Blocking the Dkk1-LRP6 interaction prevents acute amyloid-β-driven cognitive impairment

Synapse loss driven by amyloid-β (Aβ) is an early event in Alzheimer's disease (AD). Although the mechanism by which Aβ drives synapse loss remain poorly understood data indicate that a disruption of Wnt signalling plays an important part. We have shown that Aβ exerts its effects on synapses through Dickkopf-1 (Dkk1), a secreted protein that acts upon Wnt signalling via a direct interaction with the canonical Wnt pathway co-receptor proteins, LRP5 and LRP6, preventing their interaction with the receptor Frizzled. This antagonises canonical, Wnt/β-catenin, signalling and allows concomitant activation of non-canonical signalling pathways. We contend that it is the switch from canonical to non-canonical Wnt signalling activity that drives synapse loss and subsequent cognitive impairment in AD, driven by Aβ and mediated by Dkk1. Preventing the Dkk1-LRP5/6 interaction could protect synapses and cognition against Aβ by maintaining canonical Wnt signalling. To test this, we mapped the Dkk1-LRP6 interaction by peptide array and identified a small peptide able to disrupt the Dkk1-LRP6 interaction. This Dkk1-LRP6 'disruptor' peptide dose dependently restores canonical Wnt signalling in the presence of Dkk1; blocks Dkk1-driven dendritic spine loss in primary rat cortical cultures and the accompanying increase in endogenous Aβ production; and when administered intracerebroventricularly to a rat acute Aβ model, blocks Aβ-driven cognitive impairment. These data support our contention that the ability of Aβ to induce Dkk1 and the effects of Dkk1 on LRP6 are an important element in AD aetiopathology and establish Dkk1 as a therapeutic target for protecting synapse and cognition in AD.