Upregulated Ca2+ Release from the Endoplasmic Reticulum Leads to Impaired Presynaptic Function in Familial Alzheimer's Disease

Abstract

Neurotransmitter release from presynaptic terminals is primarily regulated by rapid Ca²⁺ influx through membrane-resident voltage-gated Ca²⁺ channels (VGCCs). Moreover, accumulating evidence indicates that the endoplasmic reticulum (ER) is extensively present in axonal terminals of neurons and plays a modulatory role in synaptic transmission by regulating Ca²⁺ levels. Familial Alzheimer's disease (FAD) is marked by enhanced Ca²⁺ release from the ER and downregulation of Ca²⁺ buffering proteins. However, the precise consequence of impaired Ca²⁺ signaling within the vicinity of VGCCs (active zone (AZ)) on exocytosis is poorly understood. Here, we perform in silico experiments of intracellular Ca²⁺ signaling and exocytosis in a detailed biophysical model of hippocampal synapses to investigate the effect of aberrant Ca²⁺ signaling on neurotransmitter release in FAD. Our model predicts that enhanced Ca²⁺ release from the ER increases the probability of neurotransmitter release in FAD. Moreover, over very short timescales (30-60 ms), the model exhibits activity-dependent and enhanced short-term plasticity in FAD, indicating neuronal hyperactivity-a hallmark of the disease. Similar to previous observations in AD animal models, our model reveals that during prolonged stimulation (~450 ms), pathological Ca²⁺ signaling increases depression and desynchronization with stimulus, causing affected synapses to operate unreliably. Overall, our work provides direct evidence in support of a crucial role played by altered Ca²⁺ homeostasis mediated by intracellular stores in FAD.

Keywords: Alzheimer’s; IP3R; asynchronous release; depression; endoplasmic reticulum; facilitation; neuronal calcium signaling; neurotransmitter release; short-term plasticity; synaptic transmission; synchronous release.

Figure 2

Kinetic schemes used in the model: (a) Gating kinetics of IP₃R. Four−state model representing the possible states along with corresponding transition rates; (b) Model for VGCC gating with four closed states (C1–C4) and one open Ca²⁺ conducting state (O); (c) Scheme for Ca²⁺ binding to synaptotagmin with dual Ca²⁺ sensors for fast synchronous (S with five Ca²⁺-binding sites), slow asynchronous (A with two Ca²⁺−binding sites), and spontaneous exocytosis; (d) Adapted with permission from Ref. [68]. 2009, Elsevier Inc. The overall release scheme, which includes vesicle mobilization from a reserve pool (R) to docked, unprimed pool (U), molecular priming to vesicles unattached to a Ca²⁺ channel (V), and conversion to vesicles coupled to a VGCC cluster (W). Both vesicle pools are released through the dual sensor release model. Channel−attached vesicles are strongly dependent on [$C{a}_{AZ}^{2+}$], whereas [$C{a}_{cyt}^{2+}$ ] governs the release of detached vesicles. Reaction rates along with respective references are listed in Table 6, Table 8, and Table 9.

Figure 3

Gain−of−function enhancement of IP₃R gating in FAD: The P_o (A,D), τ_o (B,E), and τ_c (C,F) of IP₃R given by the model (empty bars) and observed values in primary cortical neurons from WT (A–C) and 3xTg AD mice (filled bars) (D–F) at [$C{a}_{cyt}^{2+}$] = 1 μM and [$I{P}_3$ ] = 10 μM; (G) Sample time−traces generated by stochastically simulating a single IP₃R channel in cortical neurons from WT (left column) and AD (right column) mice at different [$C{a}_{cyt}^{2+}$ ] and [$I{P}_3$ ] values shown in the figure; (H) P_o of IP₃R as a function of [$C{a}_{cyt}^{2+}$ ] at [$I{P}_3$ ] = 0.3 μM (thin lines) and 1 μM (thick lines) in WT (solid lines) and AD-affected (dotted) neurons; (I) P_o of IP₃R as a function of [$I{P}_3$ ] at [$C{a}_{cyt}^{2+}$ ] = 0.1, 0.25, and 1 μM (the increasing value of [$C{a}_{cyt}^{2+}$ ] is represented by the thickness of the line) in WT (solid lines) and AD-affected (dotted) neurons. Experimental values shown for comparison in (A–F) are from [21].

