Abstract
The α7 nicotinic acetylcholine receptor (nAChR) subtype, unlike other neuronal nicotinic receptors, exhibits a relatively high permeability to Ca++ ions. Although Ca++ entry through this receptor subtype has been implicated in various Ca++-dependent processes in the central nervous system, little is known about how this receptor modulates mammalian intracellular Ca++ dynamics. Intracellular Ca++responses evoked by activation of the human α7 nAChRs stably expressed in HEK-293 (human embryonic kidney) cells were studied. Inward current and intracellular Ca++ transients were recorded simultaneously in response to a fast drug application system. Current recordings under whole-cell voltage-clamp and fast ratiometric intracellular Ca++ imaging acquisition were synchronized to drug pulses. The mean peak [Ca++]i observed with 100 μM (−)-nicotine was 356 ± 48 nM (n = 8). The magnitude of the intracellular Ca++ elevation corresponds to a 20% fractional current carried by Ca++ ions. The EC50 of the intracellular Ca++ responses for (−)-nicotine, (±)-epibatidine, 1,1 dimethyl-4-phenyl-piperazinium and acetylcholine were 51, 3.5, 75 and 108 μM, respectively. These EC50 values strongly correlate with those recorded for the cationic inward current through α7 nAChR. α-Bungarotoxin, methyllcaconitine or extracellular Ca++chelation ablated (−)-nicotine-evoked increase in intracellular Ca++ concentration. This study provides evidence that cation influx through the human α7 nAChR is sufficient to mediate a significant, transient, rise in intracellular Ca++ concentration.
Neuronal nAChRs belong to the family of ligand-gated multisubunit ion channels (Deneris et al., 1991; Luetje and Patrick, 1991), members of which include receptors for γ-amino-n-butyric acid, glutamate, glycine and serotonin (Karlin and Akabas, 1995). A variety of nAChR subtypes have been described in the brain, autonomic ganglia and sensory tissues (McGehee and Role, 1995). In recent years, it has become increasingly clear that such molecular diversity arises as a consequence of homomeric or heteromeric assembly of at least 11 gene products, namely, α2-α9 and β2-β4. The α2-α6 subunits in conjunction with β2-β4 subunits form heteromeric ion channels, whereas the α7-α9 subunits form homooligomeric channels when expressed in Xenopusoocytes (Couturier et al., 1990; Séguélaet al., 1993; Peng et al., 1994) or mammalian cell lines (Gopalakrishnan et al., 1995). Some α and β subunits may have a modulatory action on other nAChRs. Most of the reports have failed to show that α5 and α6as well as β3 subunits form functional receptors in conjunction with complimentary beta or alphareceptors, respectively (Boulter et al., 1987; Wada et al., 1990). It has been shown that these subunits may modify the physiology and pharmacology of functional recombinant receptors (Ramirez-Latorre et al., 1996).
In the central nervous system, a major class of nicotinic receptors defined by the high-affinity binding of α-[125I]bungarotoxin comprises predominantly, if not exclusively, the α7 subtype. Unlike other nAChR subtypes, the homomeric α7 subunit expressed in Xenopusoocytes exhibits a higher Ca++ permeability relative to monovalent cations (pCa/pmonovalent ratio: 20,Séguéla et al., 1993; 10, Bertrand et al., 1993; pBa/pmonovalent ratio: 17,Sands et al., 1993). These values are in the same order of magnitude, if not higher, than those reported for the NMDA class of glutamate receptors where pCa ++/pNa + ratio values ranging from 10.3 to 14.3 have been reported (Zarei and Dani, 1994;Castro and Albuquerque, 1995). The permeability ratios are also much higher than those determined for the muscle (α1)2β1γδ nAChRs (ratio values ranging from 0.1 to 1.0; Vernino et al., 1992; Costaet al., 1994) and the ganglionic α3-containing nAChRs (1.5, Vernino et al., 1992; Fieber and Adams, 1991) and other non-NMDA glutamate receptors (Dingeldine et al., 1992).
The physiological role(s) of Ca++ entering mammalian cellsvia the homomeric α7 nAChRs remains to be elucidated. Influx of Ca++ through the α7nAChR subtype has been suggested to be of particular importance in activating several Ca++-dependent processes including neurite growth, synaptic transmission and neurotrophic effects. Recent studies have demonstrated the involvement of α-bungarotoxin-sensitive nAChRs in (−)-nicotine-evoked Ca++ influx in ciliary ganglion neurons (Vijayaraghavan et al., 1992; Zhanget al., 1994), hippocampal neurons (Alkondon and Albuquerque, 1993) and at the synaptic junctions of the medial habenula and interpeduncular nuclei (McGehee et al., 1995). A role for this nAChR in neuromodulation/neuroprotection emerges from the observation that ligands that interact with this subtype have neuroprotective properties (De Fiebre et al., 1995; Meyeret al., 1994; Donnelly-Roberts et al., 1996).In vivo, α7 nAChR may coexist with other α-bungarotoxin-sensitive subunits (Anand et al., 1993) and this limits a direct analysis of the precise functional properties of the homomeric combination. Although the Ca++ permeability of the α7 homomer has been convincingly demonstrated by permeability ratio studies (i.e., by evaluating Ca++-dependent shift in the reversal potential with the extended Goldman-Hodgkin-Katz equation) (Séguéla et al., 1993; Bertrand et al., 1993; Sands et al., 1993), these measurements do not necessarily indicate whether the receptors can significantly elevate intracellular Ca++levels in mammalian cells (Vernino et al., 1994). Accordingly, the actual amount of Ca++ influx in mammalian cells and the kinetics of such events remain to be defined.
