(AGENPARL) - Roma, 24 Aprile 2024 - (AGENPARL) – mer 24 aprile 2024 Article
Hippocampus-to-amygdala pathway drives the
separation of remote memories of related events
Graphical abstract
Authors
Giulia Concina, Luisella Milano,
Annamaria Renna, Eugenio Manassero,
Francesca Stabile, Benedetto Sacchetti
Correspondence
In brief
Remembering closely related events
poses a brain challenge. Each memory
must retain its uniqueness while
commonalities are recalled. Concina et al.
found that hippocampal neurons guide
amygdala inhibitory cells, separating
overlapped memories. This insight aids
understanding of memory segregation,
which is compromised in disorders like
schizophrenia and post-traumatic stress
disorder.
Highlights
Hippocampus drives the pattern separation of fear memories
in the amygdala
Parvalbumin cells enable the separation of related fear
memories in the amygdala
Recent overlapped memories are transformed into
segregated remote memories over time
Concina et al., 2024, Cell Reports 43, 114151
May 28, 2024 ª 2024 The Author(s). Published by Elsevier Inc.
https://doi.org/10.1016/j.celrep.2024.114151
OPEN ACCESS
Article
Hippocampus-to-amygdala pathway
drives the separation of remote
memories of related events
Giulia Concina,1,2 Luisella Milano,1,2 Annamaria Renna,1,2 Eugenio Manassero,1 Francesca Stabile,1
and Benedetto Sacchetti1,3,*
1Rita
Levi-Montalcini Department of Neuroscience, University of Turin, 10125 Turin, Italy
authors contributed equally
3Lead contact
https://doi.org/10.1016/j.celrep.2024.114151
2These
SUMMARY
The mammalian brain can store and retrieve memories of related events as distinct memories and remember
common features of those experiences. How it computes this function remains elusive. Here, we show in rats
that recent memories of two closely timed auditory fear events share overlapping neuronal ensembles in the
basolateral amygdala (BLA) and are functionally linked. However, remote memories have reduced neuronal
overlap and are functionally independent. The activity of parvalbumin (PV)-expressing neurons in the BLA
plays a crucial role in forming separate remote memories. Chemogenetic blockade of PV preserves individual
remote memories but prevents their segregation, resulting in reciprocal associations. The hippocampus
drives this process through specific excitatory connections with BLA GABAergic interneurons. These findings provide insights into the neuronal mechanisms that minimize the overlap between distinct remote memories and enable the retrieval of related memories separately.
INTRODUCTION
An important challenge for the brain is managing memories of
distinct events that share common features or that occur in temporal proximity. The specificity of each memory should be preserved over time, and the common features of these experiences
should be remembered. How this can be achieved is yet to be
determined. Dysfunctions in the neural mechanisms underlying
the distinction across highly similar memories form the core of
several neuropsychiatric disorders, such as post-traumatic
stress disorder (PTSD).
The hippocampus plays a pivotal role in the precise encoding
and retrieval of distinct memories stemming from closely related
experiences. This crucial function is elucidated through two processes: (1) pattern separation, aimed at minimizing overlap
among pieces of information to facilitate their separate storage
and retrieval, and (2) pattern competition, which allows the recall
of stored information from partial cues.1–7 Various neuronal processes within hippocampal circuits have been identified to support these functions.7–9 Notably, recent studies indicate that
when two distinct events occur in temporal proximity, such as
within a 5–6 h interval, certain neurons activated by the initial
event become engaged or ‘‘allocated’’ to encode the subsequent event in the hippocampus.10–12 Overlapping neuronal ensembles establish connections between memories of related
events, facilitating the capture of shared features across experiences.12 A parallel mechanism has been observed in the baso-
lateral amygdala (BLA),13 a brain region that plays a crucial role
in the formation and retrieval of fearful memories. Indeed, human
studies have also demonstrated an interplay between the hippocampus and amygdala in the pattern separation of both episodic
and emotional memories.14,15
However, the mechanism by which different memories within
overlapping ensembles are preserved over time remains unknown. It is unclear whether remote memories of related events
are stored partly within overlapping neurons or alternatively
segregated into distinct neuronal populations. Further, the
mechanisms to preserve the specificity of each memory over
time are not understood. Here, we address these issues by
analyzing the neuronal and behavioral features of two temporarily related memories at both recent and remote time intervals.
RESULTS
Memories of two events in temporal proximity are
reciprocally linked at a recent time point but are
independent at a remote time point
When memories of closely timed events are allocated to shared
neuronal populations, the retrieval of one memory increases the
likelihood of retrieval of the other.10,11,13 Thus, we studied the
process evolution over time by investigating whether the retrieval
of one memory recruits neurons common to the other memory at
either recent or remote time points when two fear events occur in
temporal proximity (such as 6 h apart). To this end, we employed
Cell Reports 43, 114151, May 28, 2024 ª 2024 The Author(s). Published by Elsevier Inc.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
the Daun02 inactivation method, which allowed inhibiting the activity of recall-induced neurons without affecting the neighboring
cells.16–19 c-Fos-lacZ transgenic rats carry a transgene in which
the c-Fos promoter drives lacZ gene transcription and b-galactosidase (b-gal) expression. Induction occurs only in activated
neurons that coexpress b-gal and c-Fos. When the prodrug
Daun02 is administered post-memory recall, b-gal converts
Daun02 into daunorubicin, which induces cell death (Figure 1A).
c-Fos-lacZ transgenic rats were subjected to one auditory stimulus (conditioned stimulus [CS1]: pure tone, 15 kHz) paired with a
painful unconditioned stimulus (US). Another different tone (CS2:
pure tone, 1 kHz) was paired with the US. Rats could easily
distinguish between the two tones (Figure S1). In accordance
with previous studies,13 pairings of CS1-US and CS2-US were
separated by a brief (6-h) delay, allowing rats to learn two distinct
yet temporally related events. Seventy-two hours after the
learning session, CS1 was presented to reactivate recent memory of CS1 (CS1 memory reactivation) (Figure 1B). To selectively
trigger the reactivation of CS1 memory while minimizing interference associated with contextual cues, CS1 was presented in a
novel environment. The presentation of CS1 in this new context
resulted in a substantial increase in freezing behavior (indicative
of defensive responses) across all rats (Figures S2A). Ninety minutes later, Daun02 or vehicle was injected into the BLA
(Figures S3A and S3B). The assessment of memory retention
for CS2 (i.e., the memory not explicitly reactivated, CS2 memory
retention) and CS1 (CS1 memory retention) was conducted 3
and 5 days later, respectively, in distinct new environments. During CS2 memory retention, Daun02-treated rats exhibited lower
freezing compared to vehicle-treated rats (Figure 1C). Furthermore, in an additional group of rats, Daun02 administration in
the absence of CS1 reactivation did not impact freezing to CS2
(Figure 1C), confirming that amnesia, when present, was specifically induced by Daun02-mediated inactivation of neurons activated during CS1 memory recall. Finally, during the CS1 memory
retention test, Daun02-treated rats showed reduced freezing
(Figure 1D), aligning with the notion that Daun02 injection
impaired neurons associated with CS1 memory. Ninety minutes
after the CS1 memory retention test, rats were sacrificed, and
the number of cells expressing b-gal in the BLA was quantified.
In comparison to vehicle-treated rats, the Daun02-treated group
exhibited a reduced number of cells expressing b-gal
(Figures S4A and S4B), thereby confirming the efficacy of the
Daun02 manipulation.17–19 We further assessed the correlation
between the percentage of cells expressing b-gal and freezing
responses to CS1 within each group (Figure S4C), revealing no
significant correlation. This lack of correlation may be attributed
to the fact that defensive behavior is likely regulated by a complex brain network. Similar results were obtained by counterbalancing tones (Figure S5A).
Taken together, these data are consistent with previous findings demonstrating that when memories of two events occurred
in temporal proximity (like the 6-h group), CS1 memory recall recruited neurons necessary for CS2 memory.10–13
To ascertain that the observed effects were specifically linked
to two distinct yet temporally related events, we compared the
data obtained in Daun02-treated rats in the 6-h group with
those of two additional rat groups. In one group, rats learned
Cell Reports 43, 114151, May 28, 2024
Article
about a continuous fear event (i.e., CS1-US and CS2-US pairings occurred in a single continuous trial with a 0-min intertrial interval; 0-min group). In the second group, two temporally distant
events were presented (CS1-US and CS2-US pairings were
separated by a long 7-day interval; 7-day group). Seventy-two
hours after the learning session, CS1 was presented to reactivate CS1 recent memory in all groups (Figure S2B). Ninety minutes later, Daun02 was injected, and the retention of CS2 and
CS1 was assessed 3 and 5 days later, respectively. During
CS2 memory retention, the freezing of rats in the 0-min and
6-h groups was lower than that in 7-day group (Figure 1E).
Conversely, all groups displayed low freezing to the subsequent
CS1 memory retention test (Figure 1F) and a low percentage
of cells expressing b-gal after CS1 memory retention trial
(Figures S4B and S4C).
Next, we performed similar experiments at a remote time interval, i.e., 3 weeks after learning (Figure 1G). At this time point, CS1
was presented (CS1 memory reactivation), and all rats displayed
high freezing (Figures S2C and S2D). 90 min later, Daun02 or
vehicle was injected into the BLA (Figures S3C and S6). The
retention of CS2 and CS1 was assessed 3 and 5 days later,
respectively. Unexpectedly, during CS2 memory retention, the
freezing of Daun02-treated rats was similar to that of vehicletreated rats (Figure 1H). Conversely, during the subsequent
CS1 memory retention test, freezing of Daun02-injected rats
was lower than in vehicle-treated rats (Figure 1I). Thus, the inactivation of CS1 recall-induced neurons at remote time points did
not affect the retention of CS2 remote memories. Similar results
were obtained by counterbalancing tones (Figure S5B). Moreover, during the CS2 memory retention test, Daun02-treated
rats in the 6-h group displayed freezing higher than the 0-min
group and similar to the 7-day group (Figure 1J). All groups displayed amnesia to CS1 (Figure 1K), and a low percentage of cells
expressing b-gal after the CS1 memory retention trial (Figure S6).
