Chronic Nicotine Treatment Up-regulates Human α3β2 but Not α3β4 Acetylcholine Receptors Stably Transfected in Human Embryonic Kidney Cells*

Human nicotinic acetylcholine receptor (AChR) subtypes α3β2, α3β2α5, α3β4, and α3β4α5 were stably expressed in cells derived from the human embryonic kidney cell line 293. α3β4 AChRs were found in prominent 2-μm patches on the cell surface, whereas most α3β2 AChRs were more diffusely distributed. The functional properties of the α3 AChRs in tsA201 cells were characterized by whole cell patch clamp using both acetylcholine and nicotine as agonists. Nicotine was a partial agonist on α3β4 AChRs and nearly a full agonist on α3β2α5 AChRs. Chronic exposure of cells expressing α3β2 AChRs or α3β2α5 AChRs to nicotine or carbamylcholine increased their amount up to 24-fold but had no effect on the amount of α3β4 or α3β4α5 AChRs, i.e. the up-regulation of α3 AChRs depended on the presence of β2 but not β4 subunits in the AChRs. This was also found to be true of α3 AChRs in the human neuroblastoma SH-SY5Y. In the absence of nicotine, α3β2 AChRs were expressed at much lower levels than α3β4 AChRs, but in the presence of nicotine, the amount of α3β2 AChRs exceeded that of α3β4 AChRs. Up-regulation was seen for both total AChRs and surface AChRs. Up-regulated α3β2 AChRs were functional. The nicotinic antagonists curare and dihydro-β-erythroidine also up-regulated α3β2 AChRs, but only by 3–5-fold. The channel blocker mecamylamine did not cause up-regulation of α3β2 AChRs and inhibited up-regulation by nicotine. Our data suggest that up-regulation of α3β2 AChRs in these lines by nicotine results from both increased subunit assembly and decreased AChR turnover.

Nicotinic acetylcholine receptors (nAChRs) 1 containing ␣3 subunits are found in autonomic ganglia where they have been found in various combinations with ␤2, ␤4, and ␣5 subunits (1)(2)(3). They have also been found in brain (4), adrenal gland (5), thymus (6), respiratory epithelial cells (7), and keratinocytes (8,9). In ganglia, ␣3 AChRs play a postsynaptic role similar to that of muscle AChRs (1,2). ␣3 AChRs also have been found in various presynaptic locations in rat brain and implicated in promoting neurotransmitter release (e.g. in dopaminergic and noradrenalinergic pathways) (10 -15). We have previously studied the subunit compositions and the pharmacological properties of human ␣3 AChR subtypes using the Xenopus oocyte expression system and the human peripheral neuroblastoma cell line SH-SY5Y (16 -19). We describe here our study of human ␣3 AChRs stably expressed in a derivative of human embryonic kidney (HEK293) cells. Compared with the Xenopus oocyte expression system, stable mammalian cell lines provide several advantages. 1) We can obtain more reproducible and long lasting levels of AChR expression that are required for detailed single channel studies without the seasonal vagaries of transient expression in Xenopus oocytes. 2) We can better study the effects of chronic exposure to ligands such as nicotine than is convenient in a transient expression system like Xenopus oocytes. 3) Cell lines expressing various AChR subtypes provide a potential for drug discovery by high-throughput screening. Previously, others have reported in article or abstract form permanently transfected cell lines expressing rat ␣3␤4 (20,21) or human ␣3␤2 (20,21), ␣3␤4 (23), and ␣3␤2␣5 AChRs (24). This is the first detailed report of a matched set of four cell lines expressing human ␣3␤2, ␣3␤2␣5, ␣3␤4, or ␣3␤4␣5 AChRs.
One important application of human ␣3 AChR cell lines is to study the effect of chronic nicotine exposure. Addiction to nicotine is characterized by up-regulation of nAChRs in brain (25)(26)(27). The number of high affinity nicotine-binding sites in the brains of tobacco smokers and animals chronically given nicotine is increased up to 2-fold (26 -30). Most of this increase is in ␣4␤2 AChRs (30), but the amount of increase varies between brain regions and probably involves several AChR types (28,29). It has been hypothesized that smoking a cigarette results in a rapid bolus of nicotine that activates the mesolimbic dopaminergic system producing pleasure and reward and that nicotine in the smoker's brain slowly builds to a low steady concentration which causes both reversible desensitization and long term inactivation of some AChR subtypes as well as increases in the amount of some AChR subtypes (19,(31)(32)(33). An understanding of the effects of chronic exposure to nicotine on the amount and functional activity of various AChR subtypes might provide better insight into mechanisms of nicotine dependence, tolerance, and withdrawal, as well as the effects of medication with nicotinic drugs.
Chronic exposure to nicotine has been shown to differentially affect both the amount and function of neuronal AChR subtypes. ␣4␤2 AChRs expressed in Xenopus oocytes or a permanently transfected cell line were shown to double in amount when chronically exposed to submicromolar concentrations of nicotine (34). These concentrations of nicotine eliminated most ␣4␤2 AChR function due to permanent desensitization (19,34). On the other hand, a mixture of ␣3 AChRs expressed by the human neuroblastoma cell line SH-SY5Y increased by 600% in response to chronic exposure to very high concentrations of nicotine (18). Micromolar concentrations of nicotine blocked only a small part of ␣3 AChR function (19). It has been suggested that some of the behavioral effects of nicotine are likely to depend on the inactivation of both ␣4␤2 and ␣7 AChRs while leaving ␣3 AChRs and other subtypes available to respond to endogenous ACh-mediated signaling and to the micromolar boluses of nicotine that occur after inhaling smoke (19,35). Like mammalian cell lines expressing ␣4␤2 AChRs (34,36) or ␣7 AChRs (37), cell lines stably expressing various subtypes of ␣3 AChRs provide excellent tools for studying the mechanisms and physiological significance of nicotine up-regulation of human ␣3 AChRs. Here we provide an initial description of these cell lines, including pharmacological and electrophysiological characterization of their AChRs, and study the effects of chronic nicotine exposure on the amount of these AChRs.

