(cid:1) -Amyloid Peptide Activates (cid:2) 7 Nicotinic Acetylcholine Receptors Expressed in Xenopus Oocytes*

The (cid:2) 7 nicotinic acetylcholine receptor is highly expressed in hippocampus and in cholinergic projection neurons from the basal forebrain, structures that are particularly vulnerable to the ravages of Alzheimer’s disease. Previous work suggests that (cid:1) -amyloid peptide can interact with (cid:2) 7 nicotinic acetylcholine receptors, although the nature of this interaction has not been well characterized. To test whether (cid:1) -amyloid peptide can activate (cid:2) 7 nicotinic acetylcholine receptors, we expressed these receptors in Xenopus oocytes and performed two-electrode voltage clamp recordings, charac-terizing the response to (cid:1) -amyloid peptide 1–42 applied at concentrations ranging from 1 p M to 100 n M . In (cid:2) 7-expressing oocytes, (cid:1) -amyloid peptide 1–42 elicits inward currents at low concentrations (1–100 p M ), whereas at higher concentrations (n M ), less effective receptor activation is observed, indicative of receptor desensitization. Preincubation with the (cid:2) 7-selective agents, the antagonist methyllycaconatine, and the agonist 4-OH-GTS-21 blocked (cid:1) -amyloid peptide-induced receptor activation. (cid:1) -amyloid peptide 1–42 at low concentrations was able to activate the L250T mutant (cid:2) 7 receptor. The endogenous Ca 2 (cid:3) -activated chloride current

Alzheimer's Disease (AD) 1 is the most common of the senile dementias, the prevalence of which is increasing rapidly with a projected 14 million affected worldwide by 2025. Early on, AD presents clinically as impaired memory formation, yet despite intensive study, the mechanisms underlying AD-related memory dysfunction remain mysterious. Familial AD is associated with several risk factors, the best correlated being age and the inheritance of specific genes (mutations or allele type) that predominantly result in increased ␤-amyloid peptide (A␤) levels (1,2,33).
Although these peptides are present in the brains and cerebrospinal fluid of normal subjects at the picomolar level, sub-stantial evidence indicates that elevated A␤ is a culprit in the cognitive decline of AD (1). A␤ is generated from the amyloid precursor protein through endoproteolytic cleavage by ␤and ␥-secretases (2). In normal individuals, A␤-(XϪ40) (A␤40) comprises the majority of the A␤ population; a far smaller fraction is made up of A␤42 (1). A␤42 is highly fibrillogenic and exhibits trophic and toxic effects on neurons (3)(4)(5). The hippocampus is a locus for the earliest detected cognitive dysfunction in AD: impairment in the encoding of new episodic memories is typical of the earliest stages of AD, and the loss of episodic memory in AD is linked to medial temporal pathology inclusive of the hippocampus (6 -8). Despite intensive study, the mechanism by which elevated A␤ leads to AD-related hippocampal dysfunction remains mysterious, not to mention the lack of an understanding of the normal physiologic role for A␤ in synaptic function and signal transduction.
As yet, a receptor for A␤ that is capable of influencing synaptic plasticity in the hippocampus remains unidentified. Cholinergic connections between the hippocampus and cortical structures within the temporal lobe and the cortical cholinergic system that originates within the basal forebrain are selectively vulnerable in the course of AD (9 -11). The loss of proper functioning of these neuronal populations is thought to underlie the loss of memory in AD patients (12,13). The ␣7 nicotinic acetylcholine receptor (nAChR) is highly expressed on neurons of hippocampus and cholinergic projection neurons from the basal forebrain. A number of recent studies have convincingly demonstrated an interaction between the ␣7 nAChR and A␤ in vitro and on neurons. For instance, it has been shown that A␤42 co-immunoprecipitates with the ␣7 nAChR in samples from postmortem AD hippocampus and that ␣7 nAChR antagonists compete for A␤42 binding to heterologously expressed ␣7 nAChRs (14). Furthermore, preincubation with A␤42 antagonizes the activation of ␣7 nAChR-like currents in hippocampal neurons, and A␤42, acting through ␣7 nAChRs, can elicit extracellular signal-regulated kinase (ERK) MAPK activation in hippocampal cultures (15,16). This last observation is likely triggered by Ca 2ϩ influx; ␣7 nAChRs are highly permeable to this pluripotent second messenger (17).
