Noradrenergic activation of the basolateral amygdala maintains hippocampus-dependent accuracy of remote memory
Contributed by James L. McGaugh, July 18, 2017 (sent for review June 15, 2017; reviewed by Ivan Izquierdo and Carmen Sandi)
Significance
Emotional arousal creates lasting and vivid memories. According to the systems consolidation theory, the hippocampus has a time-limited role in memory, and retrieval of remote memories mainly relies on neocortical networks. Here we show that this systems consolidation and associated change in memory specificity constitute a dynamically regulated process that can be modified by emotional arousal status. Norepinephrine administration into the basolateral amygdala after an episodic-like training experience maintained accuracy and hippocampus dependency of remote memory. This altered systems consolidation was paralleled by time-regulated epigenetically driven transcriptional changes of memory-related genes in the hippocampus and neocortex.
Abstract
Emotional enhancement of memory by noradrenergic mechanisms is well-described, but the long-term consequences of such enhancement are poorly understood. Over time, memory traces are thought to undergo a neural reorganization, that is, a systems consolidation, during which they are, at least partly, transferred from the hippocampus to neocortical networks. This transfer is accompanied by a decrease in episodic detailedness. Here we investigated whether norepinephrine (NE) administration into the basolateral amygdala after training on an inhibitory avoidance discrimination task, comprising two distinct training contexts, alters systems consolidation dynamics to maintain episodic-like accuracy and hippocampus dependency of remote memory. At a 2-d retention test, both saline- and NE-treated rats accurately discriminated the training context in which they had received footshock. Hippocampal inactivation with muscimol before retention testing disrupted discrimination of the shock context in both treatment groups. At 28 d, saline-treated rats showed hippocampus-independent retrieval and lack of discrimination. In contrast, NE-treated rats continued to display accurate memory of the shock–context association. Hippocampal inactivation at this remote retention test blocked episodic-like accuracy and induced a general memory impairment. These findings suggest that the NE treatment altered systems consolidation dynamics by maintaining hippocampal involvement in the memory. This shift in systems consolidation was paralleled by time-regulated DNA methylation and transcriptional changes of memory-related genes, namely Reln and Pkmζ, in the hippocampus and neocortex. The findings provide evidence suggesting that consolidation of emotional memories by noradrenergic mechanisms alters systems consolidation dynamics and, as a consequence, influences the maintenance of long-term episodic-like accuracy of memory.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Emotionally arousing experiences are well-retained in memory (1, 2). Beyond their increased strength, emotionally enhanced memories are also often characterized by increased vividness and the subjective feeling of remembering (3, 4). Both animal and human research indicate that noradrenergic activation of the basolateral amygdala (BLA) enhances memory of emotionally arousing experiences by regulating neural plasticity and information storage processes in other brain regions (5–12). A majority of studies investigating the effects of noradrenergic activation of the BLA in memory have focused on episodic (declarative) or contextual tasks that involve functioning of the hippocampus (13–17). The BLA sends extensive projections to the hippocampus (18) and has a major impact on hippocampal functioning (19–21). There is extensive evidence that the memory enhancement induced by BLA activation by norepinephrine (NE) administration or emotional arousal during or shortly after learning involves hippocampal activation (20, 22–24). This activation may increase the strength of episodic-like or context-specific memories (22, 25), enabling a more distinct separation of memory traces for individual items or similar datasets (26, 27).
Thus, it is by now well-established that noradrenergic activation of the BLA enhances the initial formation of hippocampus-dependent memories. However, an understanding of how noradrenergic activation after encoding affects the later specificity or accuracy of the enhanced episodic-like memory remains elusive. Over time, memory traces are thought to undergo a neural reorganization, referred to as systems consolidation, during which they are, at least partly, transferred from the hippocampus to neocortical networks (28–32). This temporal shift in memory representation is associated with a transformation in the quality of memory (30, 33). Retrieval of remote memories mainly relies on semantic or gist-like representations in extrahippocampal structures, for example the anterior cingulate cortex (ACC), that are likely to maintain core information but few contextual details (29, 30, 33). Such dynamic shift of neural representation in systems consolidation and accompanied change in the nature of the memory are supported by time- and brain region-specific epigenetic modifications (34, 35). The underlying molecular mechanisms such as DNA methylation, for example, expression silencing by covalent modifications of DNA, provide pathways for self-perpetuating transcriptional alternations driven by environmental signals. Previous findings have indicated that such epigenetic changes of memory-related genes play a crucial role in both the formation and maintenance of memory (35–37). For example, lower methylation levels of reelin (Reln) and increased methylation levels of protein phosphatase 1 (Pp1) in the hippocampus led to successful memory retention 1 d later (36), whereas cortical expression of protein kinase M zeta (PKMζ), the N-truncated form of protein kinase C zeta, is required for maintenance of long-term memory (37). However, whether the dynamics of systems consolidation and time-dependent change in the specificity of memory are actively regulated by environmental conditions, such as emotional arousal status, has not, as yet, been investigated.
The present experiments investigated, in rats, whether noradrenergic activation of the BLA after training on an episodic-like discrimination task affects systems consolidation dynamics and associated epigenetic and transcriptional changes to maintain episodic-like accuracy and hippocampus dependency of remote memory. Such NE-induced changes in systems consolidation may represent a critical mechanism underlying the long-term vividness of emotional memories (3, 4).
Results
Noradrenergic Activation of the BLA Posttraining Maintains Accuracy of Remote Memory.
We first investigated the effect of posttraining NE infusions into the BLA on episodic-like accuracy of recent and remote memory. Male Sprague–Dawley rats were trained on a modified version of the inhibitory avoidance task, termed the inhibitory avoidance discrimination task (38), in which they explored two contextually distinct inhibitory avoidance apparatuses within a 2-min interval but footshock was delivered only in the latter context (Fig. 1A). NE (1.0 μg in 0.2 μL) or saline was microinfused bilaterally into the BLA immediately after the training session. Either 2 d (recent) or 28 d (remote) later, retention latencies were tested in the two training contexts (i.e., “shock box” and “nonshock” box) as well as in a “novel” box to assess whether rats accurately discriminated the training context in which they had received footshock.
Fig. 1.
Fig. S1.
We first determined the effect of posttraining NE infusions on memory accuracy at a 2-d retention test. Two-way ANOVA for retention latencies in the three test environments indicated significant NE (F1,38 = 5.25; P = 0.03), context (F2,38 = 45.02; P < 0.0001), and interaction effects (F2,38 = 6.66; P = 0.003). As shown in Fig. 1B, rats treated with either saline or NE in the BLA after the training session displayed significantly longer retention latencies in the shock box than in the nonshock box (Ps < 0.01) or novel box (Ps < 0.001), indicating accurate discrimination of the shock context. Moreover, NE-treated rats, relative to saline controls, had longer retention latencies in the shock box (P < 0.05), indicating increased memory strength. Delayed infusions of NE given 3 h posttraining were ineffective (Fig. 1C), supporting the view that the NE effect is dependent on a consolidation process (8).
Next, we examined, in separate groups of rats, the effect of posttraining NE infusions on memory accuracy at a 28-d retention test. Two-way ANOVA for retention latencies revealed no NE effect (F1,46 = 0.16; P = 0.69) but a significant context effect (F2,46 = 18.27; P < 0.0001) as well as a significant interaction between NE and context (F2,46 = 8.09; P = 0.001). As shown in Fig. 1D, saline-treated rats no longer accurately discriminated the shock context, as indicated by similar retention latencies in the shock box and nonshock box (P = 0.89). Unlike saline controls, NE-treated rats continued to display accurate memory of the shock–context association, as shown by significantly longer retention latencies in the shock box than in the nonshock box (P < 0.001) or novel box (P < 0.001). Moreover, NE-treated rats, relative to saline-treated controls, had longer retention latencies in the shock box (P < 0.01), indicating that the NE-induced memory enhancement was also still present 4 wk later.
Noradrenergic Activation of the BLA Posttraining Maintains Hippocampus Dependency of Remote Memory.
We next investigated whether the accurate episodic-like remote memory displayed by the NE-treated rats continues to rely on the hippocampus. For this, NE or saline was administered into the BLA immediately after training, as above, and hippocampal involvement in the expression of the memory was determined by inactivation of the dorsal hippocampus with the GABAergic receptor agonist muscimol (0.5 µg in 0.5 µL) 20 min before retention testing (39).
At 2 d, three-way ANOVA for retention latencies revealed significant NE (F1,96 = 6.37; P = 0.002) and muscimol effects (F1,96 = 12.33; P < 0.0001) but no significant interaction between these factors (F1,96 = 2.37; P = 0.19). As shown in Fig. 2A, hippocampal inactivation of both saline- and NE-treated rats disrupted the expression of accurate memory of the shock–context association, illustrated by similar retention latencies in the shock box and nonshock box (Ps ≥ 0.94). The muscimol infusion neither impaired retention latencies of saline controls (P = 0.81) nor blocked the NE-induced memory enhancement (P = 0.96), indicating that the hippocampal inactivation did not disrupt the general expression of memory strength. Moreover, the muscimol infusion did not affect discrimination of the two training contexts from the novel box (P < 0.001). When these rats were tested again 24 h after the muscimol infusion (Fig. S2A), the expression of accurate memory of the shock–context association had returned, confirming that the muscimol-induced inactivation of the hippocampus was transient and had worn off 24 h later.
