Stem cells are progenitors that maintain stemness (self-renewal) while generating diverse differentiated cell types within the tissue where they reside. Thus, the decision by stem cells to self-renew or to differentiate has profound consequence in development, homeostasis, and regeneration. Recently, the role of cancer stem cells in several types of malignancies has been demonstrated; many parallels exist between normal and cancer stem cells specifically their ability to self-renew extensively and to generate a wide variety of differentiated cell types. Thus, elucidation of the molecular mechanisms leading to self-renewal or to differentiation is beneficial for both stem and cancer biology. I have been using Drosophila neural stem cells (neuroblasts) as a model to study regulation of self-renewal vs. differentiation.
A fly neuroblast undergoes asymmetric cell division repeatedly to self-renew a neuroblast and to generate a differentiating ganglion mother cell (GMC), which divides once to give rise to two terminally differentiated progeny. Thus, asymmetric cell division allows maintenance of the overall neuroblast pool while generating a large number of diverse progeny. Intensive studies in the past 10 years have led to the discovery that several aspects of neuroblast asymmetric cell division including difference in the size of progeny cells, location of mitotic spindle relative to the cell cortex, difference in the size of centrosomes, and the axis of cell division require correct establishment of cell polarity. However, how progeny fate asymmetry is generated is largely undefined.
Post-embryonic brain neuroblasts provide an ideal system for my study because of the invariable population size of neuroblasts (100 per brain hemisphere), a highly reproducible pattern of proliferation during development, and wealth of molecular markers for unambiguous identification of cell fates. Mutations that result in failure to differentiate in which a neuroblast makes two neuroblasts are predicted to lead to >100 neuroblasts per hemisphere, whereas mutations that result in failure to self-renew in which a neuroblast makes two GMCs are predicted to lead to <100 neuroblasts per hemisphere. Since activation of my fellowship, I have achieved the following three goals. (1) I tested the proposed hypothesis by analyzing mutations at a tumor suppressor l(2)giant larvae (lgl), a G-protein regulator partner of inscuteable (pins) and a kinase atypical Protein Kinase C (aPKC), and demonstrated that these genes are essential for regulating neuroblast self-renewal (Rolls et al., Journal of Cell Biology, 2003; Lee et al., Nature, 2006). (2) By using the study of lgl, pins, and aPKC as the basis, six new mutations defective in neuroblast self-renewal were identified via genetic screening. (3) I completed characterization of one of these mutations brain tumor (brat) and demonstrated that Brat is required for GMC differentiation (Lee et al., Developmental Cell, 2006).
Atypical Protein Kinase C
Cell polarity proteins have been demonstrated to regulate several aspects of neuroblast cell division including cell size asymmetry, spindle displacement, spindle geometry, and spindle alignment. However, the mechanistic link between cell polarity proteins and fate determination is elusive. I first analyzed mutations at lgl and pins for neuroblast self-renewal defective phenotypes. Qualitative analysis using neuroblast-specific antibody markers revealed an ectopic increase in neuroblasts in lgl mutant brains and an ectopic decrease in neuroblasts in pins mutant brains compared to similarly staged wild type control. Subsequent quantification of the total number of proliferating neuroblasts confirmed that lgl mutants consistently exhibited an increase in brain neuroblasts at all stages examined whereas pins mutants initially possessed a similar number of brain neuroblasts as wild type but subsequently exhibited a progressive decrease in brain neuroblasts. This observation raised the possibility that lgl and pins mutant brain phenotypes might be due to defects in neuroblast self-renewal. To test this hypothesis, I induced positively marked genetic clones in single neuroblasts to trace the fates of all of the progeny. Single neuroblast clones in wild type brains always possessed one neuroblast and many differentiating GMCs. In contrast, single neuroblast clones in lgl mutant brains contained an average of 2.3 neuroblasts per clone with as many as six neuroblasts in one clone. Single neuroblast clones in pins mutant brains contained an average of 0.27 neuroblasts per clone; 72.6% of single neuroblast clones did not possess any identifiable neuroblasts and the rest had 1 neuroblast per clone. Taken together, I concluded that Lgl represses neuroblast self-renewal whereas Pins promotes neuroblast self-renewal.
In order to understand how Lgl and Pins might regulate neuroblast self-renewal, I analyzed the sub-cellular localization of all available polarity markers in lgl and pins single and double mutants. aPKC consistently became ectopically localized to the cell cortex in lgl and lgl, pins double mutants where ectopic neuroblast self-renewal was observed raising the possibility that ectopic cortical aPKC might mediate ectopic neuroblast self-renewal. In support of this hypothesis, ectopic neuroblast self-renewal in the lgl mutant background can be suppressed by reduction in aPKC in a gene dosage-dependent manner. Furthermore, over-expression of constitutively membrane targeted wild type aPKC but not kinase-dead aPKC was sufficient to induce a dramatic increase in neuroblasts. Combined, Lgl and Pins function to restrict aPKC kinase activity to the apical cortex of mitotic neuroblasts where the future neuroblasts are derived from, and cortical aPKC is a potent inducer of ectopic neuroblast self-renewal.
