Introduction

Top Summary Introduction Materials & Methods Results Discussion References

Nicotine, the pharmacologically active compound of tobacco leaves, mimics several characteristic effects of the natural ligand ACh in the nervous system through the activation of ACh-gated ion channels, referred as nicotinic. nAChRs belong to the superfamily of ligand-gated channels and have been shown to share a number of structural and functional homologies with other members, such as GABA_A, glycine, and serotoninergic receptors (Bertrand and Changeux, 1995; Lindstrom, 1996). It is thought that like the muscle nAChR, a single neuronal receptor results from the assembly of five subunits, each of which spans the membrane four times. Although only five genes are known to code for the two muscle nAChRs, 11 subunits have been identified for the neuronal nAChRs [2-9, 2, 3, and 4 (for a review, see Lindstrom, 1996)]. Homomers or different combinations of these neuronal nAChR subunits have been shown to be expressed in the peripheral and central nervous systems (Bertrand and Changeux, 1995). Based on the alignment of their protein sequences and physiological as well as pharmacological properties, these subunits can be subdivided further into two main groups: those that contribute to receptors that are sensitive to the snake toxin Bgt and those that do not bind this toxin. Another important distinction is that the three Bgt-sensitive subunits (7-9) can reconstitute functional homomeric receptors when expressed in a host system. The 7 subunit is expressed strongly in the brain, and a high degree of correlation is observed between Bgt labeling and in situ hybridization with 7 mRNA probes (Wevers et al., 1995).

Among the homomeric nAChRs, the 7 subunit is by far the most studied, and it frequently is used as a model of the nAChR family (for a review, see Bertrand and Changeux, 1995). Physiological characterizations showed that the receptor reconstituted with this subunit is highly permeable to Ca²⁺ with a selectivity pCa/pNa ratio of 10:1 (Bertrand et al., 1993). It later was shown that the influx of Ca²⁺ observed during receptor activation by ACh is sufficient to trigger an increase in the intracellular Ca²⁺ concentration (Delbono et al., 1997). In addition, it was shown that activation by low nicotine concentration of an Bgt-sensitive subunit in rat hippocampal cultured neurons results in an increase in the release of the glutamate neurotransmitter (Gray et al., 1996). Similarly, nicotinic agonist enhanced the release of the neurotransmitter GABA in the mouse thalamus (Léna and Changeux, 1997).

An important feature of the neuronal nAChRs is the negative or positive allosteric regulations caused by a variety of pharmacological agents (Bertrand et al., 1991b; Mulle et al., 1992; Valera et al., 1992; Galzi et al., 1996a; Buisson and Bertrand, 1998). For example, steroids have been found to inhibit neuronal nAChRs through allosteric interactions (Valera et al., 1992). This property is shared by both the heteromeric and homomeric receptors, as recently shown with the human 42 and 7 nAChRs (Buisson and Bertrand, 1998). On the contrary, it was shown that neuronal nAChRs are potentiated strongly by the extracellular Ca²⁺ concentration (Mulle et al., 1992; Galzi et al., 1996d). The combination of site-directed mutagenesis with electrophysiology recently led to the identification of a Ca²⁺ binding site in the extracellular domain of the 7 subunit (Galzi et al., 1996a).

The specific loss of nAChR binding sites observed in the brain of patients with neurodegenerative diseases such as Alzheimer's or Parkinson's suggests that these receptors may play a critical role in the evolution of these disorders (Warpman and Nordberg, 1995). Furthermore, stimulation of nAChRs in affected patients either by selective agonists or by compounds reducing the activity of acetylcholinesterase shows positive effects on the patients' overall cognitive abilities. A complementary approach would be to use positive allosteric effectors that could selectively enhance the activity of neuronal nAChRs.

Although many compounds have been shown to behave as positive allosteric modulators on the GABA_AR (for a review, see Rabow et al., 1995), much less is known about the neuronal nAChRs. The recent demonstration that the well known anthelmint IVM is a powerful allosteric modulator of worm glutamate receptors (Cully et al., 1994) and that these cloned subunits share homologies with the nAChRs prompted us to examine the effect of this compound on neuronal nAChRs.