The human α7 nAChR has recently been stably expressed in a mammalian cell line and shown to function as a α-bungarotoxin-sensitive ion channel exhibiting rapid kinetics of activation and inactivation (Gopalakrishnan et al., 1995). In this study, we have measured simultaneously membrane current and intracellular Ca++ kinetics with a combination of whole-cell patch-clamp and fura-2 imaging techniques. Our studies provide direct evidence that the recombinant homomeric α7nAChR subtype promotes a significant and sustained increase in [Ca++]i in mammalian cells upon activation by nAChR agonists including (−)-nicotine, acetylcholine, (±)-epibatidine and 1,1-dimethyl-4-phenylpiperazinium.
Methods
Cell culture.
A stable cell line of human embryonic kidney 293 cells transfected with an expression vector containing the human α7 nAChR cDNA was used in the present study. Because the human α7 cDNA originally cloned by Doucette-Stammet al. (1993) lacked a signal sequence, a rat signal sequence was included with a synthetic oligonucleotide (5′-GCAGCACTCGAGCCATGTGCGGCGGGCGGGGAGGCATCTGGCTGGCTTGGCCGCGGCGCTGCTGCACGTGTCCCTGCAAGGCGAGTTCCAGAGGAAGCTTTACAAGGA-3′) containing internal XhoI and HindIII sites (underlined). The forward and reverse complement of the oligo were annealed, digested with XhoI/HindIII and ligated into the HindIII/XhoI sites of the plasmid (pBluescript) containing the human α7 sequence. The entire cDNA was then cut with XhoI/NotI and the 5′-overhang filled in with Klenow polymerase, linked withBstXI adapters and ligated into the BstXI site of the expression vector pRcCMV (Invitrogen, San Diego, CA) containing the constitutive human cytomegalovirus promoter to obtain the plasmid pRcCMVα7. HEK293s were transfected with this plasmid by lipofectamine, and individual antibiotic-resistant colonies were picked and propagated as previously described (Gopalakrishnan et al., 1995). The cell line, referred to as K28, expresses a high density of α-[125I]bungarotoxin binding sites (B max, ∼970 fmol/mg protein) and has been employed in this study. Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 0.25 mg/ml geneticin and maintained in a cell incubator at 37°C with 5% CO2. Untransfected cells were grown in the same media, without geneticin. Cells were plated in poly-l-lysine-coated ultrathin Fisher coverslips at 105 cells/ml in 35-mm Petri dishes (Falcon 1008, Becton Dickinson, IL). Nonconfluent cells were studied for 4 to 5 days after plating.
Intracellular Ca++ imaging and electrophysiological recordings.
Coverslips were mounted in a small flowthrough Lucite chamber positioned on the stage of an Axiovert 135 microscope (Zeiss, Germany). The recording chamber was provided with a central hole to allow a direct access to the plated cells. A coverslip was fixed to the recording chamber by means of a Vaseline ring in such a way that only one coverslip was interposed between the cells and the microscope objective (100 ×, NA 1.3, Fluar, Zeiss). Continuous cell perfusion with the bathing-external solution (see below) was done with a push-pull syringe pump (WPI, Saratoga, FL).
Cells were voltage-clamped by use of the whole-cell configuration of the patch-clamp with an Axopatch-200A amplifier (Axon Instruments, Foster City, CA) by standard techniques (Hamill et al., 1981). Micropipettes were pulled from borosilicate glasses (Boralex) with a Flaming Brown micropipette puller (P97, Sutter Instrument Co., Novato, CA) to obtain electrode resistance ranging from 2 to 4 megohm. The composition of the internal solution (pipette) was: 120 mM KF, 20 mM KCl, 2.0 mM MgCl2, 0.1 mM EGTA, 10 mM HEPES, pH was adjusted to 7.4 with KOH. To be consistent with prior studies (Gopalakrishnan et al., 1995) F− was included in the internal solution. F− improved the seal formation and stability compared with other anions. Experiments using F− or Cl− as the main internal cation did not show differences on inward current kinetics and intracellular Ca++ responses. In addition, two external solutions were used. Most of the experiments were performed in a solution containing a mixture of various cations, briefly named “Na+ and Ca++ solution” (mM): NaCl, 120; KCl, 5; CaCl2, 2; MgCl2, 2; HEPES, 10;d-glucose, 25; pH adjusted to 7.4 with OHNa. For Ca++ “calibration” an extracellular solution containing 75 mM CaCl2 and 10 mM HEPES was used (Vernino et al., 1994). Whole-cell currents were acquired at 5 kHz and filtered at 2 kHz with pClamp software 6.0 (Axon Instruments) run in a personal computer. Digidata 1200 interface (Axon Instruments) and a Hewlett-Packard 1300T Optical Disk drive were used for A-D conversion and data storing, respectively.
Fast agonist delivery and solution exchange.
Agonists were delivered by puffing the content of a reservoir with a Harvard syringe pump through one barrel of a theta tube micropipette (Sutter). Theta tubes were pulled and carved with a diamond pencil until a sharp pipette tip of about 30 μm was obtained. The bathing solution was puffed through the other barrel of the theta tube. Both solutions were delivered at the same flux rate, and a very clear edge separated both solution streams. A Newport (Novato, CA) manipulator was used for theta glass tubes positioning. The sharpness of the interface between flowing solutions was assessed by perfusing the two barrels with solutions with different ionic strength such as phosphate-buffered saline and 10% phosphate-buffered saline solution, respectively. Theta tubes were mounted in a piezoelectric device (Burleigh, Fishers, NY) and activated by a voltage pulse. The activation of the drug-delivering system was synchronized with the current and intracellular Ca++transients recordings. The reliability of solution exchange was measured as an open-tip pipette (filled with 3 M KCl) response to displacements of the theta tube interface. The time for 100% solution exchange was 1 ms when the open tip was positioned 100 μm from the theta tube tip. The latency between pulse and solution exchange is attributable mainly to the solution diffusion time from the perfusion to the recording pipette (Jonas, 1995).
