Enhanced postnatal astrocytogenesis and premature specification of glial progenitors in the neocortex

Since in both Sip1fl/fl-Emx^Cre as well as Sip1fl/fl-Nex^Cre neurogenesis seems to end by E15.5, we suspected cortical progenitors at and after this stage to give rise to astrocytes, the next cell type generated in the cortex during development. Hence, we traced the fate of cortical progenitors of E15.5-E17.5 through BrdU pulse-chase experiments. We found that many cells born at E16.5 and E17.5 settled in upper layers of the cortical plate and expressed corresponding neuronal markers in wildtype cortices, whereas in Sip1 mutants, most of them ended up close to the ventricle in the region of the cingulate cortex. These cells did not express any upper layer neuronal marker. Interestingly, when we checked the expression of glial markers in the cortex at early postnatal stages, we detected atleast a five fold increase in GFAP expression in precisely the same region in Sip1 mutants (Fig9a,b,i) GFAP is considered a marker of terminal astrocyte differentiation (Dahl and Bignami, 1973). This suggested that progenitors at and after E15.5 were indeed getting committed towards an astrocytic fate, implying premature gliogenesis. We evaluated this further by

Fig9. Enhanced gliogenesis accompanied by premature specification of glial precursors in Sip1 depleted cortex. At early postnatal stages, immunostaining for astrocytic marker GFAP shows increased expression in both conditional mutants, especially in the cingulate cortex (a,b). This effect was quantified in the cingulate cortex where 5-6 times higher GFAP expression was seen in Sip1fl/fl-Emx^Cre in comparison to

the wildtype at P2 (i). Student's t-test was performed using 3 Wt and 3 mutant brains from different litters.

*p=0.002. BrdU pulsed cells from E16.5 show increased coexpression of GFAP and Olig2 in Sip1fl/fl-Nex^Cre (d,f, arrowheads) when compared to Wt (c,e). This effect has been quantified as the percentage of Olig2+

cells at P2 that are specified at E16.5 in Wt vs Sip1fl/fl-Nex^Cre (j). Student's t-test was performed using 3 Wt and 3 mutant brains. *p=0.02. Also, more cells born at E17.5 not only localize to the SVZ/IZ where most astrocytes are situated, but also express GFAP (g,h, arrowheads). The excessive astrocytes produced in Sip1 mutants are mature and not dividing , as shown by a double immunostaining for Ki67 and s100ß (k,l).

Similar to GFAP, there is an increase in the number of Olig2+ cells as well, in Sip1fl/fl-Nex^Cre (m,n).

Arrowheads indicate cells born presumably within the neocortex and not migrating from the ganglionic eminences. Error bars indicate S.E.M.

costaining BrdU labeled cells from E17.5 with GFAP at P2/P4. As expected, not only did we observe more cells born at E17.5 but also more of them expressed GFAP (Fig9g,h). We quantified this effect by double immunostaining for BrdU and Olig2. Olig2 is a bHLH transcription factor, which is not only essential for the development of motoneurons and oligodendrocytes in the spinal cord, but also plays a crucial role in deciding neuronal versus glial fate in cortical SVZ. It is known to be expressed transiently in immature astrocytes and VZ progenitors, and was shown recently to be important for white matter astrocyte development (Cai et al., 2007; Marshall et al., 2005). Our own observations in cortical slices cultured in vitro show a substantial amount of GFAP+ astrocytes coexpressing Olig2. Expectedly, we also detected increased numbers of Olig2+ cells at postnatal stages in Sip1fl/fl-Nex^Cre (Fig9m,n). From the pulse- chase experiments, it was interesting to note that in wildtype mice, the commitment of cortical progenitors towards an astrocytic lineage already begins at E15.5. Very few BrdU labeled cells of E15.5 and E16.5, however, can be seen to coexpress GFAP or Olig2 at P2 (Fig9c,e). In both Sip1fl/fl-Emx^Cre as well as Sip1fl/fl-Nex^Cre, however, a significantly higher percentage of Olig2+

glial precursors and GFAP+ astrocytes are born at E16.5 (Fig9d,f). This effect has been quantified in Fig9j. The extent of colocalisation between BrdU and GFAP is difficult to quantify due to the pattern of staining of BrdU and GFAP (Fig9e-h). Hence, we preferred to analyse BrdU/Olig2 stainings for the purpose of quantification. We also compared the percentage of cells born at E17.5 that ended up as upper layer neurons in the postnatal cortex in wildtype and Sip1fl/fl-Emx^Cre and found that this value seems to decrease in the latter (Fig8e,f). We could detect enhanced astrocytogenesis in the postnatal cingulate cortex until P17 in Sip1fl/fl-Emx^Cre .We next asked whether these excessive GFAP+ cells

Fig10. Increased proliferation in the cingulate cortex continues to early postnatal stages in the absence of Sip1. Increased expression of Ki67 close to the ventricle can be seen in both conditional knockouts at P0 (e,f), P2(a,b), until P8 (g,h). At P14, this effect is no longer detected (I,j). At P2 in Sip1fl/fl-Nex^Cre, while some Ki67+ cells express Olig2 and/or GFAP, many don’t (c,d,k,). At P8, again, both Ki67+GFAP+ as well as KI67+GFAP- cells can be seen around the ventricle in Sip1fl/fl-Emx^Cre (l). Arrowheads indicate cells expressing Ki67 and GFAP.

were dividing or differentiated astrocytes. To address this question, we performed two sets of experiments- we did double immunostainings against, 1. The astrocytic marker s100ß and the proliferation marker Ki67 (Fig9k,l), and, 2. Ki67 and GFAP (Fig10k,l). While in case of the former, colocalisation was easy to assess because of the nuclear pattern of staining of both markers, in case of the latter, we imaged z-stacks followed by analysis as maximum projections (2D and 3D). We believe that while the number of dividing astrocytes is not different between wildtype and Sip1 mutants, there are more differentiating astrocytes in the latter. It is important to note here that the appearance of GFAP astrocytes in both wildtype and Sip1 mutant cortices follows the same time course.

The first astrocytes in the cortex are not detected before P2 in either the wildtype or the mutant. In other words, though there seems to be a precocious specification of astrocytic progenitors in the mutants, their differentiation into mature astrocytes is not shifted to earlier stages of development.