ABSTRACT

Background and Purpose: Oligomeric Aβ_1-42(oAβ₄₂) exhibits agonist-like action at human α7- and α7β2-nicotinic receptors (collectively, α7*-nAChR). Aβ_1-42 and an N-terminal Aβ peptide fragment (N-Aβ fragment: Aβ_1-15/16) have been shown to modulate presynaptic Ca²⁺ and enhance hippocampus-based synaptic plasticity via α7* nAChR. Both the N-Aβ fragment and its essential core sequence, the N-Aβcore hexapeptide (Aβ_10-15), protect against Aβ-associated synapto- and neurotoxicity, also involving nAChR. Here, we investigated how oAβ₄₂, the N-Aβ fragment and N-Aβcore regulate the functional activity of α7*-nAChRs.

Experimental approach : Single-channel patch clamp recordings measured the impact of ACh, oAβ₄₂, the N-Aβ fragment and the N-Aβcore on the function of concatenated, human α7- and α7β2-containing nAChR expressed in nAChR-null SH-EP1 cells. Molecular dynamics simulations identified potential sites of interaction between the N-Aβ fragment and the orthosteric α7*-nAChR binding interfaces.

Key Results : Relative to the effects of ACh alone, oAβ₄₂ preferentially enhanced α7β2-nAChR open probability and open-dwell times. Co-application with the N-Aβcore neutralized these effects. Further, we demonstrate that the N-Aβ fragment alone, or in combination with ACh or oAβ₄₂, resulted in selective enhancement of α7-nAChR single-channel open probability and open-dwell times (compared to ACh or oAβ₄₂).

Conclusions and Implications : Our findings show the functional diversity of Aβ peptides in regulating α7*-nAChR function, with implications for a wide range of nAChR-mediated functions in AD. Single-channel recordings of the differential effects of oAβ₄₂, N-Aβ fragment and/or N-Aβcore on α7*-nAChR isoform function revealed the complexities of their interactions with α7*-nAChR, with new insights into the neuroprotective actions of these N-Aβ-derived peptides.

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disorder that robs individuals of their language, reasoning, and memory. An estimated 6.7 million Americans > 65 years of age and 20-30 million individuals worldwide are thought to suffer from AD. AD is typically confirmed via in vivo imaging or postmortem assessment of distinct neuropathological biomarkers: amyloid plaques primarily composed of Aβ_1-42), neurofibrillary tangles (filaments of hyperphosphorylated tau), and neurodegeneration in select areas of the brain (Jack et al., 2018).

Evidence now suggests that accumulation of amyloid plaques and neurofibrillary tangles contribute marginally to the development of cognitive decline in AD patients (Castellani et al., 2007; Marchesi, 2012). Instead, there is increased appreciation of the role(s) played by the elevation of soluble, oligomeric forms of Aβ_1-42(oAβ₄₂) in AD etiopathogenesis during a long prodromal phase, up to 15 years prior to diagnosis (Cline, Bicca, Viola, & Klein, 2018; Inayathullah & Teplow, 2011), particularly in regard to the development and progression of synapse dysfunction and loss. Specifically, oAβ₄₂ is released at the synapse in response to neuronal activity (Cirrito et al., 2005) and physiological levels of the neuropeptide increase presynaptic Ca²⁺and positively modulate hippocampus-based synaptic plasticity and contextual fear conditioning (Lawrence et al., 2014; Puzzo et al., 2008), indicating that Aβ normally functions as a synaptic regulator.