Intracellular Ca++ kinetics.
Intracellular Ca++ kinetics was measured under whole-cell voltage-clamp with fura-2 pentapotassium salt (Molecular Probes, Eugene, OR) as a Ca++-sensitive dye. However, concentration-response curves to different nAChR agonists were done at resting membrane potential. Membrane potential was determined in a separated group of cells in current-clamp mode. The AM form of fura-2 (2 μM) was used for experiments in intact cells. Fura-2 salt was dissolved in the pipette solution at concentrations ranging from 400 to 500 μM. Fura-2 as a ratiometric dye allowed absolute determinations of the intracellular Ca++ concentration ([Ca++]i). Fura-2 was not saturated at the concentrations used in our studies because 1 μM ionomycin induced further increase in fluorescence at 350 nm excitation wavelength. Ionomycin has been used previously to induce a massive influx of extracellular Ca++ to the cell without affecting intracellular stores (Delbono and Stefani, 1993).
For fluorescence studies, samples were illuminated with a 75 W xenon lamp. The light beam passed through excitation and emission filters (Omega Optical, Battleboro, VT) which were mounted in a filter wheel and a sliding microscope holder, respectively. In our particular experimental conditions, 340 and 380 nm excitation wavelengths were equally sensitive to changes in intracellular Ca++concentration. As increments in [Ca++] evoke a positive signal at 340 nm and a negative signal at 380 nm, we preferred the use of 340/350 ratio which facilitated calculation and illustration of changes in [Ca++]. For the acquisition of fura-2 fast image sequences, background fluorescence measurements at 340 and 350 nm (isosbestic point) wavelength were recorded. Background fluorescence was subtracted to the images acquired at 340 and 350 nm. The background subtraction and ratio calculation were done off-line and transformed into [Ca++]i (see below). The pipette fluorescence did not interfere with cell fluorescent recordings because a diaphragm was used to leave the pipette out of the microscope field.
To reduce out-of-focus information a high N.A. (1.3) objective was used (see above) and image processing through a “no-neighbors deblurring scheme” (Monk et al., 1992) was performed off-line. The time course of intracellular Ca++ transients was resolved with a frame-transfer cooled CCD PXL EEV37 camera (Photometrics, Tucson, AZ) as a photon detector. For image acquisition and processing Isee® software (Inovision, Durham, NC) was run in a SUN SPARCStation LX (Sun Microsystems, Santa Clara, CA). To improve the sampling rate, Ca++ image sequences were grabbed with the camera and light beam pathway shutters open. Thus, no mechanical delays were imposed on the recording system. Changes in intracellular Ca++concentration were sampled every 10 ms which provided appropriate spatial and time resolution. Modifications of pixel binning, array size and camera gain allowed the improvement of the signal/noise ratio for each experiment.
Calculation of intracellular Ca++concentration.
For [Ca++]i calculationsR =F 340/F 350,R min (0% saturation of the dye by Ca++) and R max (100% saturation with Ca++) were determined. R max was obtained by adding a low dilution of saponin (0.001%) to the bathing solution at the end of the experiments. R min was measured in cells preincubated for 30 min in 2 mM EGTA. [Ca++]i was calculated with the following equation: [Ca++] =KD [(R −R min)/(R max −R)], where R is the fluorescent ratio,R max is the maximum ratio value measured at high [Ca++]i and R min is the minimum ratio value measured at low [Ca++]i, KD is the dissociation constant of the Ca++ binding to fura-2 (224 nM; Grynkiewicz et al., 1985). The values forR min and R max determined by calibration were 0.2 and 3.5, respectively. For computer simulation of intracellular Ca++ transients Matlab software (MathWorks, Inc., MA) was used. As the basal [Ca++]i varied in a range from 65 to 350 nM (207 ± 82; n = 50) in transfected HEK293, only cells with a resting cytosolic Ca++ less than 100 nM were included in this study. The reason for this selection is that the effect of high intracellular [Ca++] on α7nAChR kinetics has not yet been explored.
Concentration-response curves.
Concentration-response curves were performed for four different nAChR agonist ((−)-nicotine, acetylcholine, DMPP iodide, DMPP and (±)-epibatidine) in intact cells (see above). An average of four different increasing concentrations were used in the same cell. The time course of changes in [Ca++], including the mean time for complete return to prestimulating levels, was determined. Drugs were applied for 50 ms with 1-min intervals. This interval assured a complete development of the desensitizing phase even at lower (or less desensitizing) concentrations and for a complete return of cytosolic [Ca++] to prestimulating levels. A saturating concentration of the agonist was included in each experiment. Thus, responses to lower agonist concentrations were normalized to a maximum elicited by the agonist. Six to eight cells were used for each group of experiments. Intracellular Ca++ responses were corrected for run-down according to procedures described below. Comparative efficacies were evaluated by measuring absolute maximal increase in [Ca++]. Lower concentrations of the agonists induced Ca++ signals that were a fraction of the maximum response. For plotting, Ca++ responses were normalized to the maximum at the saturating concentration of the agonist.
Materials.