We tested the hypothesis that A␤ could directly activate the ␣7 nAChR by expressing these receptors in Xenopus oocytes and performing two-electrode voltage clamp recordings following perfusion with pM to nM concentrations of non-aggregate A␤. Both A␤40 and A␤42 were capable of eliciting inward currents from ␣7 nAChR-expressing oocytes; however, the current amplitudes resulting from A␤42 receptor activation were much larger than those generated by A␤40. Inward currents elicited by A␤42 were comprised of Ca 2ϩ , were blocked by the ␣7 nAChR-selective antagonist methyllycaconatine (MLA), and were cross-desensitized by the ␣7 nAChR-selective agent 4OH-GTS-21. We obtained evidence that low concentrations of A␤42 are highly desensitizing; wash times up to 30 min were unsuc-cessful at allowing subsequent A␤42 activation. Desensitization induced by low concentrations of A␤42 did not have an observable effect on subsequent nicotine stimulation. However, prolonged exposure to a high concentration of A␤42 led to cross-desensitization of nicotine responses, suggesting the existence of more than one A␤42 binding site or an A␤42-induced conformation that differentially interferes with nicotine activation of ␣7 nAChRs. Overall, we conclude that A␤42 is a high affinity ligand for ␣7 nAChRs that is capable of gating Ca 2ϩ flux through the channel.

EXPERIMENTAL PROCEDURES
Oocyte Expression-Xenopus leavis (Nasco) oocytes were harvested and prepared for injection as described (17). Rat ␣7 nAChR cDNA or rat ␣7 nAChR cDNA containing the point mutation L250T, both contained in the cytomegalovirus expression vector pcDNAI/Amp (Invitrogen), were utilized for electrophysiology. ␣7 nAChR cDNAs were obtained from the laboratory of Dr. Jim Patrick.
Two-electrode voltage clamp was performed at ambient temperature utilizing an Axoclamp-2A amplifier. Current and voltage pipettes were filled with 3 M KCl, and resistances ranged from 0.1 to 0.5 megaohms. Oocytes were voltage-clamped at Ϫ60 mV unless otherwise stated. Oocytes with resting membrane potentials betweenϪ20 and Ϫ60 mV and voltage clamp injections of less than 100 nA were used. Solution exchange was achieved by using electromagnetic valves (type 1; General Valve Corp.). Data were collected at a rate of 1.667 and 5 kHz for 30-and 5-s drug applications, respectively. Data acquisition and analysis was performed with pCLAMP software (Axon Instruments, Inc.). Current amplitudes were determined by measuring the maximum negative deflection from baseline. Data normalized and reported as percent of nicotine response correct for variations in channel expression by individual oocytes. We observed variability in the nicotine response, a portion of which could have been due to the previously reported increase in ␣7 nAChR responses following the initial application of agonist (20); therefore oocytes received at least two control applications of nicotine separated by 5 min at the start of recording with the average of those responses used for normalization. Oocytes injected with wild-type ␣7 nAChR cDNA were tested for positive expression by performing a test perfusion with 100 M nicotine. Oocytes injected with L250T mutant ␣7 nAChR cDNA were tested for positive expression by performing a test perfusion with 1 or 10 M nicotine. Oocytes displaying inward currents of at least 100 nA in response to nicotine were used in experiments.

RESULTS
Our work utilized rat A␤42 peptide that was prepared under conditions to retard aggregation and fibrillization (21,22). We tested our A␤42 preparations for aggregation in a Congo red assay. At the concentrations used and under the incubation conditions tested, none of the A␤42 preparations significantly pelleted out of solution with Congo red dye at 2.5 or 25 M (data not shown (23)).