In the current work, we studied the effects of IVM on chick and human homomeric 7 nAChRs reconstituted in Xenopus laevis oocytes and in K-28 cells. We found that although IVM alone induces no detectable current, preapplications of IVM strongly enhance successive ACh-evoked currents. Investigation of the mechanisms underlying IVM potentiation suggests that this compound acts directly on the receptor protein as a positive allosteric effector.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Oocyte preparation and cDNA injection. X. laevis oocytes were isolated and prepared as described previously (Bertrand et al., 1991a). The oocytes were injected intranuclearly with 2 ng of expression vector cDNA and maintained at 18° in Barth's medium (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 10 mM HEPES, 0.82 mM MgSO₄·7H₂O, 0.33 mM Ca(NO₃)₂·4H₂O, 0.41 mM CaCl₂·6H₂O, pH 7.4 adjusted with NaOH) with antibiotics (20 µg/ml kanamycin, 100 units/ml penicillin, and 100 µg/ml streptomycin). To improve cell survival and minimize possible contamination, each oocyte was placed in one of the 96-well microtiter plates (Nunc, Naperville, CT).

Drugs and solutions. Drugs and chemicals were purchased from Sigma Chemical (St. Louis, MO) or Fluka Chemical (Ronkonkoma, NY). The IVM used in all the experiments was an essentially pure (>98%) form of 22,23-dihydroavermectin B1a. IVM and IVM-PO₄ were kindly provided by Drs. J. M. Schaeffer and A. Etter (Merck Research Labs, West Point, PA). IVM was dissolved in dimethylsulfoxide. Application of the vehicle alone induced no significant modification of the ACh-evoked currents. To minimize the possible contamination of endogenous Ca²⁺-activated Cl current, OR2-Ba²⁺ was used in most experiments as a bath solution (82.5 mM NaCl, 2.5 mM KCl, 5 mM HEPES, 2.5 mM BaCl₂, 1 mM MgCl₂, pH 7.4 adjusted with NaOH). All recording solutions were supplemented with 0.5 µM atropine to block possible endogenous muscarinic responses.

Mutagenesis. The chick 7 cDNA (Couturier et al., 1990) was subcloned as described previously (Revah et al., 1991) into the pBluescript KS⁺ vector to permit single-stranded DNA synthesis. Mutants were performed using an oligonucleotide-directed mutagenesis kit supplied by Amersham (Arlington Heights, IL).

Electrophysiological recordings. Perfusion solution was fed by gravity at a rate of ~6 ml/min. The oocytes were superfused continuously with OR2-Ba²⁺ (or OR2-Ca²⁺), and solution exchange was controlled by computer-driven electromagnetic valves (Type III; General Valve, Fairfield, NJ). Electrophysiological recordings were made 2-4 days after the injection using a two-electrode voltage-clamp (GENECLAMP amplifier; Axon Instruments, Foster City, CA). Electrodes made from borosilicate glass were filled with a filtered solution of 3 M KCl. Unless specified, the holding potential was 100 mV. All experiments were performed at 18°.

Cell line, culture, and recordings. Human embryonic kidney cells (293 cells) transfected with a plasmid containing the human 7 cDNA (K-28 cell line) were maintained in culture according to the method described previously (Gopalakrishnan et al., 1995). Cells were plated onto 35-mm Petri dishes 2-5 days before recording. During electrophysiological experiments, cells were placed in a medium containing 120 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 2 mM CaCl₂, and 10 mM HEPES, pH adjusted to 7.4 with NaOH. ACh-evoked currents were recorded using the whole-cell configuration of the patch-clamp technique. The intracellular solution contained 140 mM Cs-methanesulfonate, 5 mM NaCl, 2 mM MgCl₂, 10 mM BAPTA, and 10 mM HEPES, pH 7.4 adjusted with CsOH. In some experiments, 0.3 mM GTP (sodium salt) and 4 mM ATP (sodium salt) were added to the pipette medium as well as a regenerating system composed of 10 mM Na-phosphocreatine and 50 units/ml creatine phosphokinase. Pipettes were pulled from borosilicate glass and mounted on the head-stage of an AXOPATCH 200B amplifier (Axon Instruments). Drugs were applied using a liquid filament system based on a double barrel mounted on a piezoquartz actuator (Physics Instruments, Germany). In each side of the double barrel, a multitubing puffer (Bertrand et al., 1997) was inserted to allow combined applications and fast drug exchange (within a few msec).

Data analysis and computation. Data were captured on-line by an analog-to-digital converter (AT-MIO16; National Instruments, Austin, TX) and stored on the hard disk of a personal computer for later analysis. Mathematical analysis and curve fitting were performed on a Macintosh using personal software based on a mathematical interpreter. Curve fitting was done using a least-squares minimization algorithm (SIMPLEX). Dose-response curves were adjusted using the empirical Hill equation:

(1)

where y is the fraction of activated current, EC₅₀ is the concentration of half activation, n_H is the apparent cooperativity, and x is the agonist concentration.