Fig. 2.
Fig. S2.
At 28 d, three-way ANOVA for retention latencies showed significant NE (F1,84 = 14.07; P < 0.0001), muscimol (F1,84 = 11.87; P < 0.0001), and interaction effects (F1,84 = 14.29, P < 0.0001). As in the first experiment, rats administered saline into the BLA after training no longer accurately discriminated the shock context at this remote retention test (P = 0.83; shock vs. nonshock box) (Fig. 2B). Hippocampal inactivation of saline-treated rats did not significantly alter retention performance, supporting the systems consolidation hypothesis that the expression of remote generalized memory is associated with a loss of hippocampal engagement. By contrast, hippocampal inactivation of NE-treated rats blocked the expression of accurate memory of the shock–context association (P = 0.63; shock vs. nonshock box). Moreover, and unlike at the 2-d retention test, muscimol inactivation of NE-treated rats also impaired general expression of memory. As shown in Fig. 2B, retention latencies in the shock box of NE-treated rats administered muscimol before retention testing were significantly shorter than those of NE-treated rats administered saline into the hippocampus (P < 0.05), and also than those of saline-treated rats administered muscimol before the retention test (P < 0.05). Further, retention latencies in the shock box did not differ significantly from those in the novel box (P = 0.13), indicating that these rats, without an intact hippocampus, did not show any evidence of retention. When these NE-treated rats were tested again 24 h after the muscimol inactivation (Fig. S2B), the original accurate discrimination of the shock context as well as memory strength had returned.
Noradrenergic Activation of the BLA Posttraining Induces Time- and Tissue-Specific Transcriptomic Changes.
Our findings indicate that posttraining BLA activation promotes prolonged accuracy of episodic-like memory through a hippocampus-dependent mechanism. This change in consolidation mechanism could be supported by sustained changes at the molecular level, that is, modified gene expression. Therefore, we measured transcriptional changes in the hippocampus (dentate gyrus area) and ACC, brain structures previously associated with early and late phases of memory consolidation, respectively (29, 30). For this, we trained and tested another cohort of rats on the same task (Fig. 3E). To get a broad picture of time- and tissue-specific changes, we first used gene set enrichment analysis (GSEA) and focused on a memory-associated gene set (MGS; 122 genes; see SI Methods). These genes were previously reported as being differentially expressed in the rodent brain after training on an episodic-like memory task (adapted from ref. 40; Table S1). Posttraining NE administration into the BLA was associated with a time-dependent differential expression of the MGS in the hippocampus (Pnominal = 0.007; Padjusted < 0.02, global test; Table S2). Next, we analyzed the MGS genes individually and examined NE-dependent changes in gene expression at the remote time point in the two brain regions. Upon applying a Bonferroni correction, Reln was significantly up-regulated in the hippocampus and Pkmζ was down-regulated in the ACC (PBonferroni < 0.05).
Fig. 3.
Table S1.
| External gene name | Ensembl gene ID | Chromosome | Strand | Start position | End position | P value |
|---|---|---|---|---|---|---|
| Abat | ENSRNOG00000002636 | 10 | −1 | 5895469 | 6002068 | 1.5E-03 |
| Abcb6 | ENSRNOG00000018697 | 9 | −1 | 82143209 | 82151515 | 6.4E-03 |
| Ache | ENSRNOG00000050841 | 12 | −1 | 24487852 | 24491390 | NA |
| Adora1 | ENSRNOG00000003442 | 13 | −1 | 56097928 | 56131500 | 8.1E-02 |
| Adra1a | ENSRNOG00000009522 | 15 | −1 | 48203098 | 48295045 | 1.4E-02 |
| Adra1d | ENSRNOG00000021256 | 3 | −1 | 130621199 | 130637207 | 2.1E-02 |
| Atp1a1 | ENSRNOG00000030019 | 2 | −1 | 223440514 | 223469880 | NA |
| Atp2a2 | ENSRNOG00000001285 | 12 | 1 | 41435157 | 41482924 | 3.0E-02 |
| Atp2b2 | ENSRNOG00000030269 | 4 | −1 | 209001178 | 209104640 | 3.1E-03 |
| Bcl2l11 | ENSRNOG00000016551 | 3 | −1 | 126571721 | 126601829 | 1.8E-02 |
| Bdnf | ENSRNOG00000047466 | 3 | 1 | 107390648 | 107421863 | 3.3E-01 |
| C3 | ENSRNOG00000046834 | 9 | 1 | 8728465 | 8754412 | 3.4E-02 |
| C8b | ENSRNOG00000007639 | 5 | 1 | 128159874 | 128196920 | 5.6E-02 |
| CAMK2A | ENSRNOG00000018712 | 18 | 1 | 55428528 | 55529045 | NA |
| Camk4 | ENSRNOG00000020478 | 18 | 1 | 25464155 | 25680622 | 2.4E-02 |
| Casp1 | ENSRNOG00000007372 | 8 | 1 | 2627973 | 2636854 | 1.3E-01 |
| Casp6 | ENSRNOG00000009508 | 2 | 1 | 249539675 | 249551881 | 7.2E-02 |
| Ccnl1 | ENSRNOG00000011586 | 2 | −1 | 177110681 | 177122888 | 7.7E-03 |
| Ceacam1 | ENSRNOG00000020578 | 1 | −1 | 83589574 | 83605846 | 5.5E-03 |
| Chrna3 | ENSRNOG00000013829 | 8 | −1 | 58175886 | 58189161 | 4.7E-02 |
| Chrnb1 | ENSRNOG00000014698 | 10 | −1 | 56135720 | 56148237 | 2.8E-02 |
| Chrnb4 | ENSRNOG00000014427 | 8 | −1 | 58192309 | 58211019 | 2.8E-02 |
| Cit | ENSRNOG00000001143 | 12 | −1 | 48135859 | 48294184 | 5.6E-03 |
| Cntn1 | ENSRNOG00000004438 | 7 | 1 | 133085052 | 133270392 | 2.1E-02 |
| Cntn2 | ENSRNOG00000009033 | 13 | −1 | 54360710 | 54389586 | 8.6E-02 |
| Crh | ENSRNOG00000012703 | 2 | −1 | 124182919 | 124184783 | 2.0E-01 |
| Ctnnd2 | ENSRNOG00000010649 | 2 | 1 | 103069802 | 103662662 | 7.5E-03 |
| Cttnbp2 | ENSRNOG00000008440 | 4 | −1 | 42516939 | 42680592 | 3.6E-02 |
| Drd1 | ENSRNOG00000023688 | 17 | 1 | 13212535 | 13214770 | NA |
| Drd4 | ENSRNOG00000017927 | 1 | 1 | 221195844 | 221199031 | 9.6E-02 |
| Dusp5 | ENSRNOG00000014061 | 1 | 1 | 281658075 | 281671441 | 1.2E-03 |
| Egr1 | ENSRNOG00000019422 | 18 | 1 | 27369166 | 27372951 | 1.5E-01 |
| Fgf18 | ENSRNOG00000048389 | 10 | −1 | 17932880 | 17948600 | NA |
| Gabbr1 | ENSRNOG00000000774 | 20 | −1 | 3996333 | 4025427 | 8.7E-02 |
| Gabbr2 | ENSRNOG00000008431 | 5 | −1 | 66802348 | 67142203 | 6.4E-03 |
| Gabra4 | ENSRNOG00000002336 | 14 | 1 | 38966016 | 39040596 | 4.1E-01 |
| Gabra5 | ENSRNOG00000010803 | 1 | −1 | 113844283 | 113955085 | 3.2E-02 |
| Gabrb2 | ENSRNOG00000003680 | 10 | 1 | 27818095 | 28026899 | 1.9E-02 |
| Gabrg2 | ENSRNOG00000003241 | 10 | −1 | 26937103 | 27023929 | 5.1E-02 |
| Gabrp | ENSRNOG00000032417 | 10 | −1 | 18369151 | 18391029 | 2.2E-02 |
