Technical Review: Thalamocortical transcriptional gates coordinate memory stabilization
- perryt6
- 1 day ago
- 5 min read
Authors: Andrea Terceros, Celine Chen, Yujin Harada, Tim Eilers, Millennium Gebremedhin, Pierre-Jacques Hamard, Richard Koche, Roshan Sharma, and Priya Rajasethupathy
Affiliations: Rockefeller University, New York, USA
Background Introduction
Decades of research have established that transcription factors like CREB1 mediate the consolidation of short-term memories into overnight memories, a process known as synaptic consolidation. However, the molecular mechanisms underlying synaptic consolidation have remained poorly understood. This study bridges a critical gap by identifying specific transcriptional regulators operating in thalamocortical circuits that enable progressive memory stabilization through distinct temporal windows. This advances our understanding of how brain circuits coordinate molecular programs to maintain long-term memories.
Materials and Methodology
The researchers used a murine model subjected to behavioural tasks to study cellular transcriptomes used as an indicator of consolidated memories. Tasks consisted of assorted multimodal contextual clues and were presented in either high repetition (HR) or low repetition (LR) contexts. Meanwhile, optogenetic silencing and activation were both used for circuit manipulation, which in turn permitted dependence studies on cortical structures. Mice were then sacrificed at various stages of training for single-cell analysis of relevant brain tissue, with several marker genes studied with relation to consolidated memories. Identified transcriptional regulators were then subjected to region-specific gene knockouts to verify functional consequences.
Results
Behavioral Task and Circuit Dependency
Mice successfully learned to discriminate between reward and aversive contexts by day 7 of training, exhibiting anticipatory licking in both HR and LR contexts. At recent retrieval, memory for both contexts was intact, but at remote time points (day 21), only HR memories persisted while LR memories were forgotten. This differential retention occurred despite no intrinsic licking preferences, confirming learned associations rather than sensory biases. Optogenetic silencing revealed that hippocampus was required for recent but not remote memory, whereas ACC was essential for remote but not recent recall—consistent with systems consolidation theory. Among projection-specific manipulations, only ANT→ACC pathway silencing produced an isolated remote memory deficit without affecting learning or recent recall, demonstrating this circuit's specific role in memory stabilization. Conversely, optogenetic activation of ANT→ACC during training significantly enhanced LR memory retention at remote time points. These findings establish the HPC–ANT–ACC pathway as critical for progressive memory consolidation over weeks.
Transcriptional Divergence Between Consolidated and Forgotten Memories
Single-cell RNA sequencing revealed progressive transcriptional divergence between HR and LR conditions beginning as early as recent retrieval—well before behavioral differences emerged. In ANT, differential gene expression analysis identified synaptic plasticity-related modules that peaked transiently at mid-retrieval then decreased by remote time points. In ACC, histone methylation-related gene programs emerged and were sustained through remote and late-remote retrieval. Transcriptional distance (measured via Wasserstein distance of DEGs) between HR and LR increased progressively in ANT through mid-retrieval, then plateaued, whereas in ACC it continued expanding through late-remote time points. Gene Ontology enrichment revealed that ANT DEGs were enriched for neuronal synaptic plasticity, projection development, and calcium channel activity, while ACC DEGs were enriched for histone methyltransferase activity, transcription coregulator binding, and chromatin modification. These findings demonstrate that ANT and ACC engage temporally and functionally distinct molecular programs during memory stabilization, with ANT supporting early plasticity-driven changes and ACC implementing sustained chromatin-level modifications for long-term storage.
Identification of Cellular Macrostates via Pseudotime Analysis
Palantir pseudotime trajectory analysis revealed that neurons occupy phenotypic continua rather than discrete states during memory consolidation. In ANT, the trajectory bifurcated into two apex branches representing "early consolidation" and "late consolidation" macrostates. HR neurons progressively occupied these advanced states at mid and remote retrieval, whereas LR neurons remained near the pseudotime origin, suggesting failure to enter consolidation-associated cellular states. Approximately 20-30% of HR neurons reached advanced macrostates, consistent with the proportion of memory-encoding neurons identified in previous imaging studies. In ACC, a single apex macrostate emerged, with HR neurons advancing further along the trajectory than LR neurons at mid and remote time points. These cellular state transitions were specific to memory-relevant circuits; control visual cortex (V1) showed no such divergence. Fos-positive (recently active) neurons showed similar trajectory shifts, confirming that state transitions occur in behaviorally engaged populations. The identification of discrete cellular macrostates associated with memory persistence provides a conceptual framework for understanding how molecular programs coordinate memory stabilization.
