Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Neuronal nicotinic ACh receptors (nAChRs) are expressed in both the peripheral nervous system and the CNS, where they play an important role in the control of synaptic transmission (Gray et al., 1996; McGehee et al., 1995; Role and Berg, 1996). These receptors are multisubunit complexes (Anand et al., 1991; Cooper et al., 1991), and a family of genes encoding 11 neuronal subunits has been identified (2-9, 2-4; for reviews, see McGehee and Role, 1995; Sargent, 1993). The 7, 8 and 9 subunits are functional when expressed alone in Xenopus oocytes, whereas the 2, 3 and 4 subunits are functional only when expressed in pairwise combination with 2 or 4 subunits (McGehee and Role, 1995; Sargent, 1993). More recent data indicate that co-expression of 5 with heterologously expressed 42, 32 or 34 receptors can modify the pharmacological and biophysical properties of the corresponding pairwise combination (Wang et al., 1996; Ramirez-Latorre et al., 1996). Also, although functional roles for 6 and 3 subunits have not been firmly established, there now is evidence that 6 may form a functional receptor in combination with 4 (Gerzanich et al., 1997) and that the 3 subunit functions as a structural entity (Forsayeth and Kobrin, 1997).

The precise subunit composition of nAChRs in CNS neurons is a matter of considerable interest. Earlier studies suggested that the majority of immuno-isolated high-affinity binding sites for [³H]cytisine or [³H]nicotine in chick and rat brain contained the 4 and 2 subunits (Flores et al., 1991; Schoepfer et al., 1988; Whiting et al., 1987a; Whiting et al., 1987b). However, the existence of other receptor subunit combinations with either low affinity for [³H]nicotine and [³H]cytisine, or high-affinity sites in low abundance, is still likely. For example, in situ hybridization has revealed the presence of mRNA encoding 4 subunits in hippocampus, cortex, medial habenula, cerebellum and locus ceruleus (Dineley-Miller and Patrick, 1992). Although brains from transgenic mice lacking the 2 subunit lose a majority of their high-affinity binding sites for [³H]nicotine, nicotine-evoked electrophysiological responses were unaltered in neurons from the medial habenula (Picciotto et al., 1995). Additionally, the Type III current found in cultured rat hippocampal cells has been tentatively attributed to 34 receptors on the basis of a similar pharmacology observed for pairwise expression of these subunits in oocytes (Alkondon and Albuquerque, 1993). These data support the idea that certain nAChRs in the brain contain 4 subunits. Determining the functional properties of 4-containing receptors may therefore provide new insights into the role of nAChRs in the brain.

Studies examining the function of heterologously expressed nAChRs reveal a diverse pharmacological profile in receptors containing 2, 3 or 4 subunits in combination with 2 or 4 subunits (Chavez-Noriega et al., 1997; Gerzanich et al., 1995; Hussy et al., 1994; Luetje and Patrick, 1991). The pharmacological characteristics of specific nAChR subtypes also display species specificity. For example, nicotine acts as a full agonist on rat 32 receptors, whereas it is a partial agonist on chick and human 32 receptors (Chavez-Noriega et al., 1997; Hussy et al., 1994). Also, nicotine is more potent than DMPP on rat and chick 7 nAChRs (Gerzanich et al., 1993; Amar et al., 1993; Séguéla et al., 1993), whereas DMPP is more potent than nicotine on human 7 (Chavez-Noriega et al., 1997; Peng et al., 1994). Therefore, in the discovery of nAChR-directed therapeutic agents for humans, it is preferable to use human nAChRs as primary targets.

The pharmacological properties of recombinant human nAChRs are now beginning to emerge. Reports have appeared describing the characteristics of human 7 (Peng et al., 1994), as well as some limited information on 32 and 34 receptors (Gerzanich et al., 1995), expressed in Xenopus oocytes. A full characterization of the pharmacological properties of human 24, 34, 44, 22, 32, 42 and 7, expressed in Xenopus oocytes, was recently completed (Chavez-Noriega et al., 1997). To date, however, only 7 and 42 human recombinant receptors have been characterized after stable expression in mammalian cells (Gopalakrishnan et al., 1995; Buisson et al., 1996; Gopalakrishnan et al., 1996), which is a system more adaptable to drug discovery by high-throughput screening. We describe here the pharmacological characterization of human recombinant 24, 34 and 44 receptors stably expressed in HEK293 cells.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture and stable transfection of HEK293 cells. HEK293 cells (American Type Tissue Collection, Rockville, MD) were grown in Dulbecco's modified Eagle medium supplemented with 6% bovine calf serum, 50 units/ml penicillin and 50 µg/ml streptomycin in a humidified atmosphere containing 6% CO₂. The isolation and characterization of cDNAs encoding human 2, 3, 4 and 4 nAChR subunits were reported previously (Elliott et al., 1996). In order to optimize expression levels, the cDNAs encoding the 2 and 4 subunits were modified by replacing the 5 untranslated region with an efficient ribosomal binding site, 5-GCCACC-3 (Kozak, 1987). The A4B4 cell line (previously referred to as 10C4-6; Stauderman et al., 1995) was developed using expression constructs in which the cDNAs encoding the human 4 and 4 subunits were subcloned into the vector pCMV-T7, a modified version of pCMV (Clontech, Palo Alto, CA) containing a T7 promoter site. The plasmid pSV2neo (Clontech), which carries the neomycin-resistance gene, was cotransfected with the 4 and 4 plasmids in construction of this cell line. The expression constructs used to develop the A2B4 cell line (previously referred to as 13B5-12; Stauderman et al., 1995) were generated by subcloning cDNAs containing the entire coding sequences of human 2 and 4 into the pcDNA3 plasmid (Invitrogen, San Diego, CA), which carries the neomycin-resistance gene. The same 4 construct was used to develop the A3B4.2 cell line, in combination with human 3 cDNA subcloned into the vector pZeoCMV, a modified version of pcDNA3 (Invitrogen) that carries the Zeocin-resistance gene from pZeoSV (Invitrogen).

