A Functional Screen Identifies Polarity Genes Implicated in HSC Fate

Abstract

The molecular details governing self-renewal in tissue stem cells of the invertebrate systems of Drosophila Melanogaster and C. Elegans have been instructive for equivalent tissues in vertebrates. In the aforementioned invertebrates, an integral group of genes involved in cell polarity seem able to intrinsically act as or affect cell fate determinants (CFDs) during the process of stem cell asymmetric cell division (ACD). On this premise, we focused on potential polarity genes that may act as CFD during HSC self-renewal. 72 CFD candidates were chosen from a literature review that addressed mechanisms of ACD. Gene expression profiles were performed on both highly purified Long Term Repopulating-HSC populations and primary Leukemia Stem Cells. A significant number of these candidates were highly and differentially expressed. The highest ranking 60% of candidates (42 of the initial 72 genes) was then chosen for a functional in vitro to in vivo over-expression screen. The underlying theory of this screen is based on the ability of Hoxb4-induced HSCs, as compared to control vector-induced HSCs, to expand during a short in vitro culture period, together with their ability to provide significant long-term reconstitution upon transplantation after this in vitro expansion. Therefore, a positive candidate would be one that has a Hoxb4-like expansion effect on HSCs. In brief, using a 96 well plate format, 1500 CD150+48-Lin-Ly5.1+ donor derived HSCs were infected independently with each candidate, together with negative (vector alone) and positive (Hoxb4 and Nup98-Hoxa10 fusion) controls, for a total of 12 days and equal proportions of HSCs were transplanted after 5 and 12 days of in vitro culture into recipient Ly5.2+ mice. The read out measurement was donor Ly5.1+ peripheral blood reconstitution performed at monthly intervals for 5 months. At day 5 transplantations, 12 of the 42 genes had donor reconstitution above the empty vector control at 16 weeks. Of these 12 genes, only 4 retained positive long-term transplant donor reconstitution after the extra week of infection to 12 days. These 4 genes were: Ap2a2, Gpsm2, Tmod1 and Kif3a. Of these, the first 2 genes are robust candidates, having been replicated in 4 independent experiments. Interestingly, both these CFD candidates, Ap2a2 (as part of the endocytic machinery that interacts with membrane receptors) and Gpsm2 (as a G-protein signaling modulator that also influences mitotic spindle orientation) potentially provide mechanisms that allow the HSC to communicate with the niche. Ap2a2 induced HSCs in particular are able to reconstitute to levels beyond and equivalent to Hoxb4 and Nup98-HoxA10-induction, respectively. Oligoclonality (ruling out insertional mutagenesis) and multipotency from donor-derived Ly5-1+ HSCs in recipients at 20 plus weeks post-transplantation has also been performed. Endogenous Ap2a2 is localized predominantly asymmetrically in purified LTR-HSCs, as opposed to a predominant symmetrical distribution in E14 fetal liver HSCs. Initial live cell microscopy of LTR-HSCs infected with Ap2a2 fluorescent fusion proteins confirms the asymmetrical distribution, and further mechanistic insights should follow with prolonged video microscopy.