| Gfra2 | ENSRNOG00000014010 | 15 | 1 | 56280942 | 56372307 | 7.6E-03 |
| Gja1 | ENSRNOG00000000805 | 20 | 1 | 39612108 | 39624547 | 9.9E-02 |
| Grik1 | ENSRNOG00000001575 | 11 | −1 | 31426908 | 31828130 | 1.8E-02 |
| Grik2 | ENSRNOG00000000368 | 20 | −1 | 55396403 | 56122333 | 3.8E-02 |
| Grin1 | ENSRNOG00000011726 | 3 | −1 | 2489166 | 2515963 | 2.3E-02 |
| Grin2a | ENSRNOG00000033942 | 10 | 1 | 4527220 | 4936773 | 8.1E-02 |
| Hmox1 | ENSRNOG00000014117 | 19 | 1 | 25622556 | 25629372 | 3.8E-02 |
| Homer1 | ENSRNOG00000047014 | 2 | 1 | 42107429 | 42208021 | NA |
| Htr3a | ENSRNOG00000006595 | 8 | −1 | 51803897 | 51816207 | 8.2E-02 |
| Htr6 | ENSRNOG00000049761 | 5 | −1 | 161242626 | 161257707 | 9.6E-02 |
| Htt | ENSRNOG00000011073 | 14 | −1 | 81796508 | 81942093 | 5.7E-03 |
| Icam1 | ENSRNOG00000020679 | 8 | 1 | 22092135 | 22103927 | 1.1E-02 |
| Ifng | ENSRNOG00000007468 | 7 | 1 | 61333602 | 61337640 | 1.0E-01 |
| Igfbp2 | ENSRNOG00000016957 | 9 | 1 | 79888614 | 79915188 | 2.0E-01 |
| Il15 | ENSRNOG00000003439 | 19 | −1 | 34532034 | 34598725 | 1.3E-01 |
| Il1b | ENSRNOG00000004649 | 3 | 1 | 128133909 | 128140349 | 4.6E-02 |
| Il2ra | ENSRNOG00000047647 | 17 | −1 | 72202947 | 72250588 | 3.0E-01 |
| Ins1 | ENSRNOG00000012052 | 1 | 1 | 280213615 | 280214178 | 3.3E-01 |
| Insr | ENSRNOG00000029986 | 12 | −1 | 3852577 | 3989125 | 1.8E-02 |
| Itpr1 | ENSRNOG00000007104 | 4 | 1 | 204714826 | 205047900 | 2.2E-02 |
| Jak1 | ENSRNOG00000011157 | 5 | −1 | 123865518 | 123966233 | 3.0E-02 |
| Junb | ENSRNOG00000042838 | 19 | −1 | 37069075 | 37070109 | 3.0E-01 |
| Kcna5 | ENSRNOG00000019719 | 4 | −1 | 226075051 | 226076859 | 8.6E-01 |
| Kcna6 | ENSRNOG00000047736 | 4 | 1 | 232324385 | 232357644 | 2.4E-02 |
| Kcnb1 | ENSRNOG00000046949 | 3 | −1 | 170010756 | 170094516 | 1.8E-01 |
| Kcnd2 | ENSRNOG00000029610 | 4 | 1 | 48100008 | 48607643 | 1.9E-02 |
| Kcnh5 | ENSRNOG00000009542 | 6 | −1 | 107294666 | 107578241 | 4.2E-03 |
| Kcnj11 | ENSRNOG00000021128 | 1 | −1 | 103188484 | 103189656 | 2.5E-01 |
| Kcnj16 | ENSRNOG00000004713 | 10 | 1 | 99084253 | 99086021 | 7.8E-02 |
| Kcnj4 | ENSRNOG00000013869 | 7 | −1 | 120710788 | 120738030 | 1.9E-02 |
| Kcnn2 | ENSRNOG00000016675 | 18 | 1 | 38985930 | 39129190 | 1.9E-01 |
| Lsamp | ENSRNOG00000031852 | 11 | −1 | 62956510 | 63592532 | 1.0E-01 |
| Mag | ENSRNOG00000021023 | 1 | −1 | 90500839 | 90516262 | 1.3E-01 |
| Map1b | ENSRNOG00000017428 | 2 | −1 | 48837124 | 48930483 | 1.6E-03 |
| Map2 | ENSRNOG00000011841 | 9 | −1 | 73841103 | 73922677 | 4.9E-02 |
| Mapk14 | ENSRNOG00000000513 | 20 | 1 | 7968539 | 8028708 | 3.8E-03 |
| Mapk9 | ENSRNOG00000002823 | 10 | 1 | 35114480 | 35145312 | 2.7E-02 |
| Mprip | ENSRNOG00000003226 | 10 | 1 | 45775202 | 45890063 | 5.4E-03 |
| Ncan | ENSRNOG00000048036 | 16 | 1 | 20900031 | 20925162 | 3.0E-02 |
| Ncs1 | ENSRNOG00000008761 | 3 | −1 | 15914623 | 15960622 | 2.8E-02 |
| Nes | ENSRNOG00000018681 | 2 | 1 | 206747073 | 206755585 | 1.9E-02 |
| Nos2 | ENSRNOG00000049980 | 10 | −1 | 65423626 | 65455464 | NA |
| Nr1h2 | ENSRNOG00000019812 | 1 | −1 | 101619967 | 101625287 | 1.2E-01 |
| Nr3c1 | ENSRNOG00000014096 | 18 | −1 | 31408742 | 31430004 | 8.3E-02 |
| Oprl1 | ENSRNOG00000016768 | 3 | 1 | 180934268 | 180940194 | 9.6E-04 |
| Pak1 | ENSRNOG00000029784 | 1 | 1 | 168974734 | 169089649 | 2.2E-02 |
| Pik3r1 | ENSRNOG00000018903 | 2 | −1 | 50895264 | 50965217 | 4.0E-03 |
| Pld1 | ENSRNOG00000028156 | 2 | 1 | 133400609 | 133554180 | 1.0E-01 |
| Plec | ENSRNOG00000023781 | 7 | −1 | 117215919 | 117275610 | 1.2E-02 |
| Pkmζ | ENSRNOG00000015480 | 5 | −1 | 176118465 | 176228206 | 1.7E-04 |
| Ppp1cc | ENSRNOG00000001269 | 12 | −1 | 41744539 | 41761986 | 1.1E-01 |
| Ppp3ca | ENSRNOG00000009882 | 2 | 1 | 260462929 | 260735288 | 5.1E-02 |
| Rab1b | ENSRNOG00000050510 | 1 | −1 | 227413058 | 227413663 | NA |
| Rb1 | ENSRNOG00000016029 | 15 | −1 | 58804732 | 58942062 | 9.8E-02 |
| Reln | ENSRNOG00000021441 | 4 | 1 | 9349481 | 9775752 | 9.9E-05 |
| Scn9a | ENSRNOG00000006639 | 3 | −1 | 59207262 | 59286961 | 3.6E-03 |
| Slc1a1 | ENSRNOG00000014816 | 1 | 1 | 254195390 | 254275293 | 4.4E-02 |
| Slc1a3 | ENSRNOG00000016163 | 2 | −1 | 79348445 | 79422645 | 5.4E-02 |
| SLC24A2 | ENSRNOG00000008169 | 5 | −1 | 109317349 | 109564102 | NA |
| Slc32a1 | ENSRNOG00000015393 | 3 | −1 | 160500060 | 160504727 | 5.6E-02 |
| Slc6a4 | ENSRNOG00000003476 | 10 | −1 | 62851326 | 62872481 | 6.4E-02 |
| Snap25 | ENSRNOG00000006037 | 3 | 1 | 136083976 | 136269421 | 5.5E-02 |
| Sncg | ENSRNOG00000010437 | 16 | −1 | 9049951 | 9054488 | 6.1E-02 |
| Sod3 | ENSRNOG00000003869 | 14 | −1 | 61071313 | 61077045 | 1.8E-01 |
| Stat4 | ENSRNOG00000050942 | 9 | −1 | 54051157 | 54167491 | 4.2E-02 |
| Stx12 | ENSRNOG00000011804 | 5 | −1 | 154712006 | 154740248 | 1.3E-02 |
| Stx2 | ENSRNOG00000000936 | 12 | 1 | 33261860 | 33285512 | 6.2E-02 |
| Stx5 | ENSRNOG00000018847 | 1 | 1 | 231876955 | 231893829 | 4.6E-02 |
| Stx8 | ENSRNOG00000003849 | 10 | 1 | 54258230 | 54498841 | 3.7E-02 |
| Svop | ENSRNOG00000000693 | 12 | 1 | 50074360 | 50132727 | 5.4E-02 |
| Synj1 | ENSRNOG00000002051 | 11 | −1 | 34717258 | 34791435 | 2.0E-03 |
| Syt2 | ENSRNOG00000004756 | 13 | 1 | 56621946 | 56629657 | NA |
| Syt4 | ENSRNOG00000017333 | 18 | −1 | 23893019 | 23902428 | 1.3E-02 |
| Syt7 | ENSRNOG00000026432 | 1 | 1 | 233382666 | 233440845 | 8.1E-03 |
| Syt8 | ENSRNOG00000020245 | 1 | 1 | 222500771 | 222503063 | 9.3E-02 |
| Syt9 | ENSRNOG00000019613 | 1 | 1 | 178591098 | 178768239 | 6.1E-02 |
| Tacr1 | ENSRNOG00000005853 | 4 | 1 | 177782405 | 178095041 | 2.6E-02 |
| Tgfbr3 | ENSRNOG00000002093 | 14 | 1 | 3509601 | 3683814 | 6.9E-02 |
| Thra | ENSRNOG00000009066 | 10 | 1 | 86465253 | 86480926 | 8.4E-03 |
| Tph1 | ENSRNOG00000011672 | 1 | −1 | 103746878 | 103774100 | 3.8E-02 |
| Tubb5 | ENSRNOG00000000821 | 20 | 1 | 5519551 | 5523476 | 2.8E-02 |
| Xcl1 | ENSRNOG00000002964 | 13 | −1 | 88079753 | 88083196 | 3.0E-01 |
For the gene set enrichment analysis, we created a memory-associated gene set composed of genes previously implicated in memory processes (adapted from ref. 40; expanded from refs. 35, 36, 51, and 62). The final gene set consisted of 122 memory genes. Columns show MGS gene symbols, Ensembl identifiers, chromosome, strand, start and end positions, and global-test nominal P values. NA, not available.
