Reactivation-dependent transfer of fear memory between contexts requires M1 muscarinic receptor stimulation in dorsal hippocampus of male rats

Discussion

In this investigation, we introduce a novel approach to studying the transfer of learned fear to other contexts, with a specific emphasis on the importance of reactivation of the original memory and the role of M1 mAChRs in male rat. Our findings, derived from the novel FRaT task, suggest that additional contextual information can be incorporated into an existing fear memory and that this memory modification is reactivation- and time-dependent.

First, we used the FRaT task to update contextual fear memory in rats, apparently incorporating information from a neutral context presented shortly after reactivation of the original fear memory. This finding is consistent with reconsolidation theory, which posits that upon reactivation, memories temporarily enter a labile state, making them susceptible to modification (Nader et al. 2000). We have previously described a similar phenomenon for object memory updating in rats (Jardine et al. 2020). Updating of contextual fear memory with new information has been explored in other studies investigating the linking of aversive experiences across time (Jung et al. 2023; Zaki et al. 2023). Furthermore, others have demonstrated the possibility of updating an established contextual fear memory with appetitive information, suggesting the versatility and adaptability of memory processes (Ramirez et al. 2013; Haubrich et al. 2015). However, here we demonstrate updating of a contextual fear memory without altering the original conditioned response, as seen in Experiment 1 where rats were still freezing in the fear conditioning chamber on test day even if alternate contextual information had been introduced post-reactivation.

Behavior consistent with contextual fear transfer was only seen in the FRaT task when rats were exposed to the original conditioned context shortly before alternate context presentation; no updating occurred when the original context exposure was omitted or when the reactivation occurred 6 h before alternate context exposure. The reactivation- and time-dependence of this effect is significant as it delineates the specificity of the memory updating process. It suggests that the rats are not indiscriminately generalizing their fear to any context they encounter, but rather, the reactivation primes the original memory for subsequent modification. This interpretation is also supported by the relatively lower freezing levels in the different alternate context condition on test day. This distinction between memory updating and generalized fear is crucial for understanding the nuances of memory processes operating here. If the rats were simply generalizing their fear to any context through fear memory updating, it would imply a broad and nonspecific response. However, our results indicate a more targeted and specific process, whereby the reactivated original memory is modified to incorporate new information presented within a restricted period of lability (Nader et al. 2000). It is also worth noting that rats in the various conditions did not differ in terms of freezing either during the memory reactivation session or when they were placed into the alternate contexts following reactivation; this pattern reinforces our interpretation of the present results in terms of a specific memory reconsolidation-based process of fear transfer. Additionally, our results showing that freezing in the alternate context persists when tested 4 weeks later suggest a lasting, rather than temporary, effect. A shorter-term effect might indicate a less specific effect related to the recent exposure to the original context. This alternative explanation is also not supported by our failure to see enhanced freezing when rats were tested in the different alternate context condition.

It is important to note that the effect seen in the FRaT task could potentially be attributed to some sort of second-order conditioning. Because the alternate contexts are presented post-reactivation, it is possible that this contextual information becomes associated with the original context, or the response or fear state produced by reexposure to this context. For example, second-order conditioning can occur for auditory fear (Debiec et al. 2004, 2006; Payne et al. 2023). Specifically, Debiec et al. (2006) paired a previously conditioned tone (CS1) with a new tone (CS2) and showed that animals will freeze to CS2 after several such pairings. However, the procedures in Debiec et al. (2006) differed significantly from the present methods. Here, the rats were exposed to the putative CS2, the alternate context, FOLLOWING the exposure to CS1. Moreover, the fact that fear spreads to the alternate context when it is presented 20 min following exposure to the original context, suggests that the effect observed in the FRaT task is not the result of backward second-order conditioning, for which there is very little actual evidence, and which would likely require a much shorter delay between context presentations (Mowrer et al. 1988; Barnet et al. 1991). Additionally, the short duration of the reactivation session makes it less likely that the rats established a second-order association between the contexts following a single pairing. Nonetheless, it is possible that a lingering fear state produced by reexposure to the training context was associated with the alternate context during its presentation shortly after the reactivation session, and this could also explain the effects observed here. Testing the effects of extinction could possibly help to distinguish between these explanations, as extinguishing freezing to the alternate context might not affect freezing to the training context if second-order conditioning underlies the present results; this is an important question for future investigation.

