The role of <i>Satb2</i>, <i>Ctip2</i> and Fezl in cortical connectivity and the elucidation of their downstream pathways

Abstract

Satb2 has been shown to be a key postmitotic determinant of callosal projection neurons (Alcamo et al., 2008; Britanova et al., 2008). One way by which Satb2 is acting to confer such identity involves the repression of Ctip2, a transcription factor expressed in corticospinal motor neurons (CSMN) of layer V and other subcerebral projection neurons within the deep cortical layers (V and VI). In the current study we have found that Unc5h3, a receptor that mediates repulsion to netrin-1 signals and is robustly downregulated throughout the whole cortical plate of the Satb2-/-, plays an important role in guiding axons toward the midline. Unc5h3 is mostly expressed in callosal projection (Satb2+) neurons and largely excluded from the subcerebral projection (Ctip2+) neurons of the neonatal mouse cortex. Overexpression of Unc5h3 in the Satb2 deficient neocortex leads to a partial restoration of callosal projections in a stage specific manner (at E12.5 but not E14.5). Furthermore, the additional deletion of Ctip2 from the Satb2 deficient cortex, accomplished via the generation of Satb2-/-;Ctip2-/- compound mutants, results in the restoration of Unc5h3 expression and in a similar partial rescue of callosal projections. Given that Satb2 deficient neurons show an ectopic upregulation of Ctip2 and a loss of Unc5h3 expression, which is restored upon Ctip2 deletion, we conclude that Ctip2 is repressing Unc5h3. The above conclusion is further supported by the fact that Ctip2 protein can bind to putative binding sites on the Unc5h3 promoter region and repress its expression in vivo and in vitro. Moreover, the expression of DCC, another netrin-1 receptor that mediates attraction to netrin-1 signals, is upregulated in the superficial cortex of both Satb2-/- and Satb2-/-;Ctip2-/- mutants, suggesting that Satb2 is somehow repressing DCC expression. This upregulation can, at least partially, explain the lack of a full callosal restoration in the Satb2-/- both upon Unc5h3 overexpression at E12.5 (a time point when deep and superficial layers can be influenced) and after the additional deletion of Ctip2 (i.e. in the Satb2-/-;Ctip2-/-. In parallel, we have studied the effect of the deletion of Satb2 in Fezl-/- mice (Fezl is a transcription factor robustly expressed by CSMN and other subcerebral projection neurons). We have found that unlike Fezl-/- mice, that almost completely lack subcortical projections, the Satb2-/-;Fezl-/- compound mutants have partially regained some aspects of their subcerebral connectivity. However, Ctip2 expression does not seem to be fully restored in the deep layers of the compound mutants, while the ectopic Ctip2 expression in callosal UL neurons is maintained similar to what is seen in Satb2-/- mice. The above results imply that the induction of Ctip2 expression by Fezl is most likely not mediated by Satb2; at least not in the cells that express high levels of Ctip2 and project to the spinal cord (type I, layer V neurons). Additionally, the absence of Fezl from progenitor cells(i.e. in the Satb2-/-;Fezl-/-) has clearly no effect on the ectopic expression of Ctip2 in callosal neurons of Satb2 mutants.