Architectonic heterogeneity in neurons is thought to be important for equipping the mammalian cerebral cortex with an adaptable network that can organize the manifold totality of information it receives. To this end, the dendritic arbors of supragranular pyramidal neurons, even those of the same class, are known to vary substantially. This diversity of dendritic morphology appears to have a rostrocaudal configuration in some brain regions of various species. For example, in humans and non-human primates, neurons in more rostral visual association areas (e.g., V4) tend to have more complex dendritic arbors than those in the caudal primary visual cortex. A rostrocaudal configuration is not so clear in any region of the mouse, which is increasingly being used as a model for neurodevelopmental disorders that arise from dysfunctional cerebral cortical circuits. Therefore, in this study we investigated the complexity of dendritic arbors of neurons distributed throughout a broad area of the mouse cerebral cortex. We reduced selection bias by labeling neurons restricted to become supragranular pyramidal neurons using in utero electroporation. While we observed that the simple rostrocaudal position, cortical depth, or even functional region of a neuron was not directly related to its dendritic morphology, a model that instead included a caudomedial-to-rostrolateral gradient accounted for a significant amount of the observed dendritic morphological variance. In other words, rostrolateral neurons from our data set were generally more complex when compared to caudomedial neurons. Furthermore, dividing the cortex into a visual area and a non-visual area maintained the power of the relationship between caudomedial-to-rostrolateral position and dendritic complexity. Our observations therefore support the idea that dendritic morphology of mouse supragranular excitatory pyramidal neurons across much of the tangential plane of the cerebral cortex is partly shaped by a developmental gradient spanning several functional regions.

Dysfunctional dendritic arborization is a key feature of many developmental neurological disorders. Across various human brain regions, basal dendritic complexity is known to increase along a caudal-to-rostral gradient. We recently discovered that basal dendritic complexity of layer II/III cortical pyramidal neurons in the mouse increases along a caudomedial-to-rostrolateral gradient spanning multiple regions, but at the time, no molecules were known to regulate that exquisite pattern. Integrin subunits have been implicated in dendritic development, and the subunit with the strongest associations with autism spectrum disorder and intellectual disability is integrin β3 (Itgb3). In mice, global knockout of Itgb3 leads to autistic-like neuroanatomy and behavior. Here, we tested the hypothesis that Itgb3 is required for increasing dendritic complexity along the recently discovered tangential gradient among layer II/III cortical pyramidal neurons. We targeted a subset of layer II/III cortical pyramidal neurons for Itgb3 loss-of-function via Cre-loxP-mediated excision of Itgb3. We tracked the rostrocaudal and mediolateral position of the targeted neurons and reconstructed their dendritic arbors. In contrast to controls, the basal dendritic complexity of Itgb3 mutant neurons was not related to their cortical position. Basal dendritic complexity of mutant and control neurons differed because of overall changes in branch number across multiple branch orders (primary, secondary, etc.), rather than any changes in the average length at those branch orders. Furthermore, dendritic spine density was related to cortical position in control but not mutant neurons. Thus, the autism susceptibility gene Itgb3 is required for establishing a tangential pattern of basal dendritic complexity among layer II/III cortical pyramidal neurons, suggesting an early role for this molecule in the developing brain.

Cell-autonomous neuronal functions of genes can be revealed by causing loss or gain of function of a gene in a small and sparse population of neurons. To do so requires generating a mosaic in which neurons with loss or gain of function of a gene are surrounded by genetically unperturbed tissue. Here, we combine the Cre-lox recombination system with in utero electroporation in order to generate mosaic brain tissue that can be used to study the cell-autonomous function of genes in neurons. DNA constructs (available through repositories), coding for a fluorescent label and Cre recombinase, are introduced into developing cortical neurons containing genes flanked with loxP sites in the brains of mouse embryos using in utero electroporation. Additionally, we describe various adaptations to the in utero electroporation method that increase survivability and reproducibility. This method also involves establishing a titer for Cre-mediated recombination in a sparse or dense population of neurons. Histological preparations of labeled brain tissue do not require (but can be adapted to) immunohistochemistry. The constructs used guarantee that fluorescently labeled neurons carry the gene for Cre recombinase. Histological preparations allow morphological analysis of neurons through confocal imaging of dendritic and axonal arbors and dendritic spines. Because loss or gain of function is achieved in sparse mosaic tissue, this method permits the study of cell-autonomous necessity and sufficiency of gene products in vivo.

No abstract

No abstract