RNA isolation and northern blot analysis. Total cellular RNA was isolated from stably transfected cells using the RNeasy Total RNA Kit (QIAGEN, Inc., Chatsworth, CA). Then 5 µg of each RNA sample was fractionated by electrophoresis on a 1% agarose gel containing 1 M formaldehyde and transferred to Zeta-Probe (BioRad, Hercules, CA) by downward alkaline blotting (Chomczynski, 1992). Ribosomal RNAs were visualized by staining the membrane with methylene blue. Blots were sandwiched between two sheets of Whatman 3MM paper and hybridized in 250 mM sodium phosphate buffer, pH 7.2, 250 mM NaCl, 7% SDS, 1 mM EDTA and 50% formamide containing 5 × 10⁵ cpm/ml heat-denatured probe. Probes encompassing the entire coding regions of 2, 3, 4 and 4 were labeled with [³²P] using the Prime-It RMT Random Primer Kit (Stratagene, La Jolla, CA). Blots were hybridized overnight at 42°C, rinsed briefly in 1× SSPE (150 mM NaCl, 10 mM NaH₂PO₄, 1 mM EDTA, pH 7.4), 0.2% SDS and then washed three times in 0.1× SSPE, 0.1% SDS at 65°C. Blots were exposed to X-ray film (Kodak Biomax) with an intensifying screen.

Membrane preparation and immunoblot analysis. Cells were harvested from 10-cm plates and washed with PBS (140 mM NaCl, 3 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4). Washed cells were resuspended in 50 mM Tris pH 7.4, 1 mM EDTA containing a cocktail of protease inhibitors (Complete, Boehringer Mannheim, Indianapolis, IN) and homogenized with a Dounce homogenizer. The homogenate was centrifuged at 1000 × g for 5 min to remove cellular debris, and the supernatant fraction was centrifuged at 100,000 × g for 120 min to pellet the membranes. The membranes were resuspended in RIPA buffer (50 mM Tris pH 7.6, 150 mM NaCl, 0.5% deoxycholate, 1% Nonidet P-40, 1% SDS) containing the protease inhibitor cocktail. Protein concentration was determined using a Bio-Rad Protein Assay. Aliquots of membranes were stored at 70°C.

Fluorescence-based measurements of [Ca⁺⁺]_i. For the measurement of [Ca⁺⁺]_i in cell populations, cells were plated on poly-D-lysine-coated 96-well microtiter plates at a density of 2 × 10⁵ cells/well. Twenty-four hours after plating, the cells were washed with HBK (155 mM NaCl, 4.6 mM KCl, 1.2 mM MgSO₄, 1.8 mM CaCl₂, 6 mM glucose and 20 mM HEPES, pH 7.4). Washed cells were incubated with 20 µM fluo-3-acetoxymethylester (Molecular Probes Inc., Eugene, OR) for 1.5 to 2 h at 20°C. Dye not taken up by cells was removed by aspiration followed by washing with 250 µl HBK. Fluorescence measurements were performed at 0.33-s intervals using a 96-well microtiter plate-reading fluorometer (Cambridge Technology, Inc., Watertown, MA). Ten basal fluorescence readings were recorded before activation with nAChR agonists. Responses after the addition of agonists were recorded for approximately 60 s. Antagonists were tested after a preincubation period of 5 to 10 min. Maximal fluorescence (F_max) was determined after lysing the cells with 0.25% Triton X-100, and minimal fluorescence (F_min) was determined after subsequent quenching with 10 mM MnCl₂. Calculation of [Ca⁺⁺]_i was performed as described by Kao et al. (1989). Cellular responses were quantitated either by calculating the ratio of peak [Ca⁺⁺]_i after agonist addition to basal [Ca⁺⁺]_i before agonist addition or by calculating the difference between peak [Ca⁺⁺]_i and basal [Ca⁺⁺]_i. EC₅₀ and IC₅₀ values were calculated using the ratio of peak [Ca⁺⁺]_i to basal [Ca⁺⁺]_i. Curve fitting was performed by Prism software (GraphPad, Inc., San Diego, CA) using the equation for a single-site sigmoidal dose-response curve with a variable slope, where Y is the response amplitude, X is the concentration of agonist or antagonist and n is the Hill coefficient. For determination of rank orders of potency, EC₅₀ or K_b values (converted to log values) were compared by one-way ANOVA with Student-Newman-Keuls (SNK) multiple comparisons test (P < .05; SigmaStat, Jandel Scientific, Inc., San Rafael, CA). Other statistical comparisons also utilized a one-way ANOVA with SNK multiple comparisons test, as indicated in the text. Hill slopes were analyzed by Prism using a one-sample t test compared with a theoretical mean of 1.0 for agonists and of 1.0 for antagonists. The IC₅₀ values for DHE were converted into K_b values using the Leff-Dougall (Leff and Dougall, 1993) variant of the Cheng-Prusoff equation where [A] is the agonist concentration, A₅₀ is the EC₅₀ value for the agonist and n is the Hill coefficient for the agonist. EC₅₀, IC₅₀ and K_b values are presented in the text as geometric means (S.D., +S.D.).