EXPERIMENTAL PROCEDURES
cDNAs, mAbs, and Antisera-The cDNAs for human ␣3, ␤2, and ␤4 subunits were cloned in this laboratory and described in our previous report (16). The cDNA for human ␣5 was kindly provided by Dr. Francesco Clementi (see Ref. 38). The cDNA for the human ␣3 subunit was subcloned into the selective mammalian expression vector pcDNA3.1/Zeo(ϩ)(Invitrogen), which carries the Zeocin TM resistance gene. The cDNAs for human ␤2 and ␤4 subunits were subcloned into the expression vector pRc/CMV (Invitrogen), which carries the neomycin resistance gene. The cDNA for the human ␣5 subunit was subcloned into the expression vector pCEP4 (Invitrogen), which carries the hygromycin resistance gene. Monoclonal antibodies mAb210 to ␣1, ␣3, and ␣5 subunits (39), mAb290 to ␤2 subunits (40), mAb268 to ␣5 subunits (41,42), rabbit antiserum 3709 to a synthetic peptide corresponding to amino acids 348 -387 of human ␣3 subunits, and rabbit antiserum 3724 to a synthetic peptide corresponding to amino acids 387-401 of human ␤2 subunits have been described previously (16). Mouse antiserum Hub4.4 to human ␤4 subunits was raised against a fusion protein of human ␤4 subunit large cytoplasmic domain (amino acids 305-419) with bacterial glutathione S-transferase. The fusion protein was constructed in the pGEX-4T-2 vector (Amersham Pharmacia Biotech) and expressed in bacteria. It was affinity purified with a glutathione-Sepharose 4B (Amersham Pharmacia Biotech) column before being used to immunize mice. Hub4.4 does not cross-react with human AChR subunits ␣3, ␤2, or ␣5 on Western blots (data not shown). Rat antiserum to human ␣5 subunits was raised against a fusion protein of human ␣5 subunit extracellular domain and large cytoplasmic domain (43). The fusion protein was constructed in the pET-26b(ϩ) vector (Novagen) and expressed in bacteria. It was purified by gel exclusion chromatography before being used to immunize rats. The antiserum to ␣5 was absorbed with a resin-coupled fusion protein of human ␣3 AChR subunit extracellular and cytoplasmic domains to remove antibodies to conserved epitopes shared with ␣3 and other subunits.
Cell Culture and Stable Transfection of Human tsA201 Cells-Human tsA201 cells, a derivative of the human embryonic kidney cell line 293 (44), were maintained in Dulbecco's modified Eagle's medium (high glucose) (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (Hyclone), 100 units/ml penicillin, 100 g/ml streptomycin (Life Technologies, Inc.), and 2 mM L-glutamine (Life Technologies, Inc.) in a CO 2 (5%) incubator at saturating humidity. The ␣3␤2 cell lines were established by co-transfecting tsA201 cells with Hu␣3/pcDNA3.1/Zeo(ϩ) and Hu␤2/pRc/CMV using LipofectAMINE (Life Technologies, Inc.) following the manufacture's instructions. The ␣3␤4 cell lines were developed by co-transfecting tsA201 cells with Hu␣3/pcDNA3.1/Zeo(ϩ) and human␤4/pRc/CMV following the same procedure. The ␣3␤2␣5 and ␣3␤4␣5 cell lines were obtained by transfecting the established ␣3␤2 and ␣3␤4 cell lines with Hu␣5/pCEP4. For the cloning of stably transfected cell lines, 72 h after transfection, selection medium containing Geneticin (600 g/ml, Life Technologies, Inc.) and Zeocin (500 g/ml, Invitrogen) was used for the ␣3␤2 and ␣3␤4 cell lines. For the ␣3␤2␣5 and ␣3␤4␣5 cell lines, the third antibiotic drug hygromycin (200 g/ml, Boehringer Mannheim) was added to the above selection medium. The initial screening for cell colonies expressing functional ␣3␤2 and ␣3␤4 AChRs was based on a solid phase radioimmunoassay (RIA) with [ 3 H]epibatidine on mAb210-coated Immulon 4 (Dynatech) microwells. Screening for ␣3␤2␣5 and ␣3␤4␣5 cell lines was performed by both RIA and Western blot analysis with mAb268 (16) to detect the presence of the ␣5 subunit. Untransfected tsA201 cells were used as a negative control for these assays. Chronic exposure of cells to nicotine and/or other ligands (all purchased from Sigma) was performed by adding the ligand(s) into the culture medium at the relevant concentration and incubating the cells in the medium for the indicated duration of time before harvesting the cells. Effects of the protein synthesis inhibitor cycloheximide (Sigma) on nicotine-induced up-regulation were studied by adding the chemical into the culture medium with or without the presence of nicotine and incubating the cells in the medium for a certain period of time before harvesting.
Isolation, Purification, and Immunoprecipitation of AChRs from Cells-Cells were suspended from 35-mm Petri dishes with buffer A (in mM: Na 2 HPO 4 ⅐NaH 2 PO 4 , 50; pH 7.5; NaCl, 50; EDTA, 5; EGTA, 5; benzamidine, 5; iodoacetamide, 15; phenylmethylsulfonyl fluoride, 2) and pelleted by centrifugation. The pellets were rinsed 3 times with buffer A before buffer C (buffer A containing 2% Triton X-100) was used to solubilize the AChRs in the cells (4°C, 1 h). After removing cellular debris by centrifugation, the cleaned extracts were incubated in mAbcoated microtiter wells for solid phase RIA, or with mAb-coupled Actigel (Sterogen) for purifying AChRs for use in immunoblot assays, or loaded directly onto 5-ml sucrose gradients (5-20% sucrose, w/w) for sedimentation analysis (16). For immunoprecipitation of AChRs with subunitspecific antibodies, the extract was incubated with mAb or antiserum in the presence of [ 3 H]epibatidine (5 nM) for 12 h. The AChR-antibody complexes were immunoprecipitated with Zysorbin (Zymed Laboratories Inc.) for mouse antibodies or sheep anti-rat IgG for rat antibodies.
[ 3 H]Epibatidine-labeled AChRs in the pellet were quantified using liquid scintillation counting. Nonspecific precipitation was measured using either normal mouse serum or normal rat serum. For filter binding assays, the AChRs in the extract (100 l) were labeled with [ 3 H]epibatidine (5 nM) at 4°C for 12 h. Then the extract was diluted with 4 ml of 10 mM Tris-HCl, pH 7.5, 0.5% Triton X-100 before it was loaded onto Whatman GF/B filters (presoaked with 0.5% polyethyleneimine). The filters were rinsed three times with 4 ml of 10 mM Tris-HCl, pH 7.5, 0.5% Triton X-100 to remove free [ 3 H]epibatidine. [ 3 H]Epibatidinelabeled AChRs on the filter were quantified using liquid scintillation counting. Nonspecific binding was measured using extracts of nontransfected tsA201 cells.
Cell-surface Labeling with 125 I-mAb210 or mAb210 Plus Fluoresceinlabeled Goat Anti-rat IgG (F-GART)-For surface labeling with 125 I-mAb210, cells in 35-mm Petri dishes were rinsed with phosphatebuffered saline (PBS) before 125 I-mAb210 (2 nM, 0.7-2.6 ϫ 10 18 dpm/ mol, diluted in PBS, 10% fetal bovine serum) was added into the dish. Labeling took place at 25°C for 1 h. Cells were then suspended from the dish and collected in microcentrifuge tubes by centrifugation. The pellets were washed three times with PBS to remove nonspecifically bound mAbs. The bound 125 I-mAb210 was measured by ␥-counting. For surface labeling with mAb210 followed by F-GART (Kirkegaard & Perry Laboratories, Inc.), cells were cultured on collagen-coated glass coverslips. Cells were rinsed once with PBS before fixation in 2% paraformaldehyde/PBS for 20 min at 25°C. After fixation, cells were rinsed 3 times with PBS before being incubated with mAb210 (30 nM) diluted in PBS containing 1% bovine serum albumin for 1 h at 25°C. This was followed by 3 washes with PBS and incubation for 1 h at 25°C with F-GART (50 g/ml). The cells were then washed three times with PBS before being mounted in Pro-Long (Molecular Probes). Digital photomicrographs were taken with a Leica TCS 4D confocal microscope.