Thus, at remote time points, the memories of two different but
temporally related events are separate and independent, similar
to the memories of two temporally distant events, while memories of CS1 and CS2 occurring in a unique continuative event
continue to be linked.
The auditory cortex, particularly the higher-order auditory cortex, plays a key role in the storage and retrieval of remote auditory memories.20–22 Therefore, in the 6-h group, we retrieved the
CS1 remote memory and injected Daun02 into the auditory cortex (Figures S7A and S7B). During the CS2 remote memory
retention test, freezing was similar between Daun02-and
vehicle-injected rats (Figures S7C and S7D), suggesting that
CS2 memory was also unaffected by CS1 remote memory
retrieval at the cortical level.
Another feature of allocated auditory fear memories is that the
extinction of one recent memory also extinguished the other
one.13 Thus, we investigated whether memories of temporarily
related events could extinguish each other at either recent or
remote time points. Wild-type rats were trained at 0-min, 6-h,
and 7-day intervals (Figure 2A). CS1 recent memory was extinguished by repeatedly presenting CS1 in the absence of US
72 h after learning (Figure 2B). The following day, we assessed
CS2 memory retention. Freezing to CS2 decreased in the
0-min and 6-h groups but not in the 7-day group (Figure 2C).
Article
OPEN ACCESS
Figure 1. The targeted inhibition of neurons activated during CS1 memory reactivation disrupts the retention of recent memories to CS2
while leaving unaffected the retention of CS2 remote memories
(A) Schematic of Daun02-inactivation technique in c-Fos-lacZ transgenic rats.
(B) CS1-US and CS2-US occur at 0-min, 6-h, or 7-day delay. The recent memory is reactivated by representing CS1 72 h after the initial learning. CS1, conditioned
stimulus1; CS2, conditioned stiumulus2; US, unconditioned stimulus.
(C) In the 6-h group (n = 10), Daun02-injected rats display lower freezing to CS2 than those receiving vehicle (n = 10) or Daun02 without CS1 presentation (n = 8)
(H = 19.80, p < 0.0001).
(D) A similar result was observed during the subsequent presentation of CS1 (H = 17.83, p = 0.0001).
(E) Freezing to CS2 in 0-min (n = 10) and 6-h (n = 10) groups is lower than in the 7-day (n = 12) group (Kruskal-Wallis and Dunn test, H = 10.13, p = 0.0063).
(F) All three groups showed low freezing to CS1 (H = 5.75, p = 0.056).
(G) Similar experiments were repeated 3 weeks after learning to assess remote memories.
(H) At 3 weeks, in the 6-h group (n = 9), Daun02-injected rats display freezing to CS2 like the vehicle-injected rats (n = 7) (Mann-Whitney, U = 21.50, p = 0.308).
(I) On the contrary, in the presence of CS1, Daun02-injected rats display lower freezing than the vehicle-injected rats (U = 0 p = 0.0002).
(J) At 3 weeks, the 6-h group (n = 9) displays freezing to CS2 like the 7-day group (n = 8) but higher than the 0-min group (n = 9) (H = 16.76, p = 0.0002).
(K) In the presence of the CS1, all groups showed low freezing (H = 0.80, p = 0.669).
Data are means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001.
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
Article
Figure 2. The extinction of CS1 memory extinguished CS2 recent memories while leaving unaffected CS2 remote memories
(A) Schematic diagram for evaluating CS2 memory following CS1 extinction in wild-type rats at a recent time point (72 h after learning).
(B) CS1 memory was extinguished by repeatedly presenting the CS1 in wild-type rats conditioned at 0-min (n = 11), 6-h (n = 13), or 7-day (n = 12) delay.
(C) CS1 extinction decreases CS2 freezing in 0-min and 6-h groups but not in 7-day group (H = 19.42, p < 0.0001).
(D) A similar experiment was repeated at a remote time point (3 weeks after learning).
(E) Fear extinction curves obtained by repeatedly presenting CS1 at remote (3 weeks) time point in rats conditioned at 0-min (n = 16), 6-h (n = 15), or 7-day (n = 14)
delay.
(F) At 3 weeks, CS1 extinction extinguishes CS2 memory in 0-min but not in 6-h or 7-day groups (H = 20.63, p < 0.0001).
Data are means ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.
We aimed to verify that the CS2 freezing decrease was specifically attributed to the extinction of the related CS1 memory
and not to the mere repeated presentation of a tone. Thus, in
an additional group, rats learned only the CS2-US association
and were repeatedly presented with the 15-kHz tone (which
acted as the CS1 in the previous groups). This procedure did
not decrease freezing to CS2 (Figure S8A).
We repeated a similar experiment with CS1 extinction and CS2
memory tests performed 3 weeks after learning (Figure 2D). After
CS1 remote memory extinction (Figure 2E), freezing to CS2 in the
6-h group was higher than in the 0-min group and similar to that in
the 7-day group (Figure 2F). This result was not attributed to a
change in susceptibility to fear extinction at more remote time
points because the extinction curves were similar across recent
and remote time points (Figures 2B and 2E), and CS1 remote
memory was extinguished effectively (Figure S8B). We obtained
similar results by counterbalancing the tones (Figure S8C).
Thus, memories of temporarily related events were reciprocally linked at a recent time point such that the recall (or extinction) of one memory led to the recall (or extinction) of another.
However, memories of these events are functionally independent and refractory to the modification of the other at a remote
time point.
Two temporarily related but distinct events are allocated to
overlapping neuronal populations at a recent time point and to
mostly separate neuronal populations at a remote time point.
Shortly after two auditory fear events occurring closely in time,
overlapping populations of excitatory cells are allocated to both
memories in the BLA.10–13 Therefore, we examined the populations of excitatory neurons that underlie the recall of CS1 and
CS2 memories at recent and remote time points in the BLA. To
label neurons activated by each memory recall task, we injected
an inducible, activity-dependent virus cocktail of adeno-associ-
Cell Reports 43, 114151, May 28, 2024
ated virus AAV9-c-Fos-tTA and AAV9-TRE-enhanced yellow
fluorescent protein (EYFP) viruses, which labeled the neurons
expressing the immediate-early gene c-Fos in a doxycycline
(Dox)-dependent manner (Figures 3A and 3B).23,24 To specifically examine excitatory cells, brain sections were incubated
with anti-EYFP, anti-c-Fos, and anti-CamKII (a marker for excitatory cells) antibodies. Rats received CS1-US and CS2-US pairings separated for 6 h on a Dox-on diet. Seventy hours later, CS1
was presented on the Dox-off diet to label neurons activated by
CS1 recall with EYFP. Subsequently, the rats were fed a Dox-on
diet, administered CS2 (Figures 3C and 3D), and perfused. Neurons activated by CS2 memory recall were detected by immunohistochemistry for endogenous c-Fos. The presentation of CS1
resulted in a minimal percentage of activated excitatory cells expressing EYFP in rats under the Dox-on diet (approximately 2%
labeled cells, Figure S9). Conversely, in rats following the Dox-off
diet, the percentage of activated excitatory cells expressing
EYFP, including background positive cells, was notably elevated
(Figures 3E and 3G). CS2 presentation led to endogenous c-Fos
expression (Figures 3E and 3H). Critically, the percentage of
excitatory neurons activated by both CS1 and CS2 was significantly higher than that expected by chance (Figures 3E and 3I).
We repeated a similar analysis by reactivating CS1 and CS2
memories 3 weeks after learning. Freezing to CS1 and CS2
was similar at the remote and recent intervals (Figure 3D). However, significantly fewer excitatory neurons were activated only
by CS1 at the remote time point than that in the recent time point
(Figures 3F and 3G). Conversely, we observed no difference in
the percentage of neurons activated only by CS2 (Figures 3F
and 3H). Critically, fewer neurons were activated by both CS1
and CS2 at the remote versus recent time point (Figure 3I). To
further depict the differential activation of neurons between
recent and remote intervals, we analyzed the ratio of EYFP and
Article
OPEN ACCESS
Figure 3. Memories of related events demonstrate higher overlapped neuronal ensemble at recent than at remote time points
(A and B) Viruses are injected into the BLA in wild-type rats on a doxycycline (Dox) diet.
(C) Rats feed on Dox-off diet during CS1 presentation (green). Next, they feed on Dox-on diet, and CS2 is presented (red).
(D) Freezing to CS1 (left) and CS2 (right) was similar between rats tested at recent (n = 9) and remote (n = 7) time points (Mann-Whitney, CS1, U = 20, p = 0.238;
CS2, U = 16.50, p = 0.113).
(E and F) To specifically examine excitatory cells, brain sections were incubated with anti-EYFP, anti-c-Fos, and anti-CamKII (a marker for excitatory cells)
antibodies. Confocal images of excitatory neurons expressing EYFP, c-Fos, or both at recent (E) or remote (F) time points. Arrows indicate recall-activated
neurons. Scale bar, 20 mm.
(G) Neurons activated by the only CS1 presentation (EYFP+CS1) are calculated as total EYFP+ cells minus EYFP+c-Fos+ cells. EYFP+CS1 cells decrease more in
the remote (n = 7) than in the recent (n = 9) time point (Student’s t test, t(14) = 4.19, p = 0.0009).
(H) Neurons activated by the only CS2 presentation (c-Fos+CS2) are calculated as total c-Fos+ cells minus EYFP+c-Fos+ cells. c-Fos+CS2 expression is similar
across time intervals (t(14) = 0.17, p = 0.866).
(I) EYFP+c-Fos+ cells decrease more in remote than in recent time points (t(14) = 2.38, p = 0.031). The percentage of overlapped neurons is over chance levels
(dashed lines) in the recent (t(8) = 9.03, p < 0.0001) and remote (t(6) = 3.94, p = 0.0076) time points.
(J) The reactivation rate (EYFP+c-Fos+/c-Fos+) decreases more in remote than in recent time points (t(14) = 3.42, p = 0.0041). The percentage of overlapped
ensemble is over chance levels in both recent (t(8) = 14.15, p < 0.0001) and remote time points (t(6) = 5.20, p = 0.002).
