Amyloid β oligomers suppress excitatory transmitter release via presynaptic depletion of phosphatidylinositol-4,5-bisphosphate

He, Yang; Wei, Mengdi; Wu, Yan; Qin, Huaping; Li, Weinan; Ma, Xiaolin; Cheng, Jingjing; Ren, Jinshuai; Shen, Ye; Chen, Zhong; Sun, Binggui; Huang, Fu-De; Shen, Yi; Zhou, Yu-Dong

doi:10.1038/s41467-019-09114-z

Download PDF

Article
Open access
Published: 13 March 2019

Amyloid β oligomers suppress excitatory transmitter release via presynaptic depletion of phosphatidylinositol-4,5-bisphosphate

Yang He¹^na1,
Mengdi Wei¹^na1,
Yan Wu¹,
Huaping Qin¹,
Weinan Li¹,
Xiaolin Ma¹,
Jingjing Cheng¹,
Jinshuai Ren¹,
Ye Shen¹,
Zhong Chen²,
Binggui Sun ORCID: orcid.org/0000-0002-0386-6569¹,
Fu-De Huang³,
Yi Shen ORCID: orcid.org/0000-0001-5072-4734¹ &
…
Yu-Dong Zhou ORCID: orcid.org/0000-0002-6364-3869¹

Nature Communications volume 10, Article number: 1193 (2019) Cite this article

9119 Accesses
80 Citations
18 Altmetric
Metrics details

Subjects

Abstract

Amyloid β (Aβ) oligomer-induced aberrant neurotransmitter release is proposed to be a crucial early event leading to synapse dysfunction in Alzheimer’s disease (AD). In the present study, we report that the release probability (Pr) at the synapse between the Schaffer collateral (SC) and CA1 pyramidal neurons is significantly reduced at an early stage in mouse models of AD with elevated Aβ production. High nanomolar synthetic oligomeric Aβ₄₂ also suppresses Pr at the SC-CA1 synapse in wild-type mice. This Aβ-induced suppression of Pr is mainly due to an mGluR5-mediated depletion of phosphatidylinositol-4,5-bisphosphate (PIP₂) in axons. Selectively inhibiting Aβ-induced PIP₂ hydrolysis in the CA3 region of the hippocampus strongly prevents oligomeric Aβ-induced suppression of Pr at the SC-CA1 synapse and rescues synaptic and spatial learning and memory deficits in APP/PS1 mice. These results first reveal the presynaptic mGluR5-PIP₂ pathway whereby oligomeric Aβ induces early synaptic deficits in AD.

Formation of memory assemblies through the DNA-sensing TLR9 pathway

Article Open access 27 March 2024

Vladimir Jovasevic, Elizabeth M. Wood, … Jelena Radulovic

Messenger RNA transport on lysosomal vesicles maintains axonal mitochondrial homeostasis and prevents axonal degeneration

Article Open access 10 April 2024

Raffaella De Pace, Saikat Ghosh, … Juan S. Bonifacino

APOE4/4 is linked to damaging lipid droplets in Alzheimer’s disease microglia

Article Open access 13 March 2024

Michael S. Haney, Róbert Pálovics, … Tony Wyss-Coray

Introduction

The neuropathology of Alzheimer’s disease (AD) is characterized by occurrence of senile plaques containing amyloid β (Aβ) aggregates and neurofibrillary tangles formed by hyperphosphorylated tau in the brain^1,2,3. An important cellular correlate of cognitive decline in AD is synapse loss^4,5. Thus, many studies in AD focus on exploring the underlying mechanisms of neurotoxic effects of Aβ and hyperphosphorylated tau on synapse loss and neuronal death⁶. However, synaptic dysfunction may occur before synapse loss in early AD⁴. Therefore, elucidating how pathogenic Aβ and tau species alter synaptic transmission is crucial to the diagnosis and treatment of AD. In recent years, investigating how Aβ modulates synapse function in early AD has attracted great attention^7,8. The neurotoxic soluble Aβ oligomers, including the most toxic oligomeric Aβ₄₂, have been shown to alter synaptic plasticity and synaptic transmission in various AD animal models via a variety of synaptic targets of Aβ such as ionic neurotransmitter receptors, G protein-coupled receptors (GPCRs), receptor tyrosine kinases, and cellular prion proteins (PrP^C)^9,10. Although many of these Aβ targets exist in both presynaptic and postsynaptic loci¹¹, and Aβ oligomers accumulate at both sides of the excitatory synapse^12,13, majority of the studies have only examined the toxic gain of function for Aβ as a result of its interaction with the postsynaptic targets^14,15. Examinations of Aβ-induced abnormalities in synaptic transmission have nevertheless revealed presynaptic defects are often more prominent than postsynaptic abnormalities^16,17. Thus, it is pivotal to unravel the presynaptic targets of Aβ in AD.

Physiological concentration of Aβ (picomolar)¹⁸ has been shown to positively regulate synaptic transmission via upregulating the presynaptic neurotransmitter release probability (Pr)¹⁹. Low to moderate levels of Aβ may augment Pr via increasing presynaptic Ca²⁺ by promoting presynaptic amyloid precursor protein (APP) homodimerization²⁰, activating exocytotic Ca²⁺ channels²¹, and regulating presynaptic α7 nicotinic acetylcholine receptors^22,23. However, how pathological level of oligomeric Aβ (nanomolar)¹⁸ leads to presynaptic defects remains largely obscure. Controversial results exist in the literature showing pathogenic Aβ may exert negative^{16,17,24,25,26}, positive^27,28, or no^29,30 effects on neurotransmitter release. Different Aβ species and their targets, various assemblies of Aβ monomers, and duration of Aβ action may account for these disparate observations⁹. Furthermore, there is little evidence of what the presynaptic targets of pathogenic Aβ are in early AD. There are contradicting results with respect to how nanomolar Aβ oligomers regulate voltage-gated Ca²⁺ channels^25,27,31 and SNARE complex proteins^26,28 to disrupt presynaptic neurotransmitter release.

In the current study, we aim to clarify the presynaptic deficit at an excitatory hippocampal synapse in AD models and determine the presynaptic target of pathological level of oligomeric Aβ₄₂. We identified a nanomolar oligomeric Aβ₄₂-induced, presynaptic metabotropic glutamate receptor 5 (mGluR5)-mediated hydrolysis of membrane phosphatidylinositol-4,5-bisphosphate (PIP₂) underlies the diminished Pr in early AD. Postsynaptic mGluR5 has been shown to function as an Aβ receptor or co-receptor with PrP^C^32,33,34 and blocking mGluR5 reduces cognitive impairment in AD mouse models^35,36,37. Oligomeric Aβ³⁸ and apolipoprotein E4³⁹ are also known to interfere with PIP₂ metabolism and reduction of PIP₂ phosphatase synaptojanin 1 ameliorates synaptic and behavioral deficits in AD⁴⁰. Our results for the first time establish that increasing the presynaptic PIP₂ level is an effective way to improve cognition in AD.

Results

Reduced transmission in early AD involves a reduction in Pr

To determine the synaptic deficits in early AD, we first examined spine morphology and density of apical dendrites in CA1 pyramidal neurons in 6–7-month-old APP (Swe); PS1(ΔE9) (APP/PS1) mice (Fig. 1a–c). We compared the total spine density (Fig. 1c) and the density of filopodium-like, thin, stubby, and mushroom-shaped spines (Fig. 1b) in wild-type (WT) and APP/PS1 mice. We found that in APP/PS1 mice, there were no changes in the total spine density and the density of mature forms of spines in comparison to WT mice, although the filopodium-like spine density was significantly increased. The result is consistent with a previous study showing spine loss did not occur in CA1 stratum radiatum in APP/PS1 mice, whereas the number of small-headed spines was significantly increased⁴¹. We next recorded miniature excitatory postsynaptic currents (mEPSCs) in CA1 pyramidal neurons in 6–7-month-old WT and APP/PS1 mice (Fig. 1d, e). The frequency but not the amplitude of mEPSCs was significantly lower in APP/PS1 mice than in WT controls. We then recorded evoked EPSCs at the Shaffer collateral (SC) to CA1 pyramidal neuron (SC-CA1) synapse and found that the amplitude of SC-CA1 EPSCs was strongly decreased in 6–7-month-old APP/PS1 mice (Fig. 1f, g). These changes in mEPSC frequency and SC-CA1 EPSC amplitude were not observed in 4-month-old APP/PS1 mice (Supplementary Fig. 1), suggesting that the synaptic transmission deficits occur at a later stage when Aβ production is greatly elevated (Supplementary Fig. 2). Indeed, in 4-month-old APP (Swe/Flo/Lon); PS1 (M146L/L286V) mice (5xFAD) with an accelerated Aβ production (Supplementary Fig. 2), the frequency of mEPSCs in CA1 neurons and the amplitude of SC-CA1 EPSCs were greatly reduced in comparison to APP/PS1 and WT mice of the same age (Supplementary Fig. 1a-d). By contrast, 6–7-month-old PS1^M146V knock-in mice (M146V) with a low Aβ production exhibited slightly decreased mEPSC frequency and amplitude compared to their WT littermates (Supplementary Fig. 1e, f). Interestingly, SC-CA1 EPSCs were smaller in M146V than in control mice at 6–7-month of age, implying that an Aβ-independent mechanism may underlie the decreased synaptic transmission. These results suggest that decreased synaptic transmission in the hippocampus may be a hallmark in AD mouse models that show a high level of Aβ accumulation.