Table S2.
| Log FC | CI.L | CI.R | AveExpr | t | P | AdjP | B | setName |
|---|---|---|---|---|---|---|---|---|
| 2.266 | 0.662 | 3.869 | 19.414 | 2.842 | 0.007 | 0.02 | −2.202 | MemoryGeneSet |
Posttraining NE administration into the BLA induced differential expression of the memory-associated gene set specifically in the hippocampus and to remote memory (global test: 28 vs. 2 d; controlled for saline effects and changes in the ACC). Columns show differential expression log2 fold change (Log FC), upper- and lower-bound confidence intervals (CI.L and CI.R, respectively), average expression (AveExpr; standardized to the median array expression), t value (t), P value (P), adjusted P value (AdjP; SI Methods), B statistic (B), and gene set name (setName).
Previous studies related Reln to early stages of plasticity in the hippocampus (36) and Pkmζ to long-term memory maintenance in cortex (37). Therefore, our second-level analysis focused on the expression of these two genes in a pharmacological and time- and tissue-specific context. We first examined Reln and Pkmζ expression patterns separately for NE- and saline-treated animals (Fig. S3), followed by examining the specific NE-induced changes in the expression dynamics of these genes, after accounting for saline effects (Fig. 3). In the hippocampus of animals that received NE, Reln expression was stably up-regulated at both 2 and 28 d (Fig. S3A), as opposed to an only initial up-regulation at the 2-d time point in saline-treated animals (Fig. S3C). After subtracting saline effects, NE treatment was associated with a down-regulation in Reln expression in the hippocampus at 2 d (PBonferroni < 0.05) and, interestingly, an up-regulation at 28 d (PBonferroni < 0.05), thus indicating a time-regulated expression shift (P = 0.0004; Fig. 3A and Fig. S4A). In contrast, NE treatment-associated Pkmζ expression in the hippocampus was not significantly changed over time (P > 0.05; Fig. 3A and Fig. S4A; after subtracting saline effects), mostly due to similar expression dynamics in both the NE and saline groups (Fig. S3 B and D). In the ACC, Reln expression in both NE- and saline-treated animals was down-regulated at 28 d compared with 2 d (Fig. S3 A and C), while Pkmζ was time-dependently up-regulated in animals that received saline only (Fig. S3 B and D). After subtracting saline effects, NE treatment was associated with an up-regulation of both Pkmζ and Reln expression in the ACC at 2 d (PBonferroni < 0.05) but only transitionally for Pkmζ (28 vs. 2 d, P = 0.006), while Reln expression remained up-regulated (28 vs. 2 d, P > 0.05; Fig. 3B and Fig. S4B). These time- and tissue-specific expression dynamics are in line with our behavioral observations of an altered systems consolidation in NE-treated rats, marked by a sustained hippocampus dependency.
Fig. S3.
Fig. S4.
Noradrenergic Activation of the BLA Posttraining Induces Time- and Tissue-Specific DNA Methylation Changes in Reln and Pkmζ Gene Promoters.
DNA methylation is critical for memory formation (36) as well as for memory maintenance (34, 35). Therefore, we next examined whether the changes in gene expression induced by the NE activation could be driven by epigenetic mechanisms, namely DNA methylation. Indeed, mRNA measurements of DNA methyltransferases (Dnmts), a family of enzymes that catalyzes DNA methylation (36), in the hippocampus 30 min posttraining revealed up-regulated expression of de novo Dnmt3a in NE-treated rats (P < 0.01, after subtracting saline effects; P < 0.05 for Creb transcriptional factor as a control; Fig. 3F). Thus, posttraining NE activation of the BLA induced changes in de novo Dnmt3a expression that were obvious shortly after the training session, during the initial consolidation phase.
Additionally, to examine if DNA methylation could be associated with the time- and tissue-specific expression changes of memory-associated genes, we measured DNA methylation of the Reln and Pkmζ promoters at the recent and remote time points [averaged across CpGs in relevant promoter regions: chr4:9,346,306 to 9,346,547 for Reln (35) and chr5:172,751,942 to 172,752,163 for Pkmζ (41); Fig. 3 G and H]. Further, we examined temporal changes in Dnmt3a expression. In the hippocampus, Reln promoter DNA methylation showed opposite time-dependent dynamics in NE- and saline-treated rats (Fig. S5 A and C). After subtracting saline effects, NE treatment was associated with a hypermethylation of the Reln promoter in the hippocampus at 2 d but a hypomethylation at 28 d (PBonferroni < 0.05; Fig. 3C). These findings indicate that noradrenergic activation of the BLA induced a temporal shift in Reln promoter DNA methylation in the hippocampus (P = 0.0002). Furthermore, Dnmt3a expression in the hippocampus was accordingly down-regulated at 28 d (PBonferroni < 0.05, after subtracting saline effects; Fig. S6A), suggesting that the sustained changes in DNA methylation and gene expression could be actively maintained. In the ACC, DNA methylation at the Reln promoter increased over time both in NE- and saline-treated rats (Fig. S5 A and C), while it decreased at the Pkmζ promoter in animals that received saline only (Fig. S5D). DNA methylation of the Pkmζ promoter in the ACC of NE-treated rats did not show a change across time (Fig. S5B). After accounting for saline effects, NE treatment was associated with a hypomethylation of both the Reln and Pkmζ promoters in the ACC at 2 d (PBonferroni < 0.05; Fig. 3D); this effect was only transitional for Pkmζ (28 vs. 2 d, P = 0.0004), while Reln remained hypomethylated at 28 d (PBonferroni < 0.05). NE treatment did not significantly affect Dnmt3a expression levels in the ACC at either the recent or remote time points (Fig. S6B). These observed DNA methylation changes were concordant and negatively correlated with the NE treatment-associated temporal expression dynamics of Reln and Pkmζ, both in the hippocampus and ACC.
Fig. S5.
Fig. S6.
SI Methods
Subjects.
Male Sprague–Dawley rats (280 to 320 g at time of surgery) from Charles River breeding laboratories were used. They were kept individually in a temperature-controlled (22 °C) vivarium room and maintained on a 12-h:12-h light:dark regimen (0700 to 1900 hours lights on) with ad libitum access to food and water. Training and testing were performed between 1000 and 1500 hours. All experimental procedures were in compliance with European Union Directive 2010/63/EU and approved by the Institutional Animal Care and Use Committee of Radboud University, Nijmegen, The Netherlands.
Cannula Implantation.
Rats were anesthetized with an s.c. injection of ketamine (37.5 mg/kg; Alfasan) and dexmedetomidine (0.25 mg/kg; Orion) and received the nonsteroidal analgesic carprofen (4 mg/kg; Pfizer). They were positioned in a stereotaxic frame (Kopf Instruments), and two stainless steel guide cannulae (23 gauge; Component Supply/SKU Solutions) were implanted bilaterally with the cannula tips 2.0 mm above the basolateral amygdala (BLA) [15 mm long; coordinates: anteroposterior (AP): 2.8 mm posterior to Bregma; mediolateral (ML): ±5.0 mm lateral to midline; dorsoventral (DV): 6.5 mm below skull surface] (59). Some rats received additional bilateral guide cannulae implanted 1.5 mm above the dorsal hippocampus (11 mm long; coordinates: AP: −3.6 mm; ML: ±1.9 mm; DV: −2.6 mm). After surgery, the rats were administered atipamezole hydrochloride (0.25 mg/kg s.c.; Orion) to reverse anesthesia and 3 mL of sterile saline to facilitate clearance of drugs and prevent dehydration. The rats were allowed to recover for a minimum of 7 d before commencement of training, and were handled three times for 1 min each during this recovery period.
Drug and Infusion Procedure.
Norepinephrine (NE; 1.0 μg in 0.2 μL; Sigma-Aldrich) was dissolved in saline and infused into the BLA either immediately (0 h) or 3 h after the training. Drug dose and volume were based on previous findings from our laboratory (9, 42). The GABAergic receptor agonist muscimol (3-hydroxy-5-aminomethyl-isoxazole; 0.5 µg in 0.5 µL; Sigma-Aldrich) was dissolved in saline and administered into the dorsal hippocampus 20 min before retention testing (39).
Bilateral infusions of drug or an equivalent volume of saline into the BLA were made by using 30-gauge injection needles connected to 10-μL Hamilton microsyringes by polyethylene (PE-20) tubing. The injection needles protruded 2.0 mm beyond the cannula tips, and a 0.2-μL injection volume per hemisphere was infused over a period of 30 s by an automated syringe pump (Stoelting). The injection needles were retained within the cannulae for 20 s following drug infusion to maximize diffusion.
For hippocampal infusions, the infusion needles protruded 1.5 mm beyond the guide cannulae, and a 0.5-μL injection volume per hemisphere was infused over a period of 75 s. All drug solutions were freshly prepared before each experiment.
Inhibitory Avoidance Discrimination Task.