A␤ Activates ␣7 nAChRs-The ability of low concentrations of non-fibrillar A␤42 to activate ␣7 nAChRs was tested by expressing these receptors in the Xenopus oocyte expression system and performing two-electrode voltage clamp recordings. Oocytes positive for ␣7 nAChR expression, as determined with a 5-s perfusion of 100 M nicotine, were washed for 5 min with Ringer's solution and then subjected to a 5-s perfusion with 10 pM A␤42. This concentration of A␤42 led to inward currents that averaged 44 Ϯ 12 nA (Ϫ60-mV holding potential, n ϭ 8) or 176 Ϯ 76 nA (Ϫ100-mV holding potential, n ϭ 4) in amplitude (Fig. 1a). Prior activation with nicotine was not necessary to observe A␤42-induced responses. We observed receptor activation following A␤42 perfusion on oocytes that had not yet been exposed to any other ␣7 nAChR agonist. Furthermore, we did not observe measurable current responses when mock-injected oocytes were subjected to A␤42 perfusion (n ϭ 12, data not shown). We also tested whether the far more abundant form of amyloid peptide, A␤40, was capable of activating ␣7 nAChRs expressed by Xenopus oocytes. 10 pM human A␤40 generated inward currents from ␣7 nAChR-expressing oocytes. However, the peak current amplitudes were much smaller than an equivalent concentration of A␤42; the normalized A␤40 currents were less than 20% of those generated with A␤42. 10 pM A␤40 elicited 10.13 Ϯ 1.88 nA peak current, which is 1.75 Ϯ 0.64% of the 100 M nicotine response (n ϭ 8, data not shown). Finally, the reverse (human 40 -1) peptide failed to induce ␣7 nAChR activation (n ϭ 10, data not shown). These experiments demonstrate that A␤42, as well as A␤40, is a selective and high affinity agonist for ␣7 nAChRs expressed in oocytes.

FIG. 1. Low concentrations of A␤42 desensitize A␤ response; high concentrations of A␤42 desensitize nicotine (nic) response.
Xenopus oocytes held at Ϫ100 mV were assayed for ␣7 nAChR expression with 100 M nicotine, washed for 5 min with Ringer's solution, and then perfused for 5 s with 10 pM A␤42 (a). Scale bars ϭ 200 nA ϫ 1 s. At this holding potential, peak inward current amplitude averaged 176 Ϯ 76 nA, n ϭ 4. Xenopus oocytes expressing ␣7 nAChRs respond to the first application of 10 pM A␤42 but not to a second even with up to a 30-min wash with Ringer's solution (b). Low concentrations of A␤42 do not interfere with nicotine activation of the receptors. Rapid switch to nicotine after A␤ desensitization yields a nicotine-induced inward current of equivalent magnitude to the test perfusion with nicotine. Scale bars ϭ 500 nA ϫ 1 s and 100 nA ϫ 1 s, respectively. A 2-min preincubation with high concentrations of A␤42 (100 nM) cross-desensitizes ␣7 nAChRs to nicotine (c). A␤42 inhibited 53.2 Ϯ 4.9% of the nicotine response; n ϭ 11. A 5-min wash with Ringer's solution leads to recovery of the nicotine response. Scale bars ϭ 100 nA ϫ 1 s.
Next, we evaluated whether we could generate ␣7 nAChR activation with repeated application of A␤42. Repeated activation of ␣7 nAChRs with nicotine could be achieved by performing 5-min washes between drug perfusions. In contrast, following a 5-s perfusion with 10 pM A␤42 that yielded a current response, we were unable to measure a second A␤42 response (Fig. 1b). This was the case for wash times of up to 30 min. Interestingly, at this low concentration of A␤42, loss of the A␤ response did not interfere with the nicotine response. Perfusion with 100 M nicotine within 1 min of perfusion with the second application of A␤42 resulted in a nicotine response that was indistinguishable from the nicotine test application (n ϭ 5; Fig.  1b). Clearly, the activation properties of low concentrations of A␤42 and of 100 M nicotine on ␣7 nAChRs are different.