To further test if the reduced hippocampal synaptic transmission in AD mouse models is dependent on Aβ levels, we first bath applied synthetic oligomeric Aβ₄₂ at various concentrations in brain slices from WT mice. We observed that bath application of 400 nM oligomeric Aβ₄₂ for 20 min induced a significant decrease in the amplitude of SC-CA1 EPSCs, whereas 20 nM and 100 nM oligomeric Aβ₄₂ respectively exerted positive and no effects on the evoked responses (Supplementary Fig. 3). Four hundred nanometers of Aβ₄₂ fibrils, on the other hand, did not alter the amplitude of SC-CA1 EPSCs (Supplementary Fig. 3). We thus chose 400 nM Aβ₄₂ oligomers for further in vitro studies. Bath application of 400 nM oligomeric Aβ₄₂ for 20 min reduced both mEPSC amplitude and frequency in CA1 pyramidal neurons (Fig. 1h, i) and suppressed SC-CA1 EPSC amplitude (Fig. 1j, k) in WT mice. Interestingly, 400 nM oligomeric Aβ₄₂ induced a greater suppression of mEPSCs in frequency than in amplitude (Fig. 1e), suggesting that the site of action of oligomeric Aβ₄₂ was mainly presynaptic. These data indicate that 400 nM oligomeric Aβ₄₂ exerts similar effect on hippocampal synaptic transmission as in AD mouse models with a high level of Aβ accumulation.

Although there was no spine loss in 6–7-month-old APP/PS1 mice, Aβ-induced reduction of the number of active boutons may account for the decreased mEPSC frequency in these animals, as nanomolar Aβ oligomers can reduce the recycling pool and increase the resting pool of synaptic vesicles⁴². To test this hypothesis, we loaded active synaptic vesicles with FM1–43 in cultured hippocampal neurons. Compared to vehicle treatment, treatment with 400 nM oligomeric Aβ₄₂ did not change the number of active boutons in hippocampal neurons (Fig. 1l, m). These results indicate that Aβ-induced inhibition of mEPSC frequency is not due to a decrease in the number of active boutons and suggest that early synaptic deficits in AD mouse models may involve a presynaptic mechanism other than reducing the active pool of synaptic vesicles.

A decrease in mEPSC frequency may also be due to a reduction in presynaptic release Pr. We thus first investigated whether Aβ elevation enhances paired-pulse facilitation (PPF) at the SC-CA1 synapse, as the degree of PPF is inversely related to Pr. We found that PPF was significantly increased in 6–7-month-old APP/PS1 (Fig. 2a, b) and 4-month-old 5xFAD (Supplementary Fig. 4a, b) mice compared to WT controls. In 4-month-old APP/PS1 (Supplementary Fig. 4a, b) and 6–7-month-old M146V (Supplementary Fig. 4c, d) mice, however, PPF was unchanged in comparison to WT mice, which is also reported in 3-month-old APP/PS1 mice⁴³. Similarly, Aβ₄₂ oligomers induced a significant increase in PPF in WT animals (Fig. 2c, d). These results suggest that pathogenic Aβ-induced synaptic deficits are due to a reduction in Pr. To directly determine if Pr was suppressed in APP/PS1 mice and in Aβ-treated brain slices from WT mice, we used a repeated stimulation protocol to estimate the readily releasable pool (RRP) size and Pr (Fig. 2e–l). Repeated stimulation (20 Hz) revealed a significant reduction in Pr in APP/PS1 mice (Fig. 2h) and in Aβ-treated brain slices from WT mice (Fig. 2l), although the RRP size did not change (Fig. 2g, k). Taken together, these results indicate that Aβ accumulation reduces Pr in the hippocampus.

To establish that the presynaptic deficit is due to an elevated Aβ level in 6–7-month-old APP/PS1 mice, we investigated whether reducing the Aβ level could restore hippocampal synaptic transmission. We first treated 6–7-month-old APP/PS1 and WT mice with LY-411575, a γ-secretase inhibitor that is known to decrease interstitial fluid levels of Aβ^18,44, and confirmed the effectiveness of LY-411575 by the elevated levels of C-terminal fragments (CTFs) after treatment (Supplementary Fig. 5a). We did not measure Aβ levels to indicate LY-411575 activity because existing plaques in these mice are not reduced by LY-411575^44,45. Treatment with LY-411575 significantly increased CA1 mEPSC frequency in CA1 pyramidal neurons (Supplementary Fig. 5b, c) and SC-CA1 EPSC amplitude (Supplementary Fig. 5d, e) and reduced PPF at the SC-CA1 synapse (Supplementary Fig. 5f, g) in APP/PS1 mice, implicating that inhibiting Aβ generation enhanced Pr in these mice. We next examined whether directly blocking Aβ with a specific human Aβ antibody could restore the decreased Pr in APP/PS1 mice. Long-term incubating hippocampal slices from APP/PS1 mice with 6E10 (2 μg ml⁻¹, >5 h) strongly increased mEPSC frequency (Fig. 3a, b) and evoked EPSC amplitude (Fig. 3c, d) and suppressed PPF (Fig. 3e, f), indicating that blocking Aβ with 6E10 indeed restored the presynaptic defect in these mice. Although this long-term maintenance of hippocampal slices in vitro caused a rundown in the baseline transmission in both WT and APP/PS1 mice, the effect of blocking Aβ with 6E10 on restoring synaptic transmission in APP/PS1 mice to the level in WT mice was remarkable. Taken together, our results showed that elevated Aβ levels were essential for reducing Pr in APP/PS1 mice.

PIP₂ depletion by Aβ-induced activation of mGluR5 reduces Pr

One of the key mechanisms controlling Pr involves a vesicle membrane-anchoring event preceding the formation of the SNARE complex. As a crucial phosphoinositide interacting with membrane-binding proteins, PIP₂ has been shown to play an important role in vesicle release via synaptotagmin 1-PIP₂ binding, and its level in the membrane, thus, is tightly associated with Pr⁴⁶. We found that the PIP₂ level was significantly decreased in 6–7-month-old APP/PS1 mice (Supplementary Fig. 6). This is consistent with the decreased Pr in AD mouse models. To further test if Aβ can deplete PIP₂ in axons, we immunostained and biochemically measured PIP₂ in cultured hippocampal neurons in control, vehicle (DMSO)-treated, or Aβ₄₂ oligomer-treated medium (Fig. 4a–c). In comparison to control and DMSO treatments, Aβ treatment rapidly and significantly suppressed total PIP₂ (Fig. 4c) and axonal (MAP2⁻, NF⁺) and dendritic (MAP2⁺, NF⁺) PIP₂ (Fig. 4a, b). Although the anti-PIP₂ antibody we used cross-reacts with phosphatidylinositol-3,4,5-trisphosphate (PIP₃) (Supplementary Fig. 7a), we believe Aβ-induced suppression of PIP₂ is valid since PIP₃ is significantly less abundant than PIP₂⁴⁷ and the ELISA kit we used is highly specific to PIP₂ (Supplementary Fig. 7b). To test if adding PIP₂ back to Aβ-treated neurons was sufficient to restore Aβ-induced presynaptic defect, we recorded miniature excitatory autaptic currents (mEACs) from individually inhabited neurons grown on collagen-poly-D-lysine (PDL) islands (Fig. 4d). We first established that Aβ-induced suppression of mEAC frequency was similar to that of mEPSC frequency (Fig. 4e, f). We then filled the patch pipettes with an internal solution containing 200 μM diC8-PIP₂, assuming PIP₂ would be incorporated into the axon influencing neurotransmitter release. Indeed, intracellularly applied PIP₂ rescued Aβ–induced inhibition of mEAC frequency, whereas intracellularly applied PIP₂ did not affect baseline mEAC frequency (Fig. 4e, f). These results indicate that Aβ-associated PIP₂ depletion accounts for Aβ-induced presynaptic effect.

