Contribution of adult-born neurons to memory consolidation during rapid eye movement sleep

Akinobu Ohba,Masanori Sakaguchi

Introduction:Memory consolidation stabilizes newly acquired memories by integrating them into pre-existing memory networks,which is thought to occur via changes in synaptic strength. Sleep may influence memory consolidation by modifying synaptic strength through local neuronal oscillatory activity. Recently,we found that the activity of hippocampal adultborn neurons (ABNs) is critical for memory consolidation during sleep (Kumar et al.,2020). Here,we propose a hypothesis for how changes in ABN synaptic plasticity synchronized with neural oscillations may contribute to memory consolidation during sleep.

Synaptic plasticity during sleep:Sleep occurs in cycles of rapid eye movement (REM) and non-REM (NREM) sleep,with each sleep stage associated with characteristic synchronous neuronal activities. Slow oscillations from the cortex,spindles from the thalamus,and sharp wave-ripples (SWRs) from the hippocampus are prominent during NREM sleep,whereas theta oscillations from the hippocampus and pontine-geniculateoccipital waves are prominent during REM sleep (Diekelmann and Born,2010). These synchronous activities may create a permissive environment for changes in synaptic strength.

Major theories,such as the synaptic homeostasis theory,posit that experience during wakefulness induces synaptic potentiation,whereas subsequent sleep globally downscales overall synaptic strength. As a consequence,weakly connected synapses are removed while relatively strong synapses are preserved,thereby enhancing memory due to an improved signal-to-noise ratio (Diekelmann and Born,2010). This synaptic reorganization could be mediated by SWRs and slow oscillations in NREM sleep. Indeed,silencing SWRs prevents synaptic downregulation (Norimoto et al.,2018). Moreover,in NREM sleep,synaptic downscaling occurs after presynaptic activity unless it is followed by postsynaptic activity within a specific time period (González-Rueda et al.,2018).

REM sleep may also contribute to synaptic reorganization. Newly formed spines after motor learning are predominantly pruned during REM sleep,whereas remaining spines are strengthened (Li et al.,2017). Moreover,REM sleep promotes rapid spine elimination along with dendritic Ca2+spikes after auditory fear learning (Zhou et al.,2020). These findings suggest that REM sleep is important for synaptic reorganization during memory consolidation. Indeed,in REM sleep,hippocampal neuronal activity at specific phases of the theta oscillation (i.e.,phase-locked activity) plays a critical role in both the strengthening and weakening of synapses. In particular,the plasticity of perforant path synapses from the entorhinal cortex to the dentate gyrus (DG) depends on the theta phase at which input stimulation is applied (Orr et al.,2001). Indeed,memory consolidation is impaired when theta oscillation in REM sleep is blocked (Boyce et al.,2016),suggesting that neuronal activity synchronized with theta oscillation plays a critical role in synaptic plasticity during memory consolidation in REM sleep.

Role of ABNs in memory consolidation during REM sleep:The mammalian DG is unique in that it continuously generates granular neurons into adulthood. Until recently,the dynamics and roles of ABN activity during sleep were unknown. However,we showed that after contextual fear learning,the overall activity of ABNs decreases when mice are in REM sleep (Kumar et al.,2020). Remarkably,most ABNs that were sparsely active during learning reactivate during subsequent REM sleep. Furthermore,memory consolidation is impaired if the sparse activity of ABNs is silenced during REM sleep. These results provide causal evidence that the sparse activity of ABNs during REM sleep is necessary for memory consolidation,consist with a previous finding that a sparse neuronal population in the DG reactivates during memory retrieval (Sakurai et al.,2016).

The synaptic plasticity capacity of ABNs depends on their maturational stage,with enhanced plasticity observed among 4-6-week-old ABNs (Ge et al.,2007). Correspondingly,we found that silencing the activity of 4-week-old ABNs,but not 2- or 10-week-old ABNs,during REM sleep impairs memory consolidation (Kumar et al.,2020). Moreover,random (i.e.,not phase-locked) activation of 4-week-old ABNs impairs memory,similar to a previous report (Danielson et al.,2016),suggesting that precisely timed,finely tuned ABN activity is essential for their proper function. Furthermore,silencing 4-week-old ABNs results in elongation of their spine necks,which is a sign of weakening synaptic strength and/or decoupling from presynaptic activity. By contrast,no changes in the spines of fully mature granular cells (mGCs) were observed upon their silencing during REM sleep. Considering the theta phasedependent plasticity mechanism in the DG (Orr et al.,2001),our findings suggest that finely-tuned,theta phase-locked ABN activity contributes to memory consolidation during REM sleep.

