Behavioral and genetic architecture of fear conditioning and related phenotypes

Abstract

Contextual fear conditioning is a form of Pavlovian learning during which an organism learns to fear previously neutral stimuli following their close temporal presentation with an aversive stimulus. In mouse models, freezing behavior is typically used to quantify learned fear. This dependent variable is the sum of multiple processes, including associative/configural learning, fear and anxiety, and general activity. To explore phenotypic constructs underlying contextual fear conditioning and correlated behaviors, as well as factors that may contribute to individual differences in learning and mental health, we tested BXD recombinant inbred strains previously found to show extreme contextual fear conditioning phenotypes and BXD parental strains, C57BL/6J and DBA/2J, in a series of tests including locomotor, anxiety, contextual/cued fear conditioning and non-associative hippocampus-dependent learning behaviors. Hippocampal expression of two previously identified candidate genes for contextual fear conditioning was also quantified. Behavioral and gene expression data were analyzed using exploratory factor analysis (EFA), which suggested five unique constructs representing activity/anxiety/exploration, associative fear learning, anxiety, post-shock freezing, and open field activity phenotypes. Associative fear learning and expression of one candidate gene, Hacd4, clusteredas a construct withinthefactor analysis. Post-shock freezingduring fear conditioning and expression of candidate gene Ptprd emerged as another unique construct, highlighting theindependenceof freezing after footshock from other fear conditioning variables in the current dataset.EFA results additionally suggest shared phenotypic variance in adaptive murine behaviors related to anxiety, general activity, and exploration. These findings inform understanding of fear learning and underlying biological mechanisms that may interact to produce individual differences in fear- and learning-related behaviors in mice.

Keywords: Anxiety; Factor analysis; Fear conditioning; Gene expression; Hippocampus; Learning.

Figure 1.

Summarized behavioral battery timeline.

Figure 2.

Strain means on factor composite scores. Homogenous subsets (groups of means not significantly different from each other, corrected p > .05) are denoted by common lowercase letters displayed above mean bars in panels a, b, d, and e). In panel c, homogenous subsets are denoted by common lowercase letters for females or by common symbols below mean bars for males. Post-hoc correction method indicated in panel titles: Tukey = Tukey-HSD; GH = Games-Howell. Error bars express ±1 SEM.

Figure 3.

Contextual freezing strain means. Homogenous subsets (groups of means not significantly different from each other, Games-Howell corrected p > .05) are denoted by common lowercase letters displayed above mean bars. Error bars express ±1 SEM.

Figure 4.

OLM learning (preference ratio) strain means. Homogenous subsets (groups of means not significantly different from each other, Games-Howell corrected p > .05) are denoted by common lowercase letters (females) or symbols (males) displayed above mean bars. 50% threshold (dotted black line) represents chance level of learning. Error bars express ±1 SEM.

Figure 5.

Strain means on candidate gene expression in dorsal (DH) and ventral (VH) hippocampus. Homogenous subsets (groups of means not significantly different from each other, corrected p > .05) are denoted by common lowercase letters displayed above mean bars. Post-hoc correction method indicated in panel titles: Tukey = Tukey-HSD; GH = Games-Howell. ΔCT values are inversely proportional to gene expression (higher ΔCT = lower gene expression). Error bars express ±1 SEM.