Aβ-induced rapid PIP₂ depletion confirms a predominant PIP₂ hydrolysis process evoked by Aβ oligomers³⁸. Increased PLC activity, thus, may result in Aβ-induced PIP₂ depletion. To test this hypothesis, we applied PLC inhibitor U73122 in the culture medium to block PLC activity in primary hippocampal neurons before oligomeric Aβ₄₂ treatment. In the presence of U73122, Aβ₄₂ oligomers no longer exerted the inhibitory effect on PIP₂ levels in axons and dendrites (Fig. 5a, b). U73122 alone did not change PIP₂ levels in neuronal processes (Fig. 5a, b). These results indicate that Aβ-induced PIP₂ depletion is mainly due to an Aβ-triggered, PLC-mediated PIP₂ hydrolytic event. We next investigated whether blocking PLC activity prevented the presynaptic deficit induced by oligomeric Aβ₄₂. In the presence of U73122, oligomeric Aβ₄₂-induced suppression of SC-CA1 EPSCs was partially restored (Fig. 5c–e). Notably, U73122 treatment reduced oligomeric Aβ₄₂-induced inhibition of mEPSC frequency but not amplitude (Fig. 5f, g), suggesting that blocking PLC prevented Aβ-induced inhibition of Pr. Blocking Aβ-induced inhibition of Pr via inhibiting PLC was further proved by examining PPF at the SC-CA1 synapse, as a prior application of U73122 prevented Aβ-induced increase in the paired-pulse ratio (Fig. 5h, i). In addition, U73122 alone did not change the frequency and amplitude of mEPSCs in CA1 pyramidal neurons and the PPF at the SC-CA1 synapse in WT animals (Supplementary Fig. 8a-d). Furthermore, we explored which PLC isoforms were involved in Aβ-induced PIP₂ hydrolysis in neurites using the RNA-interference approach (Supplementary Fig. 9a, b). Knocking down either PLCβ1 or β4 significantly ameliorated Aβ-induced suppression of PIP₂ levels in both dendrites and axons (Supplementary Fig. 9c, d), indicating that both PLCβ1 and β4 contribute to Aβ-induced PIP₂ hydrolysis in hippocampal neurons. Taken together, these results prove that Aβ-induced elevation of PLC activity underlies PIP₂ depletion and Pr reduction.

One of the key routes leading to PLC activation is via ligand binding to Gαq-coupled GPCRs. Aβ has been shown to activate many GPCRs, including mGluR5^32,33,34, a subtype of group I metabotropic glutamate receptors (mGluRs). We found that mGluR5 was expressed in neuronal processes, although axonal mGluR5 is less abundant than dendritic mGluR5 (Fig. 6a, b). Thus, we studied whether mGluR5 mediated Aβ-induced PIP₂ depletion and the subsequent presynaptic defect. We first examined whether a group I mGluR agonist (S)−3,5-Dihydroxyphenylglycine (DHPG) (Fig. 6) exerted similar effects as Aβ oligomers (Fig. 4). Application of DHPG in the culture medium significantly decreased PIP₂ levels in axons and dendrites in primary hippocampal neurons (Fig. 6c, d). Unlike Aβ, DHPG still depleted PIP₂ in neurites in the presence of an antibody (6D11) against PrP^C, indicating that oligomeric Aβ-induced PIP₂ hydrolysis requires activation of mGluR5 with the aid of PrP^C (Supplementary Fig. 10). Functionally, DHPG inhibited mEPSC frequency in CA1 pyramidal neurons (Fig. 6e, f) and increased PPF at the SC-CA1 synapse (Fig. 6g, h), indicating that activation of group I mGluRs induced mainly a presynaptic effect. Unlike Aβ, DHPG did not alter mEPSC amplitude (Fig. 6e, f), suggesting that activation of group I mGluRs was not involved in Aβ-induced suppression of mEPSC amplitude.

We next investigated whether inhibiting mGluR5 hampered the effectiveness of oligomeric Aβ₄₂ on PIP₂ depletion and Pr reduction. In the presence of a selective mGluR5 antagonist 3-((2-Methyl-4-thiazolyl)ethynyl)pyridine (MTEP), oligomeric Aβ₄₂ no longer suppressed PIP₂ levels in neurites in primary hippocampal neurons (Fig. 6i, j). MTEP treatment partially occluded Aβ-induced inhibition of SC-CA1 EPSCs (Fig. 6k–m), but did not prevent Aβ-induced inhibition of mEPSC amplitude in CA1 pyramidal neurons (Fig. 6n, o), which was consistent with the result showing DHPG did not change mEPSC amplitude (Fig. 6e, f). Importantly, MTEP treatment prevented oligomeric Aβ₄₂-induced changes in mEPSC frequency in CA1 pyramidal neurons (Fig. 6n, o) and PPF at the SC-CA1 synapses (Fig. 6p, q), indicating that blocking mGluR5 ameliorated Aβ-induced Pr suppression. MTEP alone did not alter PIP₂ levels in neurites (Fig. 6i, j), SC-CA1 EPSCs (Fig. 6k–m), mEPSCs in CA1 pyramidal neurons (Supplementary Fig. 8e, f), and PPF at the SC-CA1 synapse (Supplementary Fig. 8g, h) in WT animals. Furthermore, long-term treatment with MTEP (>3 h) greatly increased CA1 mEPSC frequency (Supplementary Fig. 11a, b) and SC-CA1 EPSC amplitude (Supplementary Fig. 11c, d) and reduced PPF at the SC-CA1 synapse (Supplementary Fig. 11e, f) in 6–7-month-old APP/PS1 mice, although this long-term maintenance of hippocampal slices in vitro caused a rundown in the baseline transmission. Taken together, these results imply that activation of presynaptic mGluR5 by Aβ oligomers contributes to PIP₂ depletion-associated reduction in Pr.

Inhibiting presynaptic PIP₂ drop rescues Pr and memory in AD

To establish fully that oligomeric Aβ-induced depletion of presynaptic PIP₂ underlies Pr reduction, selectively inhibiting Aβ-induced hydrolysis of presynaptic PIP₂ is essential. One way to control the PIP₂ level is to manipulate the responsiveness of Gαq-coupled GPCRs. The mammalian pho eighty-five requiring 3 (Efr3) proteins can control GPCR responsiveness⁴⁸. Indeed, halving Efr3a in cultured astrocytes from Efr3a^+/− mice resulted in a drastic decrease in DHPG-induced or Aβ-induced increase in intracellular Ca²⁺ concentration ([Ca²⁺]_i) that could be completely blocked by MTEP (Supplementary Fig. 12), indicating that knocking down Efr3a was an efficient way to suppress mGluR5 responsiveness. Importantly, we found that halving Efr3a copy number in APP/PS1 mice restored the decreased PIP₂ level in the AD mice (Supplementary Fig. 6). Oligomeric Aβ treatment was no longer effective in reducing total and neurite PIP₂ in cultured hippocampal neurons from Efr3a^+/− mice (Fig. 7a–c). Furthermore, oligomeric Aβ₄₂ was less efficient to reduce SC-CA1 EPSCs in Efr3a^+/− than in WT mice (Fig. 7d–f) and no longer altered mEPSC frequency in CA1 pyramidal neurons (Fig. 7g, h) and PPF at the SC-CA1 synapse in Efr3a^+/− mice (Fig. 7i, j). Halving Efr3a copy number restored the decreased mEPSC frequency in CA1 pyramidal neurons and the upregulated PPF at the SC-CA1 synapse in APP/PS1 mice (Supplementary Fig. 13). By contrast, oligomeric Aβ₄₂ caused a more robust inhibition of mEPSC frequency in CA1 pyramidal neurons in Efr3a overexpression mice (Supplementary Fig. 14). Taken together, these results indicate that reducing Efr3a is an effective method to inhibit Aβ-induced PIP₂ depletion. Therefore, it is feasible to inhibit Aβ-induced PIP₂ hydrolysis region-specifically by creating conditional knockouts of Efr3a in the CA3 or CA1 area in mice.

We thus created conditional Efr3a knockouts in the CA3 and CA1 areas by crossing Efr3a-loxP mice to Grik4-cre and CamKIIa-cre mice, respectively (Fig. 8). In CA1-specific Efr3a conditional knockout (CA1-Efr3a cKO) mice 4–5 months of age (Fig. 8a) oligomeric Aβ₄₂ enhanced PPF at the SC-CA1 synapse (Fig. 8b, c), implicating that deleting Efr3a at the postsynaptic site did not influence Aβ-induced Pr reduction. The minimal effect of deleting Efr3a in the CA1 area on Aβ-induced inhibition of Pr at the SC-CA1 synapse was further confirmed by directly examining the RRP size and Pr using the repetitive stimulation protocol (Fig. 8d, e). In CA3-specific Efr3a conditional knockout (CA3-Efr3a cKO) mice 4–5 months of age (Fig. 8f), however, oligomeric Aβ₄₂ was no longer effective in increasing PPF (Fig. 8g, h) and inhibiting Pr (Fig. 8i, j) at the SC-CA1 synapse, indicating that knocking out Efr3a at the presynaptic site rescued Aβ-induced Pr decrease. We then investigated whether selectively deleting Efr3a in the CA1 or CA3 area in APP/PS1 mice regulated Pr at the SC-CA1 synapse. In comparison to control mice, enhanced PPF (Fig. 8k, l) and decreased Pr (Fig. 8m, n) at the SC-CA1 synapse typical of 6–7-month-old APP/PS1 mice were still observed in age-matched APP/PS1 mice deleted for Efr3a in the CA1 area. In 6–7-month-old APP/PS1 mice deleted for Efr3a in the CA3 area, however, enhanced PPF and decreased Pr at the SC-CA1 synapse were restored to the control level (Fig. 8o–r). These results demonstrate that selectively deleting Efr3a at the presynaptic site of the SC-CA1 synapse effectively rescues the reduced Pr in APP/PS1 mice. Taken together, these data prove that inhibiting Aβ-induced presynaptic PIP₂ depletion efficaciously reduces Aβ-induced Pr reduction.

