Drosophila Stem Cell Debate Upended: Prospero Keeps Neuron Lineages on Track

For nearly two decades, textbooks and review articles have told a tidy story about the fruit fly nervous system: when the chromatin-binding protein Prospero is lost, immature neuron precursors supposedly revert to stem cells and proliferate uncontrollably. A new study published in EMBO Reports dismantles that narrative with a series of careful lineage-tracking and cell-biology experiments in Drosophila embryos. The work, led by Kalpana Makhijani, Shreekant Verma, Jordan Mar and Krishna Moorthi Bhat, shows instead that progeny of neuroblasts in prospero mutants never dedifferentiate. They become cells of scrambled identity that express neuroblast, ganglion mother cell and neuronal genes simultaneously, and many of them even extend axons, but they either fail to divide or divide just once or twice extra, in a lineage-dependent manner.

The stakes of this question are considerable. In the Drosophila ventral nerve cord, roughly 30 neuroblasts per hemisegment divide asymmetrically in five sequential waves to self-renew while producing ganglion mother cells, each of which then divides terminally into two distinct neurons. The result is a nerve cord of about 320 neurons and 30 glial cells per hemisegment. Prospero is uniquely positioned to influence this program: in neuroblasts it sits in the cytoplasm, asymmetrically localized, so that it is inherited by the ganglion mother cell. There it becomes phosphorylated, enters the nucleus and binds chromatin along the entire lengths of chromosomes, where it is thought to repress neuroblast-specific genes and orchestrate the switch to neuronal fate.

Earlier studies had proposed that when Prospero is absent, ganglion mother cells ectopically express stem cell factors and dedifferentiate into neuroblast-like stem cells, dividing repeatedly. A corollary idea held that any excess cells were removed by continual, progressive cell death, which would explain why the mutant nerve cord never becomes visibly hyperplastic. The new study tested both predictions directly, and neither survived. Using the pan-neuronal marker Elav, the researchers counted neurons in 15-hour-old embryos and found 240 to 317 Elav-positive cells per hemisegment in mutants versus 216 to 306 in controls, a marginal, statistically insignificant difference. Western blot analysis of Elav protein across developmental stages likewise showed no significant changes.

Crucially, the mutant nerve cord was not just full of cells; it was full of differentiated neurons. Staining for the axonal markers BP102 and Fasciclin II revealed that commissures and connectives form in prospero mutants, albeit with severe axon guidance defects, and axons abnormally crossing the midline. The presence of axons, the authors note, is an indisputable hallmark of neuronal differentiation. If ganglion mother cells were en masse reverting to bona fide neuroblast stem cells, escaper neurons would be rare and the nerve cord hyperplastic. Neither was observed.

The apparent stem cell signature in mutants, ectopic expression of the neuroblast markers Deadpan, Worniu and Miranda in deeper cell layers, turns out to be gene de-repression rather than fate reversion. The dpn and mira promoters carry Prospero-binding sites, so without Prospero these genes are simply switched on in progeny cells. Analysis of the well-mapped GMC4-2a to RP2/sib lineage showed that ectopic Deadpan expression was randomized and did not predict division behavior: some ganglion mother cells divided normally, some failed to divide and became a neuron directly, and no extra divisions occurred in this lineage across 200 hemisegments examined. In the Eve-positive EL neuronal cluster, the normal number of 8 to 10 neurons formed per hemisegment, with 2 to 3 expressing only Eve and the rest co-expressing Deadpan, all retaining the neuronal marker Elav. Numb, whose promoter lacks Prospero-binding sites, showed no increase at all, exactly what de-repression rather than dedifferentiation would predict.