Rats were subsequently trained in two contextually distinct inhibitory avoidance apparatuses within a single training session, but footshock was delivered only in the latter context. On the retention test, they were tested in both training contexts as well as in a novel context (Fig. 1A). Each apparatus had the same geometry and consisted of a trough-shaped alley (91 cm long, 15 cm deep, 20 cm wide at the top, and 6.4 cm wide at the bottom) divided into two compartments, separated by a sliding door that opened by retracting into the floor. The starting compartment (31 cm) was made of opaque white plastic and was well-lit; the shock compartment (60 cm) was made of two dark, electrifiable metal plates and was not illuminated. The training context in which footshock was given (shock box) did not have any contextual modifications. The safe training context (nonshock box) had four vertical white stripes (2 cm wide) taped in the dark compartment together with tape placed on the floor, closing the gap between the two plates. The novel box (used on the retention test only) had two white circles (3.5-cm diameter) taped on each wall of the dark compartment, and the gap between the plates was closed with tape (Fig. S1). All three inhibitory avoidance apparatuses were located next to one another in a sound- and light-attenuated room. For a detailed description of the task, see ref. 38.
For training, rats were initially placed in the starting compartment of the nonshock box and could freely explore the apparatus for 20 s without footshock being delivered. Afterward, the rats were removed from the apparatus and, after a delay of 2 min, placed in the starting compartment of the second inhibitory avoidance apparatus (shock box). After the rat stepped completely into the dark compartment, the sliding door was closed and a single inescapable footshock (0.60 mA; 1 s) was delivered. The rats were removed from the apparatus 20 s after termination of footshock and, after drug treatment, returned to their home cages. On the retention test, either 2 or 28 d after training, they were tested, in a randomized order and without delay, in the two training contexts (i.e., shock box and nonshock box) and in a novel box they had not visited before. Previously, we have shown that the order of retention testing in the different test environments does not affect retention latencies (38). For all three boxes, the rats were placed in the starting compartment and their latency to enter the dark compartment with all four paws (maximum latency of 600 s) was recorded. Longer latencies in the shock box compared with the nonshock box were interpreted as indicating accurate memory of the shock–context association. Different groups of trained rats were used for the 2- and 28-d retention tests.
Cannula Placement Verification.
Rats were deeply anesthetized with sodium pentobarbital and perfused transcardially with 0.9% saline, followed by 4% formaldehyde. The brains were removed and immersed in 4% formaldehyde. At least 24 h before sectioning, brains were placed in a 25% sucrose solution for cryoprotection. Coronal sections of 50 µm were cut on a cryostat, mounted on gelatin-coated slides, stained with cresyl violet, and examined by light microscopy by an observer blind to drug treatment. Rats with needle tips located outside the boundaries of the BLA and/or hippocampus, or with extensive tissue damage at the target areas, were excluded from final analyses.
For verification of cannula placement of flash-frozen brains, 50-µm-thick coronal sections were cut, collected on gelatin-coated slides, and let to dry. Sections were then fixed in 100% acetone for 30 min and stained with cresyl violet and processed as mentioned above.
Tissue Collection.
Rats were administered an overdose of sodium pentobarbital 30 min after training for Dnmt mRNA measurements or immediately after retention testing for mRNA expression and DNA methylation measurements. Within 90 s after the pentobarbital injection, rats were decapitated, and the brains were rapidly removed and flash-frozen by submersion for 2 min in a beaker filled with precooled isopentane on dry ice. Flash-frozen brains were stored at −80 °C.
Brain tissue containing the hippocampus or ACC was cut into 350-µm-thick coronal slices and further dissected using a 0.75-mm brain puncher (Stoelting). Bilateral punches from the dentate gyrus area of the hippocampus (AP: −2.64 to −3.86 mm) and ACC (AP: +3.00 to +1.92 mm) were collected from three consecutives slices for a total of six punches. Punches were preserved at −20 °C for at least 16 h and later in RNAlater-ICE (Ambion Life Technologies).
Tissue Homogenization and Nucleic Acid Extraction.
Parallel DNA and RNA isolation was performed using a chaotropic lysis protocol. After removing RNAlater-ICE, tissue samples were flash-frozen in liquid nitrogen and disrupted using TissueLyser II (Qiagen) for 1 min at 28 Hz. Then, 420 µL of guanidinium thiocyanate disruption lysis buffer (4.5 M guanidinium thiocyanate, 2% N-lauroylsarcosine, 50 mM EDTA, pH 8, 25 mM Tris⋅HCl, pH 7.5, 0.1 M beta-mercaptoethanol, 0.2% antifoam A) was added. Disrupted tissue was solubilized by vortexing for 1 min at maximum speed. The lysate was then split into two equal parts for RNA and DNA isolations.
RNA isolation was performed by adding 600 µL TRI Reagent (Ambion Life Technologies) to 200 µL tissue lysate and pulse vortexing. Then, 830 µL of absolute ethanol was added (final ethanol concentration ∼51%) and the mixture was loaded on a Zymo-Spin IIC column. RNA was purified using a Direct-zol Kit (Zymo Research) following the recommended procedure, including DNase I treatment. The concentration and quality of the RNA were determined using a NanoDrop 2000 (Thermo Scientific) and RNA 6000 Nano Kit on a Bioanalyzer 2100 instrument (Agilent).
DNA isolation was first performed with the QIAamp DNA Micro Kit (Qiagen). Two hundred microliters of Buffer AL (Qiagen) and 200 µL absolute ethanol were added to 200 µL lysate and mixed by pulse vortexing. DNA was purified on a QIAamp MinElute column using the recommended protocol. An additional purification was performed using a Genomic DNA Clean & Concentrator Kit (Zymo Research) using the recommended protocol. The concentration and quality of the DNA were determined using fluorometry (Qubit dsDNA BR Assay Kit; Invitrogen) and gel electrophoresis.
Transcriptomic Analysis: Target Synthesis and GeneChip Hybridization.
Target synthesis was performed starting from 25 ng total RNA by using NuGEN kits: Ovation Pico WTA System V2 (3302-12), followed by an Encore Biotin Module (4200-12), using standard protocols. Synthesis reactions were carried out on a PCR machine (TProfessional TRIO; Biometra). Eighty-five microliters of mixture (25 ng/µL DNA) was loaded on an Affymetrix Rat Gene 2.0 ST Array (902124) and hybridized for 17 h (45 °C, 60 rpm) in a Hybridization Oven 640 (Affymetrix).
The arrays were washed and stained on a Fluidics Station 450 (Affymetrix) by using a Hybridization, Wash, and Stain Kit (Affymetrix) under the FS450_0007 protocol. The GeneChips were processed with an Affymetrix GeneChip Scanner 3000 7G. DAT images and CEL files of the microarrays were generated using an Affymetrix GeneChip Command Console.
Transcriptomic Analysis: GeneChip Data Preprocessing.
Preprocessing of the Gene ST arrays was done with the RMA workflow using the oligo R package (57). In short, background correction was performed using the convolution of signal and noise distributions. Log2-transformed values were obtained and further normalized using quantile normalization (57). Above-background expression filtering was applied using the Mas5Calls algorithm: Probe sets called as absent (cutoff P ≥ 0.05) were excluded from all arrays (Mas5Calls function from the affy R Package: https:/bioconductor.org/packages/release/bioc/html/affy.html). For annotation purposes, we used the ragene20stprobeset.db 8.2.0 library (Bioconductor release 3.3, package ragene20stprobeset.db 8.2.0). Probe set summaries were further annotated by NetAffx biological annotation (Affymetrix).
Differential expression analysis was performed using linear models, with the limma R package (60), considering treatment (two levels: NE and saline), tissue (two levels: hippocampus and ACC), and time point (two levels: 2 and 28 d). From these models, contrasts were built to assess temporal effects of posttraining NE treatment in different tissues. P values were corrected for multiple testing (for all probe sets and 13 contrasts) using the Benjamini and Hochberg (61) false discovery rate (FDR < 0.05). Additionally, a fold-change criteria ≥|1.8| was applied (median |FC| observed for all significant contrasts). Transcripts’ FCs were calculated by summarizing all of the probe sets that fulfilled the above criteria.
Gene Set Enrichment Analysis.
A memory gene set composed of 122 genes previously implicated in memory processes was created (35, 36, 40, 51, 62). For gene ranking based on their differential expression across the contrasts, we computed three different metrics from all associated probe sets: the mean difference in means, the mean of quotients in mean expression levels, and a combined P value. Then, each region was assigned a metrics-specific rank, and a combined rank was computed as the maximum value among the three metrics ranks. Next, GSEA was performed (63). Briefly, the ranking of the genes was based on combined rank test statistics. Then, a weighted Kolmogorov–Smirnov statistic was used as a basis for determining whether genes from a prespecified memory gene set were overrepresented among the top or bottom of the list (Bioconductor release 3.3, package GO.db). Next, shuffling the phenotypes generated the background distribution of the gene set statistics and the gene set analysis was repeated. Each randomly generated gene set for which its maximum deviation was higher than the original data was counted and, after n iterations, the P value was computed (10,000 permutations). The procedure was repeated until all gene sets were tested. For false discovery correction, we applied the Benjamini–Hochberg procedure (61).
Quantitative Real-Time PCR.