High Concentration of A␤42 Cross-desensitizes ␣7 nAChRs-Previous reports have demonstrated A␤42 antagonism of ␣7 nAChRs expressed by neurons (15,16). These studies utilized high concentrations (100 nM-1 M) of A␤42 and tested ␣7 nAChR responses following preincubation with the peptide. We therefore attempted to cross-desensitize the nicotine response of ␣7 nAChRs expressed by oocytes with 100 nM A␤42. A 2-min preincubation (co-application of nicotine and A␤42 did not antagonize the receptor response) with 100 nM A␤42 inhibited the responses of the receptors to subsequent 100 M nicotine application (53.2 Ϯ 4.9%; n ϭ 11; Fig. 1c). These results are consistent with the published reports that 100 nM and 1 M A␤42 antagonize ϳ60 -80% of neuronal ␣7 nAChR responses to agonist (15,16).
␣7 nAChR-specific Agents Block Receptor Activation by A␤42-We next demonstrated that the A␤42 effect could be blocked by the ␣7 nAChR-selective antagonist, MLA. Fig. 2a shows a representative experiment from nine replications in which MLA blocked both the nicotine and 10 pM A␤42 response from an oocyte expressing ␣7 nAChRs. Recovery of the nicotine response follows a 10-min washout with Ringer's solution (response to a second application of A␤42 was not attempted, but see the following GTS results). Likewise, preincubation (5 min) with the ␣7 nAChR-selective agent 4-OH-GTS-21 (3 or 30 M) cross-desensitized the ability of A␤42 to activate ␣7 nAChRs on oocytes (representative of n ϭ 13; Fig. 2b). Following positive responses to first nicotine and then to 4-OH-GTS-21, a 5-min perfusion with 4-OH-GTS-21 prevented subsequent activation of ␣7 nAChRs with 10 pM A␤42. A 10-min washout with Ringer's solution restored the A␤42-induced current. 4-OH-GTS-21 cross-desensitized nicotine responses as well (data not shown). The fact that a response to a second application of A␤42 was observed after washout suggests that 4-OH-GTS-21 binding prevents A␤42 binding and that the two binding sites overlap. These experiments further demonstrate that A␤42 is a selective agonist for ␣7 nAChRs expressed in oocytes.
A␤42 Activates the L250T ␣7 nAChR-We utilized the desensitization-resistant mutant L250T ␣7 nAChR to exploit the large amplitude currents generated by this receptor type (30,31). Following a test application of 1 M nicotine and a 5-min wash with Ringer's solution, 10 pm A␤42 was perfused for 30 s, which resulted in a substantial inward current (1606 Ϯ 280 nA, n ϭ 28; Fig. 3a). Shorter drug application times (5 s) yielded similarly shaped current responses (Fig. 3b). These data indicate that, as is true of the wild-type receptor, the L250T mutant version of the ␣7 nAChR binds and is activated by A␤42. Given the marked difference in the shape of the nicotine-induced currents versus the A␤42-induced currents, the data also suggest that A␤ is more effective at eliciting desensitization of the L250T mutant receptors than is nicotine.
Unlike wild-type ␣7 nAChRs, we were able to generate repeated responses to 10 pM A␤42 from the L250T mutant receptor following a 10-min wash between A␤ applications (Fig. 3b).
It was not the case, however, that A␤ was unable to desensitize the receptor. Repetitive application of 10 pM A␤42 (a 5-s A␤ perfusion every 1 min) induced current responses of decreasing magnitude until current responses were diminished to less than 5% of the initial current (data not shown). Under these conditions, short wash times (1 min) after A␤42 perfusion led to nicotine responses indistinguishable from the initial nicotine test response. Thus, analogous to wild-type ␣7 nAChRs, a low concentration of A␤42 (10 pM) has a desensitizing effect on L250T nAChRs that does not interfere with nicotine responses.