(2)

(3)

	Results

Top Summary Introduction Materials & Methods Results Discussion References

IVM potentiates the ACh-evoked response of the chick and human 7 nAChRs. The application of IVM alone evoked no detectable current in 176 oocytes expressing the chick 7 nAChR (currents <3 nA, Fig. 1A), even when concentrations as high as 300 µM were applied for >30 sec (data not shown). Similarly, no detectable modification of the ACh-evoked current was observed when 30 µM IVM was coapplied with ACh (Fig. 1A). In contrast, a clear increase in the ACh response was noticed after a 16-sec preapplication of IVM (30 µM, Fig. 1A). Recordings of responses to different ACh concentrations at control and after IVM preapplication (30 µM, 16 sec) revealed that this treatment caused a large increase in the subsequent ACh responses (Fig. 1B). Applications of a low ACh concentration (3 µM) evoked no detectable current in 7-expressing oocytes; however, when the same test pulse was applied subsequently to IVM, this low ACh concentration evoked a significant inward current. This result suggests that in addition to an increase in the current amplitude, IVM caused a shift toward a lower concentration of ACh dose-response curve.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   IVM potentiates ACh-evoked current of the 7 nAChR. A (Left to right), IVM alone neither induces current nor affects the ACh-evoked current when coapplied with the agonist but strongly increases the subsequent ACh response if used as a preapplication (16 sec). Dashed line, IVM preapplications. B, Effects of IVM preapplication on currents evoked by different ACh concentrations. Thin lines, responses recorded in control conditions. Thick lines, responses obtained in the same cell after 16-sec IVM preapplication. C, Normalized dose-response curves recorded at control () and after IVM preapplication (). Data obtained from seven cells were normalized to unity for the saturating values obtained in control conditions. Lines through data points, Hill equations (eq. 1) with respective values of EC50 = 216 µM, nH = 1.3 (control) and EC50 = 11 µM and nH = 2.5 (after IVM). Dashed lines, fits obtained with a two-state allosteric model (eq. 3) with respective coefficients of L = 106, KA = 3.9 µM, and KB = 56.8 µM (control) and L = 350, KA = 3.9 µM, and KB = 56.8 µM (after IVM) (five binding sites under each condition). D, Desensitization time constants obtained in a cell over the entire dose-response relationship are plotted as a function of the agonist concentration. Time constants were deduced by curve fitting using eq. 2. , Values obtained at control. , Time constant measured during IVM potentiation. (Lines through data points were drawn to aid the eye.) Dashed lines, time constant for the respective EC50 values. , Mean desensitization time constants measured at 1 mM ACh (n = 16). Inset, normalized currents evoked by 1 mM ACh in control (thin line) and after IVM (thick line) illustrating the difference in time constant. Bars above current traces, ACh application. Potentiation was obtained with preapplications of 30 µM IVM for 16 sec.

View this table: [in this window] [in a new window]   TABLE 1 Apparent ACh sensitivities of the 7 nAChRs and mutants before and after ivermectin treatment Values in parenthesis correspond to the number of cells tested in each condition. All measurements were performed at 100 mV.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Potentiation of the human 7 nAChR. A, Currents evoked by an ACh concentration, near the EC50, showing a strong increase after IVM preapplication. Thin line, control. Thick line, obtained in the same cell after IVM preapplication (30 µM, 16 sec). B, Comparison of the potentiation of the ACh-evoked current (at 30 µM ACh) of the chick and human 7 nAChR. Ratios were computed as IIVM/IC, with IIVM and IC corresponding to the current values (in nA) recorded after IVM or at control, respectively. Mean ratios were 16.5 ± 9.5 (n = 24) for chick 7 and 26.5 ± 7.2 (n = 4) for human 7.

IVM potentiation is independent of nonspecific Cl channel activation. It is well documented that 7 nAChR is highly permeable to Ca²⁺ (for a review, see Bertrand and Changeux, 1995) and that increase in the free Ca²⁺ concentration in the cytoplasm of the oocyte may trigger the activation of endogenous Cl currents that are sensitive to Ca²⁺ but much less to Ba²⁺. Because IVM has been shown to activate Cl-permeable channels (Schönrock and Bormann, 1993), we investigated whether the increase in ACh-evoked currents results from a direct action of IVM on the 7 nAChR or from a possible increase in the Cl contamination.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   IVM potentiation is independent of Cl channels. A, I-V relationships were determined by measuring the ACh-evoked current (100 µM, 2 sec) at different holding potentials. Right, time courses of the currents. To enhance the difference between the ACh-evoked current reversal potentials and Cl equilibrium, Cl was partially substituted by gluconate. B, The 7 mutant E237A is strongly potentiated by IVM. Left, ACh dose-response relationship measured before and after IVM preapplication (same protocol as in Fig. 1C). Right, corresponding currents. , Control. , Currents recorded after IVM preapplication (1 µM, 16 sec).

