Experiment
The hypothesis that environmental context (the presence or absence of stress) reveals unique genetic predictors of emotional behavior was tested in the current study. Sixty-two different BXD strains (592 BXD mice, n=3-5 per strain group), in addition to the C57BL/6J and DBA/2J parental strains (n=5 per strain group), were tested in the study.
Emotional learning and behavior was assessed by fear conditioning and performance on an elevated maze. Littermates of each strain were treated with either control or chronic variable stress (CVS). Chronic stress was established using the CVS model. Animals underwent a randomized schedule of one or two mild to moderate physiological and psychological stressors daily 5 days a week. Stressors included novel overnight home cage, hypoxia, cold exposure, open filed and constant motion exposure. All cohorts of animals were exposed to the same order and sequence of stressors.
Mice were subjected to the elevated plus maze (EPM) in which they were placed in the center of the apparatus and recorded and monitored. The measures collected included time spent in the closed arms (associated with anxiety-like behavior), time spent in the open arm (considered exploratory behavior), and distance traversed throughout the maze (assessed as a measure of locomotor activity within BXD strains).
Four fear conditioning chambers were used and conditioning was completed over 3 days. Day one mice were into the operant chamber and freezing behavior was assessed. The session included 4 exposures to a 30s tone that terminated with a 2s foot shock which elicited freezing behavior (unconditioned response). On day 2 mice were returned to the operant chamber and freezing behavior was monitored for 10 minutes. On day 3 the environmental context of the operant chamber was altered by replacing the grid floor with a smooth Plexiglass insert. Mice were then exposed to the 30s tone 9 times with random inter-tone intervals. Percent of time displaying freezing behavior was measured each day.
Stress effect was calculated as the difference in performance between control and the CVS littermates. Mixed model analyses were performed using R packages with strain, month of testing, and experimenter handling during testing as fixed variables and cohort as a random variable. Correlations between measure traits were computed via GeneNetwork for both the control and the stress populations.
QTL mapping was performed using complex trait analysis and mapping tools on the GeneNetwork website (http://www.genenetwork.org) using using NCBI Build 37 coordinates. GeneNetwork utilizes 3806 markers in 89 BXD recombinant inbred strains to link regions of the genome to differences in phenotype. Thresholds for suggestive and significant QTL are determined as likelihood ratio statistic (LRS) values associated with genome wide probabilities of 0.67 and 0.05, respectively.
Mapping was performed for three traits of emotional learning assessed during the fear conditional paradigm (freezing during training to tone and shock on day 1, freezing to environmental context on day 2 and freezing to tone on day 3) and three emotional behavioral traits measured within the elevated plus maze (locomotion or distance traversed, time spent in closed arms and time spent in open arms.
Table 5:
A significant QTL, Emlq1 (emotional learning QTL 1) mapped to Chr 13 in both the control and the CVS mapping populations between 78.1 and 97.5 Mb, with a peak LRS=22.2. The QTL influenced all 3 emotional traits: freezing to training, freezing to tone and freezing to context. The DBA/2J allele at this locus increased the effect of the trait.
A significant QTL, Emlq2 (emotional learning QTL 2) mapped Chromosome 14 in the control mapping population between 71.0 and 73.5 Mb, with a peak LRS=17.3. Emlq2 influenced freezing to tone with the C57BL/6J allele increasing the effect of the trait.
A significant QTL, Emlq3 (emotional learning QTL 3) mapped to Chromosome X in both the control and the CVS mapping populations between 56.2 and 68.3 Mb, with a peak LRS=17.0. Emlq3 influenced both freezing in training and freezing to tone with the C57BL/6J allele increasing the effects.
Suggestive QTL for freezing to training were mapped to Chr 1 and Chr 3 in both the control and CVS populations with the DBA/2J allele increasing the effect of the trait. Suggestive QTL for freezing to training were also mapped in the CVS population to Chrs 2 and 5 with the C57BL/6J allele increasing the effect of the Chr 2 locus and the DBA/2J allele increasing the effect of the Chr 5 locus.
A suggestive QTL for freezing in context mapped to Chromosome 13 with the DBA/2J allele associated with increased trait effect.
A significant QTL, Embq1 (emotional behavior QTL 1) mapped to Chromosome 4 in both the control and the CVS mapping populations between 65.5 and 95.7 Mb, with a peak LRS=16.5. Embq1 influenced locomotion in the EPM with the C57BL/6J allele increasing the effect of the trait.
A significant QTL, Sefq1 (stress effect QTL 1) mapped to Chromosome 16 between 74.9 and 79.1 Mb with a peak LRS=17.0. The C57BL/6J allele at Sefq1 increased locomotion in the EPM. A suggestive stress effect QTL influencing freezing to training mapped to Chromosome 15 with the DBA/2J allele increasing the trait effect.
Candidate genes were defined as genes within each QTL that had a human homolog and/or were cis-regulated. Genes were further assessed via literary search in PubMed for known relationship with stress, learning, anxiety, fear conditioning, and exploration. Candidate genes are listed in Table 6.
Analysis Tools