The SC-CA1 synapse plays an essential role in learning and memory. Therefore, it will be interesting to explore whether lowering Aβ-induced presynaptic PIP₂ depletion in the CA3 area improves the cognitive function in AD. To this end, we first investigated whether Aβ-induced presynaptic defect affected long-term potentiation (LTP) at the SC-CA1 synapse in WT and Efr3a conditional knockout mice (Fig. 9). A high-frequency train stimulation (3 bursts of 20 pulses at 100 Hz separated by 1.5 s) induced an apparent LTP of field excitatory postsynaptic potentials (fEPSPs) in WT mice (Fig. 9a–c), which had a robust presynaptic element as assessed by PPF (Fig. 9d–f) and was dramatically decreased in the presence of oligomeric Aβ₄₂ (Fig. 9a–c). Blocking mGluR5 with MTEP or deleting Efr3a in CA3 but not CA1 areas significantly ameliorated Aβ-induced LTP impairment (Fig. 9a–c), accompanying a decreased PPF following LTP induction (Fig. 9d–f). These results indicate that decreasing presynaptic GPCR responsiveness prevents oligomeric Aβ₄₂-induced LTP impairment.

We next studied if deleting Efr3a in the CA3 area ameliorated LTP impairment and improved cognitive function in APP/PS1 mice. As expected, the high-frequency train stimulation induced a dramatically decreased LTP in 6–7-month-old APP/PS1 mice in comparison to WT animals of the same age (Fig. 10a–c). Deleting Efr3a in the CA3 but not in the CA1 area significantly restored the diminished LTP in APP/PS1 mice (Fig. 10a–c). These data indicate that selectively knocking out presynaptic Efr3a enhanced the SC-CA1 synapse plasticity in APP/PS1 mice, suggesting that decreasing the presynaptic GPCR sensitivity may improve the cognitive function of APP/PS1 mice. By comparing the performance of 6–7-month-old APP/PS1 mice in the Morris water maze (MWM) task to that of age-matched WT animals, we were able to reliably separate APP/PS1 mice from their WT controls in the escape latency and target time plots (Fig. 10d, e). Deleting Efr3a in the CA3 but not in the CA1 area in 6–7-month-old APP/PS1 mice significantly shortened the escape latency and increased the time spent in the target quadrant to the levels of WT mice in the MWM test (Fig. 10d, e). These results implicate that deleting Efr3a in the CA3 area improves spatial learning and memory in APP/PS1 mice.

Discussion

Our work demonstrates a key presynaptic target of pathogenic Aβ in early AD and clarify Aβ-induced depletion of PIP₂ underlies Pr reduction at an excitatory synapse in the hippocampus. Aβ-induced activation of presynaptic mGluR5 depletes membrane PIP₂ in axons, which in turn disrupts neurotransmitter release. Notably, reducing Aβ-induced PIP₂ depletion in the CA3 area augments release Pr at the SC-CA1 synapse and enhances spatial learning and memory in APP/PS1 mice. As this Aβ-induced Pr reduction precedes synapse and neuronal loss, controlling the CA3 PIP₂ level may become an effective way of preventing AD progression.

We found that reduced mEPSC frequency in CA1 pyramidal neurons was a robust hallmark in 6–7-month-old APP/PS1 mice. CA1 pyramidal neurons mainly receive excitatory inputs from CA3 pyramidal neurons and layer III pyramidal neurons in the entorhinal cortex. Thus, Aβ-induced changes in mEPSC frequency in CA1 pyramidal neurons may be attributed to presynaptic alterations in both glutamatergic terminals forming synapses on CA1 pyramidal neurons. Although we are not able to rule out the involvement of the excitatory inputs from the entorhinal cortex completely, the decrease in mEPSC frequency may primary be associated with Aβ-induced suppression of glutamate release from the SC. Reducing the Aβ sensitivity in CA3 pyramidal neurons not only increased the amplitude of SC-CA1 EPSCs but also restored the diminished mEPSC frequency in CA1 pyramidal neurons (Supplementary Fig. 15). Because both areas CA3 and CA1 are critical in encoding memory sequences, our results implicate that reduced glutamate release at the SC-CA1 synapse represents a crucial early event leading to cognitive decline in AD.

Our results showed that 400 nM oligomeric Aβ₄₂ reduced glutamate release from the SC in WT mice to a level similar to that in 6–7-month-old APP/PS1 and 4-month-old 5xFAD mice. In another APP/PS1 transgenic line APP_Swe/PS1_M146L, however, there is no reduction in Pr⁴⁹. One possible explanation is that the effective concentration of oligomeric Aβ has to rise significantly at the target synapse to suppress Pr in early AD. Although in no way could we fully determine the effective concentration of locally distributed soluble Aβ, the concentration of the most toxic oligomeric Aβ₄₂ has to rise to hundreds of nanomolar level to inhibit Pr. A moderate increase of Aβ to the picomolar level¹⁹, or even 25-fold increase of Aβ₄₂ found in 4-month-old APP/PS1 mice⁵⁰, usually cause an increase in Pr at the SC-CA1 synapse, leading to enhanced basal synaptic transmission¹⁹ that we also observed with 20 nM oligomeric Aβ₄₂. A moderate increase in intracellular Aβ oligomers can also increase evoked responses by Ca²⁺-dependent insertion of GluA1 subunits⁵¹. A further increase in Aβ oligomers to the low micromolar level, however, induces mostly postsynaptic depression and loss of dendritic spines^8,9. Thus, although we are not in a position to define what the pathological level of oligomeric Aβ is⁸, we believe the early presynaptic deficit in AD requires elevation of soluble Aβ at least in the high nanomolar range. On the other hand, the facilitating effect of small increases in Aβ within a physiological range on presynaptic neurotransmitter release represents a physiological function of oligomeric Aβ¹⁹.

The probability of neurotransmitter release is tightly associated with Ca²⁺-dependent synaptic vesicle fusion in the presynaptic terminal. In contrast to enhancing Pr by physiological range of Aβ via increasing presynaptic Ca²⁺, our results implicate that high nanomolar Aβ mainly suppresses Pr via reducing PIP₂-dependent presynaptic vesicle fusion. In cultured neurons, 200 nM Aβ can also enhance Pr⁴², suggesting that Aβ may interfere with presynaptic function in a PIP₂-independent manner. However, Aβ-induced Pr suppression may not be apparent in other central synapses (e.g., the recurrent inputs to CA3 pyramidal neurons)⁵². It will be interesting to investigate if these synapses lack such a mechanism or a compensatory process engages following Aβ accumulation in future studies. Reduced Pr exists in other AD models. Conditionally knocking out presenilins 1 and 2 in CA3 pyramidal neurons significantly inhibited Ca²⁺ release from ryanodine receptors, thus reducing Pr at the SC-CA1 synapse⁵³. Our results established that maintaining the membrane PIP₂ at a relatively high level in axons is critical in preserving a proper Pr (Fig. 4). A low PIP₂ level in the brain is a hallmark for both aging and AD animals^38,54. However, in normal aging mice Pr is not altered. This could be due to a greater PIP₂ hydrolysis initiated by oligomeric Aβ in AD than that in normal aging animals with a low PLC expression level⁵⁵. PIP₂ interacts with synaptotagmin 1 and Munc13–2^46,56 to control exocytosis. Future studies are required to address the detailed mechanism underlying PIP₂-dependent Pr regulation in AD.

One of the major findings of our study is that inhibiting mGluR5 rescued the presynaptic deficit in early AD. Genetically deleting³⁶ or pharmacologically inhibiting^35,37 mGluR5 has been shown to significantly improve cognitive impairment in AD mice, as mGluR5 may function as an Aβ receptor or co-receptor with PrP^C in both APP overexpression and knock-in mouse models of AD^32,33,34. Our results clearly show that presynaptic mGluR5-mediated PIP₂ hydrolysis requires PrP^C as well. Although group I mGluRs are expressed in both presynaptic and postsynaptic loci⁵⁷, mGluR5 density is higher in dendrites than in axons (Fig. 6a, b). However, high nanomolar Aβ oligomers deplete PIP₂ to a greater level in axons than in dendrites. Although we do not know exactly what causes this differentially regulation of PIP₂ in neurites, we nevertheless speculate that this may be due to a combination of enhanced mGluR5 receptor responsiveness and downstream process efficacy in axons.