Cell culture media, fetal bovine serum and geneticin were purchased from Life Technologies, Inc. (Grand Island, NY). The following materials were purchased from sources indicated: acetylcholine chloride, (−)-nicotine hydrogen tartrate, atropine sulfate and 1,1-dimethyl-4-phenyl-piperazinium iodide were obtained from Sigma Chemical Co. (St. Louis, MO), α-bungarotoxin, MLA citrate and (±)-epibatidine dihydrochloride were purchased from Research Biochemicals International (Natick, MA). Fura-2 pentapotassium salt, fura-2 AM and Mag-fura-2 AM were purchased from Molecular Probes (Eugene, OR). Fura-2 pentapotassium salt was dissolved in the pipette solution at concentrations ranging from 400 to 500 μM. Fura-2 AM stock was prepared in 10% dimethyl sulfoxide. With acetylcholine as the agonist, atropine (2 μM) was used to block endogenous muscarinic receptors.
Results
Whole-cell inward current and intracellular Ca++transients were recorded simultaneously in HEK293 stably expressing the human α7 nAChR. A short pulse (50 ms) of 100 μM (−)-nicotine induces a fast inward current (activation time constant, τa = 1.1 ± 0.23 ms) that rapidly desensitizes (desensitization time constant, τd = 12.2 ± 2.6 ms;n = 5) in a cell voltage-clamped at −100 mV (fig.1). The (−)-nicotine-induced inward current was completely abolished by 3 nM α-bungarotoxin and 2 nM MLA (n = 7 for both antagonists). These results are in agreement with those previously reported for this cell line (Gopalakrishnan et al., 1995).
Simultaneous whole-cell inward current and intracellular Ca++ transient recorded from HEK293 stably expressing the human α7 nAChR. A short pulse (50 ms) of 100 μM (−)-nicotine elicits a fast inward current (activation time constant, T a = 1.1 ms) that rapidly desensitizes (desensitization time constant,T d = 12 ms) in a cell voltage-clamped at −100 mV. Shown is a representative inward current trace and selected fura-2 Ca++ ratio (340/350 nm) images at various time intervals (indicated by arrows). The inward current elicits an intracellular Ca++ transient as monitored with fura-2 pentapotassium salt. Pseudocolor (100 × 100 pixels) calcium ratio (340/350 nm) images (2-ms exposure) processed off-line illustrate changes in calcium concentrations at them times indicated by arrows. The maximum [Ca++]i are 69, 104, 142, 305 (peak), 223 and 192 nM for images (a–f) respectively.
Time course and magnitude of (−)-nicotine-evoked Ca++ transient.
Figure 1 also shows selected fura-2 Ca++ ratio (340/350 nm) images at various time intervals. The pattern of (−)-nicotine-evoked Ca++transient was recorded simultaneously with inward current responses from the same cell. Previous studies with the adrenal chromaffin cells have shown that the amplitude of the peak Ca++ transient increases in the negative voltage range (Zhou and Neher, 1993). Thus, hyperpolarizing holding potentials were used to better define the Ca++ responses in the present study. The intracellular Ca++ rises from within the first 10 ms after exposure to 100 μM (−)-nicotine and further elevations were observed during the desensitization phase of the current. The (−)-nicotine-evoked maximal increase in [Ca++]i was observed at the time during which the desensitization of the current was complete. The mean peak [Ca++]i observed with 100 μM (−)-nicotine was 356 ± 48 nM (n = 8). The mean peak [Ca++]i corresponded throughout the drug pulse to the cell outermost shells. Thereafter, [Ca++]i starts to decline slowly reaching 50% of the peak response within 20 ± 6 s (n= 8) after beginning of the nicotine pulse. The time course of the peak [Ca++]i changes after application of 100 μM nicotine is depicted in figure 2A. A longer time was required for returning [Ca++]i to prestimulation values (range, 0.5–1 min; n = 8). No inward current (n = 8) or intracellular Ca++ changes (n = 25) were detected in the presence of 300 to 1000 μM (−)-nicotine in untransfected cells, consistent with previous reports showing lack of nAChR-mediated responses in HEK293 (Gopalakrishanan et al., 1995). As a control for potential effects of geneticin on the kinetics of Ca++ influx, transfected cells grown in the absence of geneticin for 24 to 48 h before testing were identical in response to experiments done in the presence of geneticin (n = 15).
Relation between intracellular Ca++transient and inward current. (A) Time course of peak intracellular [Ca++] sampled every 10 ms in whole-cell voltage-clamped HEK293 expressing the human α7 nAChRs and stained with 500 μM fura-2 via the patch pipette. The horizontal bar at the bottom indicates the exposure to 100 μM (−)-nicotine application (50-ms duration). Only the first 500 ms of the recording is shown. (B) Total charge measured as the integral of the current trace (shaded area). (C) Simulation of intracellular Ca++transient assuming that influx through plasma membrane is the Ca++ source (see text).
In our experimental conditions cells were overloaded with fura-2 to override the endogenous Ca buffers. Ca++pumping/sequestration in the presence of high exogenous buffers has an intrinsic time constant of >30 s (Neher and Augustine, 1992). Because (−)-nicotine-evoked current flow is completed in about 100 ms, the effect of Ca++ pumping/sequestration on Ca++influx calculations is negligible. In this way, fluorescent emission at 510 nm is directly proportional to the rate of Ca++ influx: Δ[Ca++] (t) = −I CaΔt/(2FV) (Sala and Hernandez-Cruz, 1990), where I Ca is the transmembrane Ca++ current, F is Faraday’s constant and V is the volume. For a current of 663 nC (fig.2B); and, assuming a 20% fractional current carried by Ca++ ions, the expected peak Δ[Ca++]i is 300 nM (fig. 2C), which corresponds with experimental records (fig. 2A). Another way to verify this prediction, is through the calculation that 20% of the total charges (633 nC) is carried by Ca++ ions. Because a flow of 140 pA is equivalent to 3.5 × 10−17 mol of Ca++ for a cell of 10 to 20 μm diameter, 85% accessible volume fraction and a value of Ks (binding capacity of the endogenous buffer) (Neher and Augustine, 1992) of 40 give a Δ[Ca++]i increase of 300 to 400 nM. Mathematical predictions were corroborated by use of pure extracellular Ca++ solution (Vernino et al., 1994). In these experiments, Ca++ was the only cation allowed to carry inward current in response to (−)-nicotine pulses. The nicotine-induced current in pure Ca++ was reduced to 21 ± 2.7% (n = 7) of the control current amplitude. Control was the (−)-nicotine-induced current recorded in the same cell in the presence of the extracellular Na+ and Ca++ solution (see “Methods”).
