Amyloid Beta42 oligomers up-regulate the excitatory synapses by potentiating presynaptic release while impairing postsynaptic NMDA receptors.

KEY POINTS: NMDA receptors (NMDARs) are key molecules for controlling neuronal plasticity, learning and memory processes. Their function is impaired during Alzheimer's disease (AD) but the exact consequence on synaptic function is not yet fully identified. An important hallmark of AD onset is represented by the neuronal accumulation of Amyloid Beta42 oligomers (Abeta42) that we have recently shown to be responsible for the increased intracellular Ca2+ concentration through ryanodine receptors (RyRs). Here we characterized the effects of Abeta42 on NMDA synapses showing specific pre- and post-synaptic functional changes that lead to a potentiation of basal and synchronous NMDA synaptic transmission. These overall effects can be abolished by decreasing Ca2+ release from RyRs with specific inhibitors that we propose as new pharmacological tools for AD treatment. ABSTRACT: We have recently shown that Amyloid Beta42 oligomers (Abeta42) cause calcium dysregulation in hippocampal neurons by stimulating Ca2+ release from ryanodine receptors (RyRs) and inhibiting Ca2+ entry through NMDA receptors (NMDARs). Here, we found that Abeta42 decrease the average NMDA-activated inward current and that Ca2+ entry through NMDARs is accompanied by Ca2+ release from the stores. The overall amount of intraellular Ca2+ concentration([Ca2+ ]i ) increase during NMDA application is 50% associated with RyR opening and 50% with NMDARs activation. Addition of Abeta42 does not change this proportion. We estimated the number of NMDARs expressed in hippocampal neurons and their unitary current. We found that Abeta42 decrease the number of NMDARs without altering their unitary current. Paradoxically, the oligomer increases the size of electrically evoked eEPSCs induced by NMDARs activation. We found that this is the consequence of the increased release probability (p) of glutamate and the number of release sites (N) of NMDA synapses, while the quantal size (q) is significantly decreased as expected from the decreased number of NMDARs. An increased number of release sites induced by Abeta42 is also supported by the increased size of the ready releasable pool (RRPsyn) and by the enhanced percentage of paired pulse depression (PPD). Interestingly, the RyRs inhibitor dantrolene prevents the increase of PPD induced by Abeta42 oligomers. In conclusion, Abeta42 up-regulates NMDA synaptic responses with a mechanism involving RyRs that occurs during the early stages of Alzheimer's disease (AD) onset. This suggests that new selective modulators of RyRs may be useful for designing effective therapies to treat AD patients.

Dose-dependent suppression of hippocampal contextual memory formation, place cells, and spatial engrams by the NMDAR antagonist (R)-CPP.

We recently reported that the competitive NMDAR antagonist (R,S)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP) does not suppress NMDAR-mediated field EPSPs (fEPSPNMDA) or long-term potentiation (LTP) in vitro at concentrations that block contextual conditioning in vivo. Here we tested one possible explanation for the mismatch - that the hippocampus is relatively resistant to CPP compared to other brain structures engaged in contextual fear conditioning. Using the context pre-exposure facilitation effect (CPFE) paradigm to separate the hippocampal and extra-hippocampal components of contextual learning, we found that the active enantiomer (R)-CPP suppressed the hippocampal component with an IC50 of 3.1 mg/kg, a dose that produces brain concentrations below those required to block fEPSPNMDA or LTP. Moreover, using in-vivo calcium imaging of place cells and spatial engrams to directly assess hippocampal spatial coding, we found that (R)-CPP dose-dependently reduced the development of place cells and interfered with the formation of stable spatial engrams when it was administered prior to exposing mice to a novel context. Both effects occurred at doses that interfered with freezing to context in CPFE experiments. We conclude that (R)-CPP blocks memory formation by interfering with hippocampal function, but that it does so by modulating NMDARs at sites that are not engaged in vitro in the same manner that they are in vivo - perhaps through interneuron circuits that do not contribute to fEPSPs and are not required to elicit LTP using standard induction protocols in vitro, but are essential for successful mnemonic function in vivo.