Electrophysiology-Transfected cells were plated onto glass coverslips coated with rat tail collagen (Type 1; Collaborative Biomedical Products, Bedford, MA) at least 2 days prior to use. Currents were monitored by standard whole cell patch clamp techniques (45). Agonistcontaining solutions were applied to the cells by gravity fed fused glass tubing that was connected to multiple reservoirs mounted above the recording chamber. The recording solution was composed of the following (in mM): NaCl, 150; KCl, 5; MgCl 2 , 1; CaCl 2 , 2; HEPES, 5; pH 7.3. The recording electrodes were formed from borosilicate glass tubing and had resistances of typically 5-8 M⍀ and were filled with a solution having the following composition (in mM): cesium gluconate, 150; EGTA, 10; HEPES, 10; pH 7.2 (CsOH). Cell access resistances were typically 8 -12 M⍀ and were compensated only when peak currents were in excess of 2 nA. Cells transfected with ␣3␤2 or ␣3␣5␤2 AChRs were treated for 12 h with 100 M nicotine to increase the levels of functional AChRs followed by a minimum of a 1-h wash with normal media prior to recording. Currents were evoked by application of agonist flowing from one of the two glass perfusion tubes. Prior to application of agonist, the cell was isolated in a continuously flowing control solution from the other tube. Agonists were selected via two 6-way valves in series. Currents were activated by a 2-s application of agonist flowing from the glass tubing attached to a piezo-electric application system that was activated by an isolated pulse stimulator (A-M Systems model 2100). The speed of solution exchange was determined by clamping an open recording electrode at 0 mV and then measuring the change in the solution junction potential when moving from a normal recording solution to one that had been diluted 5-fold with deionized water. The time constant for the change in clamping current during the solution exchange was approximately 1 ms at the solution flow rates used for activating currents in these cells. Recordings were obtained with an Axopatch 1-D amplifier (Axon Instruments, Inc., Foster City, CA) with the filtered output at 2 kHz and sampled with pClamp 6.0.3 (Axon Instruments) onto a personal computer using a TL-1 DMA interface (Axon Instruments) and a Labmaster A-D converter (Axon Instruments) at 2 kHz. All currents were normalized to the peak current obtained for 300 M ACh, and concentration/response curves were fitted to a logistic equation in Origin (Version 4.1; Microcal Software, Inc., Northampton, MA). In some cases, the concentration/response relationship would peak and then decline with increasing concentrations of agonist. In the cases where the decline caused an obviously inferior fit, the reduced amplitude responses at higher concentrations were not included in the fit, i.e. for ␣3␤2␣5 both 3 mM ACh and 1 mM nicotine were not included in the fits but are shown in the figures. Representative traces were constructed by opening data files in Axograph 3.55 (Axon Instruments) and exporting data to Canvas 5.0 (Daneba Software, Inc., Miami, FL).
Nicotine Treatment and Electrophysiological Recordings in Xenopus Oocytes-Oocytes were obtained from Xenopus laevis (Nasco). Stage V-VI oocytes were selected and injected with combinations of ␣3 and ␤2 or ␣3 and ␤4 subunit cRNAs (15 ng of each subunit in a total volume of 55 nl). Currents were measured using a standard two-microelectrode voltage clamp amplifier (Oocyte Clamp OC-725; Warner Instrument Corp., Hamden, CT) as described previously (17). All recordings were digitized using MacLab software and hardware (AD Instruments, Castle Hill, Australia). Data were analyzed using Kaleidagraph (Synergy Software, Reading, PA). For nicotine treatment, the oocytes were incubated overnight at 18°C in Petri dishes containing the relevant concentration of nicotine. After overnight nicotine treatment, oocytes were placed in dishes of control saline for at least 1 h before electrophysiological recordings.

Expression of AChR Subunit Combinations in Stably
Transfected tsA201 Cells-Human embryonic kidney tsA201 cells expressing the AChR subunit combinations ␣3␤2 or ␣3␤4 were initially grown in culture medium containing Zeocin to select for the ␣3 containing plasmid and neomycin to select for the ␤2 or ␤4 containing plasmid. Cell colonies that survived for 3 weeks in the selection medium were tested for AChR expression using a solid phase RIA with [ 3 H]epibatidine as ligand (Fig. 1A). Most (Ͼ90%) of the selected colonies contained AChRs that bound [ 3 H]epibatidine. Expression levels varied. Colonies expressing higher levels of AChRs (Ͼ200 fmol/mg protein for ␣3␤2 cell lines, and Ͼ800 fmol/mg protein for ␣3␤4 cell lines) were recloned by limiting dilution in 96-well plates. Expression of ␣3 AChRs in ␣3␤2 and ␣3␤4 cell lines was stable with respect to [ 3 H]epibatidine binding for at least 3 months in continuous culture.
Our previous studies using the Xenopus oocyte expression sys- For each assay, cells from one 35-mm Petri dish (containing approximately 1 ϫ 10 6 cells with a protein content of 0.1 mg) were collected for each assay. Values represent the mean Ϯ S.E. from four dishes of each AChR subtype. B, Western blots detect ␣5 subunits in ␣3␤2␣5 and ␣3␤4␣5 cell lines. ␣3 AChRs were affinity purified with mAb290 specific for ␤2 subunits or antiserum specific for ␤4 subunits as appropriate. The immunoblot was probed with mAb268 which is specific for denatured ␣5 subunits. C, the fraction of ␣3 AChRs containing ␣5 subunits in ␣3␤2␣5, and ␣3␤4␣5 cell lines (upper), and the SH-SY5Y cell line (lower) were measured by immunoprecipitation analysis with antiserum to ␣5 subunits. [ 3 H]Epibatidine (5 nM) was used to label the AChRs for quantification. Data were normalized to the maximum amount of ␣3 AChRs precipitated with a saturating concentration of mAb210 (which recognizes both ␣3 and ␣5 subunits). A maximum of 49, 14, or Ͻ9% of the total ␣3 AChRs were bound in the presence of an excess of antiserum to ␣5 in the three lines. Values represent mean Ϯ S.E. of triplicate determinations. tem and immunoprecipitation of ␣3 AChRs incorporating epitope-tagged ␣5 subunits indicated that co-expression of equal amounts of cRNAs for ␣3, ␤2 or ␤4, and ␣5 subunits resulted in efficient incorporation of ␣5 subunits into Ͼ55% of AChRs (16).
