An RNA Interference Screen Reveals Fate Determinants Implicated in Asymmetric Cell Division as Potential Regulators of Hematopoietic Stem Cell Self-Renewal

Abstract During periods of extensive regeneration of the hematopoietic system, hematopoietic stem cells (HSC) undergo largely symmetrical self-renewal divisions, necessary to rapidly replenish the stem cell pool. Under homeostasis, however, it is likely that HSC rely more on asymmetric self-renewal divisions to retain an appropriate number of HSC while still enabling sufficient production of mature blood cells. The unequal partitioning of intrinsic fate determinants underlies the process of asymmetric stem cell division in lower organisms including Drosophila and C. elegans. The tumor suppressive function of specific determinants has been demonstrated in studies where mutation of fate determinants shown to be inhibitory to the self-renewal of one of the two daughter cells generated upon Drosophila neuroblast division, drives exclusive symmetrical stem cell divisions ultimately leading to the formation of larval brain tumors. As HSCs can not yet be definitively prospectively identified, it has been difficult to determine whether a similar segregation of such cell fate determinants underlies the asymmetric/symmetric self-renewal of these cells or whether deregulation of these determinants could also generate hematopoietic malignancies by inducing constitutive symmetric self-renewal divisions. We addressed these questions through a functional genetics approach taking advantage of systematic RNA interference to interrogate the function of polarity factors and cell fate determinants representing candidate HSC self-renewal regulators. From a list of 72 of such factors identified in the literature, 32 murine homologs were selected based on their differentially high level of expression in HSC-enriched populations. For each candidate we generated 3 unique short hairpin RNA (shRNA) encoding retroviral constructs also carrying EGFP for the purposes of following transduced cells. In a primary screen equal numbers of HSC-enriched Lin-CD150+CD48− cells were infected with the library in an arrayed 96-well format yielding an average gene transfer of 60.0 ± 3.2%. The in vivo reconstituting potential was then evaluated in a CRU assay such that identical proportions of each well were transplanted in duplicate. An average of 37.6 ± 5.1% long-term donor reconstitution was demonstrated by luciferase shRNA transduced controls. Directly following infection, the EGFP+ fraction of a portion of each well was separated by FACS to facilitate qRT-PCR determination of knockdown efficiency. Immunophenotypes, cell viability and morphology of well contents cultured an additional 7 days were also assessed. The percent of EGFP− and EGFP+ donor cell contribution was determined by flow cytometric evaluation of peripheral blood samples taken every 4 weeks for a period of 16 weeks. Genes for which shRNA vectors altered late transplant EGFP levels below or above defined thresholds were considered as hits. At present we have identified 4 genes for which shRNA-mediated depletion negatively affects repopulation but does not induce indiscriminate cell death in culture and 1 gene that may act as a self-renewal inhibitor. In one example, two shRNAs directed against the candidate EB3 showed a dramatic loss of EGFP+ cells in vivo. EB3, a member of the microtubule plus-end binding protein family, has previously described roles in the search-and-capture mechanism of spindle positioning. Interestingly, EB1, a closely related family member is also critical in directing the symmetrical as opposed to asymmetrical divisions of primitive neuroepithelial cells in Drosophila. Validation of all identified hits as well as further evaluation of their function through cell cycle, cell death and homing studies is ongoing.