Three paralogous genes for gonadotropin-releasing hormone (GnRH; gnrh1, gnrh2, and gnrh3) and GnRH receptors exist in non-mammalian vertebrates. However, there are some vertebrate species in which one or two of these paralogous genes have become non-functional during evolution. The developmental migration of GnRH neurons in the brain is evolutionarily conserved in mammals, reptiles, birds, amphibians, and jawed teleost fish. The three GnRH paralogs have specific expression patterns in the brain and originate from multiple sites. In acanthopterygian teleosts (medaka, cichlid, etc.), the preoptic area (POA)-GnRH1 and terminal nerve (TN)-GnRH3 neuronal types originate from the olfactory regions. In other fish species (zebrafish, goldfish and salmon) with only two GnRH paralogs (GnRH2 and GnRH3), the TN- and POA-GnRH3 neuronal types share the same olfactory origin. However, the developmental origin of midbrain (MB)-GnRH2 neurons is debatable between mesencephalic or neural crest site. Each GnRH system has distinctive anatomical and physiological characteristics, and functions differently. The POA-GnRH1 neurons are hypophysiotropic in nature and function in the neuroendocrine control of reproduction. The non-hypophysiotropic GnRH2/GnRH3 neurons probably play neuromodulatory roles in metabolism (MB-GnRH2) and the control of motivational state for sexual behavior (TN-GnRH3).
The fork cell and von Economo neuron, which are found in the insular cortex and/or the anterior cingulate cortex, are defined by their unique morphologies. Their shapes are not pyramidal; the fork cell has two primary apical dendrites and the von Economo neurons are spindle-shaped (bipolar). Presence of such neurons are reported only in the higher animals, especially in human and great ape, indicating that they are specific for most evolved species. Although it is likely that these neurons are involved in higher brain function, lack of results with experimental animals makes further investigation difficult. We here ask whether equivalent neurons exist in the mouse insular cortex. In human, Fezf2 has been reported to be highly expressed in these morphologically distinctive neurons and thus, we examined the detailed morphology of Fezf2-positive neurons in the mouse brain. Although von Economo-like neurons were not identified, Fezf2-positive fork cell-like neurons with two characteristic apical dendrites, were discovered. Examination with electron microscope indicated that these neurons did not embrace capillaries, rather they held another cell. We here term such neurons as holding neurons. We further observed several molecules, including neuromedin B (NMB) and gastrin releasing peptide (GRP) that are known to be localized in the fork cells and/or von Economo cells in human, were localized in the mouse insular cortex. Based on these observations, it is likely that an equivalent of the fork cell is present in the mouse.
Subplate (SP) neurons are among the earliest-born neurons in the cerebral cortex and heterogeneous in terms of gene expression. SP neurons consist mainly of projection neurons, which begin to extend their axons to specific target areas very early during development. However, the relationships between axon projection and gene expression patterns of the SP neurons, and their remnant layer 6b (L6b) neurons, are largely unknown. In this study, we analyzed the corticocortical projections of L6b/SP neurons in the mouse cortex and searched for a marker gene expressed in L6b/SP neurons that have ipsilateral inter-areal projections. Retrograde tracing experiments demonstrated that L6b/SP neurons in the primary somatosensory cortex (S1) projected to the primary motor cortex (M1) within the same cortical hemisphere at postnatal day (PD) 2 but did not show any callosal projection. This unilateral projection pattern persisted into adulthood. Our microarray analysis identified the gene encoding a β subunit of voltage-gated potassium channel (Kcnab1) as being expressed in L6b/SP. Double labeling with retrograde tracing and in situ hybridization demonstrated that Kcnab1 was expressed in the unilaterally-projecting neurons in L6b/SP. Embryonic expression was specifically detected in the SP as early as embryonic day (E) 14.5, shortly after the emergence of SP. Double immunostaining experiments revealed different degrees of co-expression of the protein product Kvβ1 with L6b/SP markers Ctgf (88%), Cplx3 (79%), and Nurr1 (58%), suggesting molecular subdivision of unilaterally-projecting L6b/SP neurons. In addition to expression in L6b/SP, scattered expression of Kcnab1 was observed during postnatal stages without layer specificity. Among splicing variants with three alternative first exons, the variant 1.1 explained all the cortical expression mentioned in this study. Together, our data suggest that L6b/SP neurons have corticocortical projections and Kcnab1 expression defines a subpopulation of L6b/SP neurons with a unilateral inter-areal projection.
Association projections from cortical pyramidal neurons connect disparate intrahemispheric cortical areas, which are implicated in higher cortical functions. The underlying developmental processes of these association projections, especially the initial phase before reaching the target areas, remain unknown. To visualize developing axons of individual neurons with association projections in the mouse neocortex, we devised a sparse labeling method that combined in utero electroporation and confocal imaging of flattened and optically cleared cortices. Using the promoter of an established callosal neuron marker gene that was expressed in over 80% of L2/3 neurons in the primary somatosensory cortex (S1) that project to the primary motor cortex (M1), we found that an association projection of a single neuron was the longest among the interstitial collaterals that branched out in L5 from the earlier-extended callosal projection. Collaterals to M1 elongated primarily within the cortical gray matter with little branching before reaching the target. Our results suggest that dual-projection neurons in S1 make a significant fraction of the association projections to M1, supporting the directed guidance mechanism in long-range corticocortical circuit formation over random projections followed by specific pruning.