In light of the above findings we propose that the
In light of the above findings, we propose that the β-gal+/NeuN+ cells and β-gal+/TH+ cells described by Shan et al. (2006) do not reflect neurogenesis and DA neurogenesis from Nestin+ NPCs, but rather expression of Nestin by mature midbrain neurons. Indeed, given Shan et al.\'s (2006) experimental design, their observations are possibly more consistent with this. They used transgenic mice (pNes-LacZ) that produce β-gal only whilst Nestin is being expressed (as opposed to our mice where Nestin expression triggers permanent expression of β-gal or eYFP). Thus they could not perform lineage tracing studies, and their β-gal+/NeuN+ and β-gal+/TH+ cells could only reflect neurogenesis from Nestin+ NPCs if the transition from Nestin expression to NeuN and TH expression (i.e. neuronal differentiation) occurs faster than degradation of the β-gal signal after Nestin expression ceases. To our knowledge the rate of β-gal degradation is not known but it is a foreign protein and we always observe it in what appear to be lysosomes (i.e. sites of protein degradation) in midbrain cells (e.g. bright spots in Fig. 4k). Presumably therefore it is degraded relatively quickly, leaving not much time for Shan et al.\'s (2006) β-gal+/NeuN+ and β-gal+/TH+ cells to arise via neurogenesis, particularly in adult midbrain where neurogenesis is generally considered to occur indolently or not at all. Apart from the β-gal+ or eYFP+ neurons present at early time-points following tamoxifen, our findings suggest other β-gal+ or eYFP+ midbrain cells are indeed derived from Nestin+ NPCs and might reflect constitutive adult midbrain neurogenesis. In this context, our findings provide the following insights. First, there are many such cells lining the midbrain Aq, particularly the dorsal and ventral poles. These cells have a simple columnar morphology with multiple cilia extending into the ventricular space. In these respects they are typical ependymal cells, a type of glial cell that is important for production, circulation, and exchange of cerebrospinal between the ventricles and the jak inhibitors parenchyma. In addition, these ependymal cells might be activated in response to injury or withdrawal of canonical Notch signaling to produce new neurons or glia, as they do in the LV (Carlén et al., 2009). Note, however that LV ependymal cells appear to be a limited resource for repair because they do not self-renew at rates sufficient to maintain their own population; i.e. they are not true stem cells (Carlén et al., 2009). This is consistent with the fact that very few of our β-gal+ or eYFP+ ependymal cells were GFAP immunoreactive (Hermann et al., 2009), and none were Sox2 immunoreactive, both of which are considered markers of true stem cells. These findings are broadly consistent with those of Hermann et al. (2009), who described a preservation of Nestin+ cells along the entire rostral-to-caudal extent of the adult mouse paraventricular region, despite a gradual reduction of stem (Nestin+/GFAP+) cells and neuroblasts (Nestin+/PSA-NCAM), and a complete absence of transit amplifying cells caudal to LV. Hermann et al. (2009) also noted Nestin+ Aq cells have a distinct midbrain phenotype, and were therefore unlikely to have migrated in from the forebrain (LV and 3V). Also, despite being non-proliferative in vivo (evidenced by the absence of BrdU incorporation), they could be readily proliferated when provided EGF and FGF2 in vitro (Hermann et al., 2009). There are also many β-gal+ or eYFP+ cells in the ventral midline region below Aq [see also Shan et al. (2006)]. These cells had a migrating morphology suggestive of migration from the Aq ventrally. Developmental DA neurogenesis follows a similar trajectory (Deierborg et al., 2008), which initially led us to think it might be recapitulated in the adult. However, β-gal+ or eYFP+ cells in the ventral midline were the same distance away from Aq along the dorso-ventral axis with time following tamoxifen, indicating they are not moving ventrally (or dorsally). A similar lack of temporal dispersion of proliferating Aq cells labelled with retrovirus was noted by Yoshimi et al. (2005a,b). It remains possible they are migrating rostally or caudally, however they did not express classical markers of new cell migration PSA-NCAM or DCX (see also (Aponso et al., 2008; Hermann et al., 2009); but see (Yoshimi et al., 2005a,b; Peng et al., 2008). β-gal+ or eYFP+ ventral midline cells were not Sox2, NeuN or GFAP immunoreactive either, suggesting they are some kind of intermediate between stem- or precursor cells and post-mitotic differentiated cells.