INSIGHTS | PERSPECTIVES
sion, but the extent to which this also occurs
in mammals is unknown.
Pilz et al. imaged adult neurogenesis in
the mammalian DG by removing the overlying neocortex to get visual access to RGCs
sparsely labeled by Achaete-scute homolog 1
(Ascl1)–driven expression of green fluorescent protein. Dividing RGCs and all their
progeny were followed by daily imaging for
up to several months. The RGCs underwent
typically two to three cell divisions with initial symmetric or asymmetric self-renewing
divisions (producing two or one RGCs, respectively), followed by a final self-depleting
symmetric division generating two nonradial
daughter cells (see the figure). This is consistent with previous clonal analysis in mouse
brain sections (9, 10), whereas others have
proposed long-term RGC self-renewal (11).
Importantly, the imaging analysis followed
Ascl1-expressing RGCs, whereas other subtypes of RGCs that do not express Ascl1 may
be more prone to long-term self-renewal.
The biggest surprise came when observing
the direct progeny of the RGCs that were ex-
pected to amplify the population by dividing
symmetrically. However, Pilz et al. demon-
strated that they also divide asymmetrically,
producing a neuronal daughter and an ap-
parently self-renewing mother cell. As they
divided much more than RGCs and with
the ability to stochastically switch between
asymmetric (with one differentiating daugh-
ter) and symmetric (self-renewing or ampli-
fying) division, their behavior may qualify
them as NSCs. Thus, TAPs are becoming
more similar to stem cells, as in other organs
(12). These findings also have repercussions
for the part of the definition that asserts that
stem cells are the origin of a lineage. Because
cells not at the top of the lineage tree, like
TAPs, meet some criteria of stem cells, this
part of the definition may also need to be
revisited, including reconsidering the unidi-
rectional aspect of lineage trees.
Another interesting observation was an
early wave of cell death during adult neurogenesis, apparently regulated by intrinsic
factors. This is deduced from the observation
that entire branches of a lineage are prone to
die and that cell death occurs independent
of location. This fascinating finding provokes
questions about what these intrinsic factors
These data also allow for further comparisons between neurogenesis in development and adulthood. No in vivo imaging
is available for neurogenesis in mammalian embryos, but in the developing zebrafish, amplification by TAPs is rare, and if
TAPs exist, they divide only once (13, 14).
However, the sequence of symmetric self-renewing divisions of RGCs followed by
asymmetric divisions generating a neuron
and an RGC is also prevailing in developmental neurogenesis. Interestingly, while
RGCs in adult neurogenesis obey this rule,
TAPs do not.
However, especially in the developing
mammalian brain, there are some regions
with a large population of TAPs, such as in
the telencephalon. In these regions, TAPs
also exhibit an intriguing behavior, with
increasingly faster cell cycles in each subsequent round of cell division, whereas RGCs
divide much slower (15). In the adult DG,
RGCs divide as fast as TAPs, which do not
change their cell cycle duration with subsequent divisions. Moreover, adult DG TAPs
switch between symmetric and asymmetric
division—a behavior unprecedented in the
nervous system so far. Thus, the behavior of
TAPs is another difference between embryonic and adult neurogenesis.
TAPs are key not only for regulating the
output of adult neurogenesis but also in development and evolution when they become
more frequent. Indeed, there is little lineage
amplification in adult zebrafish forebrain
neurogenesis, allowing the direct conversion
of RGCs to neurons that was not observed in
the adult mammalian DG. Thus, the mechanisms regulating the emergence of indirect
neurogenesis with an ever-enlarging and
stem cell–like TAP population are of prime
importance for brain expansion and for neuronal output in adult neurogenesis. Given
that such TAPs can readily replenish the stem
cells at the origin of the lineage tree in some
organs (12), such as the intestine, it is important to watch them in an injury paradigm
and be prepared for further surprises. j
1. P. Gao et al., Cell 159, 775 (2014).
2. M. Götz et al., Glia 63, 1452 (2015).
3. G.-A.Pilz et al., Science 359, 658(2018).
4. S.Falk,M.Götz, Curr. Opin. Neurobiol. 47,188(2017).
5. N. Urbán, F. Guillemot, Front. Cell Neurosci. 8, 396 (2014).
6. J. T. Gonçalves, S. T. Schafer, F. H. Gage,Cell 167, 897
7. S. Falkner et al., Nature 539, 248 (2016).
8. J. S. Barbosa et al., Science 348, 789 (2015).
9. F. Calzolari et al., Nat. Neurosci. 18, 490 (2015).
10. A.Sierra et al., Cell Stem Cell 16,488(2015).
11. M. A. Bonaguidi et al., Cell 145, 1142 (2011).
12. J. Beumer, H. Clevers, Development 143, 3639 (2016).
13. R. Mc Intosh et al ., Open Biol. 7, 160312 (2017).
14. P. Engerer etal.,EMBOJ. 36, 1134 (2017).
15. G.-A.Pilz et al., Nat. Comm. 4,2125(2013).
Pilz et al . observed
repeated divisions of TAPs,
as well as switching between
symmetric and asymmetric cell
divisions. TAPs, rather than RGCs, also
persisted long-term, prompting revision
of the previous concept of how adult
Analysis of Dxed brain samples
suggested that adult neurogenesis in
the DG involved little self-renewal of TAPs.
Transit-amplifying progenitors ( TAPs)
Radial glial cells (RGCs)
Constructing neural cell lineages
Adult neurogenesis is restricted to a few niches in the mammalian brain, including the DG. RGCs are NSCs,
the origin of neural lineage trees that are capable of asymmetric or symmetric divisions to self-renew. However,
Pilz et al. reveal that RGC progeny, nonradial cells or TAPs, can also undergo asymmetric or symmetric
divisions to self-renew or amplify the cell population. Thus, these TAPs have some stem cell characteristics.