Low Concentrations of A␤42 Activate ␣7 nAChRs, High Concentrations Desensitize ␣7 nAChRs-We tested a range of A␤42 concentrations on both wild-type and L250T ␣7 nAChRs expressed in Xenopus oocytes. Fig. 4 shows concentration response results for concentrations of A␤42 ranging from 100 fM to 100 nM (also shown in Table I). Noteworthy is the effect of high concentrations of A␤42 on receptor responses. A␤42 at nM concentrations is inhibitory to both the wild-type (Fig. 4a) and the L250 mutant (Fig. 4b) ␣7 nAChR expressed in Xenopus oocytes. In fact, the 100 nM effect on wild-type ␣7 nAChRs is not statistically significant from zero current amplitude (p ϭ 0.08, one-tailed Student's t test). Consistent with high concentrations of A␤42 inducing receptor desensitization, the concentration response relationship forms an inverted "U" shape for both wild-type and L250T ␣7 nAChRs. In fact, as the A␤42 concen- tration increases, the number of oocytes that fail to respond to A␤42 increases although they proved positive for ␣7 nAChR expression (Fig. 4c). In other words, with increasing A␤42 concentration, a larger proportion of oocytes responsive to nicotine fail to respond to A␤42 perfusion.
We also evaluated various concentrations of human A␤40 for its ability to activate the L250T ␣7 nAChR. As was observed for wild-type ␣7 nAChRs, A␤40 activated L250T receptors to a much lesser extent than an equivalent concentration of A␤42. Table II lists peak current amplitudes for 10 pM, 1 nM, and 100 nM A␤40 application to L250T ␣7 nAChRs expressed in oocytes. Crude analysis indicates that A␤40 is less efficacious than A␤42 at receptor activation.
A␤42 Activation of ␣7 nAChRs Leads to Ca 2ϩ Influx-Xenopus oocytes endogenously express a Ca 2ϩ -activated chloride channel that produces a net inward current when activated. ␣7 nAChRs are highly permeable to Ca 2ϩ (17) such that when expressed in Xenopus oocytes, receptor activation leads to activation of the endogenous Ca 2ϩ -activated chloride channel. Replacement of Ca 2ϩ ions with Ba 2ϩ ions in the Ringer's solution greatly reduces ␣7-nAChR generated currents because the outward chloride current no longer contributes to the whole cell current. We tested the possibility that A␤42 activation of ␣7 nAChRs leads to Ca 2ϩ influx by performing recordings of the L250T mutant ␣7 nAChR following agonist applications in the presence and absence of Ca 2ϩ in the perfusion buffer. Fig. 4d illustrates that, in the presence of Ca 2ϩ , L250T ␣7 nAChRs exhibit large inward currents in response to both nicotine and A␤42. In the absence of Ca 2ϩ , L250T ␣7 nAChRs have reduced nicotine and A␤42 responses, yet the current amplitudes are still rather large. In Ca 2ϩ -free Ringer's solution, the nicotine and A␤42 responses were 44.65 Ϯ 8.30% and 43.62 Ϯ 9.06%, respectively, of those in Ca 2ϩ Ringer's solution (n ϭ 7). These results suggest that: 1) currents induced by either nicotine or A␤42 are significantly contributed to by the Ca 2ϩ -activated chloride channel; 2) Ca 2ϩ influx occurs with both nicotine and A␤42 activation of ␣7 nAChRs; and 3) in the absence of Ca 2ϩ , ion flux still occurs through these receptors. DISCUSSION We have demonstrated that rat A␤42 peptide prepared in non-aggregate form directly activates rat ␣7 nAChRs expressed by Xenopus oocytes. This is a high affinity interaction; concentrations as low as 100 fM A␤42 are capable of inducing inward currents. A␤42 is competitively antagonized by MLA and crossdesensitized by 4-OH-GTS-21, both ␣7 nAChR-selective agents.