For measuring posttraining Dnmt expression changes, total RNA from hippocampus punches was isolated using the recommended TRIzol protocol (Life Technologies), including DNase treatment [1 μg RNA was treated with 2 U DNase (Sigma-Aldrich), 1 h at 37 °C]. The quality of RNA samples was verified by agarose gel electrophoresis and UV spectrophotometric analysis. cDNA was synthesized from 0.5 to 1 µg RNA by using the RevertAid First Strand cDNA Synthesis Kit standard protocol (Fermentas). qPCR was performed with the GoTaq qPCR Kit (Promega), including 2 µL of the reverse transcription reaction (1:20 dilution), 1:50 dilution of CXR (carboxy-X-rhodamine reference dye), and 200 nM each primer: for Dnmt3a, forward (FW)-5′-ATCGACGCCAAAGAAGTGTC-3′ and RV-5′-GCTATTCTGCCGTGTTCCAG-3′; for Dnmt3b, FW-5′-AAAGTCGAAGACGCACAACC-3′ and reverse (RV)-5′-CTTACCGCAGGACAGACAGC-3′; for Dnmt1, FW-5′-GGAGCAGATCGAGAAGGATG-3′ and RV-5′-CTTGCACTTCCCACACTCAG-3′; and for Creb, FW-5′-TCAGCCGGGTACTACCATTC-3′ and RV-5′-TTCAGCAGGCTGTGTAGGAA-3′. Samples were normalized to actin-β transcript using the following primers: FW-5′-CCAACTGGGACGATATGGAG-3′ and RV-5′-AACACAGCCTGGATGGCTAC-3′. Expression was quantified using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) and the following cycling conditions: 95 °C, 10 min; 40× (95 °C, 15 s; 60 °C, 1 min); 72 °C, 5 min.
Validation of expression changes of microarray significant genes from GSEA was performed starting from SPIA cDNA (Single Primer Isothermal Amplification; NuGEN). qPCR was performed using the SYBR FAST PCR Kit (KAPA Biosystems), in a 12-µL final reaction volume, using 2 µL cDNA template, on a RotorGene 6000A instrument (Corbett Research). Cycling conditions were as follows: 95 °C, 60 s; 32× (95 °C, 10 s; 10 °C, 15 s; 72 °C, 10 s); 72 °C, 5 min, followed by a melting curve analysis (61 to 95 °C, rising by 0.7 °C/3 s). Threshold cycles were determined using RotorGene software version 6.1 (Corbett Research). Ywhaz and Gapdh were selected as reference genes for normalization, with the following primer sets: Pkmζ, FW-5′-CTTAAAGGGACGGAAGATG-3′ and RV-5′-TAGATGGACTCGGCTTTC-3′; Reln, FW-5′-CTCCAGTTCAAGCTAAAC-3′ and RV-5′-CAGCATCATGTGAATACT-3′; Gapdh, FW-5′-TCACCACCATGGAGAAGGC-3′ and RV-5′-GCTAAGCAGTTGGTGGTGCA-3′; and Ywhaz, FW-5′-TTGAGCAGAAGACGGAAGGT-3′ and RV-5′-GAAGCATTGGGGATCAAGAA-3′ (Microsynth).
Expression levels were normalized using a geometric mean level of expression, and fold differences were calculated using the delta-delta-Ct method (64), qBasePlus software (Biogazelle). Differential expression analysis was performed using linear models, limma R package (60), using the same contrasts as described above for the microarray data analysis.
To account for multiple testing, we applied Bonferroni correction.
Pyrosequencing Analysis.
DNA methylation was quantified by direct bisulfite pyrosequencing (58). Five hundred nanograms of high-purity, intact DNA was used for bisulfite conversion using an EZ DNA Methylation-Gold Kit (Zymo Research) by following standard protocols. Bisulfite-converted (BSC) DNA quality and concentration were determined using an RNA 6000 Pico Kit on a Bioanalyzer 2100 instrument (Agilent Technologies) and NanoDrop 2000 (Thermo Scientific). BSC samples were normalized to 10 ng/µL.
The following primer sets were used: for Pkmζ, FW-5′-GATGTGATATTTTAAAGGTTGTTGAGTA-3′ and RV-5′[biotin]-CACATCTCTACCTCCTCATATC-3′; and for Reln, FW-5′-ATTTTGGTTTGGTGTTGAGTTTG-3′ and RV-5′[biotin]-ATATCATACATAACCACTATCCCTAAAAT-3′ (Microsynth). Promoter fragments were amplified using an AmpliTaq Gold Kit from Applied Biosystems (Life Technologies). PCR was done in 30-µL reactions containing the following: 1× PCR buffer II, 300 µM deoxynucleotide triphosphates, final 3.5 mM MgCl2, 200 µM each primer, and 20 ng BSC DNA. We used the following cycling conditions: 95 °C, 15 min; 50× (95 °C, 30 s; 55 °C, 30 s; 72 °C, 30 s); 72 °C, 10 min. PCR products were purified and sequenced using a PyroMark ID System (Biotage) following the manufacturer’s suggested protocol and sequencing primers: for Pkmζ, 5′-GGGTTTTGGTTAGTTTTTATT-3′; and for Reln, 5′-ACATACAAAAAAATAACTAACAAC-3′ (Microsynth).
Reln and Pkmζ gene promoter regions analyzed by bisulfite pyrosequencing are depicted in Fig. 3 G and H, respectively. The Reln gene region analyzed (enlarged box) is located in the Reln promoter in front of exon 1. The Pkmζ gene region analyzed (enlarged box) is located in the alternative promoter in front of exon 1′ that carries the unique 5′-PKMζ mRNA sequence. Numbering is relative to the transcription start site +1 located in exon 1′. The examined CpG sites are marked in bold. The sequencing primer is marked in italics. Relative size and chromosomal coordinates are also depicted (Upper and Lower, respectively; RGSC 6.0/rn6).
Controlling for PCR temperature bias was done with a series of calibrator samples of known methylation levels. Briefly, unmethylated standards were prepared by using two rounds of linear whole-genome amplification with an Ovation WGA System Kit (NuGEN), starting from 10 ng of DNA, as recommended by the manufacturer. Methylated standards were made using the CpG methyltransferase assay with M.SssI (New England Biolabs), starting from 2 µg of purified DNA, following the standard protocol. Bisulfite conversion of standard samples was done as described above.
All samples were analyzed in quadruplicate. Differential methylation analysis was performed using linear models, limma R package (60), using the same contrasts as described above for the microarray data analysis. DNA methylation levels were averaged across CpGs in the examined promoter region.
To account for multiple testing, we applied Bonferroni correction.
Discussion
Our findings indicate that NE administration into the BLA after inhibitory avoidance discrimination training maintained long-term hippocampus-dependent accuracy of memory. These effects were associated with time-regulated DNA methylation and transcriptional changes of memory-related genes, namely Reln and Pkmζ, in the hippocampus and neocortex. Such NE-induced changes in systems consolidation dynamics may represent a critical mechanism underlying the long-term vividness of emotional memories (3, 4).
Decades of research have provided compelling evidence that emotional arousal or brief periods of stress after encoding facilitate subsequent memory performance (2, 12). Evidence indicates that the BLA upon emotional arousal or noradrenergic activation modulates information transfer and neural plasticity mechanisms in different memory circuits to enhance the consolidation of different types of training experiences (10, 12, 42, 43). In line with the animal data, findings of human studies confirmed the key role of noradrenergic arousal in emotional memory enhancement (13, 44). Findings of human neuroimaging studies further indicated that arousal-induced activation of the amygdala enhances memory-related hippocampal activity (16, 17, 20, 22, 45). The present findings provide evidence that such posttraining noradrenergic activation also affects the later strength and accuracy of memory: NE administration into the BLA immediately after inhibitory avoidance discrimination learning maintained the strength as well as the long-term episodic-like specificity of memory. Moreover, the findings with rats given posttraining NE infusions into the BLA as well as hippocampal inactivation with muscimol during retention testing revealed this episodic-like specificity to be hippocampus-dependent at both recent and remote time points. These findings strongly suggest that noradrenergic activation of the BLA may alter the time-dependent transfer of the memory trace from the hippocampus to neocortical areas, thereby also preventing the transformation of the detailed, very specific memories into more semantic, gist-like memories.
According to the standard model of systems consolidation, memories are initially dependent on the hippocampus and are subsequently stored, in their original form, in other brain areas (28). Findings of several studies suggest that systems consolidation may be a more dynamic process that involves a transformation of the nature of the memory, including a decrease in memory specificity (30). To the extent that episodic or context-specific memories are retained, they might continue to require the hippocampus. Our findings indicate that saline control rats initially showed accurate episodic-like memory of the association of footshock with the training context, but at the 28-d retention test were unable to discriminate the shock context. Consistent with the systems consolidation theory, the results of hippocampal inactivation during retention testing indicated that discrimination at the recent time point depends on the hippocampus but that the expression of generalized memory at the remote time point no longer requires the hippocampus. Our findings further suggest that the hippocampus does not appear to play any crucial role in remembering the aversive experience or training contexts per se. Hippocampal inactivation also preserved the initial NE-induced memory enhancement, despite a lack of episodic-like accuracy. As noradrenergic activation is known to activate large-scale neural networks (46), it is likely that the NE administration into the BLA might have enhanced the initial consolidation of different aspects of the acquired information simultaneously in multiple brain regions (8). We were particularly interested in determining whether the accurate long-term memory of NE-treated rats continues to depend on the hippocampus. The evidence indicates that it does; hippocampal inactivation of these rats at the 28-d retention test blocked the expression of episodic-like accuracy of memory. Rather unexpectedly, however, hippocampal inactivation of NE-treated rats at this remote retention test, different from at the recent retention test, also resulted in a general impairment in memory. These findings thus indicate that BLA activation by posttraining NE infusions induces long-lasting accurate and strong memories by engaging the hippocampus for longer periods of time. Noradrenergic stimulation shortly after learning may not only decelerate but even reverse the transformation of memory and maintain hippocampus dependency, enabling long-term accurate memory.