Binding of [³H]epibatidine. Membranes were prepared by scraping cells from culture dishes in the growth medium and centrifuging for 10 min at 1200 × g. The cells were resuspended in an assay buffer containing 50 mM Tris pH 7.4, 120 mM NaCl, 5 mM KCl, 2 mM CaCl₂ and 1 mM MgCl₂, homogenized with a Polytron and centrifuged for 15 min at 2000 × g. The resulting pellet was frozen at 20°C until use. For saturation analysis, cell membranes (6-100 µg protein) were incubated with 5 pM to 1.0 nM [³H](±)epibatidine (total volume 2-5 ml) on ice for 2 h, followed by rapid filtration through Whatman GF/C filters presoaked in 0.5% polyethyleneimine for at least 30 min. Nonspecific binding was determined in the presence of 25 nM Epi. Protein concentrations were adjusted so that specifically bound ligand was always less than 10% of the total ligand in the assay and in most cases was less than 1%. Accordingly, in Scatchard plots the free concentration of ligand was considered to be equivalent to the total concentration.

Electrophysiological methods. Cells were plated on poly-D-lysine-coated glass coverslips at a density of 1.5 × 10⁵ cells/35-mm dish. One to two days after plating, recordings were performed with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA) using the whole-cell voltage-clamp configuration (Hamill et al., 1981). Membrane potential was held at 80 mV. The external recording solution (mammalian Ringer's solution) consisted of 160 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 11 mM glucose, 1 µM atropine and 5 mM HEPES, pH 7.3. Ringer's solution was superfused at a rate of 3.0 ml/min (110-µl recording chamber). The recording pipette solution was composed of 135 mM CsCl, 10 mM EGTA, 1 mM MgCl₂, 4 mM Mg-ATP and 10 mM HEPES, pH 7.3. Agonists were applied either by pressure ejection (200-700 ms pulses, General Valve Corp., Fairfield, NJ) from a glass micropipette (approximately 3-5 µm in diameter) positioned 20 to 60 µm above the cell or by a rapid application system (Jonas, 1995) using a triple-barrel glass pipette attached to an electromechanical switching device (piezo-electric drive, Winston Electronics, Millbrae, CA). Experiments were performed at room temperature. The current-voltage (I-V) relation of agonist-induced currents was determined by the application of 70 to 100-ms voltage ramps from 100 to +40 mV in the absence and in the presence of agonist. Ramps from +40 to 100 mV were also tested in some cells, resulting in the same I-V relation. The net agonist-induced current was obtained by subtracting the current observed in the absence of agonist from that measured in the presence of agonist.

Materials. Polyclonal antisera were purchased from Bethyl Laboratories (Montgomery, TX) and were generated by immunizing sheep with the fusion proteins expressing the extracellular domains of rat neuronal nAChRs 3, 4 and 4 described by Neff et al. (1995). Epi and DHE were purchased from Research Biochemicals International (Natick, MA). NEN Life Science Products (Boston, MA) was the supplier of [³H](±)epibatidine. All other compounds were obtained from Sigma Chemical Company (St. Louis, MO).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Identification of stable cell lines. HEK293 cells transfected with cDNAs encoding human nAChR subunits in the pairwise combinations 24, 34 and 44 were grown in selection medium for 2 weeks. On the basis of the reported Ca⁺⁺ permeability of nAChRs (Mulle et al., 1992; Trouslard et al., 1993; Vernino et al., 1994) cell colonies resistant to antibiotic(s) were tested for functional nAChRs by measuring elevations in [Ca⁺⁺]_i upon stimulation with 100 to 300 µM Nic. Colonies from each pairwise receptor combination with the largest Nic-induced [Ca⁺⁺]_i signal were subcloned by limiting dilution. Cell lines A2B4, A3B4.2 and A4B4 that displayed the most robust and consistent Nic-stimulated elevations of [Ca⁺⁺]_i (fig. 1) were selected for more detailed characterization of 24, 34 and 44 nAChRs, respectively. The magnitude of Nic-induced [Ca⁺⁺]_i signals was stable in each cell line for at least 6 months in continuous culture (data not shown).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Nicotine-induced changes of [Ca++]i in cells stably expressing human nicotinic receptor subunits. The A2B4, A4B4.2 and A4B4 cell lines were loaded with the Ca++-sensitive fluorescent dye fluo-3 as described in "Materials and Methods." Data points were measured at approximately 0.33-s intervals. Each point is the mean [Ca++]i value (nM) of four individual wells from a 96-well plate (error bars were omitted for clarity) and represents the average signal of approximately 200,000 cells. After ten measurements of basal [Ca++]i, the lid of the fluorometer was opened briefly, and nicotine (~EC90-95; hatched bar) was added manually, which accounts for the gap in each record before time 0. Shown are representative responses from each cell line. Detailed kinetic analyses of the rising and falling phases of the [Ca++]i signal were not performed, but in each case the peak [Ca++]i was attained 8 to 20 s after the addition of agonist. The agonist-induced [Ca++]i responses were dependent on the presence of external Ca++ (data not shown).