Among the classes of molecular targets for Aβ and derivatives in their specific forms are nicotinic acetylcholine receptors (nAChR), including those that contain α7 and perhaps other subunits (α7*-nAChR; *indicates the possible incorporation of other nAChR subunits) (Lukas et al., 1999; Whiteaker & George, 2023). Whereas α7*-nAChRs primarily exist as homopentamers (α7-nAChR) in the majority of brain regions, a minority also contain β2 subunits (α7β2-nAChR). These heteromeric α7β2-nAChR form functional receptors with several features distinct from those of homomeric α7-nAChR (Khiroug et al., 2002; Moretti et al., 2014; Murray et al., 2012). α7β2-nAChR are enriched in specific populations of cholinergic and non-cholinergic neurons of the basal forebrain (Azam, Winzer-Serhan, & Leslie, 2003; George et al., 2021; Thinschmidt, Frazier, King, Meyer, & Papke, 2005; Whiteaker & George, 2023). α7β2-nAChRs are particularly sensitive to functional modulation by various Aβ forms (Liu, Huang, Shen, Steffensen, & Wu, 2012; Liu et al., 2009). At the synapse, oAβ₄₂ at physiological concentrations (pM-nM) has been shown to activate presynaptic α7*-nAChRs (Lawrence et al., 2014), including α7β2-nAChR (Mehta et al., 2009), via direct agonistic interaction with the ligand binding domain (Tong et al., 2011), this signaling pathway largely accounting for the Aβ-induced regulation of presynaptic Ca²⁺.

Amyloidogenic Aβ peptides are produced from the amyloid precursor protein (APP) via variable cleavage by β- and γ-secretases, yielding the more abundant Aβ_1-40 and the more toxic Aβ_1-42, the latter being the predominant isoform in plaques. Aβ_1-42 harbors a hydrophilic, N-terminal and a hydrophobic C-terminal domain derived from APP extracellular and transmembrane domains, respectively (Vassar et al., 1999). Through an alternative, non-amyloidogenic α-secretase-mediated pathway, a 15-16 amino acid (1-DAEFRHDSGYEVHHQK-16) N-terminal fragment of Aβ_1-42 (N-Aβ fragment) is generated and is present in significant amounts in human cerebrospinal fluid (Portelius et al., 2011; Portelius et al., 2007), largely as the 15-mer as the lysine in position 16 is rapidly removed by carboxypeptidase. In rodent models, like oAβ₄₂, the N-Aβ fragment has been shown to regulate presynaptic Ca²⁺ and enhance synaptic plasticity and contextual fear conditioning at pM to nM concentrations (Lawrence et al., 2014; Portelius et al., 2011), thus also functioning as a neuromodulator. In addition, the N-Aβ fragment was found to safeguard against elevated oAβ₄₂-induced neuronal toxicity, synaptic dysregulation, compromised contextual fear memory and gliotoxicity (Forest, Taketa, Arora, Todorovic, & Nichols, 2021), involving nAChR (Lawrence et al., 2014). Through a combined structure-function and mutational analysis, it was demonstrated that the concentration-dependent neuromodulatory and protective actions of the N-Aβ fragment against Aβ toxicities are localized in a unique core hexapeptide sequence (10-YEVHHQ-15; N-Aβcore; (Forest et al., 2021).

Neuronal and network-level excitotoxicity is a feature of AD and has been reported in humans and in numerous mouse models of AD pathology (Minkeviciene et al., 2009; Palop et al., 2007; Palop & Mucke, 2009; Vossel et al., 2013). Recently, we have shown that oAβ₄₂exposure i) acutely activates both human α7- and α7β2-nAChR; ii) preferentially enhances α7β2-nAChR open-dwell times; and iii) when chronically exposed, induces hyperexcitability of basal forebrain cholinergic neurons (BFCNs) within the medial septum and horizontal diagonal band (George et al., 2021). These findings demonstrate specific cellular and intrinsic-level mechanisms through which oAβ₄₂ enhances BFCN excitability through the activation of α7β2-nAChR.

Here, we tested the hypothesis that smaller N-terminal domain fragments of Aβ may mimic (or block) actions of oAβ₄₂ on α7*-nAChR function, perhaps without altering effects of the natural neurotransmitter, acetylcholine (ACh). We defined the effects of the human N-Aβ fragment and N-Aβcore on the single-channel properties of human α7- and α7β2-nAChR heterologously expressed in human, native nAChR-null SH-EP1 cells.

MATERIALS AND METHODS

Chemicals

Standard reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA). Fresh stock drug solutions were made daily and diluted as required.