Functionally, we believe presynaptic mGluR5 may be a main target of soluble Aβ in early AD based on the following observations, although postsynaptic mGluR5 has been shown to be involved in Aβ-induced suppression of LTP and enhancement of LTD^29,30,58. First, blocking mGluR5 did not restore nanomolar oligomeric Aβ₄₂-induced suppression of mEPSC amplitude, implicating that activation of postsynaptic mGluR5 is not involved in the oligomeric Aβ-induced mild postsynaptic defect. Although the detailed mechanisms require further examination, removal of surface AMPA and NMDA receptors may account for this oligomeric Aβ-induced reduction of mEPSC amplitude^58,59. On the other hand, micromolar oligomeric Aβ induce mature spine loss via mGluR5-mediated downregulation of CaMKII activity in the postsynaptic site in APP knock-in mouse model of AD⁶⁰. Second, selectively decreasing the sensitivity of presynaptic GPCRs ameliorated oligomeric Aβ-induced suppression of Pr and a form of LTP with a notable presynaptic element, thus ruling out indirect presynaptic effects due to activation of postsynaptic mGluR5. A presynaptically-expressed SC-CA1 LTP can be induced by 200 Hz^61,62 or multiple trains of 100 Hz^63,64 tetanus and may require postsynaptically-activated Ca²⁺ influx through L-type calcium channels and NMDA receptors^61,62. Although we do not know the postsynaptic mechanism underlying this presynaptically-expressed LTP, our results establish that Aβ oligomers suppress this LTP through presynaptic mGluR5.

In the current study, we controlled membrane PIP₂ hydrolysis by manipulating the Efr3a level. It has long been known that Efr3a functions as an adaptor in the type IIIɑ phosphatidylinositol 4-kinase (PI4KIIIα) complex^65,66. Although PI4KIIIα is responsible for generating phosphatidylinositol 4-phosphate (PI4P), the precursor of PIP₂, reducing Efr3a or PI4KIIIα induces minimal changes in the membrane PIP₂ level^65,66. Knocking out the type 1γ phosphatidylinositol phosphate kinase, the main enzyme responsible for PIP₂ synthesis at synapses, however, significantly depletes membrane PIP₂, leading to decreased transmitter release⁶⁷. Efr3a may also control the responsiveness of other GPCRs that are putative Aβ receptors such as the type-1 angiotensin II receptor⁴⁸. Thus, controlling the expression level of Efr3a may regulate a series of GPCRs that are targets of Aβ, though we believe mGluR5 is the main target of pathological level of Aβ. Nevertheless, antagonizing Efr3a⁶⁸ may provide a more robust way to treat AD rather than targeting only one type of GPCR.

Methods

Animals

All procedures were carried out in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the Animal Advisory Committee at Zhejiang University. B6, APP/PS1 double-transgenic, 5xFAD, M146V, Grik4-cre, and Camk2a-creERT2 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Efr3a double-flox (Efr3a^f/f), Efr3a^+/− heterozygotes, Efr3a^f/f-Grik4-cre, Efr3a^f/f-Camk2a-creERT2, APP/PS1-Efr3a^f/f-Grik4-cre, APP/PS1-Grik4-cre, APP/PS1-Efr3a^f/f-Camk2a-creERT2, and APP/PS1-Camk2a-creERT2 animals were obtained by heterozygous mating. Tamoxifen was intraperitoneally (i.p.) injected once a day for 5 consecutive days at a dose of 100 mg kg⁻¹ to induce cre recombinase expression in the creER lines. For behavioral experiments, only male mice were used. The mouse genotypes were identified by PCR using genomic DNA from mouse tails and embryo tissues. Primers and detailed protocols are available in the Supplementary Table 1 and Supplementary Methods. LY-411575 was used to inhibit γ-secretase. The protocols of LY-411575 preparation and treatment are detailed in Supplementary Table 2 and Supplementary Methods.

Antibodies and drugs

The following commercially available antibodies were used: rabbit anti-amyloid precursor protein, C-terminal fragments (anti-CTFs), purified mouse anti-β-Amyloid, 1–16 (6E10), rabbit anti-Efr3a, mouse anti-PLCβ1, D-8, mouse anti-PLCβ4, A-8, mouse anti-β-actin, and HRP-conjugated secondary antibodies were used in Western blotting; mouse anti-PIP₂ antibody, chicken anti-neurofilament-L (anti-NF), rabbit anti-MAP2 antibody, rabbit anti-mGluR5 (extracellular), mouse anti-MAP2, purified anti-CD230 (Prion) antibody, and Alexa Fluor-conjugated secondary antibodies (488 donkey anti-rabbit, 546 donkey anti-mouse, 546 donkey anti-rabbit, 405 goat anti-rabbit IgG H&L, 488 goat anti-mouse IgM mu chain and 647 goat anti-chicken IgY H&L) were used in immunocytochemistry. The antibody information is detailed in Supplementary Methods.

Tamoxifen was dissolved in 100% ethanol; PI(4,5)P₂ diC8 was dissolved in the electrode solution; DHPG, MTEP, and U73122 were dissolved in the bath solution or culture medium. The final concentrations of DMSO did not exceed 0.1% throughout the study. The protocol of oligomeric Aβ₄₂ preparation and drug information are detailed in Supplementary Table 2 and Supplementary Methods.

Golgi staining

Golgi staining was carried out using an FD Rapid GolgiStain Kit according to the manufacturer’s instructions (see details in Supplementary Methods).

Slice recording

Briefly, mice (4- or 6–7-month-old) were anesthetized with isoflurane and decapitated, and transverse slices of hippocampus (300 μm for whole-cell recording or 350 μm for fEPSP recording) were cut with a tissue slicer (VT 1200S, Leica) in oxygenated ACSF (for whole-cell recording) or in oxygenated cutting solutions (for fEPSP recording). Whole-cell recordings were performed on CA1 pyramidal neurons^69,70. mEPSCs signals were recorded at −70 mV in ACSF containing 0.5 μM tetrodotoxin (TTX) and 10 μM bicuculline. Evoked EPSCs were elicited in the presence of 10 μM bicuculline using a bipolar stimulating electrode placed in stratum radiatum 300 μm away from the recording site. PPF experiments were carried out by delivering a pair of stimuli with an interval of 50 ms. To estimate the RRP size and release Pr, a repeated 20 Hz train stimulation protocol was used to evoke 40 EPSCs. The RRP size was calculated by linear interpolating the linear portion of the cumulative EPSC amplitude plot to virtual stimulus 0. The release Pr was calculated as the mean amplitude of the 1st EPSC during the repeated train stimulations divided by the RRP size. fEPSPs were elicited by stimulating the SC and recording with a borosilicate glass electrode filled with ACSF placed in CA1 stratum radiatum. LTP was induced by 3 bursts of 20 pulses at 100 Hz separated by 1.5 s. Detailed protocols are available in Supplementary Methods.

Cell culture

Primary hippocampal neuron cultures were prepared from embryonic day 18 (E18) mice⁶⁹. Briefly, embryos were removed from maternal mice anesthetized with isoflurane and euthanized by decapitation. Hippocampi were dissected in HBSS, followed by a digestion with 0.25% w/v trypsin. Neurons were centrifuged (1000 × g for 5 min) and resuspended in neurobasal medium containing 2% B27 serum-free supplement, 1% v/v penicillin/streptomycin (P/S), 0.5 mM glutamine, and 10 μM glutamate. Dissociated cells were then plated with appropriate densities in culture plates or dishes pre-coated with PDL. Cultures were kept at 37 °C in a 5% v/v CO₂ humidified incubator. Thereafter, one third to half of the medium was replaced twice a week (see details in Supplementary Methods).

Individually inhabited hippocampal neurons were grown on collagen-PDL islands. Briefly, 6.5 mm Transwell® inserts in 24 well plates were coated with PDL (12 h before culture), and coverslips were sprayed with island substrate solution containing 1 mg ml⁻¹ PDL and 3 mg ml⁻¹ rat tail collagen (3 h before culture). Dissociated cells were then plated at a density of 2000 cells per cm² onto coverslips in 24-well plates (for micro-island culture) or at a density of 50,000 cells per cm² in Transwell® inserts in 24-well plates (as high density neuronal feeder layer). After an adherence time of 4 h, the transwell inserts with neurons (high density) were placed into 24-well plates with neurons on coverslips (low density). Procedures for maintaining cultured islands were similar to those for primary hippocampal neuron cultures (see details in Supplementary Methods).