Validation of stable cell lines by analysis of mRNA and protein. Northern blot analysis showed that the cell lines A2B4, A3B4.2 and A4B4 expressed RNAs that hybridized to the appropriate subunit-specific probes (to 2 and 4, to 3 and 4 and to 4 and 4, respectively) and were of the predicted sizes for full-length transcripts (fig. 2). The RNAs encoding human 2, 3, 4 and 4 subunits were not detectable in HEK Neo/Zeo cells. The levels of RNA for each subunit were stable in each cell line for at least 6 months in continuous culture (data not shown).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Northern blot of total RNA isolated from cell lines A2B4, A3B4.2, A4B4 and HEK Neo/Zeo, showing the presence of RNA for human 2, 3, 4 and 4 subunits. Blots were hybridized under high stringency with DNA probes specific for the coding sequence of each subunit (see "Methods and Materials"). Hybridizing bands were detected corresponding to 2 and 4 transcripts in A2B4 cells, to 3 and 4 transcripts in A3B4.2 cells and to 4 and 4 transcripts in A4B4 cells. No RNAs hybridizing to the 2, 3, 4 or 4 probes were detected in HEK Neo/Zeo cells. Arrows indicate the predicted size of the full-length polyadenylated mRNA for the corresponding subunit. The larger transcripts, which are less prominent, are likely to contain the entire coding sequence and possibly some additional expression vector sequence.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Western blot of total membrane proteins from A2B4, A3B4.2 and A4B4 cells, showing the presence of 2, 4 and 4 subunit proteins. Membranes were isolated and the Western blot performed as described in "Materials and Methods." HEK293 cells stably transfected with plasmids containing only the neomycin and Zeocin resistance genes (HEK Neo/Zeo) were used for controls. The 2, 3, 4 and 4 subunits migrated at apparent molecular weights of 70, 64, 78 and 60 kDa, respectively. The molecular weights predicted for 2, 3, 4 and 4 are 59, 53, 67 and 52 kDa, respectively. Migration of subunit proteins at apparent molecular weights larger than predicted by the primary amino acid sequence is due to N-linked glycosylation (data not shown).

Quantitation of receptor expression by binding of [³H]epibatidine. As a means to quantitate and characterize the nAChRs expressed in each stable cell line, we measured the binding of [³H]epibatidine to cell membranes. We detected no specific binding to membranes prepared from HEK Neo/Zeo cells (data not shown). Specific and saturable binding of [³H]epibatidine was detected in membranes prepared from A2B4, A3B4.2 and A4B4 cells (fig. 4). Scatchard analysis revealed that binding was to a single, high-affinity site in each cell line (fig. 4, insets). The results summarized in table 1 show that the binding affinity was significantly greater in A2B4 cells than in A3B4.2 or A4B4 cells. B_max values ranged from 1100 ± 338 fmol/mg protein in the A2B4 cells to 3683 ± 1450 in the A4B4 cells, but no statistical differences were detected (table 1). The presence of high-affinity specific binding sites for [³H]epibatidine in A2B4, A3B4.2 and A4B4 cells is consistent with the expression of assembled heteromeric nAChRs (Wang et al., 1996).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Saturation binding of [3H]epibatidine to membranes from A2B4, A3B4.2 and A4B4 cells. Binding of [3H]epibatidine was measured as described in "Materials and Methods." The figure shows saturation isotherms for specific binding of [3H]epibatidine to membranes from cell lines A2B4 (24, top panel), A3B4.2 (34, middle panel) and A4B4 (44, bottom panel). Scatchard analyses (shown at right) revealed that binding was to a single site in each cell line, with the Kd and Bmax values summarized in table 1. The data presented are from one experiment that is representative of three performed in a similar manner.