On wild-type receptors, the range of effective A␤42 doses extends over the amount of soluble A␤ that occurs in normal (pM) as well as in AD (pM-nM) brain (1). This suggests the possibility of an A␤-␣7 nAChR interaction under normal physiologic conditions, and thus we propose that A␤ may be an endogenous ligand for ␣7 nAChRs. Furthermore, our data suggest that part of the pathology elicited by A␤ in AD may be due to aberrant activation of ␣7 nAChRs. In particular, as the ␣7 nAChR is a ligand-gated ion channel highly permeable to Ca 2ϩ (17), chronic activation of ␣7 nAChRs in AD could lead to dysregulation of Ca 2ϩ homeostasis and provide a molecular mechanism for the cholinergic dysfunction that is a hallmark of AD (9 -11).
The more prevalent form of amyloid peptide in brain, A␤40, was also able to activate ␣7 nAChRs. However, A␤40 was less effective than an equivalent concentration of A␤42 at activating the wild-type and L250T mutant ␣7 nAChRs. It should be noted, however, that human and rat A␤ sequences differ by three amino acids; the efficacy of human A␤40 might be diminished relative to rat A␤40 on rat ␣7 nAChRs. Whether there is a marked difference in the efficacy of rat versus human A␤ peptides at ␣7 nAChRs is unknown at this time. Nonetheless, our data suggest that A␤40, as well as A␤42, is capable of activating a7 nAChRs in situ.
One model we have proposed previously and that is consistent with our current findings is that hippocampus-dependent learning and memory impairments in early AD arise in part because of the increased A␤ burden and chronic activation of the ERK MAPK cascade in the hippocampus through ␣7 nAChRs (19). In support of this, we have demonstrated that elevation of A␤ in vivo using an animal model for AD (Tg2576 (24)) leads to the up-regulation of hippocampal ␣7 nAChR protein (19). ␣7 nAChR up-regulation in the hippocampus of Tg2576 animals is coincident with the manifestation of a contextual fear learning deficit (34), a hippocampus-dependent associative learning paradigm (25,26). Furthermore, increased ␣7 nAChR protein levels are detected concomitantly with dysregulation of the 42-kDa isoform of ERK MAPK (19). Considering that ERK MAPK activity is necessary for rodent fear learning, ␣7 nAChR up-regulation in hippocampus may serve as a biochemical marker for the synaptic plasticity impairments and learning and memory deficits in Tg2576 animals that result from chronic elevated A␤ (27)(28)(29).
Our observation of sustained inactivation of ␣7 nAChRs by low concentrations of A␤42 may reflect prolonged occupancy of its binding site(s) on ␣7 nAChRs. This putative long-lasting interaction is consistent with the hypothesis proposed by Wang et al. (32) that ␣7 nAChRs may seed or nucleate A␤ deposition, eventually leading to plaque formation. On the other hand, if the conformation of A␤ in the receptor-bound state is incompatible with further aggregation, ␣7 nAChRs may decrease the amount of free A␤, thus slowing A␤ aggregation until levels of A␤ rise beyond the binding capacity of ␣7 nAChRs.