Transcriptional Regulators as Gatekeepers of Macrostate Entry
Correlation analysis identified transcriptional regulators whose expression patterns defined entry into early and late consolidation macrostates. In ANT, four transcription factors emerged: CAMTA1 and MYT1L correlated with early consolidation, while MEF2C and TCF4 correlated with late consolidation. In ACC, three histone methyltransferases were identified: ASH1L, KMT2A, and PRDM2, all correlating with late consolidation. Predicted target modules of these regulators comprised 20-30% of remote DEGs, and average expression of targets co-varied with branch state, supporting their roles in orchestrating macrostate transitions. ATAC-seq confirmed increased chromatin accessibility of predicted CAMTA1, TCF4, and ASH1L modules at mid-retrieval, with differential persistence: CAMTA1 module accessibility returned to baseline by remote time, whereas TCF4 and H3K4me3 (ASH1L-associated) modules remained accessible, aligning with their proposed roles in transient versus sustained memory maintenance. These results nominate specific transcriptional regulators as molecular coordinators of progressive memory stabilization.
Causal Requirement for Sequential Transcriptional Cascade
CRISPR-Cas9-mediated knockout experiments revealed striking time- and region-specific requirements for identified transcriptional regulators. Creb1 knockout in hippocampus impaired learning and recent memory, as expected from prior literature. Remarkably, Camta1 knockout in ANT left learning intact but specifically impaired memory at mid-retrieval (days 8-14), whereas Tcf4 knockout produced isolated deficits at remote retrieval (day 21+). Myt1l and Mef2c knockouts showed no significant impairments despite module co-variation, indicating functional specificity. In ACC, Ash1l knockout produced selective remote memory deficits, while Kmt2a knockout showed trending but non-significant effects. These effects were region-specific: Camta1 knockout in ACC or Ash1l knockout in ANT produced no behavioral deficits, confirming circuit-level specificity. Fiber photometry recordings revealed that Camta1 and Tcf4 knockouts increased neural entropy in ANT and disrupted ANT-ACC functional correlations during retrieval, suggesting that these regulators coordinate inter-regional plasticity and information flow. These findings establish a critical CAMTA1→TCF4→ASH1L thalamocortical transcriptional cascade operating sequentially over days to weeks, with each regulator having temporally restricted, non-redundant roles in memory maintenance.
Mechanistic Insights from Chromatin Profiling
ChIP-seq analysis of CAMTA1 and TCF4 binding in ANT revealed that CAMTA1 targets were enriched for synaptic plasticity genes, whereas TCF4 targets included cell adhesion and structural refinement genes. When overlaid onto pseudotime trajectories, CAMTA1 targets showed enriched expression along the early consolidation branch, while TCF4 targets aligned with late consolidation, confirming their sequential engagement. H3K4me3 ChIP-seq in ACC identified temporally shifting methylation programs: at mid-retrieval, methylated genes were enriched for plasticity pathways, whereas at remote retrieval, structural and cytoskeletal genes predominated. Ash1l knockout specifically depleted H3K4me3 marks at plasticity gene promoters at mid-retrieval and structural gene promoters at remote retrieval, demonstrating time-dependent functional roles. ATAC-seq confirmed sustained chromatin accessibility of H3K4me3 target modules through remote time points, consistent with histone methylation's role in stable epigenetic memory. These mechanistic studies reveal that transcriptional regulators coordinate distinct gene programs—from transient plasticity to enduring structural changes—that collectively enable progressive memory stabilization.
Conclusion
Tercero and colleagues identified a critical molecular framework for understanding memory consolidation. By dissociating memory formation from consolidation in their studies, the authors revealed the thalamocortical circuit engages sequential waves of transcription mediated by several regulators separated by timescale: CAMTA1 for days, TCF4 for weeks, and ASH1L for longer-term storage. Each transcriptional regulator provides a temporal checkpoint before information is committed to permanent storage.
This paper utilized a PreciGenome solenoid valve for air flow regulation. Aversive air puffs were a key part of the behavioral task, so consistent and reliable actuation were critical when conducting thousands of trials.
For further details, check out the link to the article here:
Comments