Genetic, morphometric, and behavioral factors linked to the midsagittal area of the corpus callosum

Abstract

The corpus callosum is the main commissure connecting left and right cerebral hemispheres, and varies widely in size. Differences in the midsagittal area of the corpus callosum (MSACC) have been associated with a number of cognitive and behavioral phenotypes, including obsessive-compulsive disorders, psychopathy, suicidal tendencies, bipolar disorder, schizophrenia, autism, and attention deficit hyperactivity disorder. Although there is evidence to suggest that MSACC is heritable in normal human populations, there is surprisingly little evidence concerning the genetic modulation of this variation. Mice provide a potentially ideal tool to dissect the genetic modulation of MSACC. Here, we use a large genetic reference panel - the BXD recombinant inbred line - to dissect the natural variation of the MSACC. We estimated the MSACC in over 300 individuals from nearly 80 strains. We found a 4-fold difference in MSACC between individual mice, and a 2.5-fold difference among strains. MSACC is a highly heritable trait (h(2) = 0.60), and we mapped a suggestive QTL to the distal portion of Chr 14. Using sequence data and neocortical expression databases, we were able to identify eight positional and plausible biological candidate genes within this interval. Finally, we found that MSACC correlated with behavioral traits associated with anxiety and attention.

Keywords: BXD; QTL; corpus callosum; midsagittal area; mouse; neocortex.

Figure 1

Histogram of mean (±SEM) MSACC in 76 BXD strains (gray bars) and in the two parentals: DBA/2J (white bar) and C57BL/6J (black bar). Inset is frequency distribution of MSACC in all 303 subjects, illustrating a mostly normal distribution. MSACC is corrected for histological shrinkage.

Figure 2

Mapping MSACC in BXD RI strains. (A). Interval map of MSACC corrected for histological shrinkage across the entire genome. The x-axis represents the physical map of the genome; the y-axis and thick blue line provide the LRS of the association between the trait and the genotypes of markers. The two horizontal lines are the suggestive (blue) and significance (red) thresholds computed using 2000 permutations. There is a suggestive QTL mapping to the distal portion of Chr. 10 (red arrow). (B) Correlations between MSACC and brain weight (left) and age (right) indicate that these two variables significantly contribute to MSACC. Solid lines indicate linear relationship of the variable. Dotted line indicates quadratic relationship of the variables. (C) Interval map of MSACC corrected for shrinkage with the effects of age, sex, epoch, and brain weight regressed out. There is a suggestive QTL on Chr 14 (red arrow). (D) Interval map of all of Chr 14. Green line indicates contribution of DBA/2J alleles. Orange lines on x-axis represent high density SNP map. Discontinuous track along the top are the genes on this chromosome. (E) Haplotype map of all 76 BXD strains on 20 Mb QTL interval on Chr14 (77.5–97.5 Mb). Red lines indicate C57BL/6J alleles (maternal), green lines indicate DBA/2J alleles (paternal), blue lines indicate heterozygous alleles, and gray lines are unknown. Strains are arranged from smallest to largest MSACC (top to bottom).

Figure 3

Candidate gene evaluation. (A) Interval maps of gene expression from three neocortical mRNA expression databases for each of the candidate genes on the Chr 14 QTL interval. These databases represent neocortical mRNA expression at three ages – P3, P14, and P60. Numbers in red indicate mean expression level of the gene, and asterisks indicate cis-eQTLs. (B) Network graph illustrating correlations among MSACC and positional candidate gene expression from the three databases.

Figure 4

Covariation of MSACC with behavioral traits. (A) Scatterplot of MSACC correlated with principle component comprised of 12 traits measuring locomotor response to cocaine administration (Philip et al., ; black circles) and four traits measuring the same response following saline administration from the same study (white circles). Solid line is linear correlation of cocaine response, and dotted line is linear correlation of saline response. (B) Scatterplot of MSACC correlated with errors of omission on a test of attention.

Figure 5

Power analysis to determine optimal number of strains and mice/strain. (A) Power statistics for number of mice within strain as a function of effect size. Increasing the number of mice beyond 4 offers little benefit. (B) Power analysis of different numbers of RI strains as a function of percent variance. Increasing the number of RI strains increases the likelihood of detecting small effect QTLs.