Experiment
The unique interactions between gene and environment on cognitive performance were the focus of the current study. The behavioral genetics model of BXD recombinant inbred mice, derived from C57BL/6J and DBA/2J parental strains, was used to assess spatial learning and memory based on performance in the Morris water maze (MWM).
Sixty-two BXD strains (n=6-10 ofeach strain) and 9-10 mice from each parental strain were subject to 4 weeks of behavioral testing starting at 9 weeks of age. Two littermates per strain were tested in a cohort: one was designated behavioral control (BC), and the other placed through a chronic variable stress (CVS) paradigm. Control animals (n = 305) underwent behavioral testing for three weeks, while CVS animals (n = 305) were placed through the CVS paradigm for one week prior to starting the three week behavioral testing period (elevated plus maze, MWM, and fear conditioning) (Fig. 1).
Animals completed an elevated plus maze test the day prior to starting the MWM test. The CVS stressors continued throughout behavioral testing. The stressors were as follows: novel overnight housing, in which the mouse was singly housed in a novel rat cage with ad libitum access to food and water;
hypoxia, in which for 30 min the mice were placed in a low oxygen environment (812% oxygen);
open field, in which the mice were singly housed in an open cage (10.5 19 8) in a well-lit area for 30min;
cold room, in which the mice were placed at 4 C for 15 min in a cage devoid of bedding, singly housed;
and constant motion, in which the mice were placed on an orbital shaker at 100 rpm for 1 h.
Each stressor was repeated five times throughout the experiment with the exception of novel overnight housing, which was repeated three times. The combination and sequence of stressors changed each week to prevent predictability and limit habituation.
The MWM tests the spatial navigation and memory of the mouse, as measured through the latency to find the hidden platform. Performance was assessed as average latency to platform for all trials, training trials, and reversal trials as well as the number of entries into the platform area during probe. Stress-effect was calculated as the difference in performance between control and CVS littermates (CVS minus control).
The GeneNetwork suite of web tools (www.genenetwork.org) was used for QTL mapping (NCBI Build 37). GeneNetwork links differences in phenotype to genomic regions using 89 BXD recombinant inbred strains that have been assessed using 3806 genomic markers to identify suggestive and significant QTL with likelihood ratio statistics with genome-wide probabilities of 0.67 and 0.05, respectively.
Suggestive QTL unique to the behavioral control population were located on Chromosomes 3, 7, and 10 for all trials and on Chromosome 19 for all trials and during training:
A suggestive QTL mapped to Chromosome 3 for all trials with a peak LRS of 11.646 at peak marker rs3668064 with a confidence interval mapping between 7.7-10.1 Mb.
A suggestive QTL mapped to Chromosome 7 for all trials with a peak LRS of 11.841 at peak marker rs4226520 with a confidence interval mapping between 28.2-31.3 Mb.
A suggestive QTL mapped to Chromosome 10 for all trials with a peak LRS of 10.599 at peak marker D10Mit28 with a confidence interval mapping between 0.0-14.7 Mb.
A suggestive QTL mapped to Chromosome 19 for all trials with a peak LRS of 12.373 at peak marker rs3653886 with a confidence interval mapping between 33.5-38.1 Mb.
Significant QTL unique to the stressed population were identified on Chromosomes 1, 2, and 18 for all trails and training trials and a suggestive QTL mapped to Chromosome 8 for training trials:
QTL Mwmlpa1 (Morris water maze latency to platform, all, 1) mapped to Chromosome 1 for all trials with a peak LRS of 18.787 at peak marker NES13029525 with a confidence interval mapping between 171.9-175.4 Mb. The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5].
QTL Mwmlpt1 (Morris water maze latency to platform, training, 1) mapped to Chromosome 1 for training trials with a peak LRS of 22.301 at peak marker NES13029525 with a confidence interval mapping between 171.7-175.5 Mb. The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5.]
Several candidate genes within these Chr 1 loci have established relationships with stress and anxiety - Fcer1g, Ppox, Cadm3, Atp1a2.