Source of agonist-evoked intracellular Ca++.
The [Ca++]i distribution showed a nonuniform pattern across the cell stimulated with 100 μM (−)-nicotine (fig.3). Pixel intensity profile analysis along one vector across the cell shows a Ca++ gradient with higher concentrations near the cells periphery. This gradient is an indication of Ca++ influx through channels expressed on the external membrane. The expression of endogenous voltage-gated calcium channels was assessed in whole-cell voltage clamp in transfected (n = 15) and untransfected HEK293 (n = 25). No inward current was detected in cells pulsed from three different holding potentials (−50, −70 and −100 mV) to +60 mV with 10-mV increments. Cell hyperpolarization together with the lack of expression of voltage-gated calcium channels in HEK293 (n = 25) also support the cationic influx through expressed human α7 nAChRs.
Distribution of [Ca++]iin a HEK293 expressing the human α7 nAChRs. Nonuniform distribution of [Ca++]i in a HEK293 expressing the α7 nAChRs exposed to 100 μM (−)-nicotine. A three-dimensional representation of the pixel intensity distribution corresponding to image a and figure 1 (image d) is shown in b. The pixel intensity profile analysis along one vector across the cell (indicated by arrow in image a) shows the maximal pixel intensity around the cell periphery.
Addition of EGTA (5 mM) to the extracellular media ablated (−)-nicotine-evoked increase in [Ca++]i in transfected cells. In the presence of EGTA, [Ca++]i values in response to 100 μM (−)-nicotine were 63.7 ± 12.8 nM (n = 6). These values were not significantly different with respect to control (58.3 ± 14.6 nM; n = 6) where the cells were perfused with bathing solution, but were statistically significant with respect to cells perfused with 100 μM (−)-nicotine (356 ± 48;n = 8). These experiments demonstrate the requirement for extracellular Ca++ in mediating (−)-nicotine-evoked responses.
It was then investigated whether any contribution to (−)-nicotine-evoked Ca++ transient arose from intracellular organelles. Preincubation of the cells for 10 min in heparin (20 mg/ml) or ryanodine (5 μM) blockers of IP3and endoplasmic reticulum Ca++ release channel-ryanodine receptor, respectively, did not change the time course and distribution of the Ca++ response. Ryanodine exerts a direct effect by directly activating, locking a subconductance state and finally blocking the ryanodine receptor. The ryanodine effect takes about 3 min to be completed when used at a concentration of 5 μM (Rousseauet al., 1987; Delbono and Chu, 1995). This is the rationale for applying the (−)-nicotine pulse after the ryanodine blockade of the ryanodine receptor was completed. After incubation in ryanodine or heparin [Ca++]i was 372 ± 39 nM (n = 5) and 385 ± 63 (n = 5), respectively. This demonstrates that ryanodine receptor and IP3 receptor do not participate in the (−)-nicotine-evoked Ca++ transient response.
Pharmacology of agonist-evoked Ca++responses.
To examine whether antagonists with reported selectivity toward the α7 nAChR relative to the α4β2 and ganglionic type nAChRs (Gopalakrishnan et al., 1995; Wonnacott et al., 1993) could block (−)-nicotine-evoked rise in [Ca++]i, cells were preincubated with either α-bungarotoxin (3 nM) or MLA (2 nM) for 10 min. As in whole-cell voltage clamp the run-down was fast (0.5 ± 0.08 of the first response at 5 min after accessing into the intracellular compartment,n = 32), cells were preincubated in α-bungarotoxin and MLA before getting into whole-cell patch-clamp configuration. Subsequent application of 100 μM (−)-nicotine did not show statistically significant elevations in [Ca++]i was compared with control (n = 7 for both antagonists tested; fig.4). The abolishment of intracellular Ca++responses in the presence of these nAChR antagonists is further supported by the complete suppression of single-channel activity recorded in outside-out patches from transfected cells (Messi et al., 1996).
Blockade of intracellular Ca++responses by nAChR antagonists. Blockade of intracellular Ca++ mobilization elicited by 100 μM (−)-nicotine by α-bungarotoxin and MLA. Shown are mean (±S.E.M.) values of [Ca++]i in control cells (no exposure to nicotine), and in cells exposed to (−)-nicotine before (N) and after pretreatment with 3 nM α-bungarotoxin or 2 nM MLA (N + MLA). The difference between the peak [Ca++]i in nicotine-treated (100 μM) and control cells was significant (P < .001) (n = 6–8 cells for each group of experiments).
