Neurogenesis in the adult mouse brain

Most regions of the mammalian brain contain only differentiated neurons, which persist for life and are not renewed. However, there are two regions where stem cells reside and continue to produce new neurons throughout life: the ventricular-subventricular zone of the lateral ventricles, and the dentate gyrus (DG) of the hippocampus. Neural stem cell (NSC) numbers fall rapidly in the hippocampus of juvenile mice, but then stabilise as they transition to adulthood, ensuring lifelong hippocampal neurogenesis. This change in the rate of neurogenesis happens at different times in different mammals, but is highly evolutionarily conserved, underlining its importance. However, whether it occurs in humans is still disputed.

François Guillemot’s lab has been studying how the switch between developmental and adult patterns of neurogenesis is regulated, and how adult neurogenesis is maintained. In two papers during this quinquennium, they discovered that a tripartite population of NSCs — deeply dormant, resting and active — exists in adult mice, and that establishing and then balancing these three states is regulated by increasing proteasomal degradation of the transcription factor ASCL1 (Urbán et al, 2016; Harris et al, 2021).

In contrast to most adult stem cell lineages, which differentiate in response to damage, or to renew short-lived cells, the hippocampal NSCs directly contribute to the function of the tissue, producing new neurons that integrate into the existing circuitry of the hippocampus before they are mature. Due to their immaturity, these incoming cells are highly excitable with a high degree of synaptic plasticity, in contrast to the more mature neurons already present in the circuit. This means they have a disproportionate impact: their integration profoundly modifies the properties of the circuits, affecting the regulation of memory — specifically in the segregation of similar memories, and in memory consolidation and clearance — and mood, for example in the prevention of anxiety and depression.

All these functions require a remarkably small number of neurons. In all mammals studied, the number of new neurons added to the adult hippocampus is in the order of a few hundred a day in rodents and about 2000 neurons a day in macaques. These fairly small numbers are in contrast to the much larger number of neurons that are added to the dentate gyrus during development, when they build the structure and contribute to the formation of the circuits.

In the mouse, by two weeks after birth the pattern of stem cell activity and the cellular organisation of the hippocampal stem cell niche is broadly in place, and traditionally, this was considered to be the time when adult neurogenesis began. However, the transition is not sharp: there are still high levels of neuronal production and the dentate gyrus is still being built. Neuronal production gradually declines over the next few months in young adult mice to reach a low level by about 6 months of age, and this low level then continues throughout life. Much later, towards the end of life, there is a further decrease caused by ageing and stem cell dysfunction, but this is distinct from the early decrease, which marks the transition from building a structure to maintaining it.

Hippocampal NSCs sit atop the lineage hierarchy of neurogenesis, and therefore changes in NSC behaviour are likely to drive this transition to lower levels of neurogenesis. However, there has been some controversy in the field about the levels of self-renewal in the adult mouse hippocampus. Contradictory data suggested on one hand that NSCs went through a series of rapid asymmetric divisions and were then lost by differentiation, but puzzlingly, other research showed that activated NSCs were able to return to quiescence as well as proliferate and differentiate.

In 2016, Guillemot and colleagues shed some light on this discrepancy by showing that the hippocampal NSCs were in fact heterogeneous (Urbán et al, 2016): one fraction did indeed differentiate without self-renewing, but another fraction could return to quiescence and self-renew, and this fraction was responsible for the long term maintenance of the proliferating stem cell pool.

Guillemot’s lab has now resolved the conundrum entirely, showing that hippocampal NSCs increasingly return to quiescence as mice age (Harris et al, 2021), and that therefore, depending on the age of the mice studied, apparently contradictory results are possible.

Guillemot’s postdoc Lachlan Harris and colleagues looked at the behaviour of hippocampal NSCs over time, by measuring the fraction of proliferating stem cells either returning to quiescence or continuing to proliferate. They marked proliferating stem cells by injecting mice with the thymidine analogue 5-ethynyl- 2’-deoxyuridine (EdU), or alternatively by using tamoxifen induction of a Ki67- CreER^TD transgene to drive a tdTomato cre-reporter gene (which permanently labels cells that have come out of quiescence). Those cells that had historically proliferated were then analysed for Ki67 positivity to determine which were still in cycle.

Early in life, at the beginning of adult neurogenesis, NSCs were activated and could not return to quiescence, instead proliferating and rapidly differentiating. However, in older mice, the behaviour of the NSCs changed, and a large fraction of them returned to quiescence instead. These NSCs become progressively more stable and persisted longer in the stem cell niche as animals aged: by EdU-labelling proliferating NSCs in one-month and six-month old mice, and then examining the fate of the cells ten and 30 days later it was clear that they returned to quiescence and then persisted for far longer in the older animals.