Astrocyte cultures were prepared from 0 to 1-day-old (P0–1) mice⁶⁹. Cortices were dissected from 0 to 1-day-old mice and digested with 0.25% w/v trypsin in DMEM. Cells were allowed to grow for at least 7 days at 37 °C with 5% CO₂, and a complete medium change was performed every other day. At confluence after DIV8-10, cultures were shaken, and then incubated with 20 μM cytosine-1-β-D-arabinofuranosid (see details in Supplementary Methods).

FM1-43 loading and synaptic vesicle detection

Cultured neurons (DIV14) were transferred into a standard bath solution with 10 μM DNQX and 40 μM D-AP5. Neurons were then incubated with 5 μM FM1-43 in a hyperkalemic bath solution for 90 s. FM1-43 was then washout followed by adding ADVASEP-7 to reduce background fluorescence. Images were taken by a confocal laser-scanning microscope (Nikon A1). FM1-43-loaded vesicles were viewed through a 40X oil-immersion objective and images were acquired at a resolution of 1024 × 1024 pixel at RT (see details in Supplementary Methods).

Western blotting

Hippocampi were obtained and homogenized using a chilled Vibrahomogenizer (Vibra cell, SONICS) in 2 ml of RIPA buffer. The lysate was then centrifuged, and the supernatant collected for Western blot analysis. Proteins were separated on SDS-PAGE under denaturing conditions (for Efr3a, 10–15% Mini-PROTEAN TGX Gels; for Aβ and CTFs, 16.5% Tris-Tricine Gels) and transferred to polyvinylidene fluoride (PVDF) microporous membrane (Millipore). The membranes were then blocked with 5% skim milk-TBS (for Aβ) or 0.35% gelatin-TBST (for other proteins), and incubated with the primary antibodies followed by HRP-conjugated secondary antibodies. Protein bands were then visualized using the ECL western blotting detection substrate and analyzed with ImageJ software. Detailed protocols are available in Supplementary Methods.

Cultured neuron recording

Single-cell micro-island neuron cultures at DIV14 were used for recording. Neurons were voltage clamped at −70 mV with a Heka EPC 10 amplifier and mEACs were recorded at 32 °C in bath solution containing 0.5 μM TTX and 10 μM bicuculline. Individual events were counted and analyzed with MiniAnalysis software (see details in Supplementary Methods).

Lipid strip assay

PIP Strips^TM membranes were used for the anti-PIP₂ antibody specificity measurement according to the manufacturer’s instructions. Detailed protocols are available in Supplementary Methods.

ELISA PIP₂ assay

Mass ELISA Kit K-4500 from Echelon Biosciences was used to determine PIP₂ levels in hippocampi from WT and APP/PS1 mice (6–7-month-old), and in primary cultured hippocampal neurons from WT and Efr3a^+/− mice according to the manufacturer’s instructions⁵⁵. Detailed protocols are available in Supplementary Methods.

Immunocytochemistry

Immunofluorescence staining was carried out in cultured neurons at DIV14⁶⁹. Briefly, neurons were fixed and permeabilized followed by incubation with primary antibodies and the appropriate secondary antibodies. For co-staining with mGluR5, neurons were permeabilized before the secondary blocking, and the primary antibodies were used. Fluorescent images were acquired through a 60X oil-immersion objective using a Nikon A1 confocal laser-scanning microscope. Neuronal images were analyzed using MetaMorph with customized filter sets. Detailed protocols are available in Supplementary Methods.

Lentivirus-shRNA infection

Cultured hippocampal neurons at DIV7 were infected with lentivirus carrying DsRed-PLCβ1-shRNA (PLCβ1-shRNA) or DsRed-PLCβ4-shRNA (PLCβ4-shRNA) to knock down PLCβ1 or PLCβ4. Lentivirus expressing DsRed-scramble-shRNA was used as control. Neurons were then treated with Aβ or DMSO, and subsequently used for Western blotting (interference efficiency detection) or immunofluorescence staining. Detailed protocols are available in Supplementary Methods.

Ca²⁺ imaging

Changes in [Ca²⁺]_i were measured in astrocytes (WT and Efr3a^+/−) using the calcium-sensitive fluorescent dye Fluo-4. The astrocytes were washed with Krebs buffer and incubated with 4 μM Fluo-4 in Krebs buffer. Fluo-4 loaded astrocytes were excited at 488 nm and fluorescence emission was detected at 525 nm. The images [(baseline data (F₀) and treated data (F)] were taken through a 60X oil-immersion objective by a Nikon A1 confocal laser-scanning microscope. The relative Fluo-4 fluorescent signals expressed in arbitrary units (F/F₀) were analyzed for individual cells using MetaMorph with a fixed set of parameters (see details in Supplementary Methods).

MWM test

The MWM tests were performed in a circular tank filled with opaque water at 25 °C. Twenty-four hours before the acquisition test, a visible platform task was tested in each quadrant of the tank. In the hidden platform acquisition test, mice could swim freely to search for the escape platform within 60 s. The time taken to reach the platform was recorded as the escape latency. Mice were allowed to stay on the platform for 10 s after the hidden platform was found. The same animal was then released from a new insertion point 4 min after the previous trial. The experiment was repeated four times per mouse each day for 5 days. Twenty-four hours after the hidden platform acquisition test, probe trials were conducted by removing the platform. The numbers of entries into the area where the original platform was located and crossings over the original platform were recorded. The data were analyzed by the WaterMaze Software (Actimetrics, INC.). Detailed protocols are available in Supplementary Methods.

Statistics and reproducibility

GraphPad Prism (Version 5.01, Graph-Pad Software Inc.) was used for data display and statistical analysis. We did not predetermine sample sizes. We used Kolmogorov-Smirnov normality test to determine most data are Gaussian distributed. Significance is reported as P < 0.05, and data were expressed as mean ± SEM. Two-tailed Student t test, one-way ANOVA followed by a post hoc multiple comparison analysis based on the Dunnett’s Method, two-way ANOVA followed by a post hoc Bonferroni test, or two-way RM ANOVA followed by a post hoc Bonferroni test were used to determine significant levels between treatments and controls. Distributions of mEPSC amplitudes and interevent intervals were compared using Kolmogorov-Smirnov test.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files. The source data underlying Figs. 1–10 and Supplementary Figs 1–15 are provided as two Source Data files.