Amyloid peptides

Solutions of soluble, oligomeric amyloid (oAβ₄₂; Echelon Biosciences, Salt Lake City, UT) were prepared from aqueous stock solutions as previously described and shown to exist predominantly in the oligomeric state (George et al., 2021; Stine, Jungbauer, Yu, & LaDu, 2011). The N-Aβ fragment (1-DAEFRHDSGYEVHHQ-15), the N-Aβcore (10-YEVHHQ-15), and the inactive N-Aβcore hexapeptide (N-InAβcore; 10-SEVAAQ-15; (Forest et al., 2021) used as a control were prepared from aqueous stock solutions of peptides synthesized and isolated at >98% purity (American Peptide; Anaspec). Concentrations of the peptides used were based on previous studies of their effects on synaptic plasticity and neuronal intrinsic excitability (George et al., 2021; Lawrence et al., 2014).

Constructs for human concatenated α7- and α7β2-nAChR

Fully pentameric α7- and α7β2-nAChR concatemers were constructed as previously described (George et al., 2021). Human α7- and α7β2-nAChR cDNA constructs were subcloned into a high-expression mammalian vectors (pCEP4 and pCDNA 3.1+; respectively) that harbored the Kozak consensus sequence and native α7- and α7β2-nAChR signal peptide sequences (Figure 1). All α7*-nAChR constructs were engineered to express the fluorescent protein mCherry, facilitating the identification of α7*-nAChR successfully transfected SH-EP1 cells. Genes that encoded nAChR subunits were arranged in the order 5′- α7-α7-α7-α7-α7-3′ for homomeric α7-nAChR and 5′-α7-β2-α7-β2-α7-3′ for heteromeric α7β2-nAChR. Sequences of all subunits, together with their associated partial linkers, were confirmed by DNA sequencing (Thermo Fisher Scientific, Waltham, MA, USA), and the assembly of each translated pentamer was verified by restriction digest. For α7-nAChR, orthosteric acetylcholine (ACh) binding sites are formed at all α7/α7 subunit interfaces and involve the principal (+) face containing subunit loop C of one subunit and the adjacent subunit’s complementary (-) face containing loop D. As previously demonstrated for α4β2-containing nAChR (Lucero et al., 2016), α7/β2 subunit pairs of the α7β2-nAChR construct may form ACh binding pockets between the principal (+) face of the α7 subunit and the complementary (-) face of the β2 subunit. Whether there is ligand interaction at β2(+)/α7(-) interfaces is not clear. To enhance α7*-nAChR cell-surface expression, SH-EP1 cells were co-transfected with the human nAChR chaperone protein NACHO (Gu et al., 2016). Human NACHO was engineered in the mammalian expression vector pIRES (Addgene, Watertown, MA) to facilitate expression of ZsGreen1 fluorophore and to identify successfully-transfected cells. Fluorescence microscopy imaging was used both to assess efficacy of transfections and to identify cells exhibiting both mCherry and ZsGreen1 fluorescence as optimal candidates for patch clamp recording (Figure 1).

Single-channel electrophysiology of α7- and α7β2-nAChRs transiently expressed in SH-EP1 cells