Data are means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001.
c-Fos double-positive cells among c-Fos+ cells (EYFP+c-Fos+/cFos+).25 A significant difference between the recent and remote
time points was observed (Figure 3J). Despite being lower than
that at a recent time point, the percentage of overlapping
neuronal populations at a remote time point was significantly
higher than chance levels (Figures 3I and 3J), e.g., remote memories still maintained a significant population of overlapping
neurons.
Taken together, the percentage of overlapping neuronal ensembles activated by both CS1 and CS2 significantly decreased
at remote time points, where the most distinct neuronal populations underlie CS1 and CS2 memories. At the behavioral level,
the disruption of neurons activated by CS1 remote recall did
not affect CS2 memory, and CS1 extinction did not extinguish
CS2 memory. Therefore, this process reduced the interference
across related memories. Meanwhile, the neuronal overlap persisting at a remote time point may allow the maintenance of information on the shared features of related experiences.
Parvalbumin-positive interneurons within the BLA are
necessary to form separate remote memories
Next, we investigated cellular mechanisms that may underlie the
formation of independent remote memories. Upon or shortly after learning two contiguous events, the increase in the excitability of excitatory neurons could bias the encoding of a subsequent memory to neurons that encode the first memory.10–13,26
Because gamma-aminobutyric acid (GABA)-ergic interneurons
inhibit excitatory neurons, we hypothesized that inhibitory neurons may be activated to circumscribe the shared population
of excitatory neurons across related memories.
In the BLA, a large proportion of inhibitory synapses onto
excitatory neurons are formed by interneurons that express the
calcium-binding protein parvalbumin (PV).27,28 Therefore, we
analyzed the activity of PV-positive cells in the 0-min, 6-h, and
7-day groups as well as in naive rats. In the 0-min group, the
learning trial occurred in only one session, whereas the other
groups underwent two different learning sessions. Thus, in the
0-min group, the first trial was repeated twice, 6 h apart, like
that in the 6-h group. The animals were euthanized 90 min after
the last learning session, and the number of PV cells expressing
the c-Fos protein was analyzed. c-Fos expression has often
been used as a marker of neuronal activity in excitatory and
inhibitory neurons.29,30 In all conditioned groups, c-Fos expression was higher than in naive animals (Figures 4A and 4B). Critically, the number of PV cells also expressing c-Fos was higher in
the 6-h group than in the other groups (Figures 4A and 4C).
To investigate whether PV cells are involved in the formation
of separate remote memory, we injected an AAV encoding the
Cre-dependent hM4D(Gi) inhibitory designer receptor exclusively activated by designer drugs (DREADD) (AAV5-hSyn-DIOhM4D(Gi)-mCherry) into the BLA of PV-Cre rats (Figure 4D).31,32
Cre-dependent hM4D(Gi)-mCherry was specifically expressed
in PV-positive BLA neurons (Figure 4E). Another group received
the control AAV5-hSyn-DIO-mCherry. Both groups underwent
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
Article
Figure 4. PV cells in the BLA are necessary for the formation of separate remote memories
(A) Costaining of parvalbumin (PV) and c-Fos in the BLA of naive rats and those conditioned with CS1-US and CS2-US at 0-min, 6-h, and 7-day delay. Arrows
indicate PV-positive and c-Fos-positive cells. Scale bar, 20 mm.
(B) c-Fos expression is lower in naive (n = 10) rats than in other groups (one-way ANOVA and Tukey post-test, F(3,43) = 51.04, p < 0.0001). No differences are
observed among the 0-min (n = 12), 6-h (n = 12), and 7-day (n = 13) groups.
(C) PV-positive cells that also express c-Fos are higher in the 6-h group than in other groups (F(3,43) = 11.79, p < 0.0001).
(D) Expression of AAV5-hSyn-DIO-hM4D(Gi)-mCherry in BLA PV neurons of PV-Cre rat.
(E) Dual-labeled hM4D(Gi)-mCherry/PV neurons are indicated with white arrows. Scale bar, 20 mm.
(F) Experimental design.
(G) CS1 extinction decreases CS2 memory in DIO-hM4D(Gi)-mCherry rats (n = 10) compared with DIO-mCherry rats (n = 10) and DIO-hM4Di-mCherry rats
conditioned at a 7-day interval (n = 7) (Kruskal-Wallis and Dunn post hoc, H = 14.78, p = 0.0006).
(H) In the absence of CS1 extinction, CS1 and CS2 remote memories are similar between DIO-hM4D(Gi)-mCherry (n = 8) and DIO-mCherry (n = 6) rats (MannWhitney, CS1, U = 14.50, p = 0.208; CS2, U = 13.00, p = 0.165).
(I and J) Bicuculline injection into BLA decreases freezing to CS2 after CS1 extinction (n = 9) compared to saline-injected rats (n = 13) and bicuculline-injected rats
conditioned at a 7-day interval (n = 8) (H = 9.93, p = 0.007).
(legend continued on next page)
Cell Reports 43, 114151, May 28, 2024
Article
CS1-US pairing, followed by CS2-US pairing 6 h later. Clozapine-N-oxide (CNO) was injected intraperitoneally 60 min after
CS2-US learning to inactivate PV cells during the memory
consolidation process at a time point at which PV activity was
increased in the 6-h group. 3 weeks later, CS1 remote memory
was extinguished (Figure 4F). All groups demonstrated a similar
extinction curve (Figure S10A). The subsequent day, we assessed CS2 memory retention. Freezing to CS2 was lower in
the DIO-hM4D(Gi)-mCherry group than that in the DIOmCherry group (Figure 4G). In an additional group, in which
DIO-hM4D(Gi)-mCherry PV-Cre rats were conditioned to CS1US and CS2-US separated by 7 days, CS2 memory was unaffected by CS1 extinction (Figure 4G).
We repeated the previous experiment without extinguishing
the CS1 memory, i.e., by presenting only a few CS1 and assessing CS2 memory the following day. No differences were detected between groups (Figure 4H). Thus, the decrease in
freezing to CS2 in DIO-hM4D(Gi)-mCherry PV-Cre rats in the
6-h group was selectively attributed to the extinction of CS1
memory, i.e., the two memories were still reciprocally linked13
To further verify that GABAergic neurotransmission in BLA is
involved in the formation of separate remote memories of temporally related fear events, we injected bicuculline, a selective
GABAa receptor inhibitor, into the BLA (Figures S10B and
S10C) of wild-type rats 90 min after learning. Three weeks later,
CS1 memory was extinguished (Figure S10D), and CS2 memory
was assessed (Figure 4I). Compared with saline-injected rats,
freezing to CS2 decreased in the 6-h group but not in the
7-day group (Figure 4J). Importantly, bicuculline did not affect
CS1 or CS2 memory retention in the absence of CS1 extinction
(Figure 4K). Finally, bicuculline (or saline) was injected into the
BLA of c-Fos-lacZ transgenic rats 90 min after learning. Three
weeks later, we reactivated the CS1 memory and injected
Daun02 to block CS1-memory-related neurons. During the
CS2 memory retention test, rats that received bicuculline displayed a weaker freezing than those that received saline
(Figures 4I, 4L, and S10E).
These data identified an important function of GABAergic cells
within the BLA in memory processes. Chemogenetic inactivation
of PV cells or GABAergic neurotransmission blockade did not
affect the retention of individual remote memories but prevented
the segregation of each remote memory into a functionally independent memory such that the two memories remained reciprocally linked. This function was related to the formation of separate memories of temporally related events but not separate
memories of events that occurred 7 days apart.
Hippocampus regulates the segregation of remote
memories through axons terminating directly onto the
GABAergic interneurons of the BLA
The precise encoding and retrieval of distinct memories of
related events are the central functions of the hippocampus.
OPEN ACCESS
The hippocampus performs this function through a process
termed pattern separation, which allows the conversion of highly
related overlapping inputs into non-overlapping distinct outputs.1–7 This process may reduce interference across related
memories, ensuring that other related memories are not recalled
during the retrieval of a specific memory and that the specificity
of memories is maintained over time.1–7 Therefore, we investigated whether the hippocampus is involved in reducing the overlap between remote auditory fear memories and its underlying
mechanism.
c-Fos-lacZ transgenic rats were subjected to two learning
events separated by 6 h (Figure 5A). Ninety minutes later,
Daun02 (or vehicle) was administered into the dorsal hippocampus (Figures S11A) to inactivate the neurons specifically activated by these events. Three weeks later, the remote CS1 memory was extinguished (Figure 5B), and CS2 memory was
assessed. Compared to the control animals, freezing to CS2
was lower in Daun02-injected animals (Figure 5C). Animals that
received Daun02, in which CS1 was not extinguished, displayed
high freezing to CS2 (Figure 5C). Thus, inactivating the dorsal
hippocampus prevented the segregation of the two remote
memories.
Thus far, CS1-US and CS2-US pairings have occurred in two
distinct environments to differentiate between the two learning
events. Next, we sought to determine whether hippocampal
participation was related to the management of distinct contextual memories or distinct episodic fear memories regardless of
the learning environment. We performed a similar experiment
with CS1-US and CS2-US pairings in the same context (Figure 5D). During CS1 extinction trials, control rats exhibited a
slower extinction curve. Nevertheless, ultimately, both groups
effectively extinguished CS1 memory (Figure 5E). During CS2
memory retention, freezing to CS2 was lower in Daun02-injected
animals than in control animals (Figure 5F). Therefore, the dorsal
hippocampus plays a key role in ensuring that remote memories
of events in temporal proximity are stored as distinct memories
also when the two events occurred in the same context.
Next, we investigated whether the hippocampus interacts with
the BLA to form distinct remote memories. We injected a retrograde AAVrg-hSyn-hM4D(Gi)-mCherry into the BLA to label
and manipulate the hippocampal excitatory axons that terminated in the BLA (Figures 6A and 6B). The control group received
AAVrg-hSyn-mCherry. Both groups underwent two auditory fear
learning sessions separated by 6 h. Ninety minutes later, CNO
was injected into the hippocampus to inactivate the excitatory
neurons that projected to the BLA. Because most afferents to
the BLA exited the ventral hippocampus,33,34 CNO was administered to the ventral hippocampal region (Figures 6C and 6D).