References

Ballard, C. et al. Alzheimer’s disease. Lancet 377, 1019–1031 (2011).
Article Google Scholar
Hardy, J. & Selkoe, D. J. The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics. Science 297, 353–356 (2002).
Article ADS CAS Google Scholar
Polanco, J. C. et al. Amyloid-beta and tau complexity—towards improved biomarkers and targeted therapies. Nat. Rev. Neurol. 14, 22–39 (2018).
Article CAS Google Scholar
Selkoe, D. J. Alzheimer’s disease is a synaptic failure. Science 298, 789–791 (2002).
Article ADS CAS Google Scholar
Terry, R. D. et al. Physical basis of cognitive alterations in Alzheimer’s disease: synapse loss is the major correlate of cognitive impairment. Ann. Neurol. 30, 572–580 (1991).
Article CAS Google Scholar
Herms, J. & Dorostkar, M. M. Dendritic Spine Pathology in Neurodegenerative Diseases. Annu. Rev. Pathol. 11, 221–250 (2016).
Article CAS Google Scholar
Canter, R. G., Penney, J. & Tsai, L. H. The road to restoring neural circuits for the treatment of Alzheimer’s disease. Nature 539, 187–196 (2016).
Article ADS Google Scholar
Palop, J. J. & Mucke, L. Amyloid-beta-induced neuronal dysfunction in Alzheimer’s disease: from synapses toward neural networks. Nat. Neurosci. 13, 812–818 (2010).
Article CAS Google Scholar
Benilova, I., Karran, E. & De Strooper, B. The toxic Abeta oligomer and Alzheimer’s disease: an emperor in need of clothes. Nat. Neurosci. 15, 349–357 (2012).
Article CAS Google Scholar
De Strooper, B. & Karran, E. The Cellular Phase of Alzheimer’s disease. Cell 164, 603–615 (2016).
Article Google Scholar
Forner, S., Baglietto-Vargas, D., Martini, A. C., Trujillo-Estrada, L. & LaFerla, F. M. Synaptic impairment in Alzheimer’s disease: a dysregulated symphony. Trends Neurosci. 40, 347–357 (2017).
Article CAS Google Scholar
Koffie, R. M. et al. Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synapse loss near senile plaques. Proc. Natl Acad. Sci. USA 106, 4012–4017 (2009).
Article ADS CAS Google Scholar
Pickett, E. K. et al. Non-fibrillar oligomeric amyloid-beta within synapses. J. Alzheimers Dis. 53, 787–800 (2016).
Article CAS Google Scholar
Fu, A. K. & Ip, N. Y. Regulation of postsynaptic signaling in structural synaptic plasticity. Curr. Opin. Neurobiol. 45, 148–155 (2017).
Article CAS Google Scholar
Sheng, M., Sabatini, B.L. & Sudhof, T.C. Synapses and Alzheimer’s disease. Cold Spring Harb. Perspect. Biol. 4, a005777 (2012).
Parodi, J. et al. Beta-amyloid causes depletion of synaptic vesicles leading to neurotransmission failure. J. Biol. Chem. 285, 2506–2514 (2010).
Article CAS Google Scholar
Ting, J. T., Kelley, B. G., Lambert, T. J., Cook, D. G. & Sullivan, J. M. Amyloid precursor protein overexpression depresses excitatory transmission through both presynaptic and postsynaptic mechanisms. Proc. Natl Acad. Sci. USA 104, 353–358 (2007).
Article ADS CAS Google Scholar
Cirrito, J. R. et al. In vivo assessment of brain interstitial fluid with microdialysis reveals plaque-associated changes in amyloid-beta metabolism and half-life. J. Neurosci. 23, 8844–8853 (2003).
Article CAS Google Scholar
Abramov, E. et al. Amyloid-beta as a positive endogenous regulator of release probability at hippocampal synapses. Nat. Neurosci. 12, 1567–1576 (2009).
Article CAS Google Scholar
Fogel, H. et al. APP homodimers transduce an amyloid-beta-mediated increase in release probability at excitatory synapses. Cell Rep. 7, 1560–1576 (2014).
Article CAS Google Scholar
Bobich, J. A., Zheng, Q. & Campbell, A. Incubation of nerve endings with a physiological concentration of Abeta1-42 activates CaV2.2(N-Type)-voltage operated calcium channels and acutely increases glutamate and noradrenaline release. J. Alzheimers Dis. 6, 243–255 (2004).
Article CAS Google Scholar
Lazarevic, V. et al. Physiological concentrations of amyloid beta regulate recycling of synaptic vesicles via alpha7 acetylcholine receptor and CDK5/calcineurin signaling. Front. Mol. Neurosci. 10, 221 (2017).
Article Google Scholar
Puzzo, D. et al. Picomolar amyloid-beta positively modulates synaptic plasticity and memory in hippocampus. J. Neurosci. 28, 14537–14545 (2008).
Article CAS Google Scholar
Moreno, H. et al. Synaptic transmission block by presynaptic injection of oligomeric amyloid beta. Proc. Natl Acad. Sci. USA 106, 5901–5906 (2009).
Article ADS CAS Google Scholar
Nimmrich, V. et al. Amyloid beta oligomers (A beta(1-42) globulomer) suppress spontaneous synaptic activity by inhibition of P/Q-type calcium currents. J. Neurosci. 28, 788–797 (2008).
Article CAS Google Scholar
Yang, Y. et al. Amyloid-beta oligomers may impair SNARE-mediated exocytosis by direct binding to syntaxin 1a. Cell Rep. 12, 1244–1251 (2015).
Article CAS Google Scholar
Hermann, D. et al. Synthetic Abeta oligomers (Abeta(1-42) globulomer) modulate presynaptic calcium currents: prevention of Abeta-induced synaptic deficits by calcium channel blockers. Eur. J. Pharmacol. 702, 44–55 (2013).
Article CAS Google Scholar
Russell, C. L. et al. Amyloid-beta acts as a regulator of neurotransmitter release disrupting the interaction between synaptophysin and VAMP2. PLoS ONE 7, e43201 (2012).
Article ADS CAS Google Scholar
Li, S. et al. Soluble oligomers of amyloid Beta protein facilitate hippocampal long-term depression by disrupting neuronal glutamate uptake. Neuron 62, 788–801 (2009).
Article CAS Google Scholar
Shankar, G. M. et al. Amyloid-beta protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory. Nat. Med. 14, 837–842 (2008).
Article CAS Google Scholar
Mezler, M., Barghorn, S., Schoemaker, H., Gross, G. & Nimmrich, V. A beta-amyloid oligomer directly modulates P/Q-type calcium currents in Xenopus oocytes. Br. J. Pharmacol. 165, 1572–1583 (2012).
Article CAS Google Scholar
Hu, N. W. et al. mGlu5 receptors and cellular prion protein mediate amyloid-beta-facilitated synaptic long-term depression in vivo. Nat. Commun. 5, 3374 (2014).
Article Google Scholar
Renner, M. et al. Deleterious effects of amyloid beta oligomers acting as an extracellular scaffold for mGluR5. Neuron 66, 739–754 (2010).
Article CAS Google Scholar
Um, J. W. et al. Metabotropic glutamate receptor 5 is a coreceptor for Alzheimer abeta oligomer bound to cellular prion protein. Neuron 79, 887–902 (2013).
Article CAS Google Scholar
Haas, L. T. et al. Silent allosteric modulation of mGluR5 maintains glutamate signaling while rescuing Alzheimer’s mouse phenotypes. Cell Rep. 20, 76–88 (2017).
Article CAS Google Scholar
Hamilton, A., Esseltine, J. L., DeVries, R. A., Cregan, S. P. & Ferguson, S. S. Metabotropic glutamate receptor 5 knockout reduces cognitive impairment and pathogenesis in a mouse model of Alzheimer’s disease. Mol. Brain 7, 40 (2014).
Article Google Scholar
Hamilton, A. et al. Chronic pharmacological mGluR5 inhibition prevents cognitive impairment and reduces pathogenesis in an Alzheimer disease mouse model. Cell Rep. 15, 1859–1865 (2016).
Article CAS Google Scholar
Berman, D. E. et al. Oligomeric amyloid-beta peptide disrupts phosphatidylinositol-4,5-bisphosphate metabolism. Nat. Neurosci. 11, 547–554 (2008).
Article CAS Google Scholar
Zhu, L. et al. Phospholipid dysregulation contributes to ApoE4-associated cognitive deficits in Alzheimer’s disease pathogenesis. Proc. Natl Acad. Sci. USA 112, 11965–11970 (2015).
Article ADS CAS Google Scholar
McIntire, L. B. et al. Reduction of synaptojanin 1 ameliorates synaptic and behavioral impairments in a mouse model of Alzheimer’s disease. J. Neurosci. 32, 15271–15276 (2012).
Article CAS Google Scholar
Merino-Serrais, P., Knafo, S., Alonso-Nanclares, L., Fernaud-Espinosa, I. & DeFelipe, J. Layer-specific alterations to CA1 dendritic spines in a mouse model of Alzheimer’s disease. Hippocampus 21, 1037–1044 (2011).
Article CAS Google Scholar
Park, J., Jang, M. & Chang, S. Deleterious effects of soluble amyloid-beta oligomers on multiple steps of synaptic vesicle trafficking. Neurobiol. Dis. 55, 129–139 (2013).
Article CAS Google Scholar
Woo, J. A. et al. RanBP9 at the intersection between cofilin and Abeta pathologies: rescue of neurodegenerative changes by RanBP9 reduction. Cell Death Dis. 6, 1676 (2015).
Article CAS Google Scholar
Harris, J. A. et al. Transsynaptic progression of amyloid-beta-induced neuronal dysfunction within the entorhinal-hippocampal network. Neuron 68, 428–441 (2010).
Article CAS Google Scholar
Garcia-Alloza, M. et al. Existing plaques and neuritic abnormalities in APP:PS1 mice are not affected by administration of the gamma-secretase inhibitor LY-411575. Mol. Neurodegener. 4, 19 (2009).
Article Google Scholar
Lee, H. K. et al. Dynamic Ca2+-dependent stimulation of vesicle fusion by membrane-anchored synaptotagmin 1. Science 328, 760–763 (2010).
Article ADS CAS Google Scholar
Falkenburger, B. H., Jensen, J. B., Dickson, E. J., Suh, B. C. & Hille, B. Phosphoinositides: lipid regulators of membrane proteins. J. Physiol. 588, 3179–3185 (2010).
Article CAS Google Scholar
Bojjireddy, N., Guzman-Hernandez, M. L., Reinhard, N. R., Jovic, M. & Balla, T. EFR3s are palmitoylated plasma membrane proteins that control responsiveness to G-protein-coupled receptors. J. Cell Sci. 128, 118–128 (2015).
Article CAS Google Scholar
Gelman, S., Palma, J., Tombaugh, G. & Ghavami, A. Differences in synaptic dysfunction between rTg4510 and APP/PS1 mouse models of Alzheimer’s disease. J. Alzheimers Dis. 61, 195–208 (2018).
Article CAS Google Scholar
Cummings, D. M. et al. First effects of rising amyloid-beta in transgenic mouse brain: synaptic transmission and gene expression. Brain 138, 1992–2004 (2015).
Article Google Scholar
Whitcomb, D. J. et al. Intracellular oligomeric amyloid-beta rapidly regulates GluA1 subunit of AMPA receptor in the hippocampus. Sci. Rep. 5, 10934 (2015).
Article ADS CAS Google Scholar
Viana da Silva, S. et al. Early synaptic deficits in the APP/PS1 mouse model of Alzheimer’s disease involve neuronal adenosine A2A receptors. Nat. Commun. 7, 11915 (2016).
Article ADS CAS Google Scholar
Zhang, C. et al. Presenilins are essential for regulating neurotransmitter release. Nature 460, 632–636 (2009).
Article ADS CAS Google Scholar
Di Paolo, G. & Kim, T. W. Linking lipids to Alzheimer’s disease: cholesterol and beyond. Nat. Rev. Neurosci. 12, 284–296 (2011).
Article Google Scholar
Trovo, L. et al. Low hippocampal PI(4,5)P(2) contributes to reduced cognition in old mice as a result of loss of MARCKS. Nat. Neurosci. 16, 449–455 (2013).
Article CAS Google Scholar
Walter, A.M. et al. Phosphatidylinositol 4,5-bisphosphate optical uncaging potentiates exocytosis. Elife 6, e30203 (2017).
Romano, C. et al. Distribution of metabotropic glutamate receptor mGluR5 immunoreactivity in rat brain. J. Comp. Neurol. 355, 455–469 (1995).
Article CAS Google Scholar
Hsieh, H. et al. AMPAR removal underlies Abeta-induced synaptic depression and dendritic spine loss. Neuron 52, 831–843 (2006).
Article CAS Google Scholar
Snyder, E. M. et al. Regulation of NMDA receptor trafficking by amyloid-beta. Nat. Neurosci. 8, 1051–1058 (2005).
Article CAS Google Scholar
Zhang, H. et al. Neuronal store-operated calcium entry and mushroom spine loss in amyloid precursor protein knock-in mouse model of Alzheimer’s disease. J. Neurosci. 35, 13275–13286 (2015).
Article CAS Google Scholar
Bayazitov, I. T., Richardson, R. J., Fricke, R. G. & Zakharenko, S. S. Slow presynaptic and fast postsynaptic components of compound long-term potentiation. J. Neurosci. 27, 11510–11521 (2007).
Article CAS Google Scholar
Zakharenko, S. S., Zablow, L. & Siegelbaum, S. A. Visualization of changes in presynaptic function during long-term synaptic plasticity. Nat. Neurosci. 4, 711–717 (2001).
Article CAS Google Scholar
Emptage, N. J., Reid, C. A., Fine, A. & Bliss, T. V. Optical quantal analysis reveals a presynaptic component of LTP at hippocampal Schaffer-associational synapses. Neuron 38, 797–804 (2003).
Article CAS Google Scholar
Enoki, R., Hu, Y. L., Hamilton, D. & Fine, A. Expression of long-term plasticity at individual synapses in hippocampus is graded, bidirectional, and mainly presynaptic: optical quantal analysis. Neuron 62, 242–253 (2009).
Article CAS Google Scholar
Nakatsu, F. et al. PtdIns4P synthesis by PI4KIIIalpha at the plasma membrane and its impact on plasma membrane identity. J. Cell Biol. 199, 1003–1016 (2012).
Article CAS Google Scholar
Wu, X. et al. Structural insights into assembly and regulation of the plasma membrane phosphatidylinositol 4-kinase complex. Dev. Cell 28, 19–29 (2014).
Article CAS Google Scholar
Di Paolo, G. et al. Impaired PtdIns(4,5)P2 synthesis in nerve terminals produces defects in synaptic vesicle trafficking. Nature 431, 415–422 (2004).
Article ADS Google Scholar
Zhang, X. et al. Downregulation of RBO-PI4KIIIalpha facilitates Abeta42 secretion and ameliorates neural deficits in Abeta42-expressing Drosophila. J. Neurosci. 37, 4928–4941 (2017).
Article CAS Google Scholar
Shen, Y. et al. Postnatal activation of TLR4 in astrocytes promotes excitatory synaptogenesis in hippocampal neurons. J. Cell Biol. 215, 719–734 (2016).
Article CAS Google Scholar
Zhou, Y. D. et al. Arrested maturation of excitatory synapses in autosomal dominant lateral temporal lobe epilepsy. Nat. Med. 15, 1208–1214 (2009).
Article CAS Google Scholar