Kinetics of IVM action. Preliminary experiments revealed that short exposure to IVM produces little or no potentiation of the subsequent ACh-evoked currents. To examine in more detail the relationship between the duration of IVM preapplication and the subsequent current potentiation, a fixed concentration of both IVM (30 µM) and ACh (60 µM) was used, and the duration of the IVM prepulse was increased gradually from 1 to 128 sec. After each test, cells were superfused with control solution until recovery of their initial response. Fig. 4A illustrates the results obtained with three cells. These data show that potentiation follows a sigmoidal relationship and becomes significant only for preapplications of >4 sec. Maximal potentiation corresponded to ~14-fold of the initial response (at 60 µM ACh) and was reached within 2 min. However, due to technical reasons, a 16-sec IVM preapplication usually was used. This caused an average 4.1 ± 1.47-fold potentiation (at 60 µM ACh, n = 10) of the initial ACh-evoked current. At higher IVM concentrations, effects could last as long as 20 min. For lower IVM concentrations, recovery of the response amplitude and time course was observed within a shorter time (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   IVM mode of action. A, Potentiation by IVM is time dependent. Peak currents evoked by a fixed ACh pulse (60 µM, 2 sec) are plotted as a function of the duration of IVM (30 µM) preapplication. Values are mean ± standard deviation from three cells. (Sigmoidal line through data points was added to support the eye.) B, Water-soluble form of IVM phosphate potentiates the ACh-evoked currents. Traces correspond to currents evoked at control (thin line) and after a 6-sec pulse of IVM-PO4 (thick line). Bars, ACh application.

IVM effects are not mediated by the membrane. Because of the lipophyllic character of IVM, it was suggested that this compound partitions in the lipid bilayer (Campbell, 1989) and, as a consequence, modifies the receptor environment. To test this possibility, we examined the effects of the water-soluble IVM-PO₄ on ACh-evoked currents. As shown in Fig. 4B, a 6-sec prepulse of IVM-PO₄ potentiates ACh-evoked currents. A reduction in desensitization comparable to that caused by IVM also was observed. Moreover, when an equivalent concentration and application time of IVM-PO₄ and IVM were used on the same cell, the water-soluble form caused a more important potentiation than the lipophyllic compound. These results are consistent with a direct action of this compound as an allosteric effector of the 7 nAChRs.

Effects of IVM on the response of 7 nAChR to a partial agonist. To evaluate further the possibility that IVM acts as an allosteric effector of the 7 neuronal nAChRs [e.g., by stabilizing its active (open) conformation], the effect of IVM on the dose-response curve of a partial agonist was investigated. We selected DMPP, which behaves as a partial agonist of the chick 7 nAChR, evoking ~1-10th of the maximal response to ACh (average of 13.2% in four cells) (Bertrand et al., 1992; Peng et al., 1994). Fig. 5A illustrates that after IVM preapplication, DMPP becomes almost a full agonist. The ratio of DMPP to ACh-evoked current of 65 ± 6.7% was measured in five cells. In agreement with the observations presented here, both ACh and DMPP responses showed a significantly slower time course after IVM exposure. The apparent affinity increased on average by only 1.7-fold (from 14.2 to 8.5 µM, n = 4) for DMPP (instead of 16-fold, as determined for ACh; Table 1) after preapplication of IVM. Similar potentiation of partial agonists by allosteric effectors has been reported extensively in the past with allosteric enzymes (Monod et al., 1965) and discussed in the framework of nAChR mutants (Galzi et al., 1996).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Pharmacological properties of the IVM-potentiated responses. A, IVM changes the pharmacological profile of DMPP. The chick 7 partial agonist DMPP efficiency is strongly increased after IVM (30 µM, 16 sec) preapplication, as shown from both the dose-response relationship (left) and individual currents (right). Dose responses were measured and plotted as for Fig. 1C. , , Control. , , Obtained after IVM preapplications. Lines through data, correspond to the empirical Hill equations with respective EC50 values and Hill coefficients of 110 µM and nH = 1.2 (ACh control), 10 µM and nH = 2.5 (ACh after IVM), 9 µM and nH = 2.7 (DMPP control), and 7 µM and nH = 3 (DMPP after IVM). B, Blockade of the 7-E237A receptor by the competitive inhibitor MLA. ACh-evoked currents (ACh = 60 µM) recorded in a medium containing 2.5 mM Ca2+ before and after IVM preapplication (30 µM, 8 sec) were recorded first at control and after a 10-sec exposure to 10 µM MLA (thick line). This MLA exposure is sufficient to fully inhibit ACh-evoked currents under both conditions. Parallel recovery of the responses and their potentiation always was observed. The cell was held at 100 mV throughout the experiment.

