Skip to main content
European Commission logo print header

Molecular Mechanisms of Memory Persistence

Final Report Summary - MEMORY PERSISTENCE (Molecular Mechanisms of Memory Persistence)

Background and aims
One of the main goals of neuroscience is to understand how memories are acquired, how they can last a lifetime and how they fade away. Most of the research has focused on how memories are formed. However, the molecular mechanisms that make memory persist or fade away over time after it has been formed still remain largely unknown.
The formation of long-term memory relies on modification of the connections between neurons (synapses) as well as on alterations of synaptic strength. AMPA receptors (AMPARs) are the primary mediators of excitatory synaptic communication in the brain. The insertion and removal of AMPARs is widely accepted as one of the key final steps that bring about changes in synaptic strength and excitability. Recent studies demonstrate that maintaining a stable amount of AMPARs at the synapses is critical for memories to persist.
The focus of our study is to discover how the stability of AMPARs at synapses that form part of the circuits for learning and memory is regulated. This will allow us to understand how memories persist and how they are forgotten. We believe that the identification and description of these mechanisms will have very important implications for the treatments of diseases that affect memory, such as Alzheimer’s disease.
Various proteins interact with AMPARs to regulate their trafficking towards and away of the synaptic membranes. The molecules BRAG2 and PICK1 are particularly relevant for the removal of AMPARs from synapses, which reduces synaptic strength. BRAG2 is involved in the initiation of AMPAR internalization. PICK1, on the other hand, promotes the degradation of removed and internalised AMPARs and thus prevents that they are recycled back into the synaptic membrane. To address these questions we manipulated the endogenous levels of BRAG2 and PICK1 as well as their interaction with AMPARs in the dorsal hippocampus of rats. We then examined the effects of these manipulations on memory using a task that in which rats learn where objects are located in an open field. The memory for the object locations is acquired and stored in the hippocampus.
Experimental approach
In order to reduce the levels of PICK1 and BRAG2 in the hippocampus, we generated viruses (lentivirus and adeno-associated virus) that express RNA interference to BRAG2 or PICK1. These viruses are non-replicating, non-pathogenic and have low immunogenicity, rendering them safe to use in research. The RNA interference that the viruses express prevents the synthesis of the specific proteins and therefore reduces their endogenous levels. We injected the virus containing the specific RNA interference into the dorsal hippocampus of rats and assessed the levels of PICK1 or BRAG2 at different times after virus injection. Having determined the optimal dose and timing required for the protein knockdown, we assessed the role of this manipulation on memory performance. To complement this approach, we also examined the behavioural effect of disrupting the interaction between these proteins and AMPARs with acute injections of peptides. We used peptides that specifically interfere with the binding of BRAG2 and PICK1 to AMPARs.
Results and conclusions
We found that PICK1 knockdown did not affect memory acquisition and persistence, nor did it prevent the memory the loss induced by pepR845A, a peptide that leads to the rapid elimination of long-term memories. We then tested the effect of disrupting directly the interaction between PICK1 and AMPARs. To this end we injected the well-characterised peptide pepEVKI into the dorsal hippocampus. Our results showed that pepEVKI did not affect memory acquisition or maintenance, and also did not rescue the memory loss induced by pepR845A. Taken together these results suggest that PICK1-mediated AMPAR trafficking may not be involved in the loss of long-term object location memory.
BRAG2 knockdown did not affect memory acquisition when memory was tested one day after the end of training. Thus, we are currently testing if it can prevent memory loss induced by pepR845 or the natural memory loss that occurs over time in this object location task. We also addressed whether disrupting directly the interaction between AMPARs and BRAG2 with the interference peptide GluA23Y would prevent memory loss. We have previously shown that GluA23Y prevents the synaptic removal of AMPARs and rescues the effects of memory loss caused by pepR845A (Migues et al, 2014). We now found that infusion of GluA23Y into the hippocampus for 7 days after the end of training preserved long-term object location memories, preventing their natural decay over time; at the same time, GluA23Y did not affect the acquisition of location memories (Migues et al., 2016). These results suggest that the interaction of BRAG2 and AMPARs mediates the natural decay of long-term memories.
Overall, the results obtained in this study represent an important step towards the understanding of how memories can persist and how memories are lost. Even though not yet conclusive, the results have the potential to promote a new avenue of memory research that will lead to a more complete understanding of the central processes critical for memory, beyond the stages of acquisition and formation. In particular, our study identifies BRAG2 as an attractive target for future studies on memory persistence and forgetting, and for the development of treatments of cognitive disorders characterised by pathological memory loss.
References:
Migues P.V. et al., 2014 Hippocampus 24:1112-1119.
Migues P.V. et al., 2016 J Neurosci 36: 3481-3494.