Three weeks later, CS1 was extinguished (Figure 6E), and CS2
memory was assessed. Compared to AAVrg-mCherry animals,
AAVrg-hM4D(Gi)-mCherry animals froze to CS2 significantly
less (Figure 6F). We did not find any effects in the absence of
(K) In the absence of CS1 extinction, CS1 and CS2 memories are similar between bicuculline- (n = 10) and saline-injected rats (n = 8) (CS1, U = 22.00, p = 0.115;
CS2, U = 32.00, p = 0.498).
(L) In c-Fos-lacZ rats, Daun02 injection after CS1 memory recall decreases freezing to CS2 in bicuculline-injected rats (n = 6) compared with saline-injected rats
(n = 6) (U = 3.00, p = 0.0152).
Data are means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001.
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
Article
Figure 5. The dorsal hippocampus is necessary for the separation of temporally related
remote memories
(A) Experimental design.
(B) Curves of fear extinction performed 3 weeks
after learning in c-Fos-lacZ transgenic rats that
received Daun02 (n = 14) or vehicle (n = 10).
(C) CS2 freezing is lower in Daun02- than in
vehicle-injected rats and Daun02-injected rats
where CS1 is not extinguished (n = 11) (KruskalWallis and Dunn post hoc, H = 24.29, p < 0.0001).
(D) A similar experiment was repeated by presenting CS1 and CS2 in the same conditioning
context.
(E) Curves of fear extinction performed 3 weeks
after learning in c-Fos-lacZ transgenic (n = 10) and
vehicle-injected (n = 14) rats that received the CS1US and CS2-US pairings in the same conditioning
cage.
(F) In rats that received CS1-US and CS2-US
pairings in the same context, freezing to CS2 is
lower in Daun02- than in vehicle-injected rats
(Mann-Whitney, U = 25.50, p = 0.007).
Data are means ± SEM. *p < 0.05, **p < 0.01,
***p < 0.001, and ****p < 0.0001.
CS1 memory extinction (Figures 6F, S11B, and S11C), e.g., the
decrease of freezing to CS2 was related to CS1 memory extinction. Combined, these data showed that both the dorsal hippocampus and the output pathway that exit from the ventral hippocampus are necessary for the segregation of remote memories.
Axons arising from hippocampal excitatory neurons synapse
onto either excitatory or inhibitory neurons in the BLA.35 Thus,
the hippocampus might drive the separation of remote memories
by directly recruiting GABAergic interneurons to the BLA. To test
this possibility, we exploited the anterograde trans-synaptic
properties of AAV1-Cre vectors to drive Cre-dependent transgene expression in postsynaptic neuronal targets from transduced presynaptic neurons.36 We injected a Cre-dependent
inhibitory DREADD (AAV1-hDlx-DIO-KORD-mCyRFP) vector
into the BLA and a trans-synaptic AAV1-hSyn-Cre virus into
the ventral hippocampus to drive their selective recombination
in the BLA GABAergic cells receiving axons from the hippocampus (Figures 6G and 6H). Five weeks later, these viruses were
selectively expressed in GABAergic BLA cells (Figures 6I and
Cell Reports 43, 114151, May 28, 2024
6J), largely infecting PV cells (Figure 6J).
Two groups of rats were injected with
the two viruses in the hippocampus and
BLA or with DIO-KORD-mCyRFP alone
in the BLA. Both groups received salvinorin B37 into the BLA (Figures 6K and
S11D) 90 min after learning to inhibit
GABAergic cells specifically receiving
hippocampal projections. Three weeks
later, CS1 memory was extinguished (Figure 6L), and CS2 memory was assessed.
Compared to control rats, those injected
with both viruses displayed lower freezing
to CS2 (Figure 6M). Thus, direct projections from the hippocampus to the GABAergic interneurons
within the BLA are necessary for segregating remote auditory
fear memories.
DISCUSSION
This study provides insights into the neuronal processes that
reduce the overlap between distinct but temporally related memories, such that related memories can be stored and retrieved
independently.
Consistent with the memory allocation hypothesis,10–13 we
found that when two different fear events occurred closely in
time, recent memories of those events shared overlapping
neuronal populations in the BLA and were functionally linked.
However, recalling or extinguishing one memory did not affect
the recall or extinction of the other memory at later time points.
Thus, remote memories of related events are functionally independent, ensuring memory specificity over time and preserving
the integrity of one memory in case of modifications or disruption
Article
OPEN ACCESS
Figure 6. Excitatory hippocampal neurons that project to the BLA enable the separation of temporally related remote memories
(A) Experimental design.
(B) A retrograde AAVrg-hSyn-hM4D(Gi)-mCherry or a control retrograde AAVrg-hSyn-mCherry virus is infused into the BLA of wild-type rats.
(C) Clozapine-N-oxide (CNO) was injected into the ventral hippocampus. Scale bar, 1,000 mm (left). Plates were adapted from the Paxinos and Watson atlas
(right).
(D) AAVs injected into the BLA (left; scale bars, 500 and 50 mm) are retrogradely transported toward cell bodies of the ventral hippocampus (right; scale bar, 20 mm).
(E) Curves of fear extinction performed 3 weeks after learning in the wild-type rats that received the retrograde AAVrg-hSyn-hM4D(Gi)-mCherry virus (n = 6) or
AAVrg-hSyn-mCherry virus (n = 9) into the BLA.
(F) CS1 memory extinction decreases freezing to CS2 in AAVrg-hM4D(Gi)-mCherry rats compared with AAVrg-mCherry rats and AAVrg-hM4D-mCherry rats
(n = 7) where CS1 is not extinguished (H = 8.27, p = 0.016).
(G and H) To label axons arising from the hippocampus and contacting the GABAergic cells in the BLA, we injected an AAV1-hSyn-Cre trans-synaptic vector in the
hippocampus and an AAV1-hDlx-DIO-KORD-mCyRFP1 in the BLA of wild-type rats.
(I) Confocal images depict a selective recombination of AAV1-hSyn-Cre and AAV1-Dlx-DIO-KORD-mCyRFP in GAD+ and PV+ cells in the BLA. Scale bar, 20 mm.
(J) Five weeks after virus injections, 4 rats were submitted to the immunohistochemical analysis of virus expression, and the remaining control (n = 11) and AAV1hSyn-Cre- and AAV1-Dlx-DIO-KORD-mCyRFP-injected (n = 8) rats were submitted to behavioral experiments. In the BLA, nearly half of GABAergic interneurons
(GAD65/67-positive neurons) showed mCyRFP expression, and most neurons expressing mCyRFP were colabeled with GAD65/67 (left). Moreover, most
neurons expressing mCyRFP were colabeled with PV (left).
(K) In the behavioral experiments, wild-type rats received the injection of salvinorin B (SalB; 3 mM) directly into the BLA 90 min after CS2-US pairing. Example of
micrographs of the needle track in the BLA is shown. The higher magnification shows the absence of neuronal death after SalB injection. Scale bars, 200 mm.
(L) Curves of CS1 memory fear extinction performed 3 weeks after learning in both groups.
(M) After CS1 extinction, freezing to CS2 is lower in Cre+DIO-KORD-mCyRFP rats (n = 8) than in DIO-KORD-mCyRFP rats (n = 11) (U = 10.50, p = 0.0039).
Data are means ± SEM. *p < 0.05 and **p < 0.01.
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
of the other. At the neuronal level, the two remote memories
mostly recruited distinct neuronal populations; however, they
also shared a common neuronal population smaller than that
of recent memories but significantly higher than that by chance.
Common neuronal populations may capture shared features
across related experiences.12 Thus, common neuronal ensembles persisting across the two memories at remote time points
may allow the maintenance of shared features of past experiences. Therefore, remote memories of two related events can
maintain their specificity over time while retaining information
about shared features from past experiences.
Numerous studies have implicated BLA PV neurons in fear
learning.32,38–40 Consistent with these findings, we observed
an increase in activated PV-positive cells 90 min after learning
in all conditioned groups, compared with naive rats, with a higher
percentage of activated PV neurons in rats that learned two
events in temporal proximity. Interestingly, the blockade of PVcontaining neurons did not affect the retention of individual
remote memory but disrupted the segregation of the two memories. This highlights the selective role of BLA PV cells in the
consolidation of related remote fear memories. Notably, PV cells
are required for the formation of separate remote memories as
early as 90 min after learning, despite recent memories still being
functionally linked 72 h after learning. Remote memories require
cells that are generated or ‘‘tagged’’ shortly after learning.41,42
PV cells may play a role in influencing these cells to form distinct
remote memories.
Furthermore, the hippocampus sends axons that directly synapse onto GABAergic interneurons in the BLA, and blocking
these axons prevents the formation of separate remote memories. The mentioned mechanism relies specifically on the segregation of remote memories but not on the formation of individual
remote memories, highlighting an important distinction in the
neural mechanisms underlying these processes. The hippocampus plays a vital role in managing highly related experiences
through two processes: pattern separation and pattern competition.1–7 Pattern separation reduces overlap among pieces of information to facilitate separate storage of this information, while
pattern competition enables the recall of stored information from
partial cues.1–7 Our findings reveal a mechanism by which the
hippocampus performs pattern separation, i.e., by directly activating BLA GABAergic interneurons. Consistent with this idea,
human studies have also shown an interplay between the hippocampus and amygdala in the pattern separation of episodic and
emotional memories.14,15
Alterations in neural mechanisms that underlie the distinction
across highly similar memories to preserve memory specificity
from interference are at the core of neuropsychiatric disorders,
such as schizophrenia and PTSD. The hippocampus is particularly vulnerable to high-stress conditions, and, consequently,
pattern separation processes may be compromised. Additionally, decreased GABA levels are associated with PTSD. Therefore, dysfunction in the neural processes we have described
may contribute to the development of these disorders.