In order to determine how efficiently ␣5 was incorporated in the ␣3 AChRs of our cell lines, we performed immunoprecipitation assays with mAb210 (which binds both native ␣3 and ␣5 subunits) and an antiserum to bacterially expressed ␣5 (which binds to both native and denatured ␣5 subunits) to measure the fraction of ␣3␤2␣5 AChRs in the cell line. Data in Fig. 1C show that 49% of ␣3 AChRs in the cell line contained ␣5 subunits. In a similar manner we determined that only 14% of the ␣3 AChRs in the ␣3␤4␣5 cell line contained ␣5 subunits. In order to determine whether inefficient assembly with ␣5 subunits was an artifact of expression in transfected lines, the fraction of ␣5-containing AChRs in the human neuroblastoma cell line SH-SY5Y was also measured in the same way. Only about 9% of the ␣3 AChRs in the native neuronal cell line appeared to contain ␣5 subunits (Fig. 1C). Thus, it seems that in both transfected cell lines and neuroblastoma cell lines ␣5 subunits are incorporated into ␣3 AChRs less efficiently than is the case in Xenopus oocytes.
We also studied the sedimentation behavior of ␣3 AChRs expressed in the stably transfected cell lines. By comparing their sedimentation properties with those of native ␣3 AChRs from SH-SY5Y and those of human ␣3␤2 AChRs expressed in Xenopus oocytes, we found that in all four cell lines, the AChR complexes (indicated by the binding of [ 3 H]epibatidine) cosedimented with native ␣3 AChRs in the 11 S region, which corresponded to fully assembled pentamers (Fig. 2). There was no evidence of partially assembled ␣3␤2 dimers, for example, which would have been expected to form an epibatidine-binding site at their interface (46) but would have sedimented much more slowly than the ␣3␤2␣3␤2␤2 arrangement of five subunits around the central ion channel expected of native AChRs.
Transfected Cells Express Functional AChRs-All transfected cell lines that exhibited binding of [ 3 H]epibatidine responded with currents to applications of nicotine or ACh, although with differing current amplitudes and differing percentages of cells responding (Figs. 3 and 4). The subunit combination ␣3␤4 gave the most robust responses, typically reaching maximal currents of 1-5 nA, with some cells responding with currents as large as 15-20 nA. The EC 50 values for activation by ACh and nicotine for ␣3␤4 AChRs were 79 Ϯ 8 and 56 Ϯ 15 M, respectively. The Hill coefficients were 1.5 and 1.6, respectively. ACh and nicotine had equivalent efficacies in activating ␣3␤4 AChRs in these cells. The currents of the ␣3␤4 cells exhibited a slow decay during agonist application. ␣3␤2 transfected cells had insufficient expression to characterize their functional properties without nicotine-induced up-regulation (see below). Therefore, for functional studies, these cells were incubated for 12 h with 100 M nicotine followed by a minimum of a 1-h wash in normal culture media prior to recording. All functional studies for ␤2-containing AChRs expressed in tsA201 cells were preceded by nicotine exposure in this manner. The ␣3␤2 AChRs generally had smaller responses and were undetectable in a greater number of cells. The largest currents for these cells were between 1 and 1.5 nA. The EC 50 values for activation by ACh or nicotine were 209 Ϯ 26 and 70 Ϯ 6 M, respectively. The Hill coefficients were 1.7 and 1.3, respectively. For ␣3␤2 AChRs, nicotine appears to have only 60% efficacy compared with ACh. In stark contrast to the ␣3␤4 AChR responses, the decay of ␣3␤2 currents was extremely rapid and usually complete within 0.5 s, indicating a rapid desensitization rate. The need to expose the ␤2-containing cell lines to nicotine for functional studies raises the concern that the functional properties of the AChRs might be altered by nicotine exposure. To address this concern we performed a parallel series of experiments in oocytes expressing ␣3␤2 and ␣3␤4 AChRs, which demonstrated that overnight incubation of the oocytes in 100 M nicotine had no substantial effect on the rates of desensitization (which is also more rapid for ␣3␤2 AChRs expressed in oocytes (16,49)), EC 50 values for activation of the ␣3␤2 or ␣3␤4 AChRs by ACh and nicotine, or the efficacy of nicotine (which also is a partial agonist for ␣3␤2 AChRs expressed in oocytes (16,49)) (data not shown).
The EC 50 for activation of ␣3␤4␣5 AChRs by ACh was 81 Ϯ 15 M and by nicotine was 42 Ϯ 5 M. The Hill coefficients were 1.6 and 1.7, respectively. ACh and nicotine had equivalent efficacies. Thus, the pharmacological properties were essentially identical to ␣3␤4 cells. This may not be surprising given that Fig. 1 shows that only 14% of these AChRs incorporated ␣5 subunits. There was no clear evidence for a dual population of AChRs in the concentration/response relationships. The decay of the currents was indistinguishable from that observed for ␣3␤4 AChR currents. For ␣3␤2 cells transfected with the ␣5 subunit, the cells were also incubated with 100 M nicotine to increase AChR expression followed by at least a 1-h wash. The maximal currents were about 1 nA and bore a striking resemblance in time course to the parent ␣3␤2 AChR cell line responses. Despite the 49% incorporation of ␣5 subunits shown for these AChRs in Fig. 1, there was no evidence of two populations of AChRs greatly differing in their responses. The EC 50 for activation by ACh was 121 Ϯ 18 M and by nicotine was 83 Ϯ 12 M. The concentration/response relationships had Hill coefficients of 1.6 and 1.3, respectively. The efficacy of nicotine increased in the presence of ␣5 but remained less than that of ACh, probably reflecting the somewhat lower incorporation of ␣5 (49%) in the cells compared with Xenopus oocytes (72%) (16). Although the EC 50 for activation by ACh was somewhat lower than that found for the ␣3␤2 cells, there were diminishing responses at the saturation end of the concentration/response curves suggestive of agonist-mediated channel blockade that would attenuate the responses leading to an underestimation of the true EC 50 value and obscuring the relative efficacies of these agonists. Therefore, it would appear that the ␣5 subunit does not alter the apparent affinity of this AChR for activation by ACh when compared with the ␣3␤2 AChR, but the ␣5 subunit might increase the susceptibility to channel blockade by ACh.
Chronic Treatment with Nicotine Up-regulates ␣3␤2 but Not ␣3␤4 AChRs in Cell Lines-Chronic exposure of the ␣3␤2 cell line to nicotine increased the amount of ␣3␤2 AChRs up to 24-fold as measured by [ 3 H]epibatidine binding to immunoisolated solubilized AChRs (Fig. 5). The EC 50 for nicotine-induced up-regulation of ␣3␤2 AChRs was 2 Ϯ 0.3 M (n ϭ 4). This concentration is 30 -40-fold lower than the EC 50 for activation of the ␣3␤2 AChRs by nicotine (Fig. 4) but 10-fold higher than the EC 50 for up-regulation of ␣4␤2 AChRs (34) or the typical serum concentration of nicotine in a tobacco user (35). Although the maximum effect of nicotine was seen at 1000 M, at 0.2 M, which is very close to the serum concentration of nicotine for smokers (47), nicotine was able to up-regulate ␣3␤2 AChRs more than 3-fold in 7 h (Fig. 5). At concentrations as high as 1 mM, nicotine did not affect cell proliferation. The up-regulation could be seen as early as 15 min after nicotine (100 M) exposure and reached a maximum by 8 h. Nicotine also up-regulated ␣3 AChRs in ␣3␤2␣5 cell lines to the same extent (up to 22-fold) as AChRs in the ␣3␤2 cell line (Fig. 5).