Validation of functional nAChRs by electrophysiology. In voltage-clamped cells (V_h = 80 mV), application of 30 to 300 µM Nic, ACh, DMPP or Cyt elicited inward currents ranging from 13 pA to 14 nA in A2B4 cells (n = 20), from 1.1 to 20 nA in A3B4.2 cells (n = 20) and from 160 pA to 10.9 nA in A4B4 cells (n = 20). These currents desensitize rapidly in the presence of the agonist (see fig. 5). Application to A3B4.2 cells of high concentrations of DMPP (30 µM and above) produced a "rebound" inward current upon removal of the agonist from the bath (fig. 5). This rebound current was not produced, or was minimal, in response to ACh, Nic, or Cyt (data not shown for Nic or Cyt) applied at the same concentration in A3B4.2 cells. Also, the rate of decay of current elicited by DMPP (100 µM) was faster compared with that of currents elicited by Nic (300 µM), by Cyt (100 µM; data not shown) and by ACh (300 µM) (P < .05) in A3B4.2 cells. In the A2B4 and A4B4 cells, no rebound current was detected with any agonist at concentrations up to 100 µM. The rebound current in A3B4.2 cells and the accelerated rate of decay of inward current are probably associated with voltage-dependent open-channel block by DMPP, as has been shown for nicotinic agonists on native nAChRs in bovine adrenal chromaffin cells (Machonochie and Knight, 1992) and on other recombinant nAChRs (Bertrand et al., 1992).

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Whole-cell currents obtained in A2B4, A3B4.2 and A4B4 cells. Whole-cell patch-clamp recordings were performed as described in "Materials and Methods." The recordings shown are from six different cells. Agonist was applied during the period indicated by the horizontal bar above each trace. Desensitization of the responses in the continued presence of the agonist is detectable in all three cell lines. The "rebound" current produced upon removal of DMPP in the A3B4.2 cells is probably due to relief of voltage-dependent open-channel block (see text). The mean amplitudes of agonist responses are summarized in the text.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Current-voltage relation of agonist-induced currents in A2B4, A3B4.2 and A4B4 cells. A strong inward rectification was observed in all cells. The Vrev for these currents varied between 12 and +12 mV (n = 16, see text). Voltage ramps and leak subtractions were carried out as indicated in "Materials and Methods."

Pharmacological characterization of nAChR-mediated increases in [Ca⁺⁺]_i. Epi, Nic, Cyt, DMPP and Sub stimulated concentration-dependent increases in [Ca⁺⁺]_i in the A2B4, A3B4.2 and A4B4 cells. Kinetics of the changes in [Ca⁺⁺]_i typically showed a rapid rising phase that reached a peak 8 to 20 s after the addition of agonist, followed by a slower relaxation toward basal [Ca⁺⁺]_i levels (see fig. 1). The maximal amplitude of the increases in [Ca⁺⁺]_i varied among cell lines. For example, Epi stimulated a maximal increase in [Ca⁺⁺]_i of 1152 ± 244 nM (mean ± S.E.; n = 4) in A2B4 cells, of 1912 ± 690 nM (n = 4) in A3B4.2 cells and of 880 ± 229 nM (n = 4) in A4B4 cells. ACh also stimulated increases in [Ca⁺⁺]_i, but we did not characterize its effects because of concern over degradation and the necessity of including atropine to block endogenous muscarinic receptors in HEK293 cells.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Concentration-dependent effects of nAChR agonists on [Ca++]i in A2B4, A3B4.2 and A4B4 cells. Cells were stimulated with either ()nicotine (Nic, ), cytisine (Cyt, ), DMPP (), suberyldicholine (Sub, ) or (+)epibatidine (Epi, ) at the concentrations indicated. The agonist-induced increase of [Ca++]i was quantitated using fluo-3 as described in "Materials and Methods" and is plotted as a percent of the maximal peak [Ca++]i/basal [Ca++]i response to nicotine in each cell line. Each point represents the mean value (± S.E.) from 3 or 4 experiments, each performed in quadruplicate. The EC50 values, Hill coefficients and relative efficacies are summarized in table 2.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Effect of nAChR antagonists on agonist-stimulated [Ca++]i elevations in A2B4, A3B4.2 and A4B4 cells. The stimulation-induced increases of [Ca++]i were quantitated using fluo-3 as described in "Materials and Methods." Cells were incubated with either mecamylamine (Mec, ), DHE () or d-tubocurarine (d-Tubo, ) for 5 to 10 min before the addition of agonist. The agonist was 20 µM nicotine for the A2B4 cells (EC90) and was 200 µM nicotine for the A3B4.2 and A4B4 cells (EC81 and EC93, respectively). The data are plotted as percent of the control peak [Ca++]i/basal [Ca++]i response to nicotine. Each point represents the mean value (± S.E.) from 3 or 4 experiments, each performed in quadruplicate.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This is the first report describing functional and pharmacological characteristics of human recombinant 4-containing neuronal nAChRs stably expressed in a mammalian cell line. We utilized the Ca⁺⁺-permeability of nAChRs to develop an assay for functional receptors based on the measurement of [Ca⁺⁺]_i. With this method, HEK293 cells expressing functional recombinant nAChRs were identified by monitoring nicotinic agonist-induced increases in [Ca⁺⁺]_i. The agonist-evoked [Ca⁺⁺]_i signals were mediated by activation of nAChRs, as shown by blockade with nicotinic antagonists.