Effects of a competitive inhibitor on IVM potentiation. MLA is a high affinity blocking agent of the homomeric 7 nAChR that displays an IC₅₀ value for this compound in the picomolar range (Palma et al., 1996). Compelling evidence indicates that MLA competes with ACh and inhibits the response by stabilizing the receptor in a closed state (Wonnacott et al., 1993; Bertrand et al., 1997). To determine whether MLA interacts with IVM, potentiation experiments were conducted before, during, and after MLA blockade. Typical results obtained with the Ca²⁺-impermeant mutant E237A receptor are presented in Fig. 5B. These data illustrate that incubation in presence of a MLA concentration sufficiently high to suppress the ACh-evoked current also abolishes IVM potentiation (middle trace). Full recovery of both the ACh-evoked current and IVM potentiation were obtained readily after a 19-min wash (Fig. 5B, right traces). Similar data have been obtained with the wild-type receptor (data not shown).

Mutations in the channel domain alter IVM potentiation. The putative allosteric effects of IVM also were examined on chick 7 receptors mutated at two distinct locations (V251T and L247T) in the second transmembrane segment, TM2. As shown previously, these mutations produce pleiotropic effects on the receptor properties with an increase in apparent affinity, a loss of desensitization, and alteration of agonist-versus-antagonist profile (Revah et al., 1991; Devillers-Thiéry et al., 1992; Bertrand and Changeux, 1995). Furthermore, it was shown recently that in the L247T mutant, a fraction of the nAChR is spontaneously open (Bertrand et al., 1997). As illustrated in Fig. 6, IVM affected the two mutants expressed in X. laevis oocytes in a strikingly different manner. Preapplication of IVM on oocytes expressing the V251T mutant provoked a small (6.4 ± 2%, n = 7) but consistent increase in the leakage current in the absence of agonist. It is important to note that in every cell tested, this increased leak current is suppressed readily by the application of MLA (data not shown). In addition, however, a small but consistent reduction in the maximal ACh-evoked current was observed after IVM preapplication. Determination of the ACh dose-response relationship, over a broad range of agonist concentration, of this mutant before and after IVM exposure revealed a significant increase in the apparent affinity of ~5-fold (Fig. 6A, left, and Table 1). No significant modification of the time course of the ACh-evoked current was noticed after IVM exposure (Fig. 6A, right). As in the case of the wild-type receptor (Fig. 1C), dose-response curves can be fitted with the two-state allosteric model (eq. 3). A multiplying factor that was inferior to unity was, however, introduced to describe the smaller amplitude of responses recorded after IVM exposure. In contrast, when the experiments were repeated in oocytes expressing the L247T mutant, a different pattern was observed (Fig. 6B). First, IVM strongly reduced the amplitude of the ACh-evoked current at every concentration tested; second, the apparent affinity for ACh remained unchanged (Fig. 6B, left, and Table 1). Also, IVM exposure did not significantly increased the leak current. Furthermore, as in the case of the V251T, no modification of the response time course was detected for the L247T mutant (Fig. 6B, right).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   IVM potentiation distinguishes TM2 mutants. Effects of IVM preapplication on oocytes expressing the 7-V251T (A) and 7-L247T (B) mutants. Dose-response curves were obtained using the same protocol presented in Fig. 1C. Lines through data points, correspond to the Hill equations with respective EC50 values and Hill coefficients of 0.7 µM and nH = 2.4 (7-V251T control, n = 6) 0.08 µM and nH = 1.8 (7-V251T same cells, after IVM) 1.55 µM and nH = 1.8 (7-L247T control, n = 6) 1.55 µM and nH = 1.8 (7-L247T same cells, after IVM). Dashed lines, fits obtained with a two-state allosteric model (eq. 3) with respective coefficients L = 1800, KA = 0.14 µM, and KB = 3.1 µM (control) (five binding sites), whereas L = 12 (after IVM) for the V251T. Values of L = 30, KA = 0.8 µM, and KB = 3.3 µM (five binding sites) were used for the L247T both at control and after IVM. A scaling factor of 0.3 was added to take into account the lower current amplitude of L247T mutant after IVM. Right, individual currents recorded before and after IVM. Thin lines, control. Thick lines, recordings obtained after IVM preapplication (30 µM, 15 sec for 7-V251T; 10 sec for 7-L247T).