FIG. 3. Functional properties of ␣3 AChRs. Representative currents recorded for ACh concentration/response analysis of the four ␣3 cell lines. All currents were recorded at a holding potential of Ϫ60 mV, with the exception of the ␣3␤4 cells which were recorded at Ϫ30 mV to avoid the excessive current required for maintaining adequate clamp at Ϫ60 mV. Agonist application durations are indicated by the solid bars above each series of traces. AChR expression was up-regulated before this analysis by exposure of the ␣3␤2 and ␣3␤2␣5 cell lines to 100 M nicotine for 12 h followed by 1 h of rinsing before analysis. As in Xenopus oocytes expressing cRNAs (16,49), ␣3␤2 AChR responses desensitize more rapidly than do ␣3␤4 AChRs, but in both cases the kinetics are more rapid in transfected cells.
AChRs were visualized by labeling fixed cells with mAb210 followed by F-GART (Fig. 6). Without exposure to nicotine, ␣3␤2 AChRs were virtually undetectable on the cell surface, but ␣3␤2 AChRs were detectable inside permeabilized cells. Nicotine added to the culture medium at 10 M for 12 h dramatically increased ␣3␤2 AChRs on the cell surface and inside the cells. Some clusters of ␣3␤2 AChRs were formed on the cell surface. The average diameter of the clusters was about 1 m. Most ␣3␤4 AChRs were present on the surface of the cells, and incubation in nicotine did not change their amount or distribution. The clustering of ␣3␤4 AChRs on the surface membrane was striking. Their clusters were bigger (1-3 m in diameter) and much more frequent than those of ␣3␤2 AChRs.
Up-regulation by nicotine of ␣3␤2 AChRs was studied by immunoblot assay to demonstrate the increase of both ␣3 and ␤2 subunits in assembled form. ␣3␤2 AChRs were purified with mAb290-Actigel, which binds specifically to ␤2 subunits. Blots of purified ␣3␤2 AChRs were labeled with an antiserum to an Concentration/response curves for the four ␣3 cell lines. Concentration/response curves for nicotine and ACh are shown for the four cell lines. EC 50 values are given along with their Hill coefficients as determined by fitting the data to a logistic equation as described under "Experimental Procedures." Holding potentials were Ϫ30 mV for the ␣3␤4 and ␣3␤4␣5 cells, whereas ␣3␤2 and ␣3␤2␣5 cells were held at Ϫ60 mV. As in Xenopus oocytes expressing cRNAs (16,49), nicotine is a partial agonist on ␣3␤2 AChRs and nearly a full agonist on ␣3␤2␣5 AChRs. However, the potency of agonists on ␣3␤2 AChRs in transfected cells is lower than in oocytes, and the potency on ␣3␤4 AChRs is higher. ␣3-specific oligopeptide (Fig. 7, lanes 1 and 2). An increase in ␣3 subunits was detected as early as 15 min after nicotine exposure. The maximum effect was seen after about 8 h. ␣3␤2 AChRs were also immunopurified using mAb210-Actigel, which recognized ␣3 subunits, and the Western blots labeled with an antiserum to a ␤2-specific oligopeptide. The amount of ␤2 subunits also increased from an early stage of nicotine exposure as well (Fig. 7, lanes 3 and 4). Doublet bands of ␣3 or ␤2 subunits were seen on the blots more obviously when cells were treated with nicotine. This probably arises from variable glycosylation of the subunit, because only one band of ␣3 or ␤2 was seen on the blots if the purified AChRs were deglycosylated before they were loaded on the SDS-polyacrylamide gel (data not shown).
The up-regulation effect of nicotine was subunit-specific. We FIG. 7. Prolonged nicotine exposure causes increased ␣3␤2 but not ␣3␤4 AChRs. AChRs were solubilized with Triton X-100 and isolated with one set of subunit-specific antibodies to obtain assembled subunits. The blots were then probed with another set of antibodies to visualize the isolated subunits. Nicotine treatment increased the amount of ␣3 and ␤2 subunits assembled into AChR complexes in ␣3␤2 cell lines. No change was found for the amount of ␣3 and ␤4 subunits from assembled AChRs in ␣3␤4 cell lines. compared the amount of ␣3␤4 AChRs from the cell line before and after nicotine exposure by surface binding assay (Fig. 6), immunoblot assay (Fig. 7, lanes 5-8), and solid phase RIA (Fig.  11A). None of these assays detected a change even 48 h after nicotine was added to the culture medium. ␣3␤4␣5 AChRs, like ␣3␤4 AChRs, were not up-regulated by nicotine (data not shown). In the neuroblastoma cell line SH-SY5Y, there are both ␣3␤2 and ␣3␤4 AChRs (16). It is known that nicotine can up-regulate ␣3 AChRs in SH-SY5Y cells (18), but it was not clear which subtypes of ␣3 AChRs were up-regulated. In order to resolve this, we used antibodies specific for ␤2 or ␤4 subunits separately to immunoprecipitate ␣3 AChR subtypes in SH-SY5Y cells. After chronic exposure to nicotine, the amount of ␤2-containing AChRs in SH-SY5Y cells increased 3-fold, whereas the amount of ␤4-containing AChRs was not changed (Fig. 8). These observations strongly suggest that ␤2 subunits play an important role in nicotine-induced up-regulation of ␣3 AChRs in neurons.
Nicotine-induced Up-regulation of ␣3␤2 AChRs Results from Enhanced Subunit Assembly and Decreased Turnover Rate of AChRs-We tested the possibility that nicotine treatment increased the affinity of ␣3 AChRs for [ 3 H]epibatidine by comparing the binding to ␣3␤2 AChRs before and after exposure to 100 M nicotine. The K D value was not changed (107 Ϯ 18 nM versus 126 Ϯ 10 nM, n ϭ 4), which means that nicotine did not increase the affinity of detergent-solubilized immunoisolated ␣3␤2 AChRs for [ 3 H]epibatidine. We also excluded the possi-bility that nicotine exposure increased the affinity of ␣3␤2 AChRs for mAb210 and mAb290 by showing that nicotineinduced up-regulation could also be demonstrated by employing a filter binding assay using solubilized AChRs labeled with [ 3 H]epibatidine that did not require use of mAbs (data not shown). It was clear that the up-regulation resulted from increasing the amount of AChRs rather than from increasing their affinity for both mAbs.
Since up-regulation of ␣3␤2 AChRs started as early as 15 min after the cells were exposed to nicotine, it is reasonable to suggest that the proper folding and assembly of ␣3 and ␤2 subunits were enhanced by nicotine. We tested this possibility by probing for ␤2 subunits on immunoblots of ␣3␤2 cell extracts, both affinity purified by mAb210-Actigel and applied directly as crude extract (Fig. 9, upper panel). Comparing the specific signal for ␤2 subunits on the blot with and without exposing the cells to nicotine for 30 min, we found only a small increase in the total amount of ␤2 subunits in the cell, but the amount of ␤2 subunits that was assembled with ␣3 subunits greatly increased (more than 5-fold). Since the affinity of ␣3␤2 AChRs for mAb210 was not changed, the large increase in ␤2 subunits purified by the ␣3-specific mAb210-Actigel a short time after exposure to nicotine (before the total amount of ␤2 subunits increased greatly) suggests that the rate of subunit assembly into ␣3␤2 AChRs increased upon nicotine exposure.
