Connection from granular to dysgranular retrosplenial cortex is required for contextual fear memory retrieval

DE OLMOS, SOLEDAD; ERIC L. SIGWALD; PONCE, NICOLÁS ERIC; ALFREDO LORENZO

San Diego

Congreso; 2018 Annual Meeting, Society for Neuroscience; 2018

Society for Neuroscience of USA

The retrosplenial cortex (RSC) is divided in dysgranular (A30) and granular (A29) subregions. However, the functional role of each RSC subdivision in episodic memory processing remains poorly understood. In recent years, the participation of the RSC in contextual fear memory (CFM) has gained increased attention; nevertheless these studies have regarded the RSC as a uniform structure. The aim of this work was to reveal functional differences of RSC subdivisions in CFM processing. We used male rats and assessment of c-Fos and Egr-1 expression to infer neuronal activity during contextual fear conditioning. Data showed that during training and test, expression of Egr-1 dropped in A29 while Egr-1 and c-Fos rose in A30. Repeated measures ANOVA (F(5,37)= 83.967, p=0.0000) confirmed significant difference between RSC subdivisions, suggesting distinct or complementary roles for each area. Further analysis and statistical comparisons showed that similar to A29, reduced Egr-1 expression during training and test also occurred in caudomedial entorhinal cortex (CEnt). Contrarily, increases in c-Fos and Egr-1 in A30 during training and test were coincidental with medial entorhinal cortex (MEnt), dorsal CA1 field of hippocampus (CA1), central (CeA) and basolateral (BLA) nucleus of the amygdala (ANOVA (F(12,108)=75.006, p=0.0000), suggesting that A30 and these limbic structures are coupled into a distributed network supporting CFM encoding and retrieval. We later used systemic applications of MK801 to induce a non-invasive and selective elimination of neurons in layers IV-Va of A29 (A29MK801 neurons). Loss of A29MK801 neurons after CFM consolidation did not affect activation of limbic structures during test, but significantly impaired activation of A30 (ANOVA (F(18,144)=5.7599, p=0.0000) and the expression of freezing during test (ANOVA F(2,12)=34.168, p=0.00001). This observation indicates that during test, A29MK801 neurons are critically required for coupling activity of A30 with limbic structures and the retrieval of CFM. By using silver staining, immunolabeling and track-tracing methods we showed that A29MK801 are typical pyramidal neurons (likely excitatory) projecting to superficial layers of A30, and confirmed that GABAergic neurons in A29 were not affected by MK801. Collectively, our results indicate that A29MK801 neurons play a critical role coupling activation of A30 with limbic structures during fear memory recall, which is required for CFM retrieval and expression of adaptive freezing behavior.