The agonist-induced [Ca⁺⁺]_i signal in the A2B4, A3B4.2 and A4B4 cell lines derives from the entry of Ca⁺⁺ through cell surface receptors, because no agonist-induced [Ca⁺⁺]_i responses were detected in the absence of external Ca⁺⁺ (data not shown). The transient nature of the agonist-stimulated [Ca⁺⁺]_i signal may be explained by a combination of desensitization of the human nAChRs (see fig. 5 and Chavez-Noriega et al., 1997) and increased buffering of cytosolic Ca⁺⁺. Whether the [Ca⁺⁺]_i signal also involves secondary activation of calcium-induced calcium release mechanisms or involves endogenous VGCCs (Berjukow et al., 1996) is not known. However, experiments with Cd⁺⁺ to block endogenous VGCCs suggest that they do not participate in the agonist-induced elevation of [Ca⁺⁺]_i. Stetzer et al. (1996) have recently demonstrated nicotine-induced [Ca⁺⁺]_i signals in HEK293 cells that stably express both recombinant L-type Ca⁺⁺ channels and rat 34 receptors. It was suggested that the nicotine-induced [Ca⁺⁺]_i signal was amplified by functional coupling of nAChR-mediated depolarization to activation of the recombinant VGCCs. Our results indicate that an ample [Ca⁺⁺]_i signal can be achieved without functional coupling of human 4-containing nAChRs to recombinant VGCCs. The [Ca⁺⁺]_i assay provides a rapid and convenient way to study the function and pharmacology of human nAChRs.

Northern and western blot analyses demonstrated the presence of RNA and protein associated with the human 2 and 4, the human 3 and 4 and the human 4 and 4 subunits in the A2B4, A3B4.2 and A4B4 cells, respectively. The cell lines were stable for at least 6 months in continuous culture and for at least 2 years in frozen storage, as judged by stability in steady-state levels of RNA and agonist-induced [Ca⁺⁺]_i signals.

Specific binding of [³H]epibatidine, a high-affinity ligand for nAChRs (Houghtling et al., 1994), to membranes isolated from A2B4, A3B4.2 and A4B4 cells was saturable and was described by a single site, which is consistent with the presence of a homogeneous population of high-affinity receptors. The affinity for [³H]epibatidine was 4- to 5-fold higher for human 24 receptors (K_d = 42 pM) than for 34 (K_d = 230 pM) and 44 (K_d = 187 pM). These relatively high affinities are consistent with the K_i for (±)epibatidine binding to human 42 receptors stably expressed in HEK293 cells (70 pM; Gopalakrishnan et al., 1996). The similarity of affinity values suggests that it might be difficult to resolve putative 24, 34, 44 or 42 nAChR subtypes in brain solely on the basis of their relative affinities for [³H]epibatidine binding.

The K_d values for [³H]epibatidine binding were at least 200-fold lower than the EC₅₀ values for Epi-induced [Ca⁺⁺]_i signals, a result consistent with the interpretation that in equilibrium binding assays [³H]epibatidine labels a high-affinity desensitized state of nAChRs (Gerzanich et al., 1995). Interestingly, the similarity in binding affinities at 34 and 44 receptors (203 and 187 pM, respectively) was not reflected in the relative EC₅₀ values for Epi-induced [Ca⁺⁺]_i signals (151 and 38 nM, respectively). Thus the binding data would not have predicted the 4-fold selectivity in potency of Epi-induced [Ca⁺⁺]_i signals at 34 and 44 receptors. These results provide evidence that binding assays do not always reflect the functional selectivity of nicotinic agonists on subtypes of nAChRs.

The presence of functional nAChRs on the surface of the A2B4, A3B4.2 and A4B4 cells was confirmed with whole-cell patch-clamp recordings. Previous studies using Xenopus oocytes have shown that human 2, 3, 4 and 4 subunits are functional only when expressed in pairwise combinations of x4 (Elliott et al., 1996). These data are consistent with the interpretation that the functional responses detected in A2B4, A3B4.2 and A4B4 cells derive from pairwise combinations of nAChR subunits.

The maximal agonist-induced current detected in the A2B4, A3B4.2 and A4B4 cell lines (14, 20, and 10.9 nA, respectively) corresponds to the rank order of maximal Epi-induced [Ca⁺⁺]_i signals. The current-voltage relations for agonist-induced currents demonstrated strong inward rectification, a result consistent with the properties reported for native nAChRs recorded from cultured rat sympathetic neurons (Trouslard et al., 1993), bovine adrenal chromaffin cells (Nooney et al., 1992), acutely dissociated rat medial habenula neurons (Mulle et al., 1991) and chick ciliary ganglion neurons (Rathouz and Berg, 1994). Recombinant chick 42 nAChRs stably expressed in mouse fibroblasts (Whiting et al., 1991) and human 42 nAChRs stably expressed in HEK293 cells (Gopalakrishnan et al., 1996) also display similar rectifying properties.