IVM potentiates responses of the human 7 nAChR expressed in human embryonic kidney 293 cells. To further evaluate the time course of IVM action and rule out effects specific to the oocyte system, we examined the effects of IVM on K-28 cells expressing the human 7 nAChR (Gopalakrishnan et al., 1995). The application of IVM alone induced no detectable current in every cell tested (n = 15). Traces presented in Fig. 7 illustrate that preapplication of 25 µM IVM elicited a significant increase in the current evoked by low ACh concentration (30 µM). Although a clear potentiation already was observed after a 1-sec exposure, the maximal effect was obtained with 5-10-sec preapplications. Complete recovery of both amplitude and time course of the ACh-evoked currents was observed within 2-5 min, depending on the duration of the IVM prepulse. Moreover, data presented in Fig. 7 show that potentiation and recovery can be repeated several times in a same cell.

View larger version (8K): [in this window] [in a new window]   Fig. 7.   Time course of IVM potentiation and recovery in K-28 Cells. ACh-evoked currents measured in whole-cell configuration were recorded at regular intervals. The cell was maintained in voltage clamp at 100 mV and challenged with 200-msec pulses of 30 µM ACh (small horizontal bars). The application of a short IVM pulse (25 µM, arrows) induced no current deflection (not shown) but provoked a strong potentiation of the ACh-evoked currents. Complete and readily reversible, potentiation could be induced by successive IVM prepulses. Left to right (arrows), IVM pulse durations of 30-, 20-, and 5-sec, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   IVM and IVM-PO4 potentiate the human 7 nAChR in K-28 cells. A, Apparent agonist affinity of a single cell was determined using 200-msec pulses of ACh concentrations applied in a growing order first in control solution and then in the presence of 25 µM IVM prepulse. Fit of the data with the empirical Hill equation yielded an apparent affinity of 120 µM and Hill coefficient of 1.6. In the presence of IVM prepulses, these values shift to 30 µM and 1.8, respectively. B, Currents evoked by increasing ACh concentrations (3, 10, 30, 100, 300, and 1000 µM; horizontal bars) under control conditions and after IVM prepulses (same as in A) are superimposed. C, Mean potentiations induced by IVM and IVM-PO4 were 3.46 ± 0.53 and 4.75 ± 1.07, respectively. IVM prepulse was ~10 sec. D, Currents evoked by a 30 µM ACh concentration (horizontal bar) at control and after a 10-sec prepulse of IVM (or IVM-PO4).

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

The data presented in this work demonstrate that the anthelmintic drug IVM strongly potentiates the ACh-evoked current of the 7 homomeric neuronal nAChRs from both chick and human.

IVM is a semisynthetic analog of the natural compound avermectin that contains 80% 22,23-dihydroavermectin B1a and 20% of the B1b homologue (Campbell, 1989). Avermectin was isolated ~20 years ago from the soil bacterium Streptomyces avermitilis in a large scanning of natural substances with anthelmintic activity (Burg et al., 1979). In humans, IVM has been considered the drug of choice for the treatment of river blindness (for a review, see Ottesen and Campbell, 1994). River blindness (Onchocerciasis), which affects ~21 millions persons in the world, is caused by infection with the filarial nematode Onchocerca volvulus (Mahmoud, 1995). A single oral administration of IVM in a standard dose of 150 µg/kg of body weight has been shown to rapidly eliminate microfilariae (Pacqué et al., 1991) with a very low rate of adverse reactions. The mechanism of action of IVM remains to be elucidated.

Recently, an avermectin-sensitive glutamate-gated Cl channel displaying an important structural homology with the nAChR subunits was identified and cloned from Caenorhabditis elegans (Cully et al., 1994). In addition, IVM has been reported to potentiate, in an allosteric manner, ligand-gated channels activated by GABA (Sigel and Baur, 1987; Krusek and Zemková, 1994). These observations prompted us to investigate the effects of IVM on neuronal nAChRs using the functional 7 neuronal nAChR.

We found that a preapplication of micromolar concentrations of IVM markedly increased the current evoked by a subsequent application of ACh, thereby causing a 20-fold shift of the apparent affinity of ACh. Concomitantly, in every cell tested, the Hill coefficient of the dose-response curve increased. Similar IVM sensitivity and potentiation of the ACh-evoked current were observed in receptors reconstituted with either the human or chick 7 nAChR subunits.