Figure 1

Schematic of the overall multi-compartmental Ca²⁺ model. The arrowheads show the direction of the fluxes involved and the dotted half circles signify the Ca²⁺ domains around the IP₃Rs and VGCC clusters.

Figure 4

Characterization of neurotransmission in response to [$C{a}_{cyt}^{2+}$] steps: (A) Total release events obtained from a single dual−sensor fusion process after clamping [$C{a}_{cyt}^{2+}$ ] at different values; (B) Release profile following allosteric fusion in response to stepwise [$C{a}_{cyt}^{2+}$ ] clamp; (C) Regulation of the peak release rate in response to clamped [$C{a}_{cyt}^{2+}$ ] levels shows lower and right-shifted dose responses relative to the experimental data for the Calyx of Held; (D) [$C{a}_{cyt}^{2+}$ ] dependence of time−to−peak rate indicates exponentially decreasing but longer time delay to peak release when matched with data for the Calyx of Held. Experimental values shown for comparison in (C,D) are from [72].

Figure 5

ER-driven upregulation of cytosolic Ca²⁺ leads to enhanced synaptic vesicle release in FAD: (A) Neurotransmitter release rate in response to a single AP in WT and AD−affected synapse. Change in the P_r of a single synaptic vesicle (B); peak release rate (C); and the average number of vesicles released (D) as functions of the number of VGCCs; (E) Time delay of peak release rate and (F) decay time to basal release rate as functions of P_r; (G) Change in $\left[C{a}_{AZ}^{2+}\right]$ with number of VGCCs; (H) Cumulative Ca²⁺ from peak to basal level as a function of P_r.

Figure 6

FAD−associated Ca²⁺ upregulation enhances STF: (A) Release profile following paired−pulse stimulation protocol; (B) PPR is inversely related to intrinsic P_r (obtained after first pulse), and is higher in the AD−affected synapse; (C) Similar to the first pulse (Figure 3G), cumulative $\left[C{a}_{AZ}^{2+}\right]$ after the second pulse increases with the P_r and is higher in the AD−affected synapse; (D) Pr in response to the second pulse (${\mathrm{P}}_{\mathrm{r}2}$ as a function of P_r following the first pulse (${\mathrm{P}}_{\mathrm{r}1}$ Higher values indicate that the synapse responds more strongly to the subsequent stimulus in a paired−pulse protocol; (E) Decay time of release rate after second pulse also exhibits a biphasic behavior; (F) Cumulative $\left[C{a}_{AZ}^{2+}\right]$ following the second pulse reflects the biphasic behavior observed in time delay of peak−to−basal release rate in panel (E).

Figure 7

The AD−affected synapse exhibits stronger depression in response to 20 pulse stimulus train delivered at 20 Hz: Facilitation obtained from peak rate (A) and P_r (B) shows that AD pathology induces more severe depression relative to control conditions. Peak release rate (C) and P_r (D) following each AP in the train. (E) Pulse train depression is primarily governed by RRP depletion, which is more severe in synapses with AD pathology. (F) ${[C{a}^{2+}]}_{AZ}$ (top) and zoom-in (inset) showing the differences in basal ${[C{a}^{2+}]}_{AZ}$ levels. (G) Asynchronous release peaks and subsequently decays following depletion of RRP. (H) Peak synchronous release mimics the response seen in the overall release.

Figure 8

AD−related Ca²⁺ disruptions impair spike–SVR synchrony: The phase of individual release events with respect to AP in WT (A) and AD−affected (B) synapses over a wide range of P_r values. Black dots indicate the event phases corresponding to 35 VGCCs with initial P_r during pulse train (P_r1) equal to 0.14; (C) Synchrony of release event in response to the preceding pulse in WT and AD−affected synapses; (D) The magnitude of relative synchrony change from WT to AD conditions as a function of initial release probability during prolonged stimulation.