Genetic Variation in CNS Myelination and Functional Brain Connectivity in Recombinant Inbred Mice

Abstract

Myelination greatly increases the speed of action potential propagation of neurons, thereby enhancing the efficacy of inter-neuronal communication and hence, potentially, optimizing the brain's signal processing capability. The impact of genetic variation on the extent of axonal myelination and its consequences for brain functioning remain to be determined. Here we investigated this question using a genetic reference panel (GRP) of mouse BXD recombinant inbred (RI) strains, which partly model genetic diversity as observed in human populations, and which show substantial genetic differences in a variety of behaviors, including learning, memory and anxiety. We found coherent differences in the expression of myelin genes in brain tissue of RI strains of the BXD panel, with the largest differences in the hippocampus. The parental C57BL/6J (C57) and DBA/2J (DBA) strains were on opposite ends of the expression spectrum, with C57 showing higher myelin transcript expression compared with DBA. Our experiments showed accompanying differences between C57 and DBA in myelin protein composition, total myelin content, and white matter conduction velocity. Finally, the hippocampal myelin gene expression of the BXD strains correlated significantly with behavioral traits involving anxiety and/or activity. Taken together, our data indicate that genetic variation in myelin gene expression translates to differences observed in myelination, axonal conduction speed, and possibly in anxiety/activity related behaviors.

Keywords: behavior; gene expression; genetics; myelin; oligodendrocyte.

Figure 1

Clustering of gene expression data shows pronounced differences in myelin gene expression between C57BL/6J (C57), DBA/2J (DBA), and their BXD progeny. BXD microarray data were obtained from the GeneNetwork ‘Hippocampus Consortium PDNN’ dataset. Transcripts were selected and their gene expression patterns clustered using Euclidean distance and average linkage clustering. (A) The clustering pattern indicates a clear group regulation of all myelin transcripts investigated, with the parental strains C57 and DBA opposite at the most extreme ends. (B,C) Clustering of transcript data related to axonal/dendritic markers and lipid metabolism, respectively, did not reveal similar patterns of co-regulation. The number of strains is 71. The scale bar is depicting Log2 gene expression values. The intensity threshold was set to 0 for all depicted functional groups (i.e., myelin, axonal/dendritic and lipid genes).

Figure 2

Clustering of gene expression from the whole brain INIA Brain dataset shows pronounced differences in myelin gene expression between C57, DBA, and their BXD progeny. BXD microarray data for the whole brain were derived from the GeneNetwork ‘INIA Brain PDNN’ dataset. (A) Myelin transcripts were selected, and their gene expression patterns clustered using Euclidean distance and average linkage clustering. A clear clustering pattern involving all myelin transcripts was detected, with the parental strains C57 and DBA at the most extreme ends. The number of strains is 41. The scale bar is depicting Log2 gene expression values. The intensity threshold was set to 0. (B) Correlation between the INIA Brain PDNN and the independent Hippocampus Consortium PDNN database. Pearson correlation r = 0.43, p < 0.01. The number of strains is 37.

Figure 3

qPCR analysis confirms myelin gene expression differences between the parental C57 and DBA strains. (A) qPCR was performed on myelin gene transcripts in hippocampi from C57 and DBA mice. For nearly all myelin genes investigated, gene expression differences were significant, with C57 mice showing higher levels of expression. (B) qPCR analysis on myelin gene transcript in total brain tissue (minus hippocampus). DBA, n = 5; C57, n = 6. t-test (1-tailed) ** p < 0.01; * p < 0.05.

Figure 4

Differences in myelin protein expression between C57 and DBA mice in whole brain extracts. Immunoblotting was performed for selected myelin proteins on whole brain extracts of P20 (A) and P62 (adult) mice (B). Coomassie staining was used for normalization to control for differences in protein. Graphs show the quantification of myelin proteins measured in pups (P20, left panel) and adult (P62, right panel) mice, respectively. Significant expression differences represented by t-test (2-tailed) ** p < 0.01; * p < 0.05. n = 3–4 mice per group.

Figure 5

Smaller myelinated fibers in optic nerves of DBA mice. Electron microscopy (EM) analysis of optic nerves in cross-sections of either C57 or DBA, respectively, in mice at 2 months of age. Line graph: morphometric g-ratio analysis of myelinated axons in optic nerves of C57 and DBA. Bar graphs: morphometric analysis of myelinated axons showing g-ratio (averaged), axon diameter and myelin thickness for myelinated fibers. Scale bar, 2 μm. Significant expression differences: t-test (2-tailed) * p < 0.05.

Figure 6

DBA mice show, in comparison with C57, reduced signal conduction velocity for fast, myelinated fibers in the corpus callosum. A significant difference in conduction velocity was observed for the fast wave between DBA and C57, as depicted. See Figure S2 for example trace of slow conducting fibers (slow wave) and fast conduction fibers (fast wave). Significant differences are represented by t-test (2-tailed) ** p < 0.01. C57, n = 4 (14 slices); DBA, n = 5 (11 slices).

Figure 7

DBA mice show, in comparison with C57, increased anxiety levels in the dark light box. A significant difference in time spent (duration) and in the number of visits (frequency) in the light compartment of the box was observed between DBA and C57, as depicted. Significant differences are represented by t-test (2-tailed) ** p < 0.01. C57, n = 52; DBA, n = 61.

Figure 8

A summarizing view on the correlation of myelin gene expression, myelin fiber diameter and behavioral traits involving anxiety and/or activity.