To test whether up-regulation of ␣3␤2 AChRs requires protein synthesis, cycloheximide, a protein synthesis inhibitor, was used to block the synthesis of ␣3 and ␤2 subunits for 3 h before nicotine was added into the culture medium. After 90 min, the total amount of ␣3␤2 AChRs in the cells was measured by RIA (Fig. 9, lower panel). Compared with control cells, nicotine was still able to up-regulate ␣3␤2 AChRs 6-fold without new protein synthesis. This suggests that at least 25% of the up-regulation in ␣3␤2 AChRs (6-fold up-regulation without increased synthesis versus 24-fold otherwise) is due to enhanced assembly and perhaps the reduced degradation of preexisting subunits.
Previous studies in our laboratory showed that nicotine was able to decrease the turnover rate of avian ␣4␤2 AChRs stably expressed in mammalian fibroblast cells (34). To see if the up-regulation effect of nicotine for human ␣3␤2 AChRs could also be partially attributed to the stabilization effect of nicotine on ␣3␤2 AChRs in tsA201 cells, we blocked protein synthesis in the cell using cycloheximide and followed the influence of nicotine on ␣3␤2 AChRs in the cells for 24 h (Fig. 10). In the absence of nicotine, the addition of cycloheximide prevented replacement of ␣3␤2 AChRs, and the total amount decreased with a t1 ⁄2 of around 30 h. When nicotine and cycloheximide were added together, there was a rapid initial increase in ␣3␤2 AChRs due to assembly of pre-existing subunits. Western blots like those in Fig. 7 using mAb210 to ␣3 to isolate assembled AChRs and antiserum to ␤2 to reveal ␤2 on blots showed about half-maximal assembly by 30 min, about twice this amount by FIG. 9. Nicotine enhances subunit assembly of ␣3␤2 AChRs. Upper panel, immunoblots of ␤2 subunits either affinity purified with mAb210-Actigel to bind AChRs containing ␣3 subunits or in a crude extract from ␣3␤2 cells were probed with rabbit antiserum to a human ␤2 subunit synthetic peptide. This showed that nicotine caused an increase of ␤2 subunits assembled with ␣3, much more than that of total ␤2 subunits. Lower panel, protein synthesis in ␣3␤2 cells was blocked with cycloheximide (35 M) for 3 h before nicotine (100 M) was added to the culture medium. This showed that nicotine could still cause some up-regulation of ␣3␤2 AChRs in the cells without synthesis of more subunits. . In control cells, the amount of ␣3␤2 AChRs decreased nearly 50% due to turnover, whereas the amount of ␣3␤2 AChRs in nicotine-treated cells remained stable after an initial increase due to enhanced subunit assembly. Each data point represents the mean Ϯ S.E. of triplicate determinations.
3 h, and a slight further increase by 24 h (data not shown). Fig.  10 shows that by 3 h there was a 13-fold increase in ␣3␤2 AChRs. This exceeds the 5-fold increase caused by nicotine after 3 h in cycloheximide shown in Fig. 9, presumably due to turnover of about half of the unassembled ␣3 and ␤2 subunits during the 3 h after cycloheximide was added and before nicotine was introduced. Both Fig. 9 and Fig. 10 show that after nicotine is added, and the initial burst of enhanced assembly of ␣3␤2 AChRs is complete, the amounts of ␣3␤2 AChRs remain virtually constant for the duration of the experiments. Thus, in the presence of nicotine the rate of turnover of assembled AChRs is greatly decreased from the t1 ⁄2 of 30 h in control cells to an immeasurably slow t1 ⁄2 . Therefore, the 24-fold increase in ␣3␤2 AChRs induced by nicotine in cells in the absence of cycloheximide results both from increased assembly (about 13fold or about half of the total effect) and from reduced degradation (about the remaining half of the 24-fold increase).
Effects of Other Agonists and Antagonists on Up-regulation of ␣3␤2 AChRs-Nicotine is a tertiary amine that can cross cell membranes, which might permit it to influence AChR subunit maturation or assembly from within a cell. Carbamylcholine is a quaternary amine that cannot penetrate the cells and should act only through AChRs in the cell surface. When carbamylcholine was added to the culture medium of ␣3␤2 cells, the amount of ␣3␤2 AChRs was increased by 2.5-fold within 15 min. The effect of carbamylcholine reached a maximum (25fold) after about 12 h (Fig. 11A). Thus, in these cells, an agonist acting only on surface AChRs can cause up-regulation equal in extent to that caused by nicotine.
Previous studies with avian ␣4␤2 AChRs expressed in permanently transfected mouse fibroblasts showed that the competitive antagonist curare prevented the 2-fold up-regulation by nicotine (34). Curare and dihydro-␤-erythroidine (DH␤E) had little or no effect on nicotine-induced up-regulation of ␣3 AChRs in the human neuroblastoma SH-SY5Y (18). Human ␣4␤2 AChRs expressed in permanently transfected human embryonic kidney cells could be up-regulated 15-fold by nicotine (EC 50 ϭ 0.4 M) and 2-5-fold by curare (EC 50 ϭ 300 M) or DH␤E (EC 50 ϭ 107 M) (48). In this study, two competitive antagonists, curare and DH␤E, were tested for their influence on up-regulation of ␣3␤2 AChRs. Both curare and DH␤E induced increases of ␣3␤2 AChRs in tsA201 cells in the absence of nicotine (Fig. 11A), although the effect was not as dramatic as that of nicotine (a maximum of 3-fold for curare and 5-fold for DH␤E). When curare and DH␤E were co-applied with nicotine at 100-fold molar excess, the up-regulation of ␣3␤2 AChRs by nicotine was not significantly changed (Fig. 11B). It is known from our previous report (34) that at the concentrations applied, both curare and DH␤E were very effective at blocking cation flow through ␣3␤2 AChRs. This suggests that at least part of the agonist-induced up-regulation can be explained by mechanisms that do not require cation flow through ␣3␤2 AChRs.
The channel blocker mecamylamine causes up-regulation of AChRs in mouse brains (28) and of avian ␣4␤2 AChRs transfected in mouse fibroblasts (34), but not of human ␣4␤2 AChRs transfected in human embryonic kidney cells (48) or ␣3 AChRs in the human neuroblastoma SH-SY5Y (18). In our study, we found that mecamylamine did not cause up-regulation of ␣3␤2 AChRs in the cells, and it inhibited the up-regulation of ␣3␤2 AChRs by nicotine (Fig. 11, A and B). This phenomenon was observed with another channel blocker, amantadine, as well (data not shown). DISCUSSION We established stable cell lines that express four different subtypes of human ␣3 AChRs: ␣3␤2, ␣3␤2␣5, ␣3␤4, and ␣3␤4␣5. They form functional ion channels in the cell surface. We found that chronic exposure of cells expressing ␣3␤2 AChRs or ␣3␤2␣5 AChRs to nicotine or carbamylcholine up-regulated the amount of AChRs up to 24-fold but had no effect on the amount of ␣3␤4 or ␣3␤4␣5 AChRs, i.e. the up-regulation of ␣3 AChRs depends on the presence of ␤2 but not ␤4 subunits in the AChRs.