Limitations of the study
Experiments conducted with c-Fos-lacZ transgenic rats show a
certain variability in behavior. This could be due to the fact that
Cell Reports 43, 114151, May 28, 2024
Article
Daun02 might have destroyed a different percentage of neurons in different rats. To clarify this aspect, we analyzed the
expression of b-gal in the remaining neurons and found no statistically significant differences between groups. However, it
would be ideal to determine the direct correspondence between learned behavior and the number of neurons killed by
Daun02. Such a measure, however, would likely not be accurate due to several factors, including the variability in the number of neurons activated prior to Daun02 administration. It is
worth noting that despite this variability, the data obtained
with c-Fos-lacZ transgenic rats show a clear statistical difference between groups, and these results have been fully
confirmed with further experiments carried out by extinguishing
mnemonic traces in wild-type animals, which did not receive
any active compound.
Moreover, although we conducted chemogenetic experiments primarily targeting PV cells, we also utilized bicuculline—a selective inhibitor of overall GABAergic transmission—
and a viral construct that infected PV cells as well as other
GABAergic cells, such as somatostatin (SOM)-expressing cells.
Therefore, we cannot rule out the possibility that additional
GABAergic interneurons, aside from PV cells in the BLA, may
contribute to the segregation of mnemonic traces. For example,
recent findings have demonstrated that the modulation of the activity of SOM cells within the hippocampus can either promote or
inhibit the formation of new fear memories.43 A comparable
mechanism may also be at play in the BLA to segregate fear
memories. Subsequent studies are thus necessary to delve
into this aspect further.
Finally, our study showed that besides the dorsal hippocampus, the output pathway that exits from the ventral hippocampus is also necessary for the segregation of remote memories. It could therefore be that the dorsal hippocampus
processes the pattern separation of remote memories and,
through the output pathway arising from the ventral hippocampus, regulates the separation of traces in target sites,
such as the BLA. Alternatively, both dorsal and ventral regions
of the hippocampus may be involved in the pattern separation
process. Future studies will help distinguish between these
possibilities.
STAR+METHODS
Detailed methods are provided in the online version of this paper and include
the following:
KEY RESOURCES TABLE
RESOURCE AVAILABILITY
B Lead contact
B Materials availability
B Data and code availability
EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS
B Animals
B Behavioral procedures
QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.
celrep.2024.114151.
Article
ACKNOWLEDGMENTS
We thank S. Ramirez (Dept. of Psychological and Brain Sciences, Boston, MA,
USA) for the critical reading of the manuscript. We also thank S. Ramirez and S.
Grella (Dept. of Psychological and Brain Sciences, Boston, MA, USA) for the
kind gift of the AAV9-c-Fos-tTA and AAV9-TRE-EYFP viruses and G. Rainer
(University of Fribourg, Switzerland) for providing PV-Cre transgenic rats.
This work was supported by the Fondazione Cariverona (no. 2018.0780),
ricerca di Rilevante Interesse Nazionale (PRIN)’’ (20178NNRCR_002 and
2022BYZPFE) from the Italian Ministry of University and Research (MIUR).
AUTHOR CONTRIBUTIONS
G.C., A.R., and F.S. carried out and analyzed behavioral experiments; G.C.,
L.M., and E.M. performed immunohistochemical, immunofluorescent, and
confocal microscopy analyses; and B.S. devised and analyzed the experiments and wrote the manuscript. All authors discussed the results and commented the manuscript.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: May 24, 2023
Revised: February 21, 2024
Accepted: April 9, 2024
REFERENCES
1. Marr, D. (1971). Simple memory: a theory for archicortex. Phil. Trans. R.
Soc. Lond. B 262, 23–81.
2. Treves, A., and Rolls, E.T. (1994). Computational analysis of the role of the
hippocampus in memory. Hippocampus 4, 374–391.
3. McClelland, J.L., McNaughton, B.L., and O’Reilly, R.C. (1995). Why there
are complementary learning systems in the hippocampus and neocortex:
insights from the successes and failures of connectionist models of
learning and memory. Psychol. Rev. 102, 419–457.
4. O’Reilly, R.C., and Rudy, J.W. (2000). Computational principles of learning
in the neocortex and hippocampus. Hippocampus 10, 389–397.
5. Norman, K.A., and O’Reilly, R.C. (2003). Modeling hippocampal and
neocortical contributions to recognition memory: A complementarylearning-systems approach. Psychol. Rev. 110, 611–646.
6. Brady, T.F., Konkle, T., Alvarez, G.A., and Oliva, A. (2008). Visual long-term
memory has a massive storage capacity for object details. Proc. Natl.
Acad. Sci. USA 105, 14325–14329.
7. Leal, S.L., and Yassa, M.A. (2018). Integrating new findings and examining
clinical applications of pattern separation. Nat. Neurosci. 21, 163–173.
8. Anacker, C., and Hen, R. (2017). Adult hippocampal neurogenesis and
cognitive flexibility – linking memory and mood. Nat. Rev. Neurosci. 18,
335–346.
9. Cayco-Gajic, N.A., and Silver, R.A. (2019). Re-evaluating Circuit Mechanisms Underlying Pattern Separation. Neuron 101, 584–602.
10. Cai, D.J., Aharoni, D., Shuman, T., Shobe, J., Biane, J., Song, W., Wei, B.,
Veshkini, M., La-Vu, M., Lou, J., et al. (2016). A shared neural ensemble
links distinct contextual memories encoded close in time. Nature 534,
115–118.
11. Chowdhury, A., Luchetti, A., Fernandes, G., Filho, D.A., Kastellakis, G.,
Tzilivaki, A., Ramirez, E.M., Tran, M.Y., Poirazi, P., and Silva, A.J. (2022).
A locus coeruleus-dorsal CA1 dopaminergic circuit modulates memory
linking. Neuron 110, 3374–3388.e8.
12. de Sousa, A.F., Chowdhury, A., and Silva, A.J. (2021). Dimensions and
mechanisms of memory organization. Neuron 109, 2649–2662.
OPEN ACCESS
13. Rashid, A.J., Yan, C., Mercaldo, V., Hsiang, H.L.L., Park, S., Cole, C.J., De
Cristofaro, A., Yu, J., Ramakrishnan, C., Lee, S.Y., et al. (2016). Competition between engrams influences fear memory formation and recall. Science 353, 383–387.
14. Leal, S.L., Noche, J.A., Murray, E.A., and Yassa, M.A. (2017). Age-related
individual variability in memory performance is associated with amygdalahippocampal circuit function and emotional pattern separation. Neurobiol.
Aging 49, 9–19.
15. Zheng, J., Stevenson, R.F., Mander, B.A., Mnatsakanyan, L., Hsu, F.P.K.,
Vadera, S., Knight, R.T., Yassa, M.A., and Lin, J.J. (2019). Multiplexing of
Theta and Alpha Rhythms in the Amygdala-Hippocampal Circuit Supports
Pattern Separation of Emotional Information. Neuron 102, 887–898.e5.
16. Koya, E., Golden, S.A., Harvey, B.K., Guez-Barber, D.H., Berkow, A., Simmons, D.E., Bossert, J.M., Nair, S.G., Uejima, J.L., Marin, M.T., et al. (2009).
Targeted disruption of cocaine-activated nucleus accumbens neurons prevents context-specific sensitization. Nat. Neurosci. 12, 1069–1073.
17. Grosso, A., Cambiaghi, M., Renna, A., Milano, L., Roberto Merlo, G.,
Sacco, T., and Sacchetti, B. (2015). The higher order auditory cortex is
involved in the assignment of affective value to sensory stimuli. Nat. Commun. 6, 8886.
18. Grosso, A., Santoni, G., Manassero, E., Renna, A., and Sacchetti, B.
(2018). A neuronal basis for fear discrimination in the lateral amygdala.
Nat. Commun. 9, 1214.
19. Lay, B.P.P., Koya, E., Hope, B.T., Esber, G.R., and Iordanova, M.D. (2023).
The Recruitment of a Neuronal Ensemble in the Central Nucleus of the
Amygdala During the First Extinction Episode Has Persistent Effects on
Extinction Expression. Biol. Psychiatry 93, 300–308.
20. Sacco, T., and Sacchetti, B. (2010). Role of secondary sensory cortices in
emotional memory storage and retrieval in rats. Science 329, 649–656.
21. Concina, G., Renna, A., Milano, L., and Sacchetti, B. (2022). Prior fear
learning enables the rapid assimilation of new fear memories directly
22. Concina, G., Renna, A., Grosso, A., and Sacchetti, B. (2019). The auditory
cortex and the emotional valence of sounds. Neurosci. Biobehav. Rev. 98,
256–264.
23. Roy, D.S., Arons, A., Mitchell, T.I., Pignatelli, M., Ryan, T.J., and Tonegawa, S. (2016). Memory retrieval by activating engram cells in mouse
models of early Alzheimer’s disease. Nature 531, 508–512.
24. Chen, B.K., Murawski, N.J., Cincotta, C., McKissick, O., Finkelstein, A.,
Hamidi, A.B., Merfeld, E., Doucette, E., Grella, S.L., Shpokayte, M., et al.
(2019). Artificially Enhancing and Suppressing Hippocampus-Mediated
Memories. Curr. Biol. 29, 1885–1894.e4.
25. Zhang, X., Kim, J., and Tonegawa, S. (2020). Amygdala Reward Neurons
Form and Store Fear Extinction Memory. Neuron 105, 1077–1093.e7.
26. Han, J.H., Kushner, S.A., Yiu, A.P., Cole, C.J., Matynia, A., Brown, R.A.,
Neve, R.L., Guzowski, J.F., Silva, A.J., and Josselyn, S.A. (2007). Neuronal
competition and selection during memory formation. Science 316,
457–460.
27. McDonald, A.J. (1992). Cell types and intrinsic connections of the amygdala. In The Amygdala: Neurobiological Aspects of Emotion, Memory,
and Mental Dysfunction (New York: Wiley), pp. 67–96.
28. McDonald, A.J., and Betette, R.L. (2001). Parvalbumin containing neurons
in the rat basolateral amygdala: morphology and co-localization of
calbindin-D (28k). Neuroscience 102, 413–425.
29. Cohen, S.M., Ma, H., Kuchibhotla, K.V., Watson, B.O., Buzsa´ki, G.,
Froemke, R.C., and Tsien, R.W. (2016). Excitation-Transcription Coupling
in Parvalbumin-Positive Interneurons Employs a Novel CaM KinaseDependent Pathway Distinct from Excitatory Neurons. Neuron 90,
292–307.