L250T mutant ␣7 nAChRs were utilized to confirm our observations of A␤42 effects on wild-type receptors because these mutant ␣7 receptors generate much larger, longer lasting current responses (30,31). These studies again demonstrated a high affinity interaction of A␤42 with the ␣7 nAChR. Consist- ent with its desensitization-resistant properties described previously (30,31), L250T receptors can be repeatedly activated by A␤42 with washout between drug applications. However, in the absence of washout, A␤42-induced receptor desensitization occurs. In a similar manner to the effects of low concentrations of A␤ on wild-type ␣7 nAChRs, nicotine responses from L250T receptors were virtually unaffected when these receptors were concurrently desensitized to A␤42. A large proportion of the L250T whole cell current is due to secondary activation of the endogenous Ca 2ϩ -activated chloride current in Xenopus oocytes. Hence, the effects of A␤42 on L250T confirm our findings with the wild-type receptor and suggest that A␤-induced receptor activation leads to Ca 2ϩ influx. Along these lines, we have recently shown that A␤42 activates the ERK MAPK cascade through ␣7 nAChRs in hippocampal slices (19). This activation requires extracellular Ca 2ϩ and does not require action potential propagation. Thus, it is conceivable that ERK MAPK activation in hippocampus results from Ca 2ϩ influx directly through ␣7 nAChRs following A␤-activation.
One notable feature of the L250T ␣7 nAChR current response to nicotine versus A␤42 is the current profile. Whereas nicotine and other agonists for this receptor induce little desensitization during drug perfusion (30,31), at all concentrations of A␤42 tested, A␤ application leads to more rapid closure of L250T channels than 1 or 10 M nicotine. This observation may reflect desensitization or open channel block. We favor the interpretation that A␤42 induces desensitization rather than acting as an open channel blocker since during washout, we do not observe a rebound outward current as a result of channel de-block. Furthermore, the ability to fully and rapidly activate ␣7 nAChRs with nicotine following inactivation of the A␤42 response is consistent with an unblocked channel. Thus, A␤42 appears to induce L250T ␣7 nAChR channel closing in a different manner than previously tested agents.  A␤42 is highly desensitizing to wild-type receptors in that high concentration or repeated application of A␤42 inactivates the A␤ response in wild-type receptors. However, inactivation of the A␤ response with low concentrations of A␤42 (10 pM) does not interfere with subsequent nicotine responses. Consistent with our data, Liu et al. (16) did not observe significant inhibition of ␣7 nAChR responses to acetylcholine on cultured hippocampal neurons following preincubation with A␤42 at concentrations below 1 nM. These observations imply that the high affinity (low pM) A␤42 binding site and the nicotine binding site do not overlap or that the conformation induced by low concentrations of A␤42 interferes with subsequent A␤ activation but not with nicotine binding or receptor activation.
We were able to inhibit nicotine activation of ␣7 nAChRs expressed in oocytes by preincubation with 100 nM A␤42. These findings are in agreement with published reports by Petit et al. (15) and Liu et al. (16) that preincubation with high concentrations of A␤42 antagonized subsequent activation of neuronally expressed ␣7 nAChRs. These observations suggest that a second, lower affinity binding site exists for A␤42 that overlaps with the "traditional" agonist binding site. Alternatively, extended exposure to high concentrations of A␤42 may induce a conformational change in the receptor that is incompatible with subsequent activation with traditional agonists. The presence of a second, lower affinity site for A␤42 binding is supported by studies performed by Wang et al. (32) in which two K i values were detected with [ 3 H]MLA competition binding. K i values for A␤42 binding to rat ␣7 nAChRs prepared from cortex and hippocampus were 4.1 and 440 pM, respectively. Furthermore, these binding constants are compatible with our observation that 1 pM A␤42 activates ␣7 nAChRs.
The work presented here suggests that A␤/␣7 nAChR interactions may play a role in the etiology of AD. We have demonstrated that A␤ peptide functions as a ligand for ␣7 nAChRs, and we provide evidence that the binding and activation properties of A␤42 are somewhat distinct from those of agonists such as nicotine. Specifically, low concentrations of A␤42 activate the receptor and desensitize it to further activation by the peptide but not by nicotine; higher concentrations inhibit receptor activation and cross-desensitize it to nicotine. The unusual nature of A␤42-induced receptor activation and desensitization indicates that the in situ effects on ␣7 nAChRs could prove to be quite complex. In general, however, our data indicate that compounds targeted to blocking the effects of A␤ on ␣7 nAChR function may be a promising therapy for AD.