🧠 neuroscience

Memory encoding reprograms neuronal transcriptional responses via durable chromatin remodeling

This study demonstrates that in the medial prefrontal cortex, durable chromatin remodeling during remote memory storage reprograms engram neurons to alter their transcriptional responses to future experiences, thereby reducing interference between different memories.

Original authors: Ke, Y., Sun, X., Greenleaf, W. J., Schnitzer, M. J.

Published 2026-07-31

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Ke, Y., Sun, X., Greenleaf, W. J., Schnitzer, M. J.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine your brain is a massive, bustling library where every experience you have is a new book being written. For a long time, scientists thought that once a memory was "filed away," it sat there like a static book on a shelf, waiting to be read. We knew that right after you learn something, your brain cells buzz with activity, writing the first draft of that memory. But what happens next? How does a fleeting thought turn into a memory that lasts for months or even years? And how does the brain keep all these books organized without them getting mixed up? This question sits at the heart of neuroscience, the study of how our brains work. To understand this, we need to look at two main things: the "words" inside the cells (genes) and the "filing system" that decides which words get read (chromatin). Think of chromatin as the librarian who decides which books are open on the desk and which are locked in the back. If the librarian changes their mind about which books are accessible, the story the cell tells changes, even if the text inside the book stays the same. Scientists have long wondered if this librarian changes their mind slowly over time to help us remember things for the long haul, or if the librarian just sits there doing nothing after the initial learning.

A team of researchers at Stanford University decided to peek behind the curtain of this library to see how the "engram" cells—the specific neurons that hold a memory—change over time. They used a clever trick to tag these special memory-holding cells in mice with a glowing green light, allowing them to find and study just those cells weeks after a scary event (a mild foot shock in a specific box). They looked at these cells at two different times: one week and four weeks after the event. They didn't just look at the genes being read (the RNA); they also looked at the chromatin, the physical structure that controls access to those genes.

Here is what they found, and it's a bit of a plot twist. When the mice were just one week past the scary event, the memory cells looked mostly like their neighbors, with only small differences in their genes. But by four weeks, something fascinating had happened. Even if the mice hadn't been reminded of the scary event in a while, the memory cells had quietly reorganized their internal filing system. Their chromatin had changed in a way that looked more like the "blueprints" used when the brain was first developing, rather than the busy "activity" signals seen right after learning.

The most exciting part came when the researchers finally reminded the mice of the scary event. When the memory was recalled, the tagged memory cells didn't just repeat the same old song. Instead, they reacted differently than the other cells. Because of those slow, quiet changes to their chromatin over the previous weeks, the memory cells had "reprogrammed" themselves. They were now less likely to go wild with excitement when triggered again. It's as if the librarian had quietly locked away the "easy to get excited" books and opened up a different set of books that help the cell stay calm and focused.

This suggests that the brain doesn't just store memories by keeping a perfect record of the past. Instead, it actively rewrites the rules for how those memories respond to the future. By dampening the reaction of memory cells, the brain might be protecting old memories from being overwritten by new, similar experiences. It's a bit like a security system that, after a while, becomes stricter about who gets to enter, ensuring that the original guest list stays safe from intruders. The study suggests that this slow, structural change in the cell's "filing cabinet" is a key part of how we keep our memories stable and distinct, preventing them from getting jumbled together as we learn new things.

Technical Summary: Memory encoding reprograms neuronal transcriptional responses via durable chromatin remodeling

Problem Statement
Long-term memory storage in the mammalian brain, particularly within the medial prefrontal cortex (mPFC), requires mechanisms that maintain information over weeks to months. While immediate early gene activation and transient synaptic changes are well-characterized in the minutes to hours following learning, the molecular basis for the persistence of memory engrams over extended timescales remains unclear. A central biological question is how memory circuits balance the need for stability (preserving existing memories) with plasticity (storing new information). Specifically, it is unknown whether chromatin states in engram neurons merely reflect a passive record of prior neural activation or if they actively reprogram the transcriptional logic of these neurons to alter their responses to future memory retrieval events.

Methodology
The authors employed a multiomic approach to profile mouse mPFC engram neurons at two distinct time points: 7 days (early consolidation) and 28 days (remote memory) after contextual fear conditioning.