Saturation binding curves of ␣3␤2 and ␣3␤4 AChRs with and revealed no partially assembled AChRs. The expression levels for ␣3␤2 and ␣3␤4 AChRs were different, with ␣3␤4 cell lines producing more AChRs. The different levels of expression between ␣3␤2 and ␣3␤4 in tsA201 cells were also seen in transiently transfected cells by solid phase RIA; thus, this is not an artifact of the particular lines chosen. This suggests that there are intrinsic differences between ␤2 and ␤4 subunits in either the efficiencies of synthesis, maturation, or assembly with ␣3 subunits or that the rate of AChR turnover is different between ␣3␤2 and ␣3␤4 AChRs. We consider the last three possibilities as being more likely because in the absence of protein synthesis nicotine could up-regulate the amount of ␣3␤2 AChRs to a level similar to that found for ␣3␤4 AChRs.
The ␣3␤4 cells expressed high levels of functional AChRs. Responses in the early passages of the cells following recloning were typically 5-10 nA in maximal amplitude at a holding potential of Ϫ60 mV. With higher numbers of passages (Ͼ15), the responses of the cells were somewhat reduced in amplitude (1-2 nA maximum at Ϫ60 mV) but were more than sufficient for functional studies. Concentration/response studies revealed EC 50 values for activation by ACh and nicotine of 79 and 55 M. These values are lower than the values of 163 and 106 M observed when human ␣3␤4 was expressed in Xenopus oocytes (16). These values are also significantly lower than what has been reported for rat ␣3␤4 AChRs studied in transfected HEK-293 cells, but the order of potency with nicotine Ͼ ACh was the same (20). The EC 50 values for ACh or nicotine activation of ␣3␤4␣5 AChRs (81 or 42 M, respectively) were not greatly different from ␣3␤4 AChRs. This may reflect the low percentage (14%) of AChRs that contained the ␣5 subunit in ␣3␣5␤4 tsA201 cells.
Even after nicotine-induced up-regulation, ␣3␤2-transfected tsA201 cells responded less robustly than ␤4-containing AChRs to applications of either ACh or nicotine, having maximum currents of 0.5-2 nA. This may in part reflect the more rapid desensitization of ␣3␤2 AChRs. The EC 50 values were 209 and 70 M for ACh or nicotine, respectively, and nicotine had partial efficacy. ␣3␤2 AChRs expressed in Xenopus oocytes had EC 50 values of 28 and 6.8 M for ACh and nicotine, respectively, and nicotine had an efficacy of 55% (16,49). The differences in EC 50 or response kinetics between the cell lines and Xenopus oocyte expression were not due to the necessity to expose ␣3␤2 AChR lines to nicotine to get substantial expression. We found that overnight exposure of oocytes to 100 M nicotine followed by 1 h of rinsing did not increase the extent of response, change the kinetics of response, or change the EC 50 values for ACh or nicotine on either ␣3␤2 AChRs or ␣3␤4 AChRs (data not shown). Thus, differences between responses of cell lines and oocytes probably reflect a combination of cell biological differences in membrane lipid composition, levels of protein phosphorylation, and other modifications as well as technical differences that permit more rapid and uniform agonist application to small cells than to large oocytes. No published report of the pharmacological properties of human ␣3␤2 AChRs in transfected cells has appeared previously. The additional transfection of the ␣3␤2 tsA201 cells with the ␣5 subunit resulted in an EC 50 for ACh that was slightly lower than that for ␣3␤2 AChRs. The difference could reflect changes in channelgating kinetics but probably is due to agonist-induced channel block brought on by the addition of the ␣5 subunit to the channel forming region of the AChR causing a premature saturation in the concentration/response relationship. This would cause the appearance of a reduction in the EC 50 value, as we believe is the case here, but would not reflect a true change in the affinity of the AChR for ACh. We conclude that although the EC 50 values for activation of ␣3␤2 or ␣3␤2␣5 AChRs by ACh are different, the difference is probably unrelated to agonist binding. This is consistent with the model proposed previously that the ␣5 subunit would be located in a position separating the ␣3 subunits of the ␣3␤2 dimers with the agonistbinding site located between the ␣3 and ␤2 subunits (16). In this position, the ␣5 subunit would not contribute to the agonist-binding site but would contribute to the lining of the channel of the AChR and influence the permeability, conductance, and gating properties of the channel. In contrast to what was observed in AChRs expressed in Xenopus oocytes (16,49), the current decay of ␣3 AChRs was not accelerated by the coexpression of the ␣5 subunit. This might be explained by reduced incorporation of ␣5 into AChRs expressed in cell lines as compared with oocytes. In the case of ␣3␤2␣5 AChRs in the cell lines, 50% incorporation of ␣5 was sufficient to increase the efficacy of nicotine (relative to ACh) to about 80% as compared with the 100% value observed in oocytes.
Previous studies of human AChRs expressed in Xenopus oocytes indicated that as structural subunits, ␤2 and ␤4 make different contributions to the pharmacological as well as ion channel properties of ␣3 AChRs (16,17,50). Chimera constructs between rat ␤2 and ␤4 subunits proved that N-terminal extracellular domains of the two subunits are responsible (at least partially) for the different behaviors of ␣3␤2 and ␣3␤4 AChRs upon activation by agonists (51).
Differences in cytoplasmic domains between ␤2 and ␤4 subunits may also be important. It has been reported that both ␤2 and ␤4 subunits can substitute for ␤1 subunits in muscle AChRs. Only ␤4 and not ␤2, however, can associate with the muscle 43-kDa protein Rapsyn which interacts with the cytoplasmic domain to aggregate AChRs into patches and anchor them to the cytoskeleton (52,53). We observed that ␣3␤4 AChRs were aggregated into a dense array of patches on the surface of transfected cell lines, whereas ␣3␤2 AChRs were not. It may well be that some Rapsyn-like proteins are expressed in the human embryonic kidney cell line used here and account for the selective aggregation (and thus perhaps stabilization on the cell surface) of ␣3␤4 AChRs. The selective expression of some ␣3 AChRs at post-synaptic locations and others in patches at peri-synaptic regions in ciliary ganglion neurons (54) may reflect differences in composition of ␤2, ␤4, and ␣5 subunits and in their consequent ability to interact with various AChR-associated proteins and cytoskeletal elements.
We provide further evidence that ␤2 and ␤4 subunits play unique roles in ␣3 AChR subtypes by demonstrating different responses of ␣3␤2 and ␣3␤4 AChRs to chronic nicotine exposure. ␣3␤2 AChRs can be up-regulated by nicotine up to 24-fold in tsA201 cell lines, but no significant change was observed for ␣3␤4 AChRs. The same is also true when other agonists and antagonists were used to regulate the ␣3 AChRs. We have tested several different clones for each AChR subtype, with both higher and lower levels of expression, to demonstrate that the regulation of ␣3␤2, but not ␣3␤4 AChRs, by nicotinic ligands is a general phenomenon (data not shown). The upregulation of ␣3␤2 but not ␣3␤4 AChRs in SH-SY5Y cells by nicotine also argues that the subunit dependence of the upregulation response of ␣3 AChRs is not an artifact of the transfected cell lines.
