Specific Expression of E2f3 in SACs and Other Subsets of CNS Neurons
The unique effect of E2f3 in disrupting the differentiation of SACs but not other retinal neurons might be due to cell-type-specific expression or cell-type-specific activity of E2f3. Determining between these two possibilities is not easy, as E2f immunostaining in mouse tissues is problematic. We did not solve this issue for E2f1 or E2f2, but used a modified protocol [50] to successfully track E2f3 expression (Figure 5). At P0, E2f3 was detected in RPCs, consistent with a putative role in normal proliferation (Figure 5A). The signal was specific as it was absent in the E2f3 KO peripheral retina (Figure 5A). As the retina differentiated and RPC division diminished, the number of E2f3+ cells also dropped, and by P8, when division is virtually over, only a subset of post-mitotic cells in the inner retina expressed E2f3 (Figure 5A). By P18, E2f3 was also detected in two tracks in the IPL (Figure 5A and 5B), reminiscent of SAC markers such as Chat and Slc18a3 (c.f. Figures 3 and 4). This cytoplasmic E2f3 staining was also specific, as it was absent in the E2f3 KO peripheral retina of α-Cre;E2f3loxP/loxP mice (Figure 5A). Indeed, double labelling with E2f3 (red) and Chat plus Slc18a3 (green) confirmed that E2f3 is present in both SAC soma and dendrites (Figure 5B). Rb protein was also detected in the inner retina (Figure 5A), and showed a similar distribution as E2f3 in SACs (Figure 5B), and was also present in mature ganglion cells and Müller cells as reported [51]. Rb staining in SAC processes was specific as it was absent in the peripheral retina of αCre;RbloxP/loxP mice (Figure 5A). These data suggest that Rb and E2f3 colocalize in SACs and that E2f3 triggers defects in SAC differentiation because it is specifically expressed in these retinal neurons.
Figure 5  E2f3 and Rb Expression in SACs
(A) Left panels: horizontal P0, P8, and P18 retinal sections of the indicated genotypes were stained for E2f3 (red) and DAPI (blue). The arrow indicates the junction between the E2f3 null peripheral and WT central P0 retina. Note the absence of E2f3 protein in the peripheral E2f3 KO RPCs at P0 and in peripheral inner retinal neurons at P18. Far right panel: P18 retinal sections of the indicated genotypes were stained for Rb (red) and DAPI (blue). Note the absence of Rb protein in the peripheral Rb KO inner retinal neurons.
(B) WT P18 retinal sections were stained for nuclei (DAPI, blue), E2f3 (red) or Rb (red), and Chat plus Slc18a3 (green). Arrows indicate double-labelled soma. Note that the IPL processes are also double-labelled.
Scale bars are 50 μm. We also found that E2f3 is present in a specific subset of mature neurons in various brain regions (data not shown). For example, in the P20 amygdala, E2f3 colocalized with the general neuronal markers Mtap2 and Mecp2 [52], but not with Calb2, which marks a subset of neurons, or with the glial marker Gfap (data not shown). Unlike in retinal SACs, E2f3 was not coexpressed in Chat+ or Slc18a3+ cholinergic neurons located in various regions of the brain and spinal cord (data not shown). In agreement, we could not detect defects in cholinergic Rb KO neurons in the developing forebrain, but other Rb KO neurons in this region showed differentiation defects that were rescued by deleting E2f3 [53]. Together, these results suggest that the common mechanism by which Rb promotes neural differentiation is through E2f3 inhibition.