Cholinergic Homeostatic Synaptic Plasticity Drives the Progression of Aβ-Induced Changes in Neural Activity

Abstract

Homeostatic synaptic plasticity (HSP) is the ability of neurons to exert compensatory changes in response to altered neural activity. How pathologically induced activity changes are intertwined with HSP mechanisms is unclear. We show that, in cholinergic neurons from Drosophila, beta-amyloid (Aβ) peptides Aβ40 and Aβ42 both induce an increase in spontaneous activity. In a transgenic line expressing Aβ42, we observe that this early increase in spontaneous activity is followed by a dramatic reduction in spontaneous events, a progression that has been suggested to occur in cholinergic brain regions of mammalian models of Alzheimer's disease. We present evidence that the early enhancement in synaptic activity is mediated by the Drosophila α7 nicotinic acetylcholine receptor (nAChR) and that, later, Aβ42-induced inhibition of synaptic events is a consequence of Dα7-dependent HSP mechanisms induced by earlier hyperactivity. Thus, while HSP may initially be an adaptive response, it may also drive maladaptive changes and downstream pathologies.

Keywords: Alzheimer’s disease; Drosophila; beta-amyloid; homeostatic synaptic plasticity; α7 nAChR.

Figure 1.. Differential Activity Induced by Aβ42 in Identified Neurons in Young and Older Brains

(A and B) Spontaneous EPSCs recorded from antennal lobe projection neurons (PNs) in the intact brains of control (Ctrl) and elav-GAL4>>UAS-Aβ42 (Aβ42) flies at 0–2 days after eclosion (AE; A) or 9–13 days AE (B). Cumulative distributions of inter-event times and representative traces are shown. (A) In young brains, spontaneous synaptic activity was more frequent with inter-event times significantly shorter in Aβ42-expressing flies (2,439.60 ± 1182.44 ms; n = 9) compared to Ctrls (5,579.67 ± 1,804.34 ms; n = 7) (p < 0.05). (B) At 9–13 days AE, spontaneous activity was depressed with inter-event times significantly longer in Aβ42-expressing flies (2,161.49 ± 624.18 ms; n = 8) compared to Ctrls (1,256.57 ± 263.68 ms; n = 6) (p< 0.05). Scale bar, 250 ms/5 pA.

Figure 2.. Spontaneous Synaptic Activity in Primary Neurons at 5 DIV Is Enhanced by Aβ42

(A–D) sEPSCs were recorded from 5 DIV primary neurons from elav-GAL4>>UAS-Aβ42 (Aβ42) or Ctrl line (Ctrl); V_m = −70 mV. Shown are comparisons of average sEPSC inter-event duration (A), amplitude (B), and rise time, decay time, half-width, and charge transfer (D); representative traces are shown in (C). Note that activity is enhanced in neurons from Aβ42-expressing cultures, with reduced sEPSC inter-event time (A) and increased amplitude (B). For Ctrl, n = 1,609 events from 20 cells; for Aβ42, n = 2,329 events from 23 cells; *p < 0.05 (Student’s t test). Scale bar, 50 ms/5 pA.

Figure 3.. Spontaneous Synaptic Activity at 9 DIV Is Significantly Depressed in Aβ42-Expressing Cultures

(A–D) sEPSCs were recorded from primary neurons at 9 DIV from elav-GAL4>>UAS-Aβ42 (Aβ42) or Ctrl lines (Ctrl); V_m = −70 mV. Shown are comparisons of average sEPSC inter-event duration (A), amplitude (B), and rise-time, decay time, half-width, and charge transfer (D); representative traces are shown in (C). Note that activity is significantly inhibited in neurons from Aβ42-expressing cultures, with greatly increased sEPSC inter-event times (A) and decreased amplitude (B); decaytimesand half-widths of events were also reduced (D). For Ctrl, n = 5,595 events from 35 cells; for Aβ42, n = 1,353 events from 30 cells; *p < 0.05 (Student’s t test). Scale bar, 50 ms/5 pA.

Figure 4.. Aβ42 Induces an Early Increase in mEPSC Frequency, but not Amplitude, Suggesting a Pre-synaptic Change

(A and B) Comparison of mEPSC inter-event times (A) and amplitudes (B) between 5 DIV neurons from Ctrl, elav-GAL4>>UAS-Aβ42 (elav;Aβ42), and w;elav-Aβ42 cultures; V_m = −70 mV. Note that mEPSCs from both elav-GAL4>>UAS-Aβ42 and w;elav-Aβ42 lines show increased activity with significantly reduced inter-event times (822.47 ± 29.98 ms and 799.26 ± 32.24 ms, respectively) compared to Ctrl (1,983.54 ± 118.04 ms); inter-event times between elav;Aβ42 and w;elav-Aβ42 were not significantly different (ns). *p < 0.05 (Student’s t test). (C) Analyses of rise times, decay times, half-widths, and charge transfer showed no difference between Ctrl and elav;Aβ42. (D) mEPSC decay kinetics fit with a double exponential showed no differences in fast/slow time constants (tau1 and tau2) or the relative contribution of each to the decay (A₁/(A₁ + A₂) for the fast component, and A₂/(A₁ + A₂) for the slow component). (E) Representative traces of each genotype shown. For Ctrl, n = 4,437 events from 13 cells; for elav-GAL4>>UAS-Aβ42, n = 4,814 events from 8 cells; for w;elav-Aβ42, n = 3,950 from 11 cells. Scale bar, 50 ms/5 pA.