30. Madsen, H.B., Guerin, A.A., and Kim, J.H. (2017). Investigating the role of
dopamine receptor- and parvalbumin-expressing cells in extinction of
conditioned fear. Neurobiol. Learn. Mem. 145, 7–17.
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
31. Wright, A.M., Zapata, A., Hoffman, A.F., Necarsulmer, J.C., Coke, L.M.,
Svarcbahs, R., Richie, C.T., Pickel, J., Hope, B.T., Harvey, B.K., and Lupica, C.R. (2021). Effects of Withdrawal from Cocaine Self-Administration
on Rat Orbitofrontal Cortex Parvalbumin Neurons Expressing Cre recombinase: Sex-Dependent Changes in Neuronal Function and Unaltered
Serotonin Signaling. eNeuro 8, 8.
32. Yau, J.O.Y., Chaichim, C., Power, J.M., and McNally, G.P. (2021). The
Roles of Basolateral Amygdala Parvalbumin Neurons in Fear Learning.
J. Neurosci. 41, 9223–9234.
33. Ottersen, O.P. (1982). Connections of the amygdala of the rat. IV: Corticoamygdaloid and intraamygdaloid connections as studied with axonal
transport of horseradish peroxidase. J. Comp. Neurol. 205, 30–48.
34. Pitka¨nen, A., Pikkarainen, M., Nurminen, N., and Ylinen, A. (2000). Reciprocal connections between the amygdala and the hippocampal formation,
perirhinal cortex, and postrhinal cortex in rat. A review. Ann. N. Y. Acad.
Sci. 911, 369–391.
€bner, C., Bosch, D., Gall, A., Lu
€thi, A., and Ehrlich, I. (2014). Ex vivo
35. Hu
dissection of optogenetically activated mPFC and hippocampal inputs
to neurons in the basolateral amygdala: implications for fear and emotional
memory. Front. Behav. Neurosci. 8, 64.
36. Zingg, B., Chou, X.L., Zhang, Z.G., Mesik, L., Liang, F., Tao, H.W., and
Zhang, L.I. (2017). AAV-Mediated Anterograde Transsynaptic Tagging:
Mapping Corticocollicular Input-Defined Neural Pathways for Defense Behaviors. Neuron 93, 33–47.
37. Marchant, N.J., Whitaker, L.R., Bossert, J.M., Harvey, B.K., Hope, B.T.,
Kaganovsky, K., Adhikary, S., Prisinzano, T.E., Vardy, E., Roth, B.L., and
Shaham, Y. (2016). Behavioral and Physiological Effects of a Novel
Kappa-Opioid Receptor-Based DREADD in Rats. Neuropsychopharmacology 41, 402–409.
38. Windels, F., Crane, J.W., and Sah, P. (2010). Inhibition dominates the early
phase of up-states in the basolateral amygdala. J. Neurophysiol. 104,
3433–3438.
€ndemann, J., Tovote, P., Krabbe, S., Jacobson, G.A.,
39. Wolff, S.B.E., Gru
€ller, C., Herry, C., Ehrlich, I., Friedrich, R.W., Letzkus, J.J., and Lu
€thi,
A. (2014). Amygdala interneuron subtypes control fear learning through
disinhibition. Nature 509, 453–458.
Cell Reports 43, 114151, May 28, 2024
Article
40. Lucas, E.K., Jegarl, A.M., Morishita, H., and Clem, R.L. (2016). Multimodal
and site-specific plasticity of amygdala parvalbumin interneurons after
fear learning. Neuron 91, 629–643.
41. Lesburgue`res, E., Gobbo, O.L., Alaux-Cantin, S., Hambucken, A., Trifilieff,
P., and Bontempi, B. (2011). Early tagging of cortical networks is required
for the formation of enduring associative memory. Science 331, 924–928.
42. Kitamura, T., Ogawa, S.K., Roy, D.S., Okuyama, T., Morrissey, M.D.,
Smith, L.M., Redondo, R.L., and Tonegawa, S. (2017). Engrams and circuits crucial for systems consolidation of a memory. Science 356, 73–78.
43. Sharma, V., Sood, R., Khlaifia, A., Eslamizade, M.J., Hung, T.Y., Lou, D.,
Asgarihafshejani, A., Lalzar, M., Kiniry, S.J., Stokes, M.P., et al. (2020).
eIF2a controls memory consolidation via excitatory and somatostatin neurons. Nature 586, 412–416.
44. Ramirez, S., Liu, X., Lin, P.A., Suh, J., Pignatelli, M., Redondo, R.L., Ryan,
T.J., and Tonegawa, S. (2013). Creating a false memory in the hippocampus. Science 341, 387–391.
45. Paxinos, G., and Watson, C. (2007). The Rat Brain in Stereotaxic Coordinates (London: Academic Press Elsevier).
46. Ramanathan, K.R., Jin, J., Giustino, T.F., Payne, M.R., and Maren, S.
(2018). Prefrontal projections to the thalamic nucleus reuniens mediate
fear extinction. Nat. Commun. 9, 4527.
47. MacLaren, D.A.A., Browne, R.W., Shaw, J.K., Krishnan Radhakrishnan, S.,
Khare, P., Espan˜a, R.A., and Clark, S.D. (2016). Clozapine N-Oxide Administration Produces Behavioral Effects in Long-Evans Rats: Implications for
Designing DREADD Experiments. eNeuro 3.
48. Dickinson-Anson, H., and McGaugh, J.L. (1997). Bicuculline administered
into the amygdala after training blocks benzodiazepine-induced amnesia.
Brain Res. 752, 197–202.
49. Ratano, P., Everitt, B.J., and Milton, A.L. (2014). The CB1 receptor antagonist AM251 impairs reconsolidation of pavlovian fear memory in the rat
basolateral amygdala. Neuropsychopharmacology 39, 2529–2537.
50. Xiu, J., Zhang, Q., Zhou, T., Zhou, T.T., Chen, Y., and Hu, H. (2014). Visualizing an emotional valence map in the limbic forebrain by TAI-FISH. Nat.
Neurosci. 17, 1552–1559.
OPEN ACCESS
Article
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
chicken anti-Bgal
Aves Lab
chicken anti-eYFP
Invitrogen
rabbit anti-c-fos
Cell Signaling
Antibodies
mouse anti-CamKII
Merck
Cat # 05–532 RRID:AB309787
rat anti-c-fos
mouse anti-PV
Merck
Cat# P3088 RRID:AB_477329
guinea pig anti-PV
guinea pig anti-RFP
rabbit anti-GAD65/67
Abcam
Cat# ab49832 RRID: AB_880149
Alexa Fluor 488 anti-chicken
Invitrogen
Cy3 anti-rabbit
Invitrogen
Cy5 anti-mouse
Invitrogen
Cy3 anti-rat
Invitrogen
Alexa Fluor 488 anti-guinea pig
Jackson
Cy3 anti-guinea pig
Jackson
Cat# 106-165-003,
Alexa Fluor 488 anti-rabbit
Invitrogen
Merck
Cat #D9542
AAV9-c-fos-tTA
Steve Ramirez
AAV9-TRE-eYFP
Steve Ramirez
AAV5-hSyn-DIO-mCherry
Addgene
RRID: Addgene_ 50459
AAV5-hSyn-DIO-hM4D(Gi)-mCherry
Addgene
Krashes et al. (2011)
RRID: Addgene_ 44362
Bacterial and virus strains
AAVrg-hSyn- hM4D(Gi)-mCherry
Addgene
RRID: Addgene_ 50475
AAVrg-hSyn-mCherry
Addgene
RRID: Addgene_ 114472
AAV1-hDlx-DIO-KORD-mCyRFP
AAV1-hSyn-Cre
Addgene
RRID: Addgene_ 105553
Daun02
Hycultech
HY-13061-5
Doxycycline chow
Bio-Serv
F5003
bicuculline methiodide
Tocris
2503/10
Chemicals, peptides, and recombinant proteins
salvinorinB
Vinci Biochem
CAY-23582-25
Tocris
Deposited data
Raw data as supplementary file
Experimental models: Organisms/strains
WT Sprague-Dawley rats
Internal facility
c-fos-lacZ Sprague-Dawley rats
Internal facility
PV-Cre Long Evans rats
G. Rainer
LE-Tg (Pvalb-iCre)
https://imagej.net/Fiji
Prism 9
GraphPad
https://www.graphpad.com/scientific-software/prism/
Graphic State
Coulbourne
SPSS 22
Software and algorithms
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
Article
RESOURCE AVAILABILITY
Lead contact
Further information and request for resources and reagents should be directed to and will be fulfilled by the lead contact, Benedetto
Materials availability
All materials generated in this study are available from the lead contact upon request.
Data and code availability
d All data reported are shared as raw data and are included as a supplementary file.
d This paper does not report original code.
d Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon
request.
EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS
Animals
Healthy male rats (age, 65–70 days; weight, 240-350g) were housed in a plastic cage, 2–3 per cage, with food and water available ad
libitum, under a 12 h light/dark cycle (lights on at 7:00 a.m.) at a constant temperature of 22 ± 1 C. All the experiments were conducted during the light phase of the day (8 a.m.–4 p.m.). Depending on experimental demands, Sprague-Dawley rats, c-fos-lacZ transgenic rats that had been bred for 35–40 generations on a Sprague-Dawley background, or Long-Evans transgenic rats expressing
Cre recombinase under the rat parvalbumin (PV) promoter (LE-Tg (Pvalb-iCre), gifted by G. Rainer (University of Lausanne) were employed. Rats were derived from an internal animal facility breading. All the experiments were approved by the Italian Ministry of Health
(authorization no. 408/2020-PR) and by the local bioethical committee of the University of Turin.