A final vital distinction between the FRaT task and similar procedures used in Debiec et al. (2006) and Payne et al. (2023) is that in the current study, the rats were previously habituated to the alternate contexts (the putative CS2). We did this to reduce the likelihood of nonspecific fear generalization when presenting a new context shortly following the reactivation of an aversive memory. The latent inhibition literature indicates that such preexposure to a context decreases the likelihood of the context subsequently becoming associated with an aversive stimulus (Lubow and Moore 1959; Blanchard et al. 2010). Thus, previous evidence, combined with the reactivation- and time-dependence of the current effects, support the interpretation that the FRaT task reflects fear memory updating; however, future studies are needed to more confidently identify the underlying behavioral basis of these effects.

The neural mechanisms underlying the present effect remain unclear. However, linking of memories through coactivation of neuronal ensembles seems a likely explanation. Cai et al. (2016), for example, demonstrated that memories for contexts encoded closely in time (5 h) could be linked apparently via overlapping neural ensembles. Furthermore, Zaki et al. (2023) report that strong fear conditioning in one context can lead to later offline reactivation of the neural ensemble associated with the trained context as well as that for a previously presented (2 days earlier) neutral context, resulting in retrospective linking of the learned fear with that neutral context. While a similar process might be at play in the current study, the present behavioral effects are unique in that they demonstrate a critically reactivation-dependent form of prospective memory linking. Indeed, fear transfer in the FRaT task was not observed if the original contextual fear memory was not explicitly reactivated. The explicit reactivation session used here could accomplish something similar to the ensemble coactivation implicated by Zaki et al. (2023), but possibly in a more selective and prospective manner. The effect reported by Zaki et al. (2023) was only observed retrospectively and not seen when a relatively low shock protocol was used for the initial fear conditioning. Here, fear transfer occurred in a highly reactivation- and time-dependent manner. Moreover, despite both alternate contexts being preexposed 72 h before the fear conditioning session (to habituate rats to these contexts), fear only transferred to the specific alternate context presented following reexposure to the trained context. Thus, a generalized linking of fear with contexts explored within a few days of the training event did not occur; the original fear memory was only updated with the specific contextual information presented within a specific window of time following memory reactivation. There is one earlier report of apparent memory linking following reactivation, whereby fear conditioning to a second tone was enhanced when conducted shortly after presentation of a different tone previously paired with shock (Rashid et al. 2016). In contrast, however, here we provide evidence for memory reactivation- and time-dependent transfer of a fear response to an otherwise neutral context.

The fear memory updating effect reported here requires mAChR stimulation. Here, we targeted the dorsal CA1 with a selective M1 mAChR antagonist to block the destabilization of the contextual fear memory when reactivated in the original fear conditioning chamber. The fact that pirenzepine blocked reactivation-dependent transfer of fear to the alternate context is consistent with the implication of M1 mAChRs in many forms of memory destabilization (Stiver et al. 2015, 2017; Huff et al. 2022; Wideman et al. 2023; Abouelnaga et al. 2023), as well as our current hypothesis that a reconsolidation-based mechanism underlies the behavioral effect in the FRaT task. Additionally, our past work suggests that M1 mAChR activity during memory reactivation could lead to downstream activation of CaMKII and the ubiquitin-proteasome system (Stiver et al. 2017; Wideman et al. 2023), consistent with a documented role for synaptic protein degradation in memory destabilization (Lee et al. 2008; Jarome et al. 2011). The plasticity resulting from such a process, combined with the coactivation of contextual representations (ensembles) in the brain could facilitate updating of the original fear memory with the new contextual information, leading to fear transfer.

The FRaT task represents a novel animal model to facilitate the study of fear memory transfer between contexts and its behavioral and biological bases. To this end, the current study presents clear avenues for further research into the role of the cholinergic system, specifically M1 mAChRs, in fear memory updating. Elucidating such memory modification mechanisms has potential implications for better understanding and treatment of conditions like post-traumatic stress disorder (PTSD) in which maladaptive memory transfer could underlie certain debilitating symptoms. It is also important to acknowledge that only male rats have been used in this study to establish the FRaT task as a valid protocol for investigating the transfer of fear memory to alternate contexts. Future directions should investigate whether male and female rats display memory linking according to similar behavioral and neurobiological mechanisms. The present results suggest that memory linking could occur through processes involved in the reactivation and updating of the original traumatic memory and that interventions targeting this process could be effective in preventing the unfortunate spread of fear to stimuli not originally associated with the initial episode.