Currently, we are taking multiple approaches to identify the mechanism by which aPKC promote neuroblast self-renewal.
Brain Tumor and Prospero
One of the newly identified mutations from my genetic screen was a genetic allele of the tumor suppressor gene brain tumor (brat). brat mutant larvae exhibit dramatic overgrowth of the CNS leading to formation of malignant brain tumor; however, the cellular basis that contributes to brain tumor formation was never investigated. During the last phase of my post-doctoral study, I focused my effort on functional characterization of brat. Quantification of brain neuroblasts indicated that there was a continuous expansion of neuroblasts as brat mutant larvae grew, similar to lgl mutants. Unlike lgl mutants, however, a large number of small cells, presumptive GMCs and immature neurons adjacent to large neuroblasts, failed to express differentiation markers but instead maintained expression of neuroblast markers suggesting failure to commit differentiation. This observation was subsequently supported by BrdU pulse-chase labeling and confirmed by generation of mosaic mutant clones. In wild type animals, a 4-hr pulse of BrdU labeling was sufficient to positively label all large neuroblasts and 2-3 GMCs adjacent to neuroblasts that expressed differentiation markers but lacked neuroblast markers. Following a subsequent 24-hr BrdU-free chase, BrdU labeling was detected exclusively in immature neurons that maintained expression of only differentiation markers but never neuroblast markers. In brat mutant brains, a 4-hr pulse of BrdU labeling was sufficient to positively label all large neuroblasts and 2-3 smaller cells, presumptive GMCs, adjacent to neuroblasts, but these GMCs failed to express differentiation markers but maintained neuroblast marker expression. Following a 24-hr BrdU-free chase, BrdU labeling was detected in some neuroblasts and mostly in small presumptive GMCs and immature neurons, and there were fewer BrdU + cells overall. Virtually all of these BrdU + GMCs and immature neurons continued to lack expression of differentiation markers but maintain expression of neuroblast markers. With an additional 48 hours of BrdU-free chase, there was a further decline in BrdU + cells, and the majority of the BrdU + cells expressed only differentiation markers. To convincingly demonstrate that the brat mutant phenotypes were indeed due to increased self-renewal and decreased differentiation, I generated brat mutant mosaic clones in single neuroblasts. Wild type mosaic clones in single neuroblasts always possess one neuroblast and many small cells including GMCs and immature neurons that expressed exclusively differentiation markers. brat homozygous mutant mosaic clones in single neuroblasts consistently possessed multiple large neuroblasts and numerous smaller cells that were presumptive GMCs and immature neurons that lacked expression of differentiation markers but maintained expression of neuroblast markers. Taken together, Brat is required to suppress expression of neuroblast markers and to trigger expression of differentiation markers in GMCs.
To uncover the molecular link between Brat and regulation of self-renewal and differentiation, I analyzed the localization pattern of all available polarity markers in the brat mutant background. Among the apical markers tested, aPKC was found to ectopically localize to cell cortex; however, brat, aPKC double mutant brains still possessed increased neuroblasts indicating that there must be additional genes responsible for ectopic self-renewal in brat mutants. Although Miranda was shown to localize properly in brat mutant neuroblasts, Prospero (Pros), a homeodomain transcription factor and a cargo of Miranda, failed to localize to the basal cortex in mitotic neuroblasts and failed to asymmetrically segregate into GMC following completion of cytokinesis. Pros was previously shown to regulate cell fate determination and exit from cell cycle raising the possibility that defects in self-renewal and differentiation in brat mutants might be due to loss of Pros. In support of this hypothesis, virtually all cells within pros mutant mosaic clones maintained expression of neuroblast markers but lacked expression of differentiation markers.
Presently, we are taking complementary approaches to try to identify the downstream mechanism of brat and Pros.
Characterization of Other Self-Renewal Mutants
Functional study of lgl and pins provided the experimental support toward using the brain neuroblast proliferation pattern as an essay to identify genetic mutations leading to defects in self-renewal or differentiation. Using this study as a basis, a number of novel mutations defective in self-renewal or differentiation were identified through genetic screens. Currently, we are conducting functional studies of these additional self-renewal defective mutants.
My long-term goal is identify many signaling pathways expressed in both insect and vertebrate neural stem cells, and contribute significantly to our understanding of neural stem cells in birth defects (development), regenerative medicine (stem cell biology) or cancer biology.