This altered systems consolidation was paralleled by dynamic changes in DNA methylation and expression of memory-associated genes. It is very likely that the complex changes in systems consolidation depend on several molecular mechanisms. For instance, genes such as Map2 and Grin1 were also differentially regulated by the NE treatment over time in a tissue-specific manner (Table S1), supporting a role for the BLA in affecting hippocampal neurogenesis in emotional enhancement of episodic memory (47). Nevertheless, we identified here Reln and Pkmζ as the main markers. The temporal expression of these genes in the two anatomical regions, previously reported as regions important for specific and time-dependent support of memory (29, 35), was also altered, mirroring the shift in memory consolidation. The first marker, Reln, is involved in processes of neuronal migration and positioning in the developing brain (48), enhances synaptic plasticity by regulating long-term potentiation (49), and supports the formation of new synapses (50). Moreover, it has been reported that Reln expression in the hippocampus is actively, but transiently, up-regulated to facilitate the storage of new memories but returns to basal levels 24 h after fear conditioning (36). In the current study, saline-treated control rats exhibited an initial up-regulation of Reln expression in the hippocampus at 2 d, followed by a down-regulation to basal levels at 28 d, in line with the systems consolidation theory. In contrast, NE treatment of the BLA after training actively reversed this down-regulation of Reln expression in the hippocampus at the 28-d time point, thus mirroring the observed behavioral changes and supporting an enhanced involvement of the hippocampus in long-term memory maintenance. Furthermore, the noradrenergic activation also abolished the time-dependent expression changes of Reln in the ACC found in control animals, possibly supporting a diminished role of the cortex in long-term memory maintenance after NE treatment.
The second main marker, Pkmζ, has been linked with the maintenance of long-term memory in the cortex (37, 51). PKMζ is an independent catalytic domain of the full-length protein kinase C ζ and is constitutively and persistently active, without the need of a second messenger (41). Activation of this kinase enhances trafficking of AMPARs (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors) in the postsynaptic membrane and affects plasticity through modulation of late-phase long-term potentiation (52, 53). Shema et al. (54) reported that the inhibition of PKMζ in the cerebral cortex abolishes long-term memory. This is in agreement with the systems consolidation theory, and points to a cortex-associated mechanism of remote memory maintenance. However, other findings suggest that PKMζ is rather dispensable and that other kinases may play a role as well (55, 56). In the current study, saline-treated control rats exhibited a time-dependent up-regulation of Pkmζ in the ACC, supporting an increasing role of cortical PKMζ in long-term memory maintenance. In contrast, NE treatment abolished this time-dependent cortical up-regulation of Pkmζ triggered by the learning experience, further supporting an altered systems consolidation and diminished role of the cortex in the long-term maintenance of the memory trace.
Thus, these findings provide evidence that posttraining noradrenergic activation of the BLA enhances strength and long-term episodic-like specificity of memory induced by sustained hippocampus dependency. Strikingly, the remote memory remained hippocampus-dependent even 4 wk after training, suggesting a shift in systems consolidation dynamics. This shift was likely supported by epigenetically driven transcriptional changes of memory-related genes in both the hippocampus and neocortex.
Methods
Subjects.
Male Sprague–Dawley rats (Charles River) were kept individually in a temperature-controlled (22 °C) vivarium room (0700 to 1900 hours lights on). Training and testing were performed during the light cycle between 1000 and 1500 hours. All experimental procedures were in compliance with European Union Directive 2010/63/EU and approved by the Institutional Animal Care and Use Committee of Radboud University, Nijmegen, The Netherlands.
Inhibitory Avoidance Discrimination Task.
Rats were initially placed in the starting compartment of the nonshock box and could explore the apparatus for 20 s without footshock being delivered (Fig. 1A). After a delay of 2 min, they were placed in the starting compartment of a contextually different inhibitory avoidance apparatus (shock box). When the rat had stepped into the dark compartment, the sliding door was closed and a single inescapable footshock (0.60 mA; 1 s) was delivered. On the retention test, either 2 or 28 d after training, they were tested, in a randomized order, in both training contexts as well as in a novel context (38) (SI Methods).
Tissue Collection, Homogenization, and Nucleic Acid Extraction.
Brain tissue containing the hippocampus or ACC was cut into 350-µm-thick coronal slices. Bilateral punches from the dentate gyrus area of the hippocampus [anteroposterior (AP): −2.64 to −3.86 mm] and ACC (AP: +3.00 to +1.92 mm) were collected from three consecutives slices to a total of six punches. For transcriptomic and methylomic analysis, we pooled individual animals’ tissue punches. Parallel DNA and RNA isolation was performed using a chaotropic lysis protocol (SI Methods).
Transcriptomic Analysis.
Genome-wide expression changes were measured with the Rat Gene 2.0 ST Array (Affymetrix; 902124) by using standard protocols (57) (SI Methods).
Gene Set Enrichment Analysis.
First, we created a memory-associated gene set composed of 122 genes previously implicated in memory (SI Methods). GSEA was performed by using a standard weighted Kolmogorov–Smirnov statistic. Finally, we tested individual genes from the MGS across significant comparisons and applied Bonferroni correction to account for multiple testing (SI Methods).
Quantitative PCR.
Posttraining DNA methyltransferase expression measurements as well as validation of expression changes from transcriptomic analysis were performed by quantitative PCR. All reactions were performed in triplicate (SI Methods).
Pyrosequencing Analysis.
DNA methylation of the previously reported relevant promoter regions of rat Reln [chr4:9,346,306 to 9,346,547; RGSC 6.0/rn6 (35); Fig. 3G] and Pkmζ [chr5:172,751,942 to 172,752,163; RGSC (Rat Genome Sequencing Consortium/Rattus norvegicus) 6.0/rn6 (41); Fig. 3H] genes was quantified by direct bisulfite pyrosequencing (58). All reactions were performed in quadruplicate (SI Methods).
Statistics.
Inhibitory avoidance retention latencies were analyzed with two- or three-way ANOVAs with retention latencies in the different test contexts as repeated measures. Post hoc comparisons used unpaired and paired t tests. For transcriptome and DNA methylation analysis, we analyzed three biological replicates formed by pooling tissue punches from 6 to 10 independent animals for each tissue/treatment combination. Bar plots show log fold change (FC) differences in expression and methylation between compared groups. Dnmt mRNA levels were analyzed with unpaired t tests. For all comparisons, P < 0.05 was accepted as statistically significant. If not noted otherwise, Bonferroni correction was applied to account for multiple testing.
Acknowledgments
This research was supported by a Radboud University Topfund (to B.R.) and funds of the University of Basel (to A.P. and D.J.-F.d.Q.).
Supporting Information
Supporting Information (PDF)
- Download
- 491.55 KB
References
1
JL McGaugh Memory and Emotion: The Making of Lasting Memories (Columbia Univ Press, New York, 2003).
2
JL McGaugh, Making lasting memories: Remembering the significant. Proc Natl Acad Sci USA 110, 10402–10407 (2013).
3
JM Talarico, DC Rubin, Confidence, not consistency, characterizes flashbulb memories. Psychol Sci 14, 455–461 (2003).
4
T Sharot, MR Delgado, EA Phelps, How emotion enhances the feeling of remembering. Nat Neurosci 7, 1376–1380 (2004).
5
I Izquierdo, et al., Neurotransmitter receptors involved in post-training memory processing by the amygdala, medial septum, and hippocampus of the rat. Behav Neural Biol 58, 16–26 (1992).
6
JL McGaugh, Make mild moments memorable: Add a little arousal. Trends Cogn Sci 10, 345–347 (2006).
7
R Hurlemann, et al., Human amygdala reactivity is diminished by the β-noradrenergic antagonist propranolol. Psychol Med 40, 1839–1848 (2010).
8
B Roozendaal, JL McGaugh, Memory modulation. Behav Neurosci 125, 797–824 (2011).
9
A Barsegyan, JL McGaugh, B Roozendaal, Noradrenergic activation of the basolateral amygdala modulates the consolidation of object-in-context recognition memory. Front Behav Neurosci 8, 160 (2014).
10
I Izquierdo, CRG Furini, JC Myskiw, Fear memory. Physiol Rev 96, 695–750 (2016).
11
RT LaLumiere, JL McGaugh, CK McIntyre, Emotional modulation of learning and memory: Pharmacological implications. Pharmacol Rev 69, 236–255 (2017).
12
B Roozendaal, EJ Hermans, Norepinephrine effects on the encoding and consolidation of emotional memory: Improving synergy between animal and human studies. Curr Opin Behav Sci 14, 115–122 (2017).
13
L Cahill, B Prins, M Weber, JL McGaugh, Beta-adrenergic activation and memory for emotional events. Nature 371, 702–704 (1994).
