FOXG1 modulates LINE1 activity in the developing embryonic neocortex

Foxg1 is an ancient transcription factor mastering telencephalic development. mLINE1 is a large retrotransposon family contributing to plasticity of neuronal genome. All mLINE1's share putative Foxg1 binding elements. Moreover, it is known that mLINE1 transcription is upregulated in postnatal glutamatergic progenitors, whereas Foxg1 declines in embryonic progenitors of the same lineage. Based on that, we predicted that Foxg1 might limit mLINE1 expression (and activity). Here we tested such prediction, with particular attention to its cellular and molecular features. First, as expected, we found that mLINE1-mRNA encoded by all three retro-transposition competent subfamilies (A, Gf and Tf) was increased in Foxg1-loss-of-function (-LOF) mouse pups. Next, to model articulation of Foxg1-dependent LINE1 regulation, we developed a dedicated set of in vitro preparations, representing early-, mid- and late-phases of neuronogenesis. By profiling this set, we detected a progressive increase of LINE1-mRNA levels from neural stem cells (NSCs) up to neurons (Ns). Then, taking advantage of these preparations, we manipulated Foxg1 levels at different stages of the neuronogenic progression, according to specific neural cell type restrictions. We mapped changes in Foxg1 protein levels evoked of these manipulations, and we evaluated LINE1-mRNA levels originating from them. Integrated analysis of results pointed to selective mLINE1 repression by Foxg1 in neuronal progenitors (NPs) and Ns. To assess if Foxg1 modulation of mLINE1-mRNA takes place via direct gene trans-repression, we evaluated Foxg1 enrichment at mLINE1 loci via ChIP. In wild type cultures, Foxg1 was enriched along the whole body of mLINE1's belonging to all three subfamilies. However, this enrichment was restricted to mid-neuronogenic cultures, and not detectable in early-neuronogenic ones, ruling out Foxg1 modulation of mLINEs in NSCs. Intriguingly, such enrichment was more pronounced upon artificial Foxg1 overexpression, suggesting that Foxg1 may physiologically tune mLINE1 transcription. Furthermore, we found that Foxg1 overexpression in mid-neuronogenic cultures elicited a generalized decrease in transcription-activating marks (H3K4me3) and an increase of repressing ones (H3K9me3). These phenomena occurred at all mLINE1 subfamilies loci, suggesting that Foxg1 largely impacts on mLINE1 transcription via modifications of the corresponding epigenetic landscape. Next, we moved to Foxg1 control (if any) of cellular mLINE1-DNA content. First, we found that this content was upregulated by about 25% in Tubb3+, post-mitotic wild type neurons compared to their ancestors, in a retro-transcription-dependent way. Next, paradoxically, we also discovered that Foxg1 down-regulation elicited a co-linear reduction in mLINE1-DNA content, suggesting that Foxg1 is strictly needed to sustain physiological amplification of such DNA. This phenomenon was observed in vivo as well as in vitro. Puzzingly, Foxg1 overexpression led to variable outcomes, increased, unchanged or decreased mLINE1-DNA content, depending on different promoters driving it. Differential sensitivity of mLINE1 transcription and retro-transcription to Foxg1 levels could account for these phenomena. Finally, mechanisms underlying Foxg1 impact on mLINE1 DNA content were not yet investigated, however, we got evidence that direct Foxg1 protein/mLINE1-mRNA interaction could underlie it.