Brain-wide screen of prelimbic cortex inputs reveals a functional shift during early fear memory consolidation

Curation statements for this article:
  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This paper, of interest to neuroscientists in the field of memory engrams, presents novel insights to understand the complex functional network connected with the prefrontal cortex that shape memory-related neuronal ensembles and modulate memory formation in a time-dependent manner. The large data set due to the systematic approach yielded transparent, well analyzed, and clearly presented data. Still, the conclusions require additional support and extended discussion.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

This article has been Reviewed by the following groups

Read the full article See related articles

Abstract

Memory formation and storage rely on multiple interconnected brain areas, the contribution of which varies during memory consolidation. The medial prefrontal cortex, in particular the prelimbic cortex (PL), was traditionally found to be involved in remote memory storage, but recent evidence points toward its implication in early consolidation as well. Nevertheless, the inputs to the PL governing these dynamics remain unknown. Here, we first performed a brain-wide, rabies-based retrograde tracing screen of PL engram cells activated during contextual fear memory formation in male mice to identify relevant PL input regions. Next, we assessed the specific activity pattern of these inputs across different phases of memory consolidation, from fear memory encoding to recent and remote memory recall. Using projection-specific chemogenetic inhibition, we then tested their functional role in memory consolidation, which revealed a hitherto unknown contribution of claustrum to PL inputs at encoding, and of insular cortex to PL inputs at recent memory recall. Both of these inputs further impacted how PL engram cells were reactivated at memory recall, testifying to their relevance for establishing a memory trace in the PL. Collectively, these data identify a spatiotemporal shift in PL inputs important for early memory consolidation, and thereby help to refine the working model of memory formation.

Article activity feed

  1. Evaluation Summary:

    This paper, of interest to neuroscientists in the field of memory engrams, presents novel insights to understand the complex functional network connected with the prefrontal cortex that shape memory-related neuronal ensembles and modulate memory formation in a time-dependent manner. The large data set due to the systematic approach yielded transparent, well analyzed, and clearly presented data. Still, the conclusions require additional support and extended discussion.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

  2. Reviewer #1 (Public Review):

    In this study, the authors used tracing/c-fos tagging/chemogenetic methods to perform a brain-wide screen of engram cells in the prelimbic cortex (PL) and identified specific activity patterns of these inputs across different phases of fear memory consolidation. The magnitude of the data set per se is remarkable. The data are well controlled and adequately analyzed. Projection-specific chemogenetic inhibition confirmed and extended previous studies by exposing the contribution of the claustrum and insula to PL inputs during encoding and recall of recent memory, respectively.

  3. Reviewer #2 (Public Review):

    While systems memory consolidation theory tends to categorize regions that are involved in memory acquisition and early recall vs regions involved at a later time point, recent evidence showed that regions known to be active at a remote time point, like the medial prefrontal cortex, may be involved also at an earlier stage. In this manuscript from Dixsaut and Gräff, they aim to dissect mPFC engrams connectivity and identify differences between early and later involvement. They found interesting and novel results that are essential to further unravel the role and time-dependent implication of different pathways linked to the mPFC in regulating memory consolidation. Interestingly here they focus specifically on engram cells instead of targeting all mPFC cells. This aspect is significant since these are cells that are believed to have a crucial role in memory acquisition and its following reactivation.

    In particular, they analyzed several connecting regions that send projections to PL engram cells. Some (BLA, CLA and EC) active at encoding, some (RSP and INS) at recent memory recall. They further targeted these cells and inhibited them using chemogenetic manipulation during encoding (BLA,CLA and EC cells projecting to PL) or recent memory recall (RSP and INS cells projecting to PL). Via a well-designed and technically impressive analysis of this complex subnetwork linked to PL, the authors found a time-dependent implication of different projections bringing a new perspective on the role of PL engram cells in memory consolidation.

    Overall, this manuscript represents a novel discovery that can bring to a constructive re-evaluation of the mPFC role in memory consolidation.

  4. Reviewer #3 (Public Review):

    Activity in the prelimbic cortex following an aversive event is necessary for subsequent remote retrieval of the consolidated fear memory weeks later, but the circuits supporting the gradual reorganization of memory into cortical circuits are not well understood. In this study, Dixsaut and Graff examined how certain subcortical and cortical inputs to the prelimbic cortex contribute to the establishment of memory-related neuronal ensembles at encoding and two retrieval timepoints. The results confirm recent reports that certain PL inputs, such as the basolateral amygdala and entorhinal cortex contribute to the consolidation of remote contextual fear memories, but also identify novel contributions for the claustrum and insula during encoding and retrieval respectively.

    The strengths of this paper include the systematic, data-driven approach to identifying and testing relevant inputs. The use of TRAP2 mice offers advantages over previous cfos-based tagging approaches and is a powerful approach to map only those inputs that directly innervate experience-activated prelimbic cells. Through this approach, several expected and new inputs were revealed. The authors then followed up on two of these novel inputs to test their role in memory encoding and retrieval. The data reporting and presentation are transparent and thorough. While the technical and analytic aspects of this study are very strong, some core claims of this study are not fully supported by the data.

    The core claims:

    1. That neuronal ensembles after CFC training represent an associative engram. The comparison is a context exposure no-shock group, which doesn't control for an aversive experience independent of learning. Moreover, the PL is also implicated in the consolidation of a latent context memory, as shown with the context pre-exposure effect (e.g., work by Stanton and colleagues: Heroux et al., 2017; Robinson-Drummer et al., 2017). The cfos activation patterns in this study may thus represent contextual memory formation and/or contextual fear memory formation, but may also reflect aversive reactivity or attentional arousal. The observation of an ensemble reactivation-to-freezing correlation only in CNO groups (Figure 5) could reflect a contribution of these inputs to non-associative features, such that the removal of these inputs refines the relationship of the reactivated ensemble to behavior.
    2. That the indicated inputs are "necessary" for consolidation. First, the reduction in freezing as a consequence of input inactivation is relatively modest, demonstrating a modulatory impact rather than necessity, and one consistent with a potential role in non-associative or aversive valence aspects of memory. Second, the lack of effect of input removal on the ensemble re-activation itself by the removal of CLA or INS inputs does not support a necessary role in ensemble formation or reactivation. Post-training inactivation of CLA-PL also did not affect engram formation, which is a timepoint where PL ensemble activation contributes to CFC consolidation and remote engram formation (DeNardo et al., 2019). These observations in the current study suggest that CLA input to PL modifies subsequent retrieval but does not support a necessity for this pathway in consolidation.
    3. Related, the interpretation that INS-PL affects consolidation is not strongly supported since the manipulation and assessment of behavior occurred in the same retrieval session, suggesting an effect on retrieval, not consolidation. Because systems consolidation requires prefrontal activity within the first 1-2 weeks of training regardless of retrieval, determining which inputs to PL after training affected the development of remote ensembles would provide a stronger case for consolidation of relevant inputs.