A hypothesis for ABN-mediated memory consolidation in REM sleep:As a stable contextual fear memory trace can be encoded by mGCs in the DG,theta phaselocked ABN activity could promote memory consolidation by interacting with mGCs. Indeed,ABNs inhibit mGC activity via direct monosynaptic connections when a small number of ABNs are activated through inputs relevant to contextual information (Luna et al.,2019). We hypothesize that the DG contains both theta phase-locked and nonphase-locked ABNs (Figure 1A). Learning induces the formation of new spines on 4-week-old ABN dendrites via enhanced synaptic plasticity. This may increase nonphase-locked ABN activity,triggering global spine elimination and reducing overall ABN activity in REM sleep after learning (Kumar et al.,2020). However,theta phase-locked activity stabilizes spines in a subset of ABNs that remain sparsely active in REM sleep (Figure 1B). This sparse ABN activity inhibits mGC activity in a theta phase-specific manner,allowing connected mGCs to activate at the reciprocal theta phase (Figure 1C). This concurrent theta phase-locked activity of mGCs modifies their synaptic plasticity (Orr et al.,2001),allowing them to establish a memory trace (Figure 1D). This hypothesis is consistent with a previous observation that the theta phase preference of hippocampal neuron activity during REM sleep changes after context learning (Mizuseki et al.,2011).

Figure 1｜ABNs may influence mGC theta phase-locked activity during memory consolidation in REM sleep.

Testing this hypothesis would require the examination of ABN activity in relation to theta oscillation. This could be achieved using two approaches: (1) determining the effect of manipulating ABN activity at a specific theta phase during memory consolidation and/or (2) verifying whether ABNs fire at a particular theta phase. Regarding the first approach,combining optogenetics with a closed-loop feedback system might allow for phase-specific manipulation. Regarding the second approach,as genetically encoded Ca2+sensors cannot achieve the time resolution necessary to directly assess theta phaselocked neural activity,a method capable of examining ABN activity with higher temporal resolution is required. Moreover,as memoryencoding ABNs could exhibit specific activity patterns,it would be necessary to stimulate ABNs by mimicking their naturally occurring activity patternin vivo,which might be possible using a patterned light stimulation technique (Marshel et al.,2019).

Conclusion:The sparse,theta phase-locked activity of ABNs could produce changes in their synaptic strength and promote fear memory consolidation by influencing mGC activity. Future research on the mechanism by which ABN synaptic plasticity contributes to memory consolidation would pave the way toward developing new therapeutic strategies for memory disorders.

We thank K.G. Akers (Shiffman Medical Library,Wayne State University,Detroit,MI,USA) for comments on the manuscript. We thank M. Sakurai (International Institute for Integrative Sleep Medicine (WPI-IIIS),University of Tsukuba,Tsukuba,Ibaraki,Japan) for secretarial support.

This work was partially supported by grants from the World Premier International Research Center Initiative from MEXT,JST CREST grant #JPMJCR1655,JSPS KAKENHI grants #16K18359,15F15408,15H01276,15K18332,26115502,25116530,JP16H06280,and 20H03552,Takeda Science Foundation,Shimadzu Science Foundation,Kanae Foundation,Research Foundation for Opto-Science and Technology,Ichiro Kanehara Foundation,Kato Memorial Bioscience Foundation,Japan Foundation for Applied Enzymology,Senshin Medical Research Foundation,Life Science Foundation of Japan,Uehara Memorial Foundation,Brain Science Foundation,Kowa Life Science Foundation,Inamori Research Grants Program,and GSK Japan to MS.

Akinobu Ohba,Masanori Sakaguchi*

International Institute for Integrative Sleep Medicine (WPI-IIIS),University of Tsukuba,Tsukuba,Ibaraki,Japan

*Correspondence to:Masanori Sakaguchi,MD,PhD,masanori.sakaguchi@gmail.com. https://orcid.org/0000-0002-7211-9452 (Masanori Sakaguchi)

Date of submission:October 31,2020

Date of decision:December 18,2020

Date of acceptance:April 13,2021

Date of web publication:July 8,2021

https://doi.org/10.4103/1673-5374.317966

How to cite this article:Ohba A,Sakaguchi M (2022) Contribution of adult-born neurons to memory consolidation during rapid eye movement sleep. Neural Regen Res 17(2):307-308.

Copyright license agreement:The Copyright License Agreement has been signed by both authors before publication.

Plagiarism check:Checked twice by iThenticate.

Peer review:Externally peer reviewed.

Open access statement:This is an open access journal,and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License,which allows others to remix,tweak,and build upon the work non-commercially,as long as appropriate credit is given and the new creations are licensed under the identical terms.

Open peer reviewers:Tsz Kin Ng,The Chinese University of Hong Kong,China; Kui Xie,University of Auckland,New Zealand.

Additional file:Open peer review reports 1 and 2.