In the adult hippocampus of mammals, including humans, new neurons are continuously being produced from a resident population of stem cells. This process, known as adult neurogenesis, is controlled by a complex genetic programme. In addition, this system is remarkable among stem cell niches because of its regulation by extrinsic environmental stimuli. Considerable effort has been invested in recent years in the discovery of regulatory genes and proteins—yet relatively little has been done to assemble all this knowledge into functional molecular pathways. I believe that only by investigating the system as a whole can we identify vulnerabilities which might lead to dysfunction. An understanding of information flow through the molecular networks will potentially allow us to influence the path of stem cells through their fate space—and thus present targets for intervention in the case of disease.
My work so far has largely centered on the use of genetic reference panels to identify genes that modulate adult neurogenesis. An extensive dataset of hippocampal transcript expression in the BXD panel consisting of 99 mouse strains (Overall et al., 2009) has provided a core resource for the investigation of co-expression networks—an approach which has been both enlightening in itself as well as being the basis for a broader multiscalar research program. The idea that whole-genome scale datasets can enable us to move away from traditional ‘one-gene-at-a-time’ quantitative trait locus (QTL) mapping to analyses where genes are considered within their molecular environment has grabbed my attention and driven me to explore the methods further. The realisation that not only gene expression, but also expression of physiological phenotypes can be integrated into network models has inspired sevaral studies looking at the molecular neighbourhood of several core neurogenesis traits. I have also been investigating modifications to standard QTL mapping methods to incorporate transcript expression networks and expression QTLs (eQTLs) in a two-step approach to identify genomic loci associated with highly complex phenotypes. This work has yielded some surprising results at the cell biological level which are currently being actively pursued.
Modelling precursor cell biology in vitro
These quantitative genetic techniques have also been applied to isolated precursor cells grown in vitro as adherent cultures. From cultures derived from different mouse strains we could not only collect phenotypes such as proliferation rate, but also correlate these to transcript expression profiles. The resulting systems genetics resource enabled us to use QTL mapping to narrow down potential candidate genes based on the presence of a cis-eQTL. Using further cultures from this renewable resource and taking advantage of the ease of manipulation of the adherent culture model, we could demonstrate that the polymorphic gene Lrp6 is one component modulating strain differences in precursor cell proliferation (Kannan et al., 2016). This cell culture model has also been characterised more deeply at the molecular level which has resulted in a database of expression at different time points during differentiation. This has allowed us to identify genes associated with different phases of precursor maturation at an extremely fine temporal resolution. It is also an important starting point for the investigation of intrinsic pathways driving exit from the cell cycle and differentiation.
Environmental regulation of the stem cell niche
In addition to cell-intrinsic properties, regulation from the cells’ environment is also an important control on growth and differentiation. To better understand what components of the niche can affect precursor cell proliferation and neurogenesis, we have investigated gene expression changes associated with two different states of proliferation activity. The rate of proliferation can be easily experimentally altered by housing the mice in cages containing running wheels—the number of proliferating cells increases substantially after physical activity. I have shown that the dynamics of this phenomenon differ in the mouse strains C57BL/6 and DBA/2, the parental strains of the BXD panel, indicating that genes modulating precursor cell proliferation segregate in these strains (Overall et al., 2013). This model has be further used to search for transcripts exhibiting differential expression in a gene by environment interaction study.
Integrative systems biology
I am also extremely interested in the synthesis of diverse datasets (Overall, 2017) which includes making use of resources in the public domain (Overall et al., 2015). I believe it is important that the hard-won results of research are analysed to their fullest potential—and this includes reanalysing datasets outside of their original context. To this end, I have initiated a project for the adult neurogenesis community which collates all the literature on gene expression and function into a searchable database and ontology (MANGO; Overall et al., 2012).