14
B Roozendaal, BT Nguyen, AE Power, JL McGaugh, Basolateral amygdala noradrenergic influence enables enhancement of memory consolidation induced by hippocampal glucocorticoid receptor activation. Proc Natl Acad Sci USA 96, 11642–11647 (1999).
15
SM Southwick, et al., Relationship of enhanced norepinephrine activity during memory consolidation to enhanced long-term memory in humans. Am J Psychiatry 159, 1420–1422 (2002).
16
EA Kensinger, S Corkin, Memory enhancement for emotional words: Are emotional words more vividly remembered than neutral words? Mem Cognit 31, 1169–1180 (2003).
17
BA Strange, RJ Dolan, Beta-adrenergic modulation of emotional memory-evoked human amygdala and hippocampal responses. Proc Natl Acad Sci USA 101, 11454–11458 (2004).
18
M Pikkarainen, S Rönkkö, V Savander, R Insausti, A Pitkänen, Projections from the lateral, basal, and accessory basal nuclei of the amygdala to the hippocampal formation in rat. J Comp Neurol 403, 229–260 (1999).
19
EA Phelps, Human emotion and memory: Interactions of the amygdala and hippocampal complex. Curr Opin Neurobiol 14, 198–202 (2004).
20
M Fastenrath, et al., Dynamic modulation of amygdala-hippocampal connectivity by emotional arousal. J Neurosci 34, 13935–13947 (2014).
21
D Rei, et al., Basolateral amygdala bidirectionally modulates stress-induced hippocampal learning and memory deficits through a p25/Cdk5-dependent pathway. Proc Natl Acad Sci USA 112, 7291–7296 (2015).
22
MP Richardson, BA Strange, RJ Dolan, Encoding of emotional memories depends on amygdala and hippocampus and their interactions. Nat Neurosci 7, 278–285 (2004).
23
CK McIntyre, et al., Memory-influencing intra-basolateral amygdala drug infusions modulate expression of Arc protein in the hippocampus. Proc Natl Acad Sci USA 102, 10718–10723 (2005).
24
ES Lovitz, LT Thompson, Memory-enhancing intra-basolateral amygdala clenbuterol infusion reduces post-burst afterhyperpolarizations in hippocampal CA1 pyramidal neurons following inhibitory avoidance learning. Neurobiol Learn Mem 119, 34–41 (2015).
25
SK Segal, SM Stark, D Kattan, CE Stark, MA Yassa, Norepinephrine-mediated emotional arousal facilitates subsequent pattern separation. Neurobiol Learn Mem 97, 465–469 (2012).
26
T Smeets, et al., Stress selectively and lastingly promotes learning of context-related high arousing information. Psychoneuroendocrinology 34, 1152–1161 (2009).
27
US Wiemers, MM Sauvage, D Schoofs, TC Hamacher-Dang, OT Wolf, What we remember from a stressful episode. Psychoneuroendocrinology 38, 2268–2277 (2013).
28
LR Squire, P Alvarez, Retrograde amnesia and memory consolidation: A neurobiological perspective. Curr Opin Neurobiol 5, 169–177 (1995).
29
PW Frankland, B Bontempi, LE Talton, L Kaczmarek, AJ Silva, The involvement of the anterior cingulate cortex in remote contextual fear memory. Science 304, 881–883 (2004).
30
G Winocur, M Moscovitch, Memory transformation and systems consolidation. J Int Neuropsychol Soc 17, 766–780 (2011).
31
M Moscovitch, R Cabeza, G Winocur, L Nadel, Episodic memory and beyond: The hippocampus and neocortex in transformation. Annu Rev Psychol 67, 105–134 (2016).
32
T Kitamura, et al., Engrams and circuits crucial for systems consolidation of a memory. Science 356, 73–78 (2017).
33
BJ Wiltgen, AJ Silva, Memory for context becomes less specific with time. Learn Mem 14, 313–317 (2007).
34
JJ Day, JD Sweatt, DNA methylation and memory formation. Nat Neurosci 13, 1319–1323 (2010).
35
CA Miller, et al., Cortical DNA methylation maintains remote memory. Nat Neurosci 13, 664–666 (2010).
36
CA Miller, JD Sweatt, Covalent modification of DNA regulates memory formation. Neuron 53, 857–869 (2007).
37
R Shema, et al., Enhancement of consolidated long-term memory by overexpression of protein kinase Mzeta in the neocortex. Science 331, 1207–1210 (2011).
38
E Atucha, B Roozendaal, The inhibitory avoidance discrimination task to investigate accuracy of memory. Front Behav Neurosci 9, 60 (2015).
39
W Holt, S Maren, Muscimol inactivation of the dorsal hippocampus impairs contextual retrieval of fear memory. J Neurosci 19, 9054–9062 (1999).
40
S Cavallaro, V D’Agata, P Manickam, F Dufour, DL Alkon, Memory-specific temporal profiles of gene expression in the hippocampus. Proc Natl Acad Sci USA 99, 16279–16284 (2002).
41
AI Hernández, et al., Protein kinase M zeta synthesis from a brain mRNA encoding an independent protein kinase C zeta catalytic domain. Implications for the molecular mechanism of memory. J Biol Chem 278, 40305–40316 (2003).
42
B Roozendaal, NA Castello, G Vedana, A Barsegyan, JL McGaugh, Noradrenergic activation of the basolateral amygdala modulates consolidation of object recognition memory. Neurobiol Learn Mem 90, 576–579 (2008).
43
R Paz, D Paré, Physiological basis for emotional modulation of memory circuits by the amygdala. Curr Opin Neurobiol 23, 381–386 (2013).
44
AH van Stegeren, et al., Noradrenaline mediates amygdala activation in men and women during encoding of emotional material. Neuroimage 24, 898–909 (2005).
45
L Cahill, et al., Amygdala activity at encoding correlated with long-term, free recall of emotional information. Proc Natl Acad Sci USA 93, 8016–8021 (1996).
46
EJ Hermans, et al., Stress-related noradrenergic activity prompts large-scale neural network reconfiguration. Science 334, 1151–1153 (2011).
47
JE Castro, et al., Role of the amygdala in antidepressant effects on hippocampal cell proliferation and survival and on depression-like behavior in the rat. PLoS One 5, e8618 (2010).
48
F Tissir, AM Goffinet, Reelin and brain development. Nat Rev Neurosci 4, 496–505 (2003).
49
U Beffert, et al., Modulation of synaptic plasticity and memory by Reelin involves differential splicing of the lipoprotein receptor Apoer2. Neuron 47, 567–579 (2005).
50
S Niu, O Yabut, G D’Arcangelo, The Reelin signaling pathway promotes dendritic spine development in hippocampal neurons. J Neurosci 28, 10339–10348 (2008).
51
TC Sacktor, How does PKMζ maintain long-term memory? Nat Rev Neurosci 12, 9–15 (2011).
52
E Pastalkova, et al., Storage of spatial information by the maintenance mechanism of LTP. Science 313, 1141–1144 (2006).
53
Y Yao, et al., PKM zeta maintains late long-term potentiation by N-ethylmaleimide-sensitive factor/GluR2-dependent trafficking of postsynaptic AMPA receptors. J Neurosci 28, 7820–7827 (2008).
54
R Shema, TC Sacktor, Y Dudai, Rapid erasure of long-term memory associations in the cortex by an inhibitor of PKM zeta. Science 317, 951–953 (2007).
55
AM Lee, et al., Prkcz null mice show normal learning and memory. Nature 493, 416–419 (2013).
56
LJ Volk, JL Bachman, R Johnson, Y Yu, RL Huganir, PKM-ζ is not required for hippocampal synaptic plasticity, learning and memory. Nature 493, 420–423 (2013).
57
BS Carvalho, RA Irizarry, A framework for oligonucleotide microarray preprocessing. Bioinformatics 26, 2363–2367 (2010).
58
J Tost, IG Gut, DNA methylation analysis by pyrosequencing. Nat Protoc 2, 2265–2275 (2007).
59
G Paxinos, C Watson The Rat Brain in Stereotaxic Coordinates (Academic, 6th Ed, San Diego, 2007).
60
ME Ritchie, et al., limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43, e47 (2015).
61
Y Benjamini, Y Hochberg, Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B 57, 289–300 (1995).
62
IC Weaver, et al., Epigenetic programming by maternal behavior. Nat Neurosci 7, 847–854 (2004).
63
A Subramanian, et al., Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102, 15545–15550 (2005).
64
MW Pfaffl, A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29, e45 (2001).
Information & Authors
Information
Published in
Classifications
Submission history
Published online: August 8, 2017
Published in issue: August 22, 2017
Keywords
Acknowledgments
This research was supported by a Radboud University Topfund (to B.R.) and funds of the University of Basel (to A.P. and D.J.-F.d.Q.).
Authors
Competing Interests
The authors declare no conflict of interest.
Metrics & Citations
Metrics
Altmetrics
Citations
Cite this article
Noradrenergic activation of the basolateral amygdala maintains hippocampus-dependent accuracy of remote memory, Proc. Natl. Acad. Sci. U.S.A.
114 (34) 9176-9181,
https://doi.org/10.1073/pnas.1710819114
(2017).
Copied!
Copying failed.
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.