QTL Mwmlpa2 (Morris water maze latency to platform, all, 2) mapped to Chromosome 2 for all trials with a peak LRS of 16.054 at peak marker CEL-2_135876979 with a confidence interval mapping between 131.6-139.6 Mb. The significant LRS score is associated with the DBA/2J genotype at this locus. [Fig 4., Fig 5., Table 5.]
QTL Mwmlpt2 (Morris water maze latency to platform, training, 2) mapped to Chromosome 2 for training trials with a peak LRS of 18.463 at peak marker CEL-2_135876979 with a confidence interval mapping between 131.6-139.2 Mb. The significant LRS score is associated with the DBA/2J genotype at this locus. [Fig 4., Fig 5., Table 5.]
Within the genomic region of the two Chromosome 2 QTL are genes Plcb1 and Plcb4; both are associated with anxiety.
A suggestive QTL mapped to Chromosome 8 for training trials with a peak LRS of 11.12 at peak marker rs3666069 with a confidence interval mapping between 93.3-98.7 Mb.
QTL Mwmlpa3 (Morris water maze latency to platform, all, 3) mapped to Chromosome 18 for all trials with a peak LRS of 19.218 at peak marker rs6358426 with a confidence interval mapping between 16.1-24.2 Mb. The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5.]
QTL Mwmlpt3 (Morris water maze latency to platform, training, 3) mapped to Chromosome 18 for training trials with a peak LRS of 22.497 at peak rs6358426 with a confidence interval mapping between 16.1-24.1 Mb. The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5.]
The genetic region for the two Chr 18 QTL contain Cdh2 and Dsc1, each contributing to cadherin's role in synapse formation and maintaining neuronal circuits. The cis-regulated gene Ttr is also found within this region.
Two peaks were found in both populations for average latency to platform across all trails and during training, on Chromosomes 5 and 10:
A suggestive QTL mapped to Chromosome 5 for all trials in the behavioral control population with a peak LRS of 12.589 at peak marker rs3656989 with a confidence interval mapping between 63.3-66.8 Mb.
A suggestive QTL mapped to Chromosome 5 for training trials in the behavioral control population with a peak LRS of 14.415 at peak marker rs3656989 with a confidence interval mapping between 63.4-69 Mb.
Significant QTL Mwmlpa4 (Morris water maze latency to platform, all, 4) mapped to Chromosome 5 for all trials in the chronic stress population with a peak LRS of 17.743 at peak marker rs3657916 with a confidence interval mapping between 60.6-66.8 Mb. The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5.]
Significant QTL Mwmlpt4 (Morris water maze latency to platform, training, 4) mapped to Chromosome 5 for training trials in the chronic stress population with a peak LRS of 22.376 at peak marker rs3657916 with a confidence interval mapping between 60.6-68.0 Mb.
The significant LRS is associated with the C57BL/6J genotype at this locus. [Fig 4., Fig 5., Table 5.]
A suggestive QTL mapped to Chromosome 10 for all trials in the behavioral control population with a peak LRS of 11.158 at peak marker rs13480657 with a confidence interval mapping between 73.0-78.8 Mb.
A suggestive QTL mapped to Chromosome 10 for training trials in the behavioral control population with a peak LRS of 10.711 at peak marker rs13480653 with a confidence interval mapping between 73.2-79.8 Mb.
A suggestive QTL mapped to Chromosome 10 for all trials in the chronic stress population with a peak LRS of 14.21 at peak marker rs13480650 with a confidence interval mapping between 68.8-78.9 Mb.
A suggestive QTL mapped to Chromosome 10 for training trials in the chronic stress population with a peak LRS of 13.901 at peak marker rs13480653 with a confidence interval mapping between 72.3-78.5 Mb.
Genes within each significant and suggestive QTL's confidence interval were assessed for human homologues and cis-regulation to achieve a list of candidate genes - Tables 6 and 7.
Analysis Tools