Although the time course of Ca++ signals were defined in cells loaded with fluorescent dyes under whole-cell voltage-clamp conditions, concentration-response curves were carried out in intact cells caused by the run-down of responses after nAChR activation under whole-cell recording conditions (Lester and Dani, 1994). This permitted complete concentration responses for each agonist to be performed in the same cell. The resting membrane potential measured in a separate group of cells under current-clamp mode was −50 ± 3.8 mV (n = 25). The magnitude of the run-down of the intracellular Ca++ responses greatly differed in voltage-clamped whole-cells and in intact cells. In the latter, paired stimulations with 100 μM (−)-nicotine with variable intervals were used to determine the magnitude of the run-down of the Ca++signal. In dose-response studies the intervals between agonist applications were 2 min. Thus, controls applying the same (−)-nicotine concentration three to four times with 2-min intervals were performed. The ratio between the amplitudes of the second peak Ca++response over the first response after four (−)-nicotine applications, or 10 min after the first drug application, was 0.79 ± 0.07 (n = 18). Because run-down was linear as a function of time and independent of the agonist type (n = 10) and concentration (n = 16), corrections of Ca++signals were applied according to the time at which the response was elicited. It has been suggested that the addition of ATP, phosphocreatine and creatine phosphokinase to the pipette solution minimizes run-down in whole-cell voltage-clamp recordings (Alkondonet al., 1994). In our hands, this procedure did not ameliorate run-down (n = 15). Figure 5shows fura-2 fluorescence digital images after application of varying concentrations of (−)-nicotine and (±)-epibatidine in HEK293 stably expressing the α7 nAChR. Both (−)-nicotine and (±)-epibatidine elicited a concentration-dependent increase in [Ca++]i. The half-effective concentration (EC50) for (−)-nicotine was 51 ± 4 μM (n = 8) with maximal response observed at 100 μM (−)-nicotine (fig. 6A). Although the EC50value for (±)-epibatidine (3.5 ± 0.2 μM; n = 5) was approximately 14-fold lower than that of (−)-nicotine, the efficacy of these compounds did not differ significantly. Two other agonists tested for Ca++ influx responses were acetylcholine and 1,1 dimethyl-4-phenyl-piperazinium (fig. 6A). Comparative efficacies of nAChR agonists were evaluated by measuring absolute maximal increase in [Ca++]. Lower concentrations of the agonists induced Ca++ signals that were a fraction of the maximum. For plotting, Ca++ responses were normalized to a maximum or saturating concentration of the agonist. Intracellular Ca++ values in intact cells were lower than in whole-cell voltage-clamp recordings which can be attributed to differences in membrane potential. Maximum responses to (−)-nicotine, acetylcholine, DMPP and (±)-epibatidine were (mean ± S.E.M., in nM): 308 ± 27, 294 ± 32, 317 ± 33 and 311 ± 28 (n = 4–6 cells), respectively. The EC50values for acetylcholine (108 μM) and DMPP (75 μM) closely correlate with the EC50 values for evoking whole-cell current responses (fig. 6B). For the analysis included in figure 6B, EC50 for current activation are from Gopalakrishnanet al. (1995), except for (±)-epibatidine. (±)-Epibatidine-induced currents were studied under whole-cell voltage clamp following the same procedures described for (−)-nicotine. In these conditions the EC50 for (±)-epibatidine was 3 μM.
Comparison of the effect of (−)-nicotine and (±)-epibatidine on intracellular Ca++ responses. Intracellular Ca++ responses to nAChR agonists in HEK293 stably expressing α7 nAChRs. Fura-2 fluorescent digital images were recorded after application of varying concentrations of the two nAChR agonists, (−)-nicotine (left panel) and (±)-epibatidine (right panel) to the same cell (n = 5 for each agonist).
Concentration dependence of agonist-evoked Ca++ responses. (A) Shown are concentration-response curves to acetylcholine, (−)-nicotine, (±)-epibatidine and 1,1-dimethyl-4-piperazinium iodide in HEK293 stably expressing the human α7 nAChRs. Various concentrations were tested in individual intact cells and responses plotted as a fraction of the maximal intracellular Ca++ response. Intracellular Ca++ signals were normalized to maximal responses for each agonist. Data (mean ± S.E.M.; n = 4–6) were fitted to a Hill equation. The EC50 values for evoking Ca++ influx by (±)-epibatidine, (−)-nicotine, DMPP and acetylcholine were 3.5, 51, 75 and 108 μM respectively. (B) Correlation between the EC50 values for current activation and intracellular Ca++ responses.
Discussion
The definition of the physiological role(s) of various neuronal nAChR subtypes has, in recent years, been aided by the molecular cloning and expression of the various nAChR subunits and/or subunit combinations in mammalian cells (Whiting et al., 1991;Pereira et al., 1994; Puchacz et al., 1994). The human α7 nAChR has recently been stably expressed in a mammalian cell line, HEK-293, and shown to function as a homo-oligomeric ion channel exhibiting rapid kinetics of activation and inactivation of channel current, sensitive to α-bungarotoxin (Gopalakrishnan et al., 1995). This study documents, for the first time, that Ca++ permeation through the recombinant human α7 nAChRs promotes a significant transient rise in intracellular Ca++ concentration in mammalian cells. The dynamics of α7 nAChR-evoked Ca++ response has also been characterized and directly correlated with the pharmacological profile previously defined for this nAChR subtype (Gopalakrishnan et al., 1995).
Time course of intracellular Ca++transients.
Simultaneous cationic inward current and intracellular Ca++ recordings under whole-cell voltage-clamp conditions demonstrated a fast Ca++ transient that reaches a peak at the end of the desensitizing phase of the current. Such Ca++ transient and current kinetic values are similar to those reported previously in neurons expressing a diversity of nAChR subtypes (Zhou and Neher, 1993; Rathouz and Berg, 1994; Verninoet al., 1994). The decline of the [Ca++]i to base-line values is a relatively slow process in this expression. This kinetics probably differs from neurons which exhibit cell-to-cell variation in Ca++ buffer capacity. Our present studies confirm, as previously suggested, that α-bungarotoxin-sensitive nicotinic receptors elevate intracellular free Ca++ in mammalian cells (Vijayaraghavan et al., 1992). Studies with the recombinant human α7subunit stably expressed in HEK293 differ from those observed in chick ganglion neurons in terms of a much higher sensitivity of the latter to nicotine (0.1–1.0 μM), slower time course for the Ca++response (2–4 s after nicotine application, although current duration was less than 500 ms) and diminished sensitivity to α-bungarotoxin blockade at higher nicotine concentrations. Whether these differences reflect the presence of other subunits contributing to the effects found in chick ganglion neurons, as recently reported by Pugh et al. (1995), or caused by differences in the amino acid sequences of the chick and human α7 subunit remain to be determined.
