|Memory > Ca2+ in Memory Consolidation, Erasure, Extinction and Retrieval|
Ca2+ is involved in most cellular processes, even for the processes that are antagonistic to each other, such as long-term potentiation and long-term depression (Chapter 18). This chapter summarizes the roles of Ca2+ in memory consolidation, erasure, extinction and retrieval.
Memory consolidation involves synaptogenesis within potentiated dendritic branches (Chapter 26). The large amount of Ca2+ ions accompanied with the NMDA plateau in the potentiated dendritic branch may create dendritic filopodia, which serve as the "tags" to guide axonal branching toward the potentiated dendritic branch (Chapter 28). Ca2+ can also activate Myosin II to drive CABT complexes that bind stable F-actin from the spine to the shaft where they may direct dendritic elongation and branching (Chapter 27). The elongation of a dendritic branch allows the memory unit to contain more synapses, thereby strengthening the memory unit. The dendritic branching increases the number of branches (memory units), thus expanding memory capacity.
After the prototype of a synapse between a filopodium and a new axon branch is established, spine maturation from the filopodium is regulated primarily by the transcription factor, NF-κB, which targets many genes involved in spinogenesis and synaptic strengthening (de la Fuente et al., 2015; Engelmann and Haenold, 2016).
NF-κB can be upregulated by metabotropic glutamate receptor subtype 5 (mGluR5) (O'Riordan et al., 2006; Wang and Zhuo, 2012), via the PLC/PKC pathway (Figure 36-1). Since NMDA plateau is generated by a substantial glutamate pond, the potentiation of a dendritic branch should also activate mGluR5. Abundant evidence implicates mGluR5 in addiction (Mihov and Hasler, 2016), consistent with the notion that addiction arises from excessive memory consolidation (Chapter 30).
The Ca2+ influx through GluN2B-containing NMDARs may activate calcineurin to dephosphorylate the GluA1 subunit of AMPARs at S845 (Chapter 18), thereby disrupting the association between AMPARs and the postsynaptic density (PSD). The dissociated AMPARs then move laterally in the spine membrane toward endocytic zones adjacent to the PSD. After endocytosis, AMPAR-containing endocytic vesicles are transported along F-actin from the membrane to the endosome for degradation (Hanley, 2014). This process requires the motor protein myosin VI which can be activated by Ca2+ (Batters et al., 2016). Thus, activation of both calcineurin and Myosin VI could cause AMPAR internalization, resulting in memory erasure.
When calcineurin is activated, it may activate cofilin to depolymerize F-actin (Wang et al., 2005), causing CABT to switch binding partner from the dynamic F-actin to GluN2B-containing NMDARs, consequently resulting in NMDAR extinction (Chapter 20). In addition, calcineurin may activate the transcription factor, nuclear factor of activated T-cells (NFAT), which targets the key player in long-term memory extinction: BDNF (Chapter 19 and Chapter 22).
It has been well-documented that protein kinase A (PKA) plays a crucial role in memory retrieval, possibly by recovering NMDAR extinction (Chapter 13). Although PKA activity is basically determined by the binding of cAMP, some types of adenylyl cyclase (e.g., AC1 and AC8) that catalyzes the conversion from ATP to cAMP are Ca2+-dependent (Wong et al., 1999; Wang et al., 2003). Therefore, the intracellular Ca2+ from various sources may activate PKA to retrieve memory.
Acetylcholine (ACh) is an example that employs Ca2+ to enhance PKA activity (Chapter 25). By contrast, norepinephrine (NE) can regulate PKA activity by directly stimulating the production of cAMP, without involving Ca2+ (Chapter 24). Since indirect production of cAMP via Ca2+ may cause delay, NE is superior to ACh for pacing the UP and DOWN states of slow oscillations (Chapter 33; Chapter 34; Chapter 35). This also explains why ACh is virtually absent during slow wave sleep.
In Figure 36-1, we see that activation of mGluR5 can produce inositol-1,4,5-trisphosphate (IP3), which may bind to its receptor (IP3R) and trigger the release of Ca2+ from intracellular stores (endoplasmic reticulum). During slow wave sleep, mGluR5 is uncoupled from IP3R (Chapter 37), preventing the release of Ca2+ from intracellular stores. This may ensure that the slow oscillation is controlled by the NE-releasing neurons projected from locus coeruleus.
Author: Frank Lee