SH-EP1 cells were passaged and maintained as described previously (Peng et al., 1999). Briefly, wild-type SH-EP1 (native nAChR-null) were transfected using Qiagen’s Effectene transfection kit (Qiagen, Hilden, Germany). Single-channel α7- and α7β2-nAChR–mediated currents were recorded (Axopatch 200B amplifier; Molecular Devices) from successfully co-transfected SH-EP1 cells at 22^oC under a cell-attached configuration as previously described (George et al., 2021). Patches were clamped at a transmembrane potential of −100 mV. Patch pipettes were fabricated from thick-walled borosilicate glass (1.12 mm; WPI, Sarasota, FL, USA), and tips were micro-forged to a final resistance of 13 mΩ. Concentrations used for oAβ₄₂, N-Aβcore or N-Aβ fragment adhered to previously published methodologies (Forest et al., 2018; George et al., 2021; Lawrence et al., 2014). To elicit single-channel events, pipettes were filled with artificial cerebral spinal fluid (ACSF + 1.5 µM atropine) solution that contained: (i) ACh alone (300 µM), (ii) oAβ₄₂ alone (100 nM), (iii) N-Aβcore alone (1 µM), (iv) N-Aβ fragment alone (1µM), (v) ACh (300 µM) + N-Aβcore (1 µM), (vi) ACh (300 µM) + N-Aβ fragment (1 µM), (vii) oAβ₄₂ (100 nM) + N-Aβcore (1 µM), or 8) oAβ₄₂ (100 nM) + N-Aβ fragment (1 µM). The ACh concentration corresponded to the EC₅₀ value for each construct as determined by two-electrode voltage clamp (TEVC) (Moretti et al., 2014) and was used to ensure collection of sufficient open channel events for analysis without producing significant overlap of unitary events or open channel blockade of receptors.

For quality control, patches with seal qualities <10 GΩ were immediately discarded. Single-channel recordings were filtered at 5 kHz and sampled at 50 kHz by using pClamp 10.4 (Molecular Devices). Under these conditions, no single-channel events were observed in recordings from untransfected SH-EP1 cells, from cells using ligand-free patch pipettes, or from cells transfected with α7*-nAChR cDNA constructs using patch pipettes containing the N-InAβcore control.

All single-channel recordings were analyzed using QuB software [(v1.4.0.132; http://www.qub.buffalo.edu; (Qin, Auerbach, & Sachs, 2000b)], including for preprocessing and determining closed-dwell times (Supplementary Figure S1), open- and closed-dwell times within bursts, single-channel amplitudes, open probability, and burst duration. Recorded traces were baseline corrected and single-channel events were idealized according to a half-amplitude, threshold-crossing criterion. Single-channel amplitudes, as well as forward and backward gating rates, were derived from the idealized trace by fitting the raw data to simple closed–open (C↔O) kinetic models. At the single-channel level, α7- and α7β2-nAChR sojourns occur as brief, isolated openings with intermittent bursts/clusters of channel opening events with longer-lived closed intervals interspersed between bursts of channel activity, which reflected times when all potentially activatable channels in the patch are closed or desensitized (Grosman & Auerbach, 2001). Additional closed states were added to the C↔O model until the log-likelihood score failed to improve by >10 U (Qin, Auerbach, & Sachs, 2000a). The opening and closing rate constants were estimated by using maximum interval likelihood, which optimizes rate constants of a user-defined kinetic model according to the interval durations detected by the half-amplitude, threshold-crossing criterion. Components with time constants less than the selected dead time (0.2 ms, representing twice the value of the filter rise time) were thereby eliminated from consideration during fitting.

As previously described (George et al., 2021), we defined the main amplitude state for α7- and α7β2-nAChR constructs used in this study by using the α7*-specific PAM PNU-120596 to isolate the main conductance state for both α7- and α7β2-nAChR. Stability plots of single-channel amplitudes were generated to determine the quality of each single-channel recording over time and between patches and to calculate the mean amplitude of single events within bursts for each patch. Amplitude stability plots were generated from single-channel events elicited in the presence of ACh & PNU-120596 or oAβ₄₂& PNU-120596 for α7-containing (linked and concatenated) nAChR and the mean single-channel amplitude was defined for both α7- and α7β2-containing nAChR. Only the single-channel events that matched this pre-defined amplitude state were selected for further analyses.

Single-channel bursts corresponding to these precise amplitudes were segregated from isolated openings, and only bursts were used for single-channel analysis (highlighted in Figure S1). Bursts of single-channel activity were defined as series of openings separated by closures shorter than the minimum inter-burst closed duration [or T_crit; (Grosman & Auerbach, 2000)] and from other such episodes by prolonged channel desensitization Chakrapani, Bailey, and Auerbach (2004). For all groups tested, T_crit was calculated by using QuB software. Bursts that contained overlapping currents, which indicate two simultaneously active channels, were rare, and were discarded from the analysis. The advantages of using burst analysis has been described (Ciuraszkiewicz et al., 2013), as it increases the probability that adjacent openings arise from the same receptor.