The NSCs that proliferated and then returned to quiescence were distinguishable from fully dormant stem cells that had never left quiescence: they could reactivate much more readily and contributed more to the proliferating stem cell pool, in contrast to the dormant stem cells, which became more deeply quiescent with age. The contribution of these so-called resting NSCs increases over time, such that by six to 18 months of age, the resting cells are the major contributors to NSC proliferation, despite the fact that they comprise only 1.5%- 3% of the total NSC pool. These findings support a novel model where over time, hippocampal NSC proliferation is generated increasingly from resting NSCs that have the capacity to shuttle between active and quiescent states, delaying NSC depletion and ensuring long-term maintenance of the NSC pool.

In parallel with this work, Guillemot’s University of Heidelberg collaborators Thomas Stiehl and Anna Marciniak- Czochra developed a mathematical model of hippocampal NSC population dynamics over time, and showed that the model was a good fit to the biological data if based on just two time-dependent variables: that resting NSCs have much higher activation rates than dormant NSCs, and that dormant NSCs progress into deeper quiescence with time. This suggests that the transition from developmental to adult neurogenesis and the long-term maintenance of adult neurogenesis can be entirely explained by age-dependent changes in the behaviour of the three populations described; invoking other as-yet uncharacterised properties is unnecessary.

The transcriptional changes underlying these different states of quiescence were examined using a transgenic approach, whereby proliferating NSCs and their progeny were irreversibly marked in mice aged one, two or six to eight months old, enriched by flow sorting, then subjected to single-cell RNA-seq. The trajectory inference tool Slingshot (which can infer cell lineages using single cell sequencing data from complex communities of heterogeneous cells) was used to plot the different stem cells along a continuum from quiescence to proliferation. Resting cells, which as expected were far fewer in number than dormant or active cells, could be differentiated from dormant cells by their expression of a transgenic Ki67- driven proliferation marker, and were in an intermediate position in the trajectory relative to the dormant and proliferating populations.

The resting cells were a bona fide quiescent population, as measured by multiple transcriptional criteria, but could be distinguished from more deeply dormant cells by lower levels of expression of quiescence markers such as ApoE and Clu, and higher expression levels of genes such as the ribosome biogenesis gene Rps17. Together, these and other transcriptional data Indicated the resting cells were in a unique physiological state — poised quiescence — which had features of both quiescence and activity.

What is the molecular mechanism driving the switch from developmental to adult neurogenesis? A strong candidate was the transcription factor ASCL1, which Guillemot’s group had previously shown is required for the activation of quiescent NSCs (Andersen et al, 2014). Postdoc Noelia Urbán and coworkers (Urbán et al, 2016) demonstrated that reducing ASCL1 levels permits proliferating hippocampal NSCs to return to quiescence. They also discovered that ASCL1 is post-translationally regulated by the E3-ubiquitin ligase HUWE1, which promotes ASCL1’s degradation and is also required for a return to the resting state, and therefore the maintenance of adult neurogenesis: five-month-old HUWE1 knockout mice had an eight-fold reduction in the numbers of proliferating NSCs, despite overall NSC numbers being unaffected, meaning that adult neurogenesis had essentially stopped.

Extending this work, Lachlan Harris showed that there was a progressive but dramatic reduction in the fraction of NSCs expressing ASCL1 as young mice aged. He then used a hypomorphic allele of ASCL1 to demonstrate that reduction of ASCL1 expression by a third was sufficient to mimic the change in stem cell behaviour that would normally occur later in life in wild-type mice: in young mice expressing the ASCL1 hypomorph, active NSCs were able to return to the resting state, and there was a decrease in the rate of activation of dormant cells, resulting in a slowing down of the early depletion of the stem cell population (Harris et al, 2021).

In mice wild-type for ASCL1, the progressive decrease in ASCL1 and the expansion of the resting cell population as mice aged could be correlated with a progressive increase in HUWE1 activity over time: Huwe1 transcripts were present in higher numbers in stem cells from older mice, and HUWE1 activity increased with age. This reciprocal relationship strongly suggests that HUWE1 is the effector of the ASCL1-driven switch to adult neurogenesis. How the hippocampal stem cell niche regulates Huwe1 itself is currently under investigation in the Guillemot lab.

These results demonstrate how a series of early, progressive and coordinated changes in hippocampal NSC properties function to preserve proliferation beyond the juvenile period in mice. Whether similar mechanisms are present in humans to extend neurogenesis beyond childhood is an intriguing possibility.