Row 5 lineages produced slightly fewer GFP-positive cells in mutants, not more, while the fraction of progeny ectopically expressing Deadpan rose significantly. Wg staining confirmed no extra Wg-positive neuroblasts appeared at any stage. Rows 6 and 7 showed a modest increase, 22 plus-or-minus 1 Engrailed-positive cells per hemisegment in controls versus 29 plus-or-minus 0.6 in mutants, consistent with one or two extra divisions in only a few lineages, alongside widespread ectopic Deadpan. The authors conclude that roughly half the neuroblasts in a hemisegment, spanning rows 5 through 7, show no excessive proliferation.

The mechanistic thread running through these cell fate defects is cyclin E, the master regulator of cell cycle entry. In older-stage mutant embryos, cyclin E protein was strongly upregulated, confirmed with two independent antibodies, and RT-qPCR showed the increase occurs at the transcriptional level. The picture that emerges is dosage- and timing-dependent: when Prospero is lost, cyclin E is de-repressed, and lineages susceptible to elevated cyclin E within a narrow developmental window may divide once or twice extra, while precursor cells whose cyclin E levels fall too low fail to divide altogether. This explains both the limited cell loss and the limited cell gain observed across different lineages, a pattern fundamentally incompatible with runaway stem cell proliferation.

The cell death hypothesis fared no better. Using Acridine Orange staining, TUNEL assays and immunostaining for cleaved Drosophila Caspase-1, the definitive executioner marker, the researchers found no significant differences between mutants and controls at 10, 12 and 15 hours post-fertilization, apart from a small, statistically significant increase in Dcp1-positive cells only at the latest 15-hour stage, including within tracked row 5 lineages. Cell death, in other words, is not balancing out extra divisions to maintain nerve cord architecture, contrary to the earlier proposal.

What the mutants do display is a striking mitotic block. Staining for phosphorylated Histone H3, a marker of chromosome condensation, revealed far more PH3-positive cells in mutants at every stage examined: 23.2 per hemisegment at 10 hours post-fertilization versus 12.7 in controls, and 16.1 versus 2.9 at 15 hours. Combined PH3 and DAPI staining showed that mutant cells accumulate specifically in prophase and metaphase, with no corresponding increase in anaphase-telophase figures, indicating that cells enter mitosis but stall before chromosome segregation. These arrested cells were Dcp1-negative, so they are not dying, and the authors speculate that ectopic cyclin E may keep them alive. Prospero, they suggest, is likely involved in mitotic checkpoint regulation, perhaps through its control of checkpoint genes.

The authors are careful to frame their findings as embryo-specific. In larval brains, prospero mutant clones do overproliferate into tumor-like masses of cells that lack the neuronal marker Elav, so the embryonic ventral nerve cord operates under different rules, with neurons differentiating on schedule even as their identities are scrambled. Within the embryo, the team also rules out an alternative explanation based on residual maternal prospero mRNA, noting that maternal protein does not persist beyond 4 hours post-fertilization and that lineages formed well after this point still show the mixed-identity phenotype. Taken together, the study recasts Prospero not as a guardian against dedifferentiation but as a large-scale regulator of gene expression and mitotic progression, one whose loss plunges neuronal precursors into a crisis of misspecification, where cells co-express conflicting fate programs, project axons, and divide too little, too late, or once too often.

Subject of Research: The role of the chromatin-binding protein Prospero in regulating cyclin E, mitosis and cell identity during Drosophila neurogenesis

Article Title: Prospero prevents lineage-specific loss and gain by regulating cyclin E and mitosis during Drosophila neurogenesis

Article References: Makhijani, K., Verma, S., Mar, J., & Bhat, K. M. (2026). Prospero prevents lineage-specific loss and gain by regulating cyclin E and mitosis during Drosophila neurogenesis. EMBO Reports. https://doi.org/10.1038/s44319-026-00947-6

Image Credits: AI Generated

DOI: 10.1038/s44319-026-00947-6

Keywords: Drosophila, neurogenesis, Prospero, neuroblasts, ganglion mother cells, cyclin E, mitosis, cell fate, chromatin-binding protein, stem cells, cell division, embryonic nerve cord