The [Ca⁺⁺]_i assay was used to examine the pharmacological profile of the human recombinant 24, 34 and 44 receptors stably expressed in HEK293 cells. Some of the nicotinic agonists displayed marked bell-shaped concentration-response relationships. This behavior cannot be explained by Ca⁺⁺-dependent inactivation of the nAChRs at higher [Ca⁺⁺]_i concentrations, because it was observed with agonists that stimulated both relatively high and relatively low [Ca⁺⁺]_i signals, e.g., Sub and DMPP in the A2B4 cells. It is more likely that the bell-shaped curves result from either rapid desensitization or open-channel block that occurs at higher agonist concentrations. For example, Marshall et al. (1991) found that Sub produced a bell-shaped concentration-response curve on nAChRs at the frog neuromuscular junction, with the downward side of the curve resulting from open-channel block of the receptors. The "rebound" current detected upon washout of DMPP in the A3B4.2 cells (fig. 5) is consistent with relief from open-channel block. More detailed electrophysiological measurements in the stable cell lines may resolve this issue.

The relative potencies of five nicotinic agonists showed that each nAChR subtype possessed a unique agonist profile. The rank order of potency was Epi > Cyt > Sub = Nic = DMPP at 24, Epi > DMPP = Cyt = Nic = Sub at 34 and Epi > Cyt = Sub > Nic > DMPP at 44 receptors. Absolute EC₅₀ value and efficacy were also influenced by receptor subtype. For example, the EC₅₀ values for Cyt were at least 50- and 60-fold lower on 44 and 24 receptors, respectively, than on 34 receptors. Also, Sub was as efficacious as Nic on 24 receptors but was less efficacious than Nic on 44 receptors. These data support the conclusion that human 2, 3 and 4 subunits contribute to the potency, selectivity and efficacy of nicotinic agonists at 4-containing receptors.

Using statistical comparisons of EC₅₀ values, we found that the relative potency of DMPP = Cyt = Nic at human 34 receptors expressed in the A3B4.2 cell line is similar to that at rat 34 receptors transiently (DMPP = Cyt = Nic; Wong et al., 1995) or stably (Cyt  Nic; Stetzer et al., 1996) expressed in HEK293 cells. Other comparisons reveal that Cyt is as efficacious as Nic at transiently expressed rat 34 receptors (Wong et al., 1995) and at human 34 receptors stably expressed in the A3B4.2 cell line. However, DMPP was less efficacious than Nic at rat 34 receptors (Wong et al., 1995), whereas it was as efficacious as Nic at human 34 receptors. Unfortunately, efficacy data were not reported for rat 34 receptors stably expressed in HEK293 cells (Stetzer et al., 1996). Another difference is in the Hill slopes for Cyt, DMPP and Nic. These values were greater than unity at rat 34 receptors expressed in HEK293 cells (Wong et al., 1995; Stetzer et al., 1996), but at human 34 in A3B4.2 cells the Hill slopes for DMPP and Nic were not different from unity, and that for Cyt was less than unity. Different methodologies, or species differences, might explain the differences in efficacy and Hill slope between rat and human 34 nAChRs expressed in HEK293 cells.

A panel of human 4-containing nAChRs have recently been examined by transient expression in Xenopus oocytes (Chavez-Noriega et al., 1997). The EC₅₀ values for Nic and DMPP at 24 receptors expressed in oocytes (21 and 23 µM, respectively; Chavez-Noriega et al., 1997) were close to their values at 24 receptors stably expressed in the A2B4 cells (9.9 and 12.2 µM, respectively). The efficacies of DMPP and Cyt were also lower than that of Nic in oocytes expressing 24 receptors, and a similar tendency in efficacy was determined for these agonists in the A2B4 cells (although the value for DMPP did not quite achieve statistical significance compared with Nic; P = .07). On the other hand, the EC₅₀ value for Cyt was much lower in the A2B4 cells (0.44 µM) than in the oocytes (39 µM), resulting in a different rank order of potency (Cyt > Nic = DMPP in A2B4 cells, DMPP  Nic > Cyt in oocytes). The reason for this difference remains unclear (but see below for possible explanations).

At human 34 receptors stably expressed in the A3B4.2 cells, the rank order of potency of DMPP = Cyt = Nic differed slightly from the rank order of DMPP > Cyt  Nic at human 34 receptors expressed in Xenopus oocytes (Chavez-Noriega et al., 1997). However, the EC₅₀ values for Epi, DMPP, Cyt and Nic in the A3B4.2 cells (151 nM, 12, 26 and 40 µM, respectively) were within 3-fold of the corresponding values in oocytes (73 nM, 19, 72 and 80 µM; Gerzanich et al., 1995, Chavez-Noriega et al., 1997). The efficacy profile also revealed that Cyt and DMPP were as efficacious as Nic in stimulating [Ca⁺⁺]_i increases in the A3B4.2 cells, whereas Cyt was less efficacious than Nic in oocytes (Chavez-Noriega et al., 1997). Aside from the differences detected with Cyt, there is reasonable agreement in activity among Epi, Nic and DMPP at human 34 receptors expressed in the A3B4.2 cells and Xenopus oocytes.