The agonist-evoked [Ca++]i kinetic responses reported herein differ from those previously described in cells expressing multiple nAChR subunits including medial habenula neurons (Mulle et al., 1992), adrenal chromaffin cells and BC3H1 cells (Vernino et al., 1994). In these studies, a smaller and longer-lasting (−)-nicotine-evoked intracellular Ca++ response was reported. This could be attributed to a slower desensitization of the response and lower Ca++ permeability of the various nAChR subtypes expressed in these cells. Although the duration between agonist-evoked activation and desensitization of the current was no longer than 100 ms in our study, the higher permeability of the α7 nAChR for Ca++ than for monovalent cations is sufficient to elicit a significant increase in [Ca++]i. Further experiments determining Ca++ transients in different extracellular Ca++ concentrations will allow estimation of the fraction of current carried by Ca++ ions and the significance of the extracellular space in determining the magnitude of intracellular Ca++ elevations (Vernino et al., 1992, 1994; Amador and Dani, 1995).
Magnitude of intracellular Ca++elevations.
Our study demonstrates that Ca++permeation through the recombinant α7 nAChRs elicits a significant rise in [Ca++]i in HEK293. The relative increase is the same order of magnitude as observed for the α3βx nAChR subtype (expressed in chromaffin cells; Δ[Ca++]i = 50–100 nM at −50 mV), ATP and NMDA receptors (Zhou and Neher, 1993; Schneggenburger et al., 1993; Rogers and Dani, 1995). However, the increase in [Ca++]i is much lower than those reported for glutamate receptors expressed in HEK293 (Burnashev et al., 1995). The (−)-nicotine-evoked intracellular Ca++transient measured in this work is consistent with theoretical predictions based on the fraction of the inward current carried by Ca++ ions and simulations of intracellular Ca++transients resulting from Ca++ influx through the plasma membrane. A 20% fractional current carried by Ca++ ions is also consistent with direct measurements of Ca++ flux with use of pure Ca++ extracellular solution.
Pharmacology of intracellular Ca++responses.
An excellent correlation was observed between the pharmacological profiles of (−)-nicotine, acetylcholine, DMPP and (±)-epibatidine, obtained by measuring whole-cell currents and intracellular Ca++ responses (fig. 6B). The 1:1 correlation between agonist-induced intracellular Ca++ responses and whole-cell currents indicates that the measured intracellular Ca++ transient is a result of agonist-induced activation of the channel and is not influenced by mechanisms of Ca++-induced Ca++ release from internal stores or Ca++-activated secondary conductance. Ca++-induced Ca++ release was ruled out by using ryanodine receptor and IP3 receptor antagonists. In these experiments (−)-nicotine induced similar intracellular Ca++ responses than in the absence of the antagonists. A low electrochemical gradient for Cl− (F− is the main internal anion) rules out the possibility to elicit an outward Ca++-dependent Cl− current. Obviously, this mechanism can not contribute to measured intracellular Ca++signals. Further, the blockade of both current and [Ca++]i responses by nanomolar concentrations of α-bungarotoxin and MLA provides conclusive evidence that (−)-nicotine-evoked responses are mediated via an interaction with α7 nAChRs.
Source of (−)-nicotine-evoked intracellular Ca++ transient.
Further support for the conclusion that (−)-nicotine-evoked rise in [Ca++]i in transfected HEK293 is mediated primarily via the α7 nAChRs arises from the following set of observations. The spatial [Ca++]i distribution within the cell revealed by pixel intensity profile analysis (fig. 3) indicates that the Ca++ concentration is higher at the cell periphery than in central areas and that the gradient between outer and inner regions is maintained during Ca++ influx which subsequently dissipates. Second, the lack of effects of heparin and ryanodine, blockers of inositol (Deneris et al., 1991; McGehee and Role, 1995; Lindstrom et al., 1995) triphosphate receptor and ryanodine receptor, respectively, on (−)-nicotine-evoked Ca++ transients, indicates that these intracellular Ca++-mobilizing pathways do not contribute directly to the observed responses (Thayer and Miller, 1990; Zhou and Neher, 1993;Schneggenburger et al., 1993). Third, the lack of endogenous expression of NMDA and voltage-gated Ca++ channels in HEK293 (Burnashev et al., 1995; Williams et al., 1992), together with the fact that intracellular Ca++transients were recorded at a hyperpolarizing potential (−100 mV), eliminates any potential contribution of voltage-gated or other Ca++-permeable ion channels to nAChR agonist-evoked Ca++ response.
The protocol used in our study also favors a higher open probability of the α7 nAChR which has been shown to increase steeply at negative voltages by whole-cell measurements (Galzi et al., 1992; Zhang et al., 1994; De Fiebre et al., 1995) and by direct single-channel recordings at a wide range of voltages (−150 to +50 mV) in outside-out and cell-attached patch-clamp configurations (Messi et al., 1996). Finally, although the measured cytosolic Ca++ depends on mechanisms of ion extrusion, sequestration and buffering (for a review, see ref. 45), the episode of agonist-evoked Ca++ influx in these cells is short enough (<100 ms) that the contribution of Ca++sequestration or pumping across the plasma membrane during Ca++ influx are negligible (see above) (Zhou and Neher, 1993).
Relevance to cellular function.