Behavioral procedures
Auditory fear learning tasks
Rats were gently taken from their home cage and carried from the housing room to the soundproofed room. Once there, animals were
rods grid (1 cm in diameter, spaced 1.5 cm apart) connected to a shock delivery set-up. Rats were left undisturbed for 1 min. After
this time, conditioned stimuli consisting of a pure tone of 15-kHz of frequency (15 s of duration each, 80 dB, 36 s inter-trials interval,
ITI) were administered and acted as CS1 in most experiments. The last 1 s of each tone was paired with a painful unconditioned stimulus (US, 0.5 mA, 1 s). Immediately (0-min group), 6 h (6-h group) or 7 days (7-day group) later, rats were submitted to another round of
fear learning.21 They were presented with conditioned stimuli acting as CS2 and consisting of pure tones of 1-kHz of frequency (8s, 80
dB, 22s ITI). The last 1 s of each tone was paired with a painful unconditioned stimulus (US, 0.5 mA, 1 s).
In the 0-min group, CS1-US and CS2-US pairings occurred consecutively within the same conditioning cage (context A). Rats underwent 4 CS1-US pairings immediately followed by 3 CS2-US pairings, resulting in a total of 7 CS-US pairings. In cases where the
experimental requirements dictated (as in the experiments depicted in Figures 4A–4C), the 7 CS-US pairings were repeated twice
with a 6-h interval. In the 6-h and 7-day groups, rats experienced 7 CS1-US and 7 CS2-US. The two pairings occurred in a distinct
conditioning cage (context B) to further differentiate the two events. The two pairings were separated by either a 6-h interval in the 6-h
group or a 7-day interval in the 7-day group. The cage consisted of a skinner box module.17–21 Animals were carried in two different
buckets to the two conditioning chambers. At the end of the conditioning sessions, they were brought back to their home cage.
In the counterbalanced experiments, CS1 consisted of pure tones of 1-kHz of frequency, 8 s of duration each, 80 dB, 22s ITI, and
CS2 consisted of pure tones of 15-kHz, 15s, 80dB, 36s ITI.
Memory tests and extinction trials
Animals were transported singularly from the facility to the experimental rooms within small transparent buckets. Depending on the
experimental demand, 72 h or 3 weeks after auditory fear learning trials, rats were submitted to:
CS1 memory reactivation and CS2 memory test in the c-fos-lacZ transgenic rats (Figure 1, and 4L)
In c-fos-lacZ transgenic rats, CS1 memory was reactivated at recent or remote time points by presenting 4 CS1 (pure tone of
15-kHz, 15 s, 80 dB, 36 s ITI) identical to those employed during CS1-US pairing. To selectively trigger the reactivation of CS1 memory while minimizing interference associated with contextual cues,17–21 CS1 was presented in a novel environment (context C). Ninety
min later, rats received Daun02 or vehicle administration. CS2 memory was tested 3 days later in a totally new cage (context D), consisting of a transparent plastic cage with a black painted side. The cage was enclosed within a sound-attenuating box equipped with
an exhaust fan, which eliminated odorized air from the enclosure and provided background noise of 60 dB. Animals were allowed to
explore the cage for 5 min during the habituation session. The following day, after 2 min of free exploration, rats were presented with 7
Cell Reports 43, 114151, May 28, 2024
Article
OPEN ACCESS
CS2 (1 kHz, 8 s, 22 ITI) identical to those employed during CS2-US pairing. Three days after CS2 memory retention test, CS1 memory
black painted stripes on the walls.
CS1 memory extinction and CS2 memory retention test in Sprague-Dawley rats (Figures 2 and 4I–4K, and 6), c-fos-lacZ transgenic
rats (Figure 5), and PV-Cre rats (Figures 4F–4H)
To extinguish CS1 memory, rats were put in a new apparatus consisting of a transparent plastic cage with a black painted side and
enclosed within a sound-attenuating box equipped with an exhaust fan, which eliminated odorized air from the enclosure and provided background noise of 60 dB. Rats were left undisturbed for 2 min. Then, 35 CS1 identical to those employed during CS1-US
pairing (e.g., a pure tone of 15-kHz, 15 s, 80 dB) were delivered at 50 s ITI in the absence of any USs, for a total duration of
30 min. The subsequent day, CS2 memory retention was tested by presenting 7 CS2 (pure tones of 1-kHz, 8s, 80 dB, 22 s ITI) identical
to those employed during CS2-US pairing.
In the experiments where CS1 memory extinction was replaced by CS1 memory test, rats were presented only with 4 CS1 (15-kHz,
15 s, 80 dB, 36 s ITI).
CS1 and CS2 memory test in Sprague-Dawley rats injected with the virus cocktail of AAV9-c-fos-tTA and AAV9-TRE-eYFP viruses
(Figure 3)
To label neurons activated during CS1 and CS2 memory recall, we injected an inducible, activity-dependent virus cocktail of AAV9c-fos-tTA and AAV9-TRE-eYFP viruses, which labeled neurons expressing the immediate-early gene c-fos in a doxycycline (Dox)dependent manner.23,24 Rats were placed on a diet containing 40 mg/kg doxycycline (Dox) for 5 days before receiving viruses’ injections with access to food and water ad libitum. Rats were kept on this diet up to 48 h before CS1 memory test, which occurred
at recent or remote time points. Fourth-8 h before CS1 memory test, rats were habituated for 5 min to the two new cages where CS1
and CS2 memories will be tested. One cage consisted of a transparent plastic cage with a black painted side enclosed within a
black-painted stripes on the walls.
Six hours after the habituation sessions, the Dox-containing diet was replaced with standard rats’ chow (ad libitum) to open a time
window of activity-dependent labeling.24,44 Fourth-8 h later, CS1 memory was retrieved by presenting 4 CS1 (15-kHz, 15 s, 80 dB,
36 s ITI) identical to those employed during CS1-US pairing. Following behavioral tagging, rats were returned to their home cage and
were put again on the Dox diet. Twenty-four hours later, CS2 memory was retrieved by presenting 7 CS2 (pure tones of 1-kHz, 8s, 80
dB, 22 s ITI) identical to those employed during CS2-US pairing. Ninety min later, rats were intracardially perfused.
Freezing measure. In all experimental procedures, the assessment of the fear memory retention was determined as a freezing
response, analyzed as the complete absence of somatic mobility except for respiratory movements. For each animal, the amount
of time (in seconds) spent in freezing was measured offline by two independent observers who were blinded to the animal groups.
Surgical procedures
To administer active compounds or viruses in the target brain sites, rats were anesthetized with isoflurane: the induction was performed at 4% [v/v] in 2 L/min medical air and extended to a continuous exposure at 2% [v/v]) when rats were mounted in the stereotaxic apparatus.
Each rat was operated separately. Animals belonging to the active compounds or saline/vehicle groups were manipulated in an
interleaved way. An incision of the skull was made, and small burr holes were drilled to allow the penetration of a 28-gauge infusion
needle. A 10 mL Hamilton syringe mounted on an infusion pump was used to deliver substances. The substances were bilaterally
injected at a rate of 0.3 mL/min at the following stereotaxic coordinates taken from Paxinos and Watson atlas45:
Basolateral amygdala (BLA): AP: 2.4 L: ±5.5 DV: 8.3 and AP: 3.4 L: ±5.5 DV: 8.3.
Secondary Auditory Cortex (Te2): AP: 5,8 L: ±7,2 DV: 6.0 and AP: 6,8 L: ±7,2 DV: 6.0.
Dorsal Hippocampus: AP: 2.8 L: ±1.6 DV: 3.5 and AP: 4.2 L: ±2.6 DV: 3.5.
Ventral Hippocampus: AP: 5.0 L: ±5.0 DV: 6.0 and AP: 6.0 L: ±5.0 DV: 6.0.
The needle was left in place for 3 min in the case of the active compounds’ injection or 5 min in the case of the viruses’ injection. The
incision was then closed with stainless steel wound clips. The animal was given a subcutaneous (s.c.) injection of the analgesic/antiinflammatory ketoprofen (2 mg/kg body weight) and kept warm and under observation until recovery from anesthesia.
At the end of the experiments, needle track placement was verified in Nissl-stained sections. The sections were histologically verified under a microscope magnified at 2.53.
The following specific surgical procedures were employed depending on the experimental demand:
Daun02 injection into BLA (Figures 1 and 4L), Te2 (Figure S7), Dorsal Hippocampus (Figure 5)
For the treatment of c-fos-lacZ transgenic rats, Daun02 was dissolved to a final concentration of 5 mg/ml in a solution of 10%
DMSO, 6% Tween-80, and 84% phosphate-buffered saline.17,18 As in these studies,17,18 a volume of 1mL of Daun02 or vehicle
was injected per site in BLA, Te2, or Dorsal Hippocampus at the above stereotaxic coordinates.
AAV9-c-fos-tTA and AAV9-TRE-eYFP viruses’ injection into BLA (Figure 3)
To label neurons activated by memory processes, a cocktail (1:1 ratio) of AAV9-c-fos-tTA (3 x 10^13 GC/mL) and AAV9-TRE-eYFP
(1.2 x 10^13 GC/mL)23,24 was injected into BLA at the above stereotaxic coordinates at the volume of 0.8 mL per site. Viruses were
gifted by S. Ramirez (Boston University).
AAV5-hSyn-DIO-hM4D(Gi)-mCherry injection into BLA of PV-Cre rats and CNO i.p. injections (Figures 4A–4H)
Cell Reports 43, 114151, May 28, 2024
OPEN ACCESS
Article
To block parvalbumin (PV)-expressing cells within BLA, the AAV5-hSyn-DIO-hM4D(Gi)-mCherry (2.4 x 10^13 GC/mL) and control
virus (AAV5-hSyn-DIO-mCherry, 2.1 x 10^13 GC/mL)(Addgene) were injected into BLA at the above stereotaxic coordinates at
the volume of 0.8 mL per site. Fear learning trials occurred three weeks following viruses’ injections. Clozapine-N-Oxide (CNO)
(3 mg/kg, in saline solution) was injected intraperitoneally (i.p.) 60 min after CS2-US pairing. The dosage46 and time of injection47
were chosen based on previous studies in rats. In rats, levels of CNO peak at 30 min and decline to low levels by 6 h after injection.47
Bicuculline injection into BLA (Figures 4I–4L)
To block GABAergic neurotransmission, GABAA antagonist ( )-bicuculline methiodide was dissolved in saline at the dosage of
50 ng per 0.5 mL per side. The dosage was chosen based on previous studies48,49 showing no effects on fear behavior and fear memory by itself. According to those studies,48,49 a volume of 0.5 mL or saline was injected in BLA, at the stereotaxic coordinate of AP:
2.8 L: ±5.5 DV: 8.0.
AAVrg-hSyn-hM4D(Gi)-mCherry vector injection into BLA and CNO injection into the ventral hippocampus (Figures 6A–6F)
To inactivate hippocampal excitatory terminals in the BLA, rats were injected with the retrograde AAVrg-hSyn-hM4D(Gi)-mCherry
vector (2.4 x 10^13 GC/mL) into BLA at the volume of 0.8 mL per site at the above stereotaxic coordinates. A control group received the
AAVrg-hSyn-mCherry virus (2.0 x 10^13 GC/mL). All vectors were purchased from Addgene. Five weeks later, rats were submitted to
fear learning. Ninety min later, CNO was administered into the ventral hippocampus of PV-Cre rats at the volume of 0.8 mL per site at
the above stereotaxic coordinates.