Figure 5.. Aβ42 Induces a Later Decrease in mEPSC Frequency, as well as a Change in mEPSC Decay Kinetics

(A and B) Comparison of mEPSC inter-event times (A) and amplitudes (B) between 9 DIV neurons from Ctrl, elav-GAL4>>UAS-Aβ42 (elav;Aβ42), and w;elav-Aβ42 cultures; V_m = −70 mV. Note that mEPSCs from both elav-GAL4>>UAS-Aβ42 and w;elav-Aβ42 lines show decreased activity with significantly enhanced interevent times (1,788.09 ± 78.71 ms and 1,373.52 ± 72.64 ms, respectively) compared to Ctrl (989.51 ± 51.29 ms). (C) Analyses of rise times, decay times, half-widths, and charge transfer showed a reduced decay time (90%–37%) in elav;Aβ42 neurons compared to age-matched Ctrl neurons. (D) mEPSC decay kinetics fit with a double exponential showed no differences in fast or slow time constants (tau1 and tau2), but the relative contribution of the fast component (A₁/(A₁ + A₂)) was reduced in elav;Aβ42 neurons, and the relative contribution of the slow component (A₂/(A₁ + A₂)) was increased. This results in a sharpened current without a change in charge transfer. (E) Representative traces of each genotype shown. For Ctrl, n = 5,824 events from 11 cells; for elav-GAL4>>UAS-Aβ42, n = 1,995 events from 13 cells; for w;elav-Aβ42, n = 1,879 from 11 cells. Scale bar, 50 ms/5 pA. *p < 0.05 (Student’s t test).

Figure 6.. Aβ42 Co-localizes with Dα7-EGFP in Neurons, and Acute Application of Aβ42 Induces a Dα7-Dependent Enhancement in Synaptic Activity in 5 DIV Neurons

(A) Primary elav-GAL4>>UAS-Dα7-EGFP cultures were treated from 0 to 5 DIV with 500 nM Aβ42. Cultures were then fixed and immunostained for anti-GFP (green) and anti-Aβ42 (red). Shown are representative optical sections through typically round neuronal cell bodies that grow in small clusters (top) and neuronal processes; merged images show overlap (yellow) between Dα7-EGFP and Aβ42 localization in membrane of somas and processes. Scale bars, 2 μm. (B) Representative spontaneous EPSCs and quantification of inter-event times are shown after mock or exogenous application of Aβ40 or Aβ42 peptide to Ctrl or Dα7^PΔEY6 mutant (Dα7^−/−) cultures at 5 DIV; V_m = −70 mV. In Ctrl neurons, 2.5 nM Aβ40 or Aβ42 peptide resulted in a decrease in inter-event time (93.67 ± 0.74 ms and 121.11 ± 1.51 ms, respectively; n = 6 cells) compared to mock-treated neurons (255.13 ± 4.61 ms; n = 9 cells). At higher concentration (10 nM), Aβ42, but not Aβ40, induced a significant increase in interevent time (708.17 ± 48.64 ms; n = 10 cells). In Dα7^−/−, 2.5 nM Aβ42 failed to induce an increase in spontaneous activity (1,050.80 ± 41.01 ms; n = 8 cells) compared to mock-treated (948.32 ± 41.46 ms; n = 9 cells). *p < 0.05 (Student’s t test) between indicated groups.

Figure 7.. Increasing Early Cholinergic Activity Is Sufficient to Induce Synaptic Inhibition in Wild-Type, and Blocking Early Activity in Aβ42-Cultures Prevents the Development of Synaptic Inhibition

Representative spontaneous EPSCs (V_m = −70 mV) and quantification of inter-event times are shown (A–D). (A) The acetylcholine esterase inhibitor, ambenonium (Amb; 1 μM), was applied to wild-type cultures from 4 to 6 DIV, and then sEPSCs were recorded at 9 DIV; mock-treated cultures underwent the same mechanical perturbations as Amb-treated cultures. sEPSCs in Amb-treated neurons exhibited a greatly reduced frequency, with increased inter-event times (12,227.07 ± 2546.88 ms; n = 7 cells) compared to mock treated (333.63 ± 6.33 ms; n = 8 cells). (B) Cultures from the Dα7^PΔEY6 null mutant were mock or Amb treated from 4 to 6 DIV; at 9 DIV, sEPSCs were recorded. Amb treatment did not increase interevent times as in wild-type; rather, a relative decrease in inter-event times was observed (mock treated, 846.75 ± 33.88 ms, n = 11 cells; Amb treated, 663.20 ± 20.81 ms, n = 10 cells). (C) elav-GAL4>>UAS-Aβ42 (elav; Aβ42) cultures were treated with curare (20 μM) from 4 to 6 DIV to inhibit the increase in cholinergic activity induced by Aβ42. sEPSCs recorded at 9 DIV showed that curare-treated neurons did not exhibit the synaptic silencing induced in mock-treated elav;Aβ42 neurons (mock treated, 2,617.98 ± 394.46 ms, n = 8 cells; Aβ42-treated, 230.28 ± 5.00 ms, n = 9 cells). (D) Aβ42 peptide at 2.5 nM, which was shown in Figure 6B to exhibit an increase in spontaneous activity (decreased inter-event times), was applied to Ctrl and Dα7^PΔEY6 (Dα7^−/−) mutant neurons at for 30–60 min at 4 DIV. sEPSCs were then recorded at 9 DIV. Aβ42 peptide application at 4 DIV induced significant synaptic depression at 9 DIV in wild-type neurons, with greatly enhanced inter-event times (mock exposure, 830.84 ± 48.96 ms, n = 13 cells; Aβ42-exposure, 2,832.99 ± 300.78 ms, n = 8 cells), but not in Dα7^PΔEY6 mutant neurons (mock-exposure, 886.64 ± 51.41 ms, n = 8 cells; Aβ42-exposure, 703.29 ± 35.44 ms, n = 8 cells). Scale bar, 50 ms /5 pA. *p < 0.05 (Student’s t test).