None of the IVM effects can be accounted for by an indirect activation of Ca²⁺-activated Cl currents. Furthermore, given the important potentiation of the water-soluble form of IVM (IVM-PO₄), it is unlikely that IVM effects result from perturbation of the lipid bilayer organization of the membrane.

Based on the repeatability of the IVM potentiation observed on the 7 ACh-evoked currents both in X. laevis oocytes and in K-28 cells, it seems unlikely that IVM potentiation can be explained by the incorporation in the plasma membrane of new nAChR proteins from an internal store. As a correlate, this hypothesis would include the assumption that such a freshly expressed pool must be withdrawn from the plasma membrane every time IVM is removed. Finally, this hypothesis could not explain either the modification of the kinetics of the ACh-evoked responses or changes in the pharmacological profile.

The most likely interpretation of the presented results is that IVM acts on the 7 nAChR as a positive allosteric effector of the neuronal nAChR. This interpretation is reinforced by the observation that a modification of the pharmacological profile accompanies IVM potentiation. For example, DMPP, which behaves as a partial agonist of the chick 7 nAChR (Bertrand et al., 1992; Peng et al., 1994), becomes almost a full agonist after IVM preapplication, as anticipated for a positive allosteric effector that would stabilize the active open state. A similar conclusion was reached in previous attempts to model the kinetic properties of the 7 nAChR mutants on the basis of allosteric mechanisms (Galzi et al., 1996b).

Moreover, site-directed mutagenesis experiments carried out within the TM2 channel domain previously revealed that substitutions of Leu247 or Val251 by a threonine profoundly modify 7 nAChR properties (Revah et al., 1991; Devillers-Thiéry et al., 1992). Both mutations cause an increase in the apparent affinity to ACh and a loss of desensitization. They have been interpreted in terms of differential alteration of the states and conformational equilibria of the nAChR: a selective permeabilization of the desensitized state for L247T and a modification of the equilibrium constant L between basal and active states for V251T (Galzi et al., 1996b). The positions of these two mutations within the channel domain are shown in Fig. 9A.

View larger version (27K): [in this window] [in a new window]   Fig. 9.   Schematic of the nAChR and allosteric reactions. A, Side view of a schematic representation of the nAChR inserted in the membrane. Right, putative protein arrangement of one of the ACh binding site and the channel pore. Main and complementary loops participating in the formation of the ACh binding site are illustrated. Channel inset, point mutations (V251T and L247T). B, Allosteric reaction scheme including four states is shown where B is basal (closed) state, A is active (open) state, I is intermediate (closed) state, and D is desensitized closed state. L1 and L2 are the respective isomerization constants for the transitions from the B-A and/or B-D states. C, Dose-response curves computed from a two-state allosteric model (eq. 3) corresponding to the activation from B to A with different L values comparable to those predicted for (right to left) the wild-type and V251T and V251T after IVM preapplication. KA = 0.14 µM, KB = 3.1 µM, and five binding sites.

Consistent with the notion that IVM potentiates nicotinic receptor responses by shifting the conformational equilibrium in favor of the active state, it still potentiates the V251T mutant but no longer affects the L247T mutant (Fig. 9B). Furthermore, in oocytes expressing the V251T mutant, the current leakage is increased by IVM, and this current can be blocked by MLA. This indicates that on the V251T mutant, IVM behaves as a positive allosteric effector.

These experimental data can be interpreted in terms of a minimal two-state allosteric model (Monod et al., 1965). The equilibrium constant (L value) describing the equilibrium between the basal (B) and active (A) states can be set to 12 (for the V251T, see Fig. 6) in the presence of IVM to account for ~8% of the maximal ACh-evoked current. Fig. 9C shows that changing only the L value may result in a slight agonistic effect of IVM on the V251T. Indeed, a high value of L, as may be the case for the wild-type receptor, would lead to a low apparent affinity receptor with no spontaneous or IVM-evoked, channel opening. The V251T mutation then would lead to a significant reduction in L (~1000), thus leading to an increased apparent affinity for agonist together with higher cooperativity and higher response amplitude (Galzi et al., 1996b). The action of a positive allosteric effector such as IVM, according to Rubin and Changeux (1966), would correspond to a reduction in the L value and lead to detectable leak currents in the absence of acetylcholine and higher apparent affinity and response cooperativity in the presence of acetylcholine.