Genetic Trapping: Using a Fos-TRAP2; Sun1-GFP mouse line, neurons activated during fear conditioning were genetically labeled with a nuclear-targeted GFP.
Experimental Design: Mice were divided into groups based on time post-conditioning (7 or 28 days) and recall status (re-exposed to the context 60 minutes prior to tissue collection vs. no recall). Nuclei were isolated from the mPFC and sorted into GFP+ (engram) and GFP- (non-engram) populations.
Single-Cell Multiomics: The authors performed joint single-nucleus RNA sequencing (snRNA-seq) and ATAC sequencing (snATAC-seq) using the 10x Genomics Chromium platform, generating profiles for 17,669 high-quality nuclei.
Computational Framework: To disentangle complex regulatory effects, the authors defined four regulatory quantities based on linear contrasts of the experimental conditions:
1. $\Delta N$ (No-recall engram): Differences between engram and non-engram neurons in the absence of recall.
2. $\Delta R$ (Recall response): Differences between recall and no-recall conditions.
3. $M$ (Engram modulation): The difference in recall response ( $\Delta R$ ) between engram and non-engram neurons ( $\Delta \Delta R$ ).
4. $C$ (Modulation consolidation): The temporal shift in modulation ( $M$ ) between 7 and 28 days ( $\Delta \Delta \Delta R$ ).
Decoder Analysis: Multivariate logistic decoders were trained on latent semantic indexing (LSI) embeddings of distal chromatin peaks and principal components of gene expression to identify regulatory axes distinguishing engram states and recall responses. Motif enrichment analyses were performed to link chromatin accessibility to transcription factor (TF) activity.

Key Results

Distinct Transcriptional Signatures: In the absence of recall, engram neurons showed only modest transcriptional differences ( $\Delta N$ ) at 7 and 28 days. However, upon memory recall, engram neurons exhibited distinct transcriptional programs ( $M$ ) compared to non-engram neurons of the same cell type. This modulation was significantly larger than the no-recall differences, indicating that the engram state is defined less by a static resting signature and more by a reconfigured response to reactivation.
Chromatin Metaplasticity: While gene expression differences in the no-recall state were modest at 28 days, chromatin accessibility decoders robustly distinguished engram from non-engram neurons. This "distal chromatin engram signature" emerged over the 4-week consolidation period.
Shift in Regulatory Logic:
- At 7 days, chromatin signatures in engram neurons were enriched for activity-dependent motifs (e.g., AP-1/FOS/JUN), consistent with recent neural activation.
- By 28 days, AP-1 signatures receded. The 28-day engram chromatin state was instead enriched for motifs associated with developmental programs and cell-state regulation (termed "Persistent TFs" or P-TFs), including homeobox, POU/TALE, NFAT/SOX, and MEF2 families.
Functional Consequences of Remodeling: The chromatin changes at 28 days predicted a dampening of transcriptional responses to future excitation. Specifically, engram neurons showed:
- Attenuated Excitability: A compression of the dynamic range for recall-induced expression of voltage-gated ion channel genes.
- Regulatory Shift: Preferential modulation of regulatory genes (chromatin regulators, RNA splicing, ubiquitination) over direct plasticity effectors.
- Synaptic Restraint: Upregulation of genes linked to synaptic inhibition and downregulation of structural plasticity genes.
Developmental Reuse: The distal chromatin peaks distinguishing 28-day engram neurons overlapped significantly with regulatory elements active in the embryonic cortex (E13.5–E18.5), suggesting that memory consolidation recruits conserved developmental regulatory machinery to establish a durable, stable state.
Cell-Type Specificity: Excitatory (glutamatergic) and inhibitory (GABAergic) neurons utilized distinct strategies. Glutamatergic engrams were characterized by strong modulation of recall responses ( $M$ ), whereas GABAergic engrams showed stronger constitutive differences in the no-recall state ( $\Delta N$ ). However, both cell classes shared broad chromatin and transcriptional engram signatures.

Significance and Claims
The paper proposes that chromatin structural changes serve as a "slow-timescale component of neural computation" that reduces interference between different memory representations. Rather than merely preserving a molecular record of past learning, the chromatin architecture of engram neurons undergoes a maturation process over weeks that "repurposes conserved regulatory machinery" to dampen the extent to which these neurons engage in further information storage.

The authors define this phenomenon as chromatin metaplasticity: a higher-order form of plasticity where experience alters the cell's ability to induce plasticity in the future. By shifting the transcriptional control logic of engram neurons toward a state that attenuates responses to new excitation (while preserving the capacity for memory recall), the brain stabilizes remote memories against the destabilizing effects of memory interference. The study concludes that the state of a neuron should be defined to include not just synaptic weights but also its chromatin state, which dictates its responsiveness to future inputs.

More like this