Download references

Acknowledgements

This work was supported by the Major Research Program from the State Ministry of Science and Technology of China (2013CB530902), the National Natural Science Foundation of China grants (81770839, 81571125, 81571088, 91132712, 81300979, 81821091, 81521062, and 81221003), the 111 project, and the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2017PT31038 and 2018PT31041). We thank the excellent technical assistant of the Imaging Facility at Zhejiang University School of Medicine.

Author information

These authors contributed equally: Yang He, Mengdi Wei.

Authors and Affiliations

Department of Neurobiology, and Department of Ophthalmology of the First Affiliated Hospital, NHC and CAMS Key Laboratory of Medical Neurobiology, Zhejiang University School of Medicine, Hangzhou, 310058, China
Yang He, Mengdi Wei, Yan Wu, Huaping Qin, Weinan Li, Xiaolin Ma, Jingjing Cheng, Jinshuai Ren, Ye Shen, Binggui Sun, Yi Shen & Yu-Dong Zhou
Department of Pharmacology, NHC and CAMS Key Laboratory of Medical Neurobiology, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, Zhejiang, 310058, China
Zhong Chen
Center for Stem Cell and NanoMedicine, Laboratory for System Biology, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai, 201210, China
Fu-De Huang

Authors

Yang He
View author publications
You can also search for this author in PubMed Google Scholar
Mengdi Wei
View author publications
You can also search for this author in PubMed Google Scholar
Yan Wu
View author publications
You can also search for this author in PubMed Google Scholar
Huaping Qin
View author publications
You can also search for this author in PubMed Google Scholar
Weinan Li
View author publications
You can also search for this author in PubMed Google Scholar
Xiaolin Ma
View author publications
You can also search for this author in PubMed Google Scholar
Jingjing Cheng
View author publications
You can also search for this author in PubMed Google Scholar
Jinshuai Ren
View author publications
You can also search for this author in PubMed Google Scholar
Ye Shen
View author publications
You can also search for this author in PubMed Google Scholar
Zhong Chen
View author publications
You can also search for this author in PubMed Google Scholar
Binggui Sun
View author publications
You can also search for this author in PubMed Google Scholar
Fu-De Huang
View author publications
You can also search for this author in PubMed Google Scholar
Yi Shen
View author publications
You can also search for this author in PubMed Google Scholar
Yu-Dong Zhou
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Y.-D.Z., Yi S. and F.-D.H. designed the study. Y.-D.Z., Yi S., Y.H. and M.W. wrote the paper. Yi S., Y.H., M.W., Y.W., X.M. and W.L. analyzed the data. Y.H. and X.M. performed electrophysiology experiments. Y.H. performed synaptic vesicle detection and Ca²⁺ imaging experiments. M.W. and Y.W. did behavioral studies. M.W., H.Q., Y.W., J.C. and W.L. did immunostaining and biochemical studies. F.-D.H. engineered Efr3a transgenic mice. B.S., J.R., Z.C. and Ye S. contributed intellectually to the manuscript.

Corresponding authors

Correspondence to Fu-De Huang, Yi Shen or Yu-Dong Zhou.

Ethics declarations

Competing interests

F.-D.H. has shareholding of a company possessing part of the intellectual property raised in this study. The remaining authors declare no competing interests.

Additional information

Journal Peer Review Information: Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data for Main figures 1–10

Source data for Supplementary figures 1–15

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

He, Y., Wei, M., Wu, Y. et al. Amyloid β oligomers suppress excitatory transmitter release via presynaptic depletion of phosphatidylinositol-4,5-bisphosphate. Nat Commun 10, 1193 (2019). https://doi.org/10.1038/s41467-019-09114-z

Download citation

Received: 12 March 2018
Accepted: 21 February 2019
Published: 13 March 2019
DOI: https://doi.org/10.1038/s41467-019-09114-z

This article is cited by

Characterization of Retinal VIP-Amacrine Cell Development During the Critical Period
- Xuhong Zhang
- Xiaoyu Wang
- Jianping Tong
Cellular and Molecular Neurobiology (2024)
Toll-Like Receptor 4 Deficiency Ameliorates Propofol-Induced Impairments of Cognitive Function and Synaptic Plasticity in Young Mice
- Qiao-Ding Dai
- Kang-Song Wu
- Li-Da Su
Molecular Neurobiology (2024)
Asymmetric dysregulation of glutamate dynamics across the synaptic cleft in a mouse model of Alzheimer’s disease
- Kyle J. Brymer
- Emily P. Hurley
- Matthew P. Parsons
Acta Neuropathologica Communications (2023)
Mechanisms and pathology of protein misfolding and aggregation
- Nikolaos Louros
- Joost Schymkowitz
- Frederic Rousseau
Nature Reviews Molecular Cell Biology (2023)
Endocytosis and Alzheimer’s disease
- Łukasz Zadka
- Marta Sochocka
- Jerzy Leszek
GeroScience (2023)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.