Modulation of learning and memory by the genetic disruption of circadian oscillator populations

Abstract

While a rich literature has documented that the efficiency of learning and memory varies across circadian time, a close survey of that literature reveals extensive heterogeneity in the time of day (TOD) when peak cognitive performance occurs. Moreover, most previous experiments in rodents have not focused on the question of discriminating which memory processes (e.g., working memory, memory acquisition, or retrieval) are modulated by the TOD. Here, we use assays of contextual fear conditioning and spontaneous alternation in WT (C57Bl/6 J) mice to survey circadian modulation of hippocampal-dependent memory at multiple timescales - including working memory (seconds to a few minutes), intermediate-term memory (a delay of thirty minutes), and acquisition and retrieval of long-term memory (a delay of two days). Further, in order to test the relative contributions of circadian timing mechanisms to the modulation of memory, a parallel set of studies were performed in mice lacking clock timing mechanisms. These transgenic mice lacked the essential circadian gene Bmal1, either globally (Bmal1 null) or locally (floxed Bmal1 mice, which lack Bmal1 in excitatory forebrain neurons, e.g. cortical and hippocampal neurons). Here, we show that in WT mice, retrieval (but not working memory, intermediate-term memory, or acquisition of long-term memory) is modulated by TOD. However, transgenic mouse models lacking Bmal1 - both globally, and only in forebrain excitatory neurons - show deficits regardless of the memory process tested (and lack circadian modulation of retrieval). These results provide new clarity regarding the impact of the TOD on hippocampal-dependent memory and support the key role of hippocampal and cortical circadian oscillations in circadian gating of cognition.

Keywords: Bmal1; Circadian; Hippocampus; Learning; Memory.

Fig. 1.

Experimental design of behavioral assays. A) Key: white and black rectangles indicate, respectively, light and dark conditions under a 12 h light / 12 h dark illumination schedule. Light gray and dark gray rectangles indicate, respectively, subjective day and subjective night under total darkness (DD). Arrows indicate training times, while triangles indicate testing times. B) Working memory (spontaneous alternation) and intermediate-term memory (CFC) assays. Animals were dark adapted on day 1. On day 2, they underwent either the spontaneous alternation test, or fear conditioning with testing following a 30-min delay. C) Assay to test the TOD impact on acquisition of the long-term memory. Animals were dark adapted on day 1. On day 2, they underwent fear conditioning at either CT4 or CT16. Two days later (day 4), memory was tested at CT10. D) Timeline to test the TOD effects on retrieval of the long-term memory. Animals were dark-adapted on day 1. On day 2, they underwent fear conditioning at CT10. Two days later (day 4), memory was tested at either CT4 or CT16.

Fig. 2.

Spontaneous alternation. A) Spontaneous alternation of WT, Bmal1 cKO, and Bmal1 gKO mice at CT4 or CT16 on the y-maze (N, respectively, = 31, 31, 11, 12, 17, 15). B) Total entries of WT, Bmal1 cKO, and Bmal1 gKO mice at CT4 or CT16 on the y-maze (N, respectively, = 31, 31, 11, 12, 17, 15). *, p < 0.05; n.s., p > 0.05; bars indicate genotype comparisons; no TOD differences were found.

Fig. 3.

Contextual fear conditioning – baseline freezing and intermediate-term memory. A) Baseline freezing of WT, Bmal1 cKO, and Bmal1 gKO mice in the context before fear conditioning at CT4 or CT16 (N, respectively, = 14, 14, 13 14, 5, 10). B) Freezing response of WT, Bmal1 cKO, and Bmal1 gKO mice following conditioning at CT4 or CT16, with a 30-min delay between training and retrieval (N, respectively, = 18, 16, 14, 15, 7, 10). C) Freezing response of WT mice with or without the CaMKII-CRE gene following conditioning at CT4 with a 30-min delay between training and retrieval (N, respectively, = 9, 9). *, p < 0.05; n.s., p > 0.05; bars indicate genotype comparisons; no TOD differences found.

Fig. 4.

Contextual fear conditioning – long-term memory. A) Freezing response of WT, Bmal1 cKO, and Bmal1 gKO mice (here, CT indicates time of acquisition; all animals tested at CT10; N, respectively, = 17, 16, 13, 14, 6, 10). B) Freezing response of WT, Bmal1 cKO, and Bmal1 gKO mice (here, CT indicates time of retrieval; all animals underwent fear conditioning at CT10; N, respectively, = 13, 12, 9, 23, 8, 8). *, p < 0.05; n.s., p > 0.05; long bars indicate genotype comparisons, short bars indicate TOD comparisons; no TOD difference seen in panel A; for panel B, the letters A, B indicate genotype differences at CT4, and C indicates no genotype differences at CT16.