Such an interpretation also is in agreement with the data obtained on the L247T mutant receptor. Indeed, according to the interpretation proposed previously (Revah et al., 1991; Galzi et al., 1996b), the channel of the L247T mutant would become open in the desensitized conformation, and the equilibrium constants between the conformational states would not be altered. Thus, if IVM reduces the equilibrium constant between the basal (B) and active (A) states, its effect would become detectable only at the high agonist concentrations required to stabilize the active state and should not be observed at the concentrations that stabilize the desensitized but conducting state.

These data therefore document and demonstrate further that the V251T and L247T mutations differentially alter the properties of the allosteric states and conformational equilibria of the 7 nAChR, and, as a consequence, their responsiveness to IVM potentiation.

The high affinity displayed by the 7 nAChR for IVM and its water-soluble form IVM-PO₄ supports the view that this receptor carries a specific site that recognizes this type of pharmacological ligand. Experiments performed in a normal versus a low Ca²⁺ concentration showed that IVM causes a similar increase in apparent affinity under the two conditions (data not shown). Therefore, it seems unlikely that IVM enhances nAChR responses through the Ca²⁺ binding sites. Alternatively, IVM and steroids might modulate receptor activity through the same or overlapping sites. However, steroids have been shown to inhibit rather than to activate the 7 nAChR (Buisson and Bertrand, 1998), leaving the possibility that these two sites are distinct.

The slow time course observed for both the onset and recovery of IVM effects on X. laevis oocytes can reflect either a slow kinetic of action or a restriction of diffusion of this compound across the vitelline membrane. The time course of potentiation and recovery observed in K-28 cells suggests that IVM diffusion might be a limiting factor in the oocyte system.

The existence of a specific site allowing the binding of an allosteric effector does not necessarily imply the existence of an endogenous natural ligand, as shown for the bezafibrate site on hemoglobin (Perutz et al., 1986). Thus, although IVM may interact with a specific domain of the protein, this drug may cause an allosteric potentiation comparable to that produced by a physiological but yet unknown compound that binds to another site.

At this stage, the possibility that IVM indirectly potentiates nAChRs through the activation of a second messenger system cannot be excluded totally. Experiments with K-28 cells expressing the human 7 nAChR (Gopalakrishnan et al., 1995) revealed that IVM potentiation is not restricted to X. laevis oocyte but can be observed equally in another preparation. Thus, if IVM is acting through a specific receptor in the cell membrane, such a receptor should be expressed by cells as different as the oocyte and human embryonic kidney cells. Moreover, the effect of IVM on the human 7 nAChR expressed in K-28 cells was not modified by the addition of GTP plus an ATP-regenerating system into the pipette solution (see Materials and Methods). Potentiation also was independent of the addition of the chelating agent BAPTA in the intracellular medium. Thus, it is very unlikely that IVM effects are mediated by a second messenger system such as intracellular Ca²⁺ signaling, G protein pathway, or a kinase/phosphatase pathway.

Given the slow time course of agonist application in the oocytes and the fast desensitization of the 7 nAChR, it can be argued that partial masking of the ACh-evoked response occurs in this system. However, this interpretation does not apply for experiments with the K-28 cells, in which agonist application is performed within milliseconds. Furthermore, although this hypothesis could have taken into account the current increase observed at a high ACh concentration, it does not explain adequately the important shift in apparent ACh affinity observed for low nondesensitizing agonist concentrations. Finally, a strong argument in favor of the allosteric modulation is the modification of the pharmacological profile observed for the partial agonist DMPP, which cannot be explained on the basis of desensitization.

The strong potentiation caused by IVM on 7 nAChR seems to be reminiscent of the allosteric modulation of GABA_AR by benzodiazepines or barbiturates (for a review, see Rabow et al., 1995). Benzodiazepines were shown to modify the probability of channel opening, whereas barbiturates increased the mean open time (Puia et al., 1990). From these observations, it was concluded that benzodiazepines act by modifying the allosteric isomerization coefficient L, whereas barbiturates are thought to alter the transition from liganded closed state to the open state (Rabow et al., 1995). In this context, the effects caused by IVM resemble those of benzodiazepines on the GABA_AR and are consistent with a reduction in the L coefficient.

In conclusion, we propose that IVM behaves as an allosteric effector that binds to a specific site on the 7 nAChR that is distinct from either the Ca²⁺ (or steroid) regulatory sites. Studies of pharmacological agents that act as allosteric modulators of the nAChRs may lead to the design of new types of compounds that could compensate the deleterious effects of neurodegenerative disorders such as Alzheimer's or Parkinson's diseases.