For human 44 receptors stably expressed in the A4B4 cells, the rank order of potency of Cyt > Nic > DMPP was identical to that for 44 receptors expressed in Xenopus oocytes (Chavez-Noriega et al., 1997). The EC₅₀ values for Cyt, Nic and DMPP were similar in A4B4 cells (0.52, 6.7 and 18 µM, respectively) and oocytes (0.9, 5 and 19 µM, respectively). Also, the efficacies of Cyt and DMPP were less than that of Nic in both the A4B4 cells and the oocytes expressing 44 receptors. Thus the pharmacological profiles of Cyt, Nic and DMPP were the same for human 44 receptors expressed in HEK293 cells or in Xenopus oocytes. The pharmacological differences detected between oocytes and stable cell lines expressing human 24 or 34 receptors may result from the divergent methodologies employed, from differences in subunit stoichiometries or from intrinsic differences between the two expression systems.

The antagonists d-tubocurarine and mecamylamine displayed no selectivity in potency among 24, 34 and 44 receptors stably expressed in HEK293 cells. That the Hill slope for d-tubocurarine was significantly less than 1.0 on 44 receptors may reflect a noncompetitive mechanism of action (Chavez-Noriega et al., 1996, Mulle et al., 1991). In contrast, the competitive antagonist DHE displayed subtype selectivity, which is consistent with results reported previously with rat or human 24, 34 and 44 receptors expressed in Xenopus oocytes (Harvey et al., 1996; Harvey and Luetje, 1996; Chavez-Noriega et al., 1997). At human 24, 34 and 44 receptors expressed in Xenopus oocytes (Chavez-Noriega et al., 1997), the K_b values for DHE were 3.61, 13.8 and 0.01 µM, respectively. The K_b value determined for 34 receptors expressed in oocytes agrees with the K_b value determined in the A3B4.2 cells (K_b = 9.0 µM), but the K_b values at 24 and 44 receptors expressed in oocytes lack agreement with the corresponding cell lines. These discrepancies may be related to the lower concentrations of ACh (30 µM, EC₁₆) or Nic (10 µM, EC₇₀) used in the oocyte study compared with the higher concentrations of nicotine used here (EC_80-93). High concentrations of nicotinic agonists have been associated with channel-blocking activity (De Fiebre et al., 1995; Machonochie and Knight, 1992; Sine and Steinbach, 1984), which could complicate the inhibitory effects of DHE. In an observation consistent with this hypothesis, preliminary evidence indicates that the K_b value for DHE determined in A4B4 cells at 10 µM Nic (EC₆₀) is 72 nM (data not shown), a value closer to that determined in oocytes (Chavez-Noriega et al., 1997). Additionally, it is possible that the K_b for competitive antagonists may change depending on the agonist employed (Brabet et al., 1995).

Recent data indicate a significantly greater number of nAChR subunit combinations in the CNS and the peripheral nervous system than were implicated previously by biochemical studies that assigned dominant roles to 42 receptors in the CNS and 34 receptors in the peripheral nervous system (McGehee and Role, 1995). Work is now in progress that may make it possible to associate nicotinic responses in neurons with specific nAChR subtypes. For example, although Cyt was reported to be a partial agonist at stimulating [³H]-dopamine release in rat striatal slices or synaptosomes (El-Bizri and Clarke, 1994; Sacaan et al., 1995), Grady et al. (1992) have shown that Cyt acts as a full agonist in evoking [³H]-dopamine release from mouse striatal synaptosomes. The latter result suggests the involvement of 4, because 4 combined with 2, 3 or 4 was shown to be more responsive to Cyt than 2 expressed with the same  subunits (Luetje and Patrick, 1991). Additionally, Marks et al. (1993) reported the rank order of potency for agonist-induced Rb⁺ efflux from mouse thalamic synaptosomes as Cyt (0.09 µM) > Nic (0.18 µM) > ACh (0.54 µM)  DMPP (0.64 µM). This order does not correlate with the reported rank order of potency for rat or human 42 (Chavez-Noriega et al., 1997; Luetje and Patrick, 1991), but it does correlate with human 44 expressed in oocytes and the A4B4 cell line. Clearly, stable expression of human recombinant nAChRs containing 4 subunits will prove useful in elucidating the composition of native nAChRs.

In conclusion, the results presented in this report show the A2B4, A3B4.2 and A4B4 cell lines to be powerful tools for examining the properties of human recombinant 24, 34 and 44 nAChRs, respectively. The advantage they offer in the identification of subtype-selective compounds will be valuable both for developing potential therapeutic agents and in probing the physiological function of nAChRs.