Genetic modulation of striatal volume by loci on Chrs 6 and 17 in BXD recombinant inbred mice

Abstract

Natural variation in the absolute and relative size of different parts of the human brain is substantial, with a range that often exceeds a factor of 2. Much of this variation is generated by the cumulative effects of sets of unknown gene variants that modulate the proliferation, growth and death of neurons and glial cells. Discovering and testing the functions of these genes should contribute significantly to our understanding of differences in brain development, behavior and disease susceptibility. We have exploited a large population of genetically well-characterized strains of mice (BXD recombinant inbred strains) to map gene variants that influence the volume of the dorsal striatum (caudate-putamen without nucleus accumbens). We used unbiased methods to estimate volumes bilaterally in a sex-balanced sample taken from the Mouse Brain Library (www.mbl.org). We generated a matched microarray data set to efficiently evaluate candidate genes (www.genenetwork.org). As in humans, volume of the striatum is highly heritable, with greater than twofold differences among strains. We mapped a locus that modulates striatal volume on chromosome (Chr) 6 at 88 +/- 5 Mb. We also uncovered an epistatic interaction between loci on Chr 6 and Chr 17 that modulates striatal volume. Using bioinformatic tools and the corresponding expression database, we have identified positional candidates in these quantitative trait locus intervals.

Figure 1

Mean ± SEM striatal volume in BXD RI lines (gray bars) and their parental strains, C57BL/6J (black bars) and DBA/2J (white bars).

Figure 2

Mapping striatal volume. A. LRS scores for striatal volume across the entire genome. The x-axis represents the physical map of this chromosome; 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 1000 permutations. There are QTLs (red arrows) on the proximal end of Chr 2, the distal end of Chr 6, the central region of Chr 8, and the proximal end of Chr 11. B. LRS for residual striatal volume (regressing out the effects of age, sex, plane of section, epoch, and non-striatal brain weight) reveals a significant QTL on the distal end of Chr 6. The LRS for the QTL on Chr 8 is now diminished (see text for explanation), and new QTLs on the distal end of Chr 1 and the proximal end of Chr 17 are revealed. C. LRS map of all of Chr 6. Orange lines on x-axis represent high density SNP map. Discontinuous track along the top are the genes on this chromosome. D. A 10 Mb interval bordering the QTL on Chr 6.

Figure 3

Pair-scan correlations demonstrate interactions between two QTLs for striatal volume. A. Pair-scan analysis of residual striatal volume across the genome. The upper left half of the plot highlights any epistatic interactions (corresponding to the column labeled “LRS Interact”). The lower right half provides a summary of LRS of the full model, representing cumulative effects of linear and non-linear terms (column labeled “LRS Full”). A significant interaction between QTLs on Chr 6 and 17 (white circle) is shown. B. Enlargement of intersection of Chrs 6 and 17 as seen in A, illustrating region of significance. C. Histogram illustrating the effect on adjusted striatal volume of carrying either the parental (D2) or maternal (B6) or both alleles at the Chr 6 and Chr 17 intervals. Having D2 alleles at both intervals significantly increases adjusted striatal volume when compared to all other allelic combinations.

Figure 4

Scatterplots illustrating correlation of adjusted striatal volume with traits from the BXD Phenotype Database. Pearson product moment correlations (in box) of adjusted striatal volume with (A) transferrin saturation percent (Trait 10822), (B) the volume of the internal granule cell layer (IGL) of the cerebellum ITrait ID 10006), (C) the number of striatal cholinergic neurons (Trait 10106), and (D) the number of convulsions following nitrous oxide (NO) withdrawal (Trait 10027). *P <.05, **P <.01.

Figure 5

Expression of potential candidate genes, and hypothesized gene networks. A. Interval map for Htra2 gene expression (see Fig 2 for explanation of the graph). This Chr 6 gene significantly modulates gene expression in the Chr 17 QTL interval (red arrow). B. Interval map for the Chr 6 gene Nfu1, which modulates its own gene expression, as well as gene expression on Chrs 11 and in the QTL interval on Chr 17. C and D. Interval maps for Glo1 and Ddah2, respectively, which are genes on the Chr 17 QTL interval that significantly modulate their own expression. These cis-QTLs were verified at the probe level, and are not likely due to undetected SNPs. E. Hypothetical gene networks modulating striatal volume based on gene function (left) and gene expression (right). Arrowed dotted lines indicate QTLs and circular arrowed lines indicate cis-QTLs, i.e., genes that modulate expression in the same chromosomal interval. In the case of both trans- and cis-QTLS, red lines represent contribution of the C57BL/6J and green the DBA/2J haplotype. Solid red lines represent positive correlations and solid green lines represent negative correlations, which are based on gene expression. Solid black line indicates missense SNP. See text for further explanation.