Molecular Dynamic Simulations

The α7-nAChR homopentamer model was built based on the cryo-electron microscopy (cryo-EM) structure from the Protein Data Bank (PDBID:7KOO). We removed the transmembrane and cytoplasmic domains and kept the extracellular domain only. As there is no pentameric structure of the α7β2-nAChR stoichiometry, we extracted β2 subunits from the cryo-EM structure of the α4β2-nAChR stoichiometry (PDBID:6CNK) and replaced two α7 subunits in the α7 homopentamer to generate the model of the α7β2-nAChR heteropentamer. We used Auto Dock Vina (Eberhardt, Santos-Martins, Tillack, & Forli, 2021) to dock the N-Aβ fragment onto one of five α7+/α7- interfaces in the α7-nAChR homopentamer (Figure 6, top panels) and one of two α7+/β2- interfaces in α7β2-nAChR heteropentamer (Figure 6, bottom panels). The search space is defined to be a cube with a side length of 50 Å surrounding the midpoint from loop C of the (+) subunit interface to loops D and E of the (-) subunit interface. For each of the α7- and α7β2-nAChR stoichiometries, we selected the top 10 (out of 100) binding poses predicted by Auto Dock Vina and performed molecular dynamics simulations. The modeled nAChR/N-Aβ fragment complex structures were solvated in the TIP3P water model (Jorgensen, Chandrasekhar, Buckner, & Madura, 1986) and neutralized with K⁺ and Cl^- at 0.15 M. The van der Waals interactions were smoothly switched off over 10–12 Å using a force-based switching function and the long-range electrostatic interactions were calculated using the particle-mesh Ewald method (York, Wlodawer, Pedersen, & Darden, 1994). Each system was first equilibrated in the canonical ensemble (NVT) for 125 ps. The production run was performed in the isothermal–isobaric ensemble (NPT) at 310.15K and 1 bar with hydrogen mass repartitioning and 4 femtosecond time steps . The production run was checked every 50 ns for 400 ns and it was terminated if the N-Aβ fragment completely moved out of the binding site. All simulations were performed with OpenMM package and CHARMM36(m) force field . The simulation systems and input files were generated by CHARMM-GUI Solution Builder, and visualization was done using VMD (Humphrey, Dalke, & Schulten, 1996).

Data analysis for single-channel recordings

Single-channel amplitudes, closed-dwell times, and open-dwell times were determined from individual patches, and tau (τ) values for each exponential fit were averaged across multiple patches. Averaged single-channel amplitudes, closed-dwell times, open probabilities (P_open), and open-dwell times for α7- or α7β2-nAChR (elicited with N-Aβcore or N-Aβ fragment) were compared to results obtained in the presence or absence of ACh or oAβ₄₂. For each group, single-channel open-dwell durations were pooled from multiple patches and histograms were generated by using Qub software. All open-dwell time duration histograms were best fit with two exponentials. Histograms of closed-dwell time durations within bursts (Supplementary Figure S2) were best fit with a single exponential. Each individual exponential and their respective time constants (τ) for burst duration were calculated using Qub software. Time constants and proportions of each exponential (i.e., short and long burst durations) were compared between α7- and α7β2-nAChR.

Data are presented as means ± SEM overlaid with individual data points, except when error estimates are calculated for gaussian or exponential distributions, in which case single-channel open and closed times are represented as histograms with the best-fit value ± SEM. Data points for each figure represent individual patches pooled from multiple cells transfected on different experimental days. For all tests, an α level of <0.05 was considered significant. Single-channel data were analyzed using a two-tailed Student’s t test when comparing the effects of N-Aβcore or N-Aβ fragment on each α7*-nAChR or by one-way or two-way ANOVA and Dunnet’s or Tukey’s multiple comparison test when comparing the means across experimental conditions for each nAChR construct (GraphPad Prism; San Diego, CA).

RESULTS