Nicotine-induced up-regulation of ␣3␤2 AChRs in tsA201 cells can be detected as early as 15 min after the AChRs have been exposed to the drug. We suggest that the quick increase of ␣3␤2 AChRs results from enhanced assembly of AChRs from existing stocks of subunits in the cell rather than from increased de novo synthesis. This hypothesis is supported by the experiments depicted in Fig. 9. Nicotine treatment during the first 30 min is shown to induce a large increase of assembled ␤2-containing AChRs (more than 5-fold) but not of total ␤2 subunits in the cell. The amount of ␣3␤2 AChRs (measured by [ 3 H]epibatidine binding) in the cells was increased (6-fold) upon nicotine exposure even without protein synthesis. In SH-SY5Y cells, neither increased mRNA synthesis for ␣3 AChR subunits nor decreased turnover rate of ␣3 AChRs was detected upon nicotine treatment (18). Thus, we suggest that nicotineinduced up-regulation of ␣3 AChRs in SH-SY5Y cells can also be explained, at least partially, by enhanced subunit maturation or assembly. Consistent with what has been found for ␣4␤2 AChRs in M10 cells, where nicotine caused a decrease in the rate of AChR turnover (34), we also noted that ␣3␤2 AChRs in tsA201 cells were stabilized by nicotine for at least 24 h. Thus, when exposed to nicotine in the tissue culture medium, at least two mechanisms are working in the cells to cause an increase in ␣3␤2 AChRs.
Phosphorylation of nAChRs has been implicated in the process of AChR up-regulation (46,47,55,56). cAMP, through activation of protein kinase A, has been shown to up-regulate muscle type nAChRs by increasing the efficiency of subunit assembly and preventing AChR degradation (55,56). In studies not shown, we found that H-7, a protein kinase inhibitor, can block at least half of the up-regulation effect of nicotine on ␣3␤2 AChRs in tsA201 cells. Whether H-7 blocks the up-regulation effect of nicotine on ␣3␤2 AChRs by a direct effect on AChRs or on some protein which interacts with them is unknown.
There are both common aspects and differences between human ␣3 AChRs in the transfected tsA201 cells and the neuroblastoma cell line SH-SY5Y with respect to regulation by ligands. ␣3␤2 AChRs in both cell lines increased upon exposure to nicotine and carbamylcholine, which means that in both cases, the binding of agonists to ␣3␤2 AChRs on the cell surface is enough to trigger the up-regulation process. But the maximum increase of ␣3␤2 AChRs induced by nicotine or carbamylcholine is less in SH-SY5Y cells compared with tsA201 cells (6versus 24-fold). Most of the increased AChRs in tsA201 cells were detected on the cell surface, whereas those from SH-SY5Y cells were all inside the cell (however, there was evidence that the increased AChRs had been transiently on the cell surface before they were internalized (18)). It may be that SH-SY5Y cells lack an AChR-associated protein similar to Rapsyn which can stabilize ␣3␤2 AChRs on the cell surface. Chronic treatment of the cells with the competitive antagonist curare or DH␤E did not change the amount of ␣3 AChRs in SH-SY5Y cells but did cause an increase of ␣3␤2 AChRs in tsA201 cells. The channel blocker mecamylamine did not cause nicotineinduced up-regulation of ␣3 AChRs in SH-SY5Y cells but did inhibit the up-regulation of ␣3␤2 AChRs by nicotine. Supposing the up-regulation is mediated through signaling proteins inside the cells, the intrinsic differences among those proteins from two cell types with distinct origins might influence the upregulation of the AChRs. The differential influence of various host cells has also been observed in the study of nicotineinduced up-regulation of ␣4␤2 AChRs. ␣4␤2 AChRs in rat brains and in stably transfected M10 cells can be up-regulated by chronic nicotine exposure to a maximum of 2-fold (34,57), whereas ␣4␤2 AChRs stably transfected in human embryonic kidney (HEK) cells can be maximally up-regulated by nicotine up to 15-fold (48). The competitive antagonist curare attenuates nicotine-induced up-regulation of ␣4␤2 AChRs in M10 cells (34), but it can up-regulate ␣4␤2 AChRs in HEK cells by 3-fold (48). The channel blocker mecamylamine causes upregulation of ␣4␤2 AChRs both in rat brain and in M10 cells (28,34), but it does not evoke a significant change in amount of ␣4␤2 AChRs in HEK cells, neither does it alter the up-regulation elicited by nicotine (48).
In the peripheral nervous system, most of the ␣3 AChRs appear to be composed of ␣3 and ␤4 subunits (1-3), whereas in brain, ␤2-containing AChRs have been found to be critical for enhancing striatal dopamine release and consequent nicotine addiction (15,58). Chronic nicotine treatment has been reported to increase dopamine release from the terminals in response to subsequent nicotine challenge (59). According to our study, ␣3␤2 AChRs in transfected cells can be up-regulated more than 3-fold within 7 h by nicotine at concentration as low as 0.2 M, a serum concentration of nicotine typical for smokers (47). Although at similar nicotine concentrations, ␣4␤2 AChRs were also found to be up-regulated both in rat brain (around 66%) (58), and in a permanently transfected cell line (around 70%) (34), functional analyses demonstrated that virtually all ␣4␤2 AChRs were inactivated whereas 80% of ␣3 AChRs were still active (19). Thus, we suggest that in a smoker's brain, most of the ␣4␤2 AChRs are permanently desensitized by long term nicotine exposure, but ␣3␤2 AChRs may be both increased in amount and retain function. This may be especially important because, at least in mice, nicotine addiction has been shown to depend on ␤2 containing AChRs that stimulate release of dopamine from striatal neurons (58). On the other hand, scanning different areas of rat brain for changes in [ 3 H]nicotine-binding sites after chronic nicotine exposure showed that not all the nicotine-binding sites were up-regulated, suggesting regional heterogeneity in brain AChR composition (60). Considering that there are AChRs in the brain that are up-regulated by nicotinic ligands little (e.g. ␣7 AChRs (18, 60)) or not at all (e.g. ␣3␤4 AChRs according to our study), it is not surprising that certain regions in brain do not respond to nicotine by upregulation of their nicotine-binding sites. It is even possible that the same AChR subtype, e.g. ␣3␤2 AChRs, expressed in different neurons may exhibit different responses to nicotine, because ␣3 AChRs (18) or ␣4␤2 AChRs (34,48) expressed in different cell types show different degrees of up-regulation. Exploring the mechanism of nicotine-induced up-regulation of human ␣3 AChRs may provide useful insights on nicotine addiction, as well as on development of nicotinic therapeutic agents for patients with Alzheimer's disease, Parkinson's disease, Tourette's syndrome, schizophrenia, or chronic pain (61).