It is known that alterations in [Ca++]i play a pivotal role in modulating cell-to-cell and intracellular signaling mechanisms in neurons of the central and peripheral nervous system (Kater et al., 1988;Kyrozis et al., 1995). Neuronal Ca++ transients resulting from activation of voltage-gated Ca++ channels and NMDA class of glutamate receptors have been widely studied (Burnashev et al., 1995; Kyrozis et al., 1995). More recently, Ca++ permeability of acetylcholine-activated currents in regions of the central nervous system have received considerable attention and it has become increasingly clear that such responses are mediated by distinct nAChR subtypes. Studies in cultured hippocampal neurons and ciliary ganglion neurons have demonstrated the existence of α-bungarotoxin-sensitive nAChRs with high Ca++ permeability, whose activation elicits a rapidly activating and rapidly desensitizing current to nAChR agonists (Vijayaraghavan et al., 1992; Castro and Albuquerque, 1995;Zhang et al., 1994). In hippocampal neurons, the relative Ca++ permeability values of the α-bungarotoxin-sensitive nAChR are similar to those of the NMDA class of glutamate receptors [pCa++/pCs+ of 6 vs. 10 for the NMDA channel (Castro and Albuquerque, 1995)] and much higher than those reported for the muscle-type nAChRs expressed in BC3H1 cells (0.2; Vernino et al., 1992) and other α-bungarotoxin-insensitive nAChRs expressed in parasympathetic ganglia (1.5; Fieber and Adams, 1991); PC12 cells (2.5; Sands and Barish, 1991) or adrenal chromaffin cells (1.5; Vernino et al., 1992). Studies with the recombinant rat or chick α7 subunit transiently expressed in Xenopusoocytes also reveal a remarkably high relative permeability to Ca++ ions with pCa/pmonovalentvalues ranging from 10 to 20 (Séguéla et al., 1993; Bertrand et al., 1993; Sands et al., 1993). This high Ca++/Na+ permeability of the α7 homo-oligomeric channel compared with the muscle (α1)2βγδ or other neuronal subtypes, including the widely distributed α4β2, is suggestive of a role for the α7 nAChR in mediating a variety of neuronal Ca++-dependent processes.
Activation of nAChRs has been associated with several Ca++-dependent processes including protein kinase C activation (Messing et al., 1989), facilitation of ACh release (Vizi and Somogyi, 1989), neurite outgrowth (Pugh and Berg, 1994) and retraction of growth cone phylopodia (Chan and Quik, 1993). More recently, a role for nAChRs in modulating glutaminergic transmission has been reported (McGehee et al., 1995). A potential role for nAChRs in neurodegenerative processes is suggested by recent findings that nAChR ligands such as (−)-nicotine and ABT-418 are neuroprotective in in vitro models of cytotoxicity (Akaike, 1994; Donnelly-Roberts et al., 1996). In fact, the effects of ABT-418 and (−)-nicotine are blocked by α-bungarotoxin which suggests the participation of α-bungarotoxin-sensitive nAChRs in mediating such effects.
The relatively high Ca++ permeability, the widespread distribution of α7-containing nAChRs in the central nervous system especially in hippocampus, limbic cortex and thalamus, the pre- and postsynaptic localization and the ability to be activated at potentials more negative than those required for activation of voltage-gated and other ligand-gated Ca++-permeable channels are notable features that endow distinctive functional role(s) for the α7 nAChR subtype (Wong and Gallagher, 1991; Mulleet al., 1992; Vijayaraghavan et al., 1992). Previous studies have shown that sustained subtoxic elevations in Ca++ can promote survival of cells that would otherwise undergo programmed cell death (Franklin and Johnson, 1992). It is noteworthy that the magnitude of rise in [Ca++]i elicited by (−)-nicotine in these cells is to a level (∼350 nM) that is not excitotoxic under normal conditions, but rather promotes neuronal growth and survival even in the absence of neurotrophic factors.
The activation of the α7 nAChR and resultant Ca++ influx at resting or hyperpolarizing conditions may play a role in neuronal plasticity and in mediating the cellular responses to excessive glutamate stimulation (McGehee et al., 1995). Should α7 nAChR-activated current density suffice to depolarize the extrasynaptic membrane (Horch and Sargent, 1995), additional Ca++ influx may be initiated by activation of voltage-gated calcium channels (Vijayaraghavan et al., 1992; Rathouz and Berg,1994). In summary, activation of human α7 nAChR has significant effects on intracellular Ca++ dynamics which could modulate unique cellular signaling processes.
Acknowledgments
We are grateful to Dr. Stephen Arneric for critically reading the manuscript and for helpful suggestions.
Footnotes
-
Send reprint requests to: Dr. Osvaldo Delbono, Dept. of Physiology and Pharmacology, Bowman Gray School of Medicine, Medical Center Boulevard, Winston-Salem, NC 27157.
-
↵1 Research in the laboratory of O.D. was supported by grants from the National Institutes of Health 2-P60AG10484, T-32-AG00182 and K01 AG00692 and from the Muscular Dystrophy Association (U.S.A.).
- Abbreviations:
- HEK293
- human embryonic kidney 293 cells
- nAChR
- nicotinic acetylcholine receptor
- EGTA
- ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid
- NMDA
- N-methyl-d-aspartate
- IP3
- inositol 1,4,5-triphosphate
- MLA
- methyllycaconitine
- DMPP
- 1,1 dimethyl-4-phenyl-piperazinium
- ABT-418
- (S)-3-methyl-5-(1-methyl-2-pyrrolidinyl) isoxazole
- HEPES
- N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid]
- PBS
- phosphate-buffered saline
- Received May 2, 1996.
- Accepted September 3, 1996.
- The American Society for Pharmacology and Experimental Therapeutics