AAV1-hDlx-DIO-KORD-mCyRFP and AAV1-hSyn-Cre injection and SalvinorinB injection (Figures 6G–6M)
To manipulate inhibitory interneurons of BLA that are directly contacted by hippocampal axons, we injected an inhibitory Credependent inhibitory DREADD (AAV1-hDlx-DIO-KORD-mCyRFP) vector (University of Zurich, UZH, VVF), in BLA (4.8 x 10^12 GC/
mL) at the volume of 0.8 mL per site at the above stereotaxic coordinates. In the meantime, a trans-synaptic Cre (AAV1-hSyn-Cre)
(2.1 x 10^13 GC) virus (Addgene), was injected in the ventral hippocampus at the volume of 1.0 mL per site at the above stereotaxic
coordinates. The control group received only the AAV1-hDlx-DIO-KORD-mCyRFP) vector into the BLA but not the AAV1-hSyn-Cre
virus into the ventral hippocampus. Five weeks later, rats were conditioned to the two auditory fear events. Ninety min later, they
received salvinorinB (SalB)37 into BLA to inhibit GABAergic cells specifically receiving hippocampal projections. SalB was dissolved
in saline (3 mM), and directly injected into BLA 90 min after CS2-US learning. SalB was injected into BLA at a volume of 0.5 mL per site
with the following stereotaxic coordinates: AP: 2.8 L: ±5.5 DV: 8.0.
Immunohistochemistry
Rats were deeply anesthetized and perfused intracardially with 4% PAF. Brains were dissected, stored overnight at 4 C, and finally
transferred to 30% sucrose. Coronal sections (35 mm) were cut on a cryostat and collected in PBS and incubated in a blocking solution for 1 h at RT.
To analyze the expression of the b-gal as an indicator of neuronal activity-induced activation and verify the effectiveness of the
Daun02 procedure, sections were incubated in primary chicken anti-Beta Gal (1:1000, Aves Lab) antibodies in the blocking solution
overnight at RT. Subsequently, sections were washed with PBS and incubated for 1 h at RT with secondary fluorescent Alexa Fluor
488 anti-chicken (1:1000, Invitrogen) antibody diluted in PBS.
To analyze cells activated by CS1 or CS2 or both memories, sections were incubated in primary chicken anti-eYFP (1:5000,
Invitrogen), rabbit anti-c-fos (1:2000, Cell Signaling), mouse antibody anti-CamKII (1:500 dilution, Merck) antibodies in the blocking
solution overnight at RT. Subsequently, sections were washed with PBS and incubated for 1 h at RT with secondary fluorescent Alexa
Fluor 488 anti-chicken (1:1000, Invitrogen), Cy3 anti-rabbit (1:1000, Invitrogen) and Cy5 anti-mouse (1:1000, Invitrogen) antibodies
diluted in PBS.
To analyze c-fos and Parvalbumin (PV) positive cells, sections were incubated with rat anti-c-fos (1:2000, SySy) and mouse anti-PV
(1:2000, Merck) antibodies and subsequently with secondary Cy3 anti-rat (1:1000, Invitrogen) and Cy5 anti-mouse (1:1000, Invitrogen) antibodies diluted in PBS. All sections were then stained with DAPI, and cover slipped.
To analyze the diffusion of AAV5-DIO-hM4D(Gi)-mCherry and AAVrg-hM4D(Gi)-mCherry vectors (and their AAV controls), mCherry
expression was visualized without any amplification. In PV-Cre rats expressing DIO-hM4D(Gi)-mCherry, sections were incubated
with guinea pig anti-PV (1:2000, SySy) antibody and Alexa Fluor 488 anti-guinea pig (1:1000, Jackson) secondary antibody.
In the KORD-dependent chemogenetic approach, AAV1-hSyn-Cre and AAV1-Dlx-DIO-KORD-mCyRFP1 recombination in BLA
neurons receiving direct projections from the ventral hippocampus was examined through mCyRFP1 amplification. Sections were
incubated in a blocking solution for 1 h at RT and then, in primary polyclonal guinea pig anti-RFP (1:500, SySy), and rabbit antiGAD65/67 (1:1000 dilution, Abcam), and mouse anti-PV (1:2000, Merck) in the blocking solution overnight at RT. Subsequently, sections were washed with PBS and incubated for 1 h at RT with secondary fluorescent Cy3 anti-guinea pig (1:1000, Jackson), Alexa
Fluor 488 anti-rabbit (1:1000, Invitrogen) and Cy5-anti mouse (1:1000, Invitrogen) antibodies diluted in PBS. All sections were stained
with DAPI, and cover slipped.
Microscopy and cell counts
Cellular markers, viral fluorophores of chemogenetic and control vectors were visualized by a Zeiss Airyscan confocal microscope.
Tissues were imaged using four lasers (405, 488, 568, and 640 nm), each corresponding this time to the peak emission spectrum for
DAPI (Nissl stain for cell nuclei); Alexa Fluor 488; mCherry, mCyRFP, Cy3; and Cy5, respectively.
In most cases (analysis of cells infected AAV9-c-fos-tTA and AAV9-TRE-eYFP, AAV5-hSyn-DIO- hM4D(Gi)-mCherry and its AAV
control, c-fos, and PV quantification, analysis of retrograde labeled hippocampal cells and KORD-expressing BLA neurons) images
Cell Reports 43, 114151, May 28, 2024
Article
OPEN ACCESS
were acquired at a 403 objective using a stack of 15 sections, spaced 1.5 mm apart (159 mm square; zoom fraction, 1.0). For AAVrghSyn-hM4D(Gi)-mCherry, micrographs of BLA were acquired as mosaic images, with each individual image acquired as a z stack of
10 sections, spaced 1.5 mm apart (159 mm square; zoom fraction, 1.0) by using a 403 objective.
Cell counts were performed in the BLA region at the anteroposterior (AP) coordinates ranging from 2.4 to 3.5 mm from the
bregma.45 Only putative neurons were included in the analysis, and glial cells, identified from their small size (5mm diameter)
and bright, uniform nuclear counterstaining, were excluded.
eYFP+, c-fos+ and eYFP+c-fos+ cells count
The number of excitatory neurons expressing eYFP (eYFP+), c-fos (c-fos+) or both (eYFP and c-fos overlap; eYFP+c-fos+) were
normalized on DAPI-positive cells. Neurons activated by the CS1 presentation minus those activated also during the CS2 presentation were calculated as total eYFP+cells minus eYFP+c-fos+cells and were referred to as neurons activated only by CS1 (eYFP+CS1).
Neurons activated by the only CS2 presentation were calculated as total c-fos+cells minus eYFP+c-fos+cells and were referred to as
c-fos+CS2 neurons.
The chance level for each neuron to be eYFP and c-fos double-positive was calculated as [(eYFP+c-fos?/DAPI) 3 (eYFP?c-fos+/
DAPI)],10,18,24,50 and the predicted overlap was: [(eYFP+c-fos?/DAPI) 3 (eYFP?c-fos+/DAPI)] 3 DAPI.
The reactivation ratio (RRI) was calculated as the ratio of eYFP and c-fos double-positive cells (eYFP+c-fos+) among all c-fos positive cells: eYFP+c-fos+/c-fos+cells. Thus, the RRI predicted by chance was equal to the predicted overlap divided by all c-fos+
cells18,50.
The observed overlap and the overlap expected by chance, as well as the predicted and the observed RRI of each group, were
compared using a Student’s two-tailed paired t test. To test the difference between groups, we used a Student’s two-tailed unpaired
t test.
c-fos+, PV+ and c-fos+PV+ cells count
The number of cells expressing c-fos, PV, and double-positive (c-fos and PV overlap) were quantified for each animal. c-fos positive
cells were normalized on the total number of DAPI cells, while double-positive cells (c-fos and PV overlap) were normalized on the
total number of PV cells. Data collected from each section were then averaged to produce the mean of each animal and the results
were statistically compared.
QUANTIFICATION AND STATISTICAL ANALYSIS
Animals from the same brood were a priori randomly assigned to each experimental group, in a weight-balanced manner. The sample
size estimation for each group was based on effect size calculation, on similar works in the field and on the previous laboratory experience. Inclusion/exclusion criteria were based on histological assessments. All experiments were conducted in at least two biological replicates.
The Kolmogorov-Smirnov test was used to determine whether the data were normally distributed. When data were normally
distributed, data from two groups were compared using two-tailed unpaired Student’s t-tests. Multiple-group comparisons were assessed using a one-way ANOVA with Tukey’s post-hoc test.
When data were non-normally distributed, they were analyzed using the Mann-Whitney test to test the differences between two
different groups. Multiple-group comparisons were assessed using Kruskal-Wallis test with Dunn post-hoc test.
To determine whether the data met the assumptions of the statistical approach, we rejected the null hypothesis at the p < 0.05 level.
All statistical analyses were performed using SPSS Statistics 22 (IBM). The statistical parameters (i.e., the exact value of n for each
experimental group of animals, SEM, the statistical test, and the exact p value) were reported in the legends. All data were presented
as mean ± SEM.
Cell Reports 43, 114151, May 28, 2024