
All nAChR isotypes share the property of being desensitized after prolonged exposure to agonists (Conti-Fine et al., 1994;Changeux, 1995; Lindstrom, 1995; Albuquerque et al., 1997). Nicotine is present in high concentrations in the blood of tobacco users (up to 70 ng/ml in heavy smokers; Russel et al., 1980). Continued exposure to nicotine should desensitize the nAChRs of endothelial cells and make them unable to respond to the endogenous ACh. If the endothelial nAChRs play a role in maintaining the integrity of the monolayer lining the blood vessels, then chronic exposure to nicotine may affect this function. This effect may be related to the atherosclerotic lesions often seen in the arteries of chronic tobacco users.
Footnotes
-
Send reprint requests to: Bianca M. Conti-Fine, Department of Biochemistry, University of Minnesota, 1479 Gortner Ave., St. Paul, MN 55108.
-
↵1 Supported by the NIDA program project grants DA05698 and DA08131 (to B.M.C.-F.), and the USPHS grant NS 21296 (to E.X.A.).
-
↵2 These authors contributed equally to this study.
-
↵3 Affiliation: Institute of Biophysics Carlos Chagas Filho, and Department of Basic and Clinical Pharmacology, Federal University of Rio de Janeiro, Rio de Janeiro, RJ 21944, Brazil.
-
↵4 Previously known as Bianca M. Conti-Tronconi.
- Abbreviations:
- AChR
- acetylcholine receptor
- AnTX
- (+)-anatoxin-a
- DiI-Ac-LDL
- 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindo-carbocyanine perchlorate-acetylated low-density lipoprotein
- Nic
- nicotine
- nAChR
- nicotinic acetylcholine receptor
- Received December 22, 1997.
- Accepted April 30, 1998.
- The American Society for Pharmacology and Experimental Therapeutics