Deletion of Sfrp1 in Sip1 conditional knockouts does not rescue the neocortical defects seen in these mutants

Previous work done in the lab involved an extensive study of the expression of the Wnt antagonist Sfrp1 in wildtype and Sip1 mutant cortices (Miquelajauregui et al., 2007). It was found that not only is it expressed in a pattern complementary to that of Sip1 in wildtype cortices at embryonic to early postnatal stages, but that it is also dramatically upregulated all over the cortex at E17.5 in the absence of Sip1 (Miquelajauregui et al., 2007). Since this phenotype coincided with the increased proliferation of astrocytic precursors in Sip1 mutants, we hypothesised that Sfrp1, a secreted molecule, signals back to VZ/SVZ progenitors and triggers either proliferation of existing glial precursors or induces a cell fate switch towards an astrocytic lineage. We performed several experiments to confirm this hypothesis. First of all, we bred both Sip1 conditional mutants with Sfrp1 total knockouts to obtain double mutants that lacked both Sip1 as well as a functional copy of Sfrp1 allele. This was followed by analysis of deep layer populations, and postnatal astrocytogenesis and proliferation. The rationale behind this approach was that in Sip1 mutants, in the absence of Sfrp1, no upregulation would be possible and hence all the phenotypic effects mediated by it would be rescued. However, we detected no such rescue.

We began our analysis by confirming deletion of Sfrp1 in these mutants. In the absence of a good antibody against Sfrp1, we performed in situ hybridisation to confirm deletion of Sfrp1 mRNA in the double mutants. As shown, Sfrp1 transcripts seem to have reduced in comparison to wildtype and Sip1fl/fl-Emx^Cre (Fig20a-c). We then immunostained P8

Fig20. Ablation of Sfrp1 in Sip1 conditional mutants does not rescue the mutant phenotype. A probe targeted against the deleted exon of Sfrp1 gene was used to detect the level of expression in Wt, Sip1fl/fl-Emx^Cre , and Sip1fl/fl-Emx^Cre Sfrp1-/- ,by in situ hybridization at P8 (a-c). The reduction in Sfrp1 transcript levels can be seen in the double mutant. Immunostaining for GFAP and Ki67 at P8 clearly shows no rescue of Sip1 mutant phenotype in the double mutants (d-f). The reduction of Ctip2+ cells at E16.5 in Sip1fl/fl-Emx^Cre remains in the double mutant as well (g-i). Overexpression of Sfrp1 (alongwith GFP reporter) in the cingulate cortex by in utero electroporation at E15.5, followed by analysis at P0 shows no colocalisation between GFP and GFAP (j), and hence, possibly no influence of Sfrp1 on cortical gliogenesis.

cortices for GFAP and Ki67. As expected, there was an increased expression of both in Sip1fl/fl-Emx^Cre especially in the region of the cingulate cortex. This enhanced expression was observed in the double mutants as well, implying clearly that deletion of Sfrp1 could not rescue the gliogenic phenotype of Sip1 mutants (Fig20d-f). Although the upregulation of Sfrp1 in the neocortex at early embryonic stages was not obvious, we nevertheless checked for expression of deep layer markers in the double mutants at E16.5. Again, the reduction in deep layer neuronal populations seemed as strong as in Sip1 mutants

(Fig20g-i). Finally, to test possible cell autonomous as well as non- cell autonomous effects of overexpressing Sfrp1 (since Sfrp1 is a secreted molecule) in the cingulate cortex, especially in context of gliogenesis, we adopted an in utero approach. Here, in collaboration with Dr. Federico Calegari (Max Planck Institute of Cell Biology and Genetics, Dresden), Sfrp1 overexpressing constructs were electroporated into E15.5 forebrain, which were then analysed at P0 for GFAP expression. The constructs encode Sfrp1 under the CMV promoter, and also contain GFP coding sequence under the SV40 promoter. We found GFP expression close to the ventricle, in migrating cells as well as some cells in the cortical plate (Fig20j). However, we found no differences in GFAP expression between the electroporated and control hemispheres, and also no colocalisation between GFP and GFAP (Fig20j). This observation suggests that Sfrp1 might not influence cortical astrocytogenesis, neither within the same cell nor on neighbouring cells. However, it is also likely that P0 is too early a stage to check GFAP expression; an alternative would be to use an earlier expressed marker like Olig2.

DISCUSSION

In an attempt to explore the puzzling and complex dynamics of cortex development, we set out to investigate the functions of a transcriptional repressor expressed at very high levels in the developing neocortex almost exclusively in its differentiating field. From a subtractive hybridisation based screening, Sip1 was identified as a gene differentially expressed between early and late stages of mouse corticogenesis, and hence selected for further studies. Since Sip1 deficient mice have been shown to die at E9.5 with major anomalies in neurulation and neural crest migration, we opted to generate and analyse cortex- specific conditional knockouts. We have shown that in the neocortex, strong expression of Sip1 is confined predominantly to postmitotic cells, starting at E12.5 until early postnatal stages. Most, though not all, upper and deep layer neurons express Sip1. At P8, Sip1 transcripts were detected primarily in the postmitotic cingulate and medial cortex, while Sip1 protein could be detected in very few cells scattered throughout the neocortex.

We have also shown that E17.5 onwards, the expression of Sip1 in the cingulate cortex is stronger in layers 5-6 than in layers 2-4.

4.1 Sip1 controls sequential cell fate switch in cortical progenitors during the course