Inactivation of both Foxo and reaper promotes long-term adult neurogenesis in Drosophila

Abstract

Adult neurogenesis occurs in specific locations in the brains of many animals, including some insects, and relies on mitotic neural stem cells. In mammals, the regenerative capacity of most of the adult nervous system is extremely limited, possibly because of the absence of neural stem cells. Here we show that the absence of adult neurogenesis in Drosophila results from the elimination of neural stem cells (neuroblasts) during development. Prior to their elimination, their growth and proliferation slows because of decreased insulin/PI3 kinase signaling, resulting in nuclear localization of Foxo. These small neuroblasts are typically eliminated by caspase-dependent cell death, and not exclusively by terminal differentiation as has been proposed. Eliminating Foxo, together with inhibition of reaper family proapoptotic genes, promotes long-term survival of neuroblasts and sustains neurogenesis in the adult mushroom body (mb), the center for learning and memory in Drosophila. Foxo likely activates autophagic cell death, because simultaneous inhibition of ATG1 (autophagy-specific gene 1) and apoptosis also promotes long-term mb neuroblast survival. mb neurons generated in adults incorporate into the existing mb neuropil, suggesting that their identity and neuronal pathfinding cues are both intact. Thus, inhibition of the pathways that normally function to eliminate neural stem cells during development enables adult neurogenesis.

Figure 1. Neuroblast growth becomes restricted during pupal development

(A,D) Z projection of the brain (optic lobe, OL, and central brain, CB) and ventral nerve cord (VNC) from wild type (WT) animals (pcna:eGFP transgenic) at 15 (A) or 48 (D) hours after pupal formation (APF). Neuroblasts express Deadpan (Dpn, red) and the E2F reporter pcna:eGFP (GFP, green). (D) Only the four mushroom body (mb) neuroblasts (nbs) located on the dorsal CB surface persist late (arrowheads in A,D). Dotted box denotes area of CB magnified in larger inset with one of the four mb neuroblasts in small inset. Scale bar (A) equals 50µm. (B,E) Quantitation of CB, (including the mb neuroblasts) (B) and mb (E) neuroblast number and number of mitotic neuroblasts based on PH3 (phospho-histone H3) per brain lobe over time. n=number of brain lobes scored per time point. (C) Distribution of CB neuroblast cell size over time. The average diameter of all CB neuroblasts from three brain lobes (3 animals) was measured for each of 3 time points. (F) Quantitation of the average diameter (diam.) of mb neuroblasts over time. Number of mb neuroblasts measured for each time point shown as white number in black columns. *P-value <0.001. Columns represent mean, ±std.dev in this and all subsequent figures. ALH, after larval hatching.

Figure 2. Reaper-family pro-apoptotic regulators and caspase activation are required for the proper timing of neuroblast cell death

(A–C) Wild type (WT) mb neuroblasts undergo cell death (B,C). Developmental time and genotype are listed above panels and markers are listed within panels in this and all subsequent figures. (A–C) Top two rows, single channel images with colored overlay in bottom row, showing co-localization of cell death and neuroblast markers (B,C). Mb neuroblast (brackets) at early pupal stages (A), days prior to mb neuroblast cell death (B,C). DNA fragmentation (C, TUNEL) and nuclear envelope breakdown, as indicated by an absence of Lamin staining (B), are hallmarks of cell death (compare with Lamin staining of a mb neuroblast with an intact nuclear envelope, B, inset). (D–F, H–J) Z-projections of the dorsal, anterior CB surface. Aberrantly persisting neuroblasts are marked with white arrows and mb neuroblasts are marked with arrowheads. In rpr⁻ and RHG miRNA adults, the aberrantly persisting neuroblasts are the mb neuroblasts. Only one of the 4 mb neuroblasts is shown in panel F. (H–J) The mb calyx is outlined by the white dash. (G,K) Quantitation of Dpn-expressing CB or mb neuroblast number per brain lobe over time. n=number of brain lobes scored at each time point. worniuGAL4 is used to induce UAS transgene expression (F,G,J,K). Scale bar (A,D) equals 10µm. see also FigureS1,S2.

Figure 3. Increased levels of nuclear Foxo restrict long-term survival of mb neuroblasts

(A–D) Subcellular localization of Foxo in neuroblasts. All neuroblasts imaged under same settings for direct comparison. Neuroblasts are marked with black brackets and their nuclei are outlined in red. (E) Quantitation of the ratio of cytoplasmic to nuclear Foxo protein, based on fluorescence pixel intensities. White number in black columns equals number of neuroblasts scored for each genotype/time. (F) Quantitation of mb neuroblast number over time. (mean, std deviation). Dp60 overexpression (o/e) inhibits PI3K signaling leading to premature mb neuroblast cell death (F,G), while foxo mutant mb neuroblasts delay death (F,H). n equals number of brain lobes scored for each genotype at each time point (F). (G,H) Dying mb neuroblasts lack a nuclear envelope. Compare Lamin staining with inset in Figure2B. Top panels are single image and bottom row is overlay. (I,J) Z-projections of the dorsal brain surface showing persisting adult mb neuroblasts (arrowheads) near the mb calyx (white outline). (K) Quantitation of mb neuroblast number over time. Numbers in columns equals number of brain lobes scored for each genotype/time. worniuGAL4 is used to induce UAS transgene expression (B,E,F,G,I–K). Scale bar (A,G,I) equals 10µm. see also FigureS3.

Figure 4. Long-term surviving mb neuroblasts proliferate and generate new adult mb neurons

(A) Summary of mb neuroblast cell size over time. Genotypes color-coded at top for panels a–c. Each dash (A) or diamond (B,C) represents one neuroblast. (A) *P-value <0.003, **P-value <0.0003. (B,C) Number of BrdU progeny generated in 24 hours relative to mb neuroblast cell size. Foxo restricts mb neuroblast proliferation rate and cell size. Mb neuroblast cell size and proliferation rate correlate in 2-week-old RHG miRNA, foxo⁻ adults (C) but not in 3–5 day adults (B). R²=coefficient of determination for a linear correlation. (D–H) Mb neuroblasts (white brackets) and progeny (numbered) following a 24 hour BrdU treatment. (H) Only 14 of 18 progeny are shown. (I,K) Month-old adult mb neuroblasts (I, bracket) generate new mb neurons, cell bodies (I, outline) and axons (K). Inset is double-labeled neuroblast, mCD8:GFP (green) and Dpn (red). genotype: worniuGAL4,UASmCD8:GFP,UAS-RHGmiRNA,foxo⁻/foxo⁻ (K) Reconstruction of the mushroom body (red) with overlay of newborn adult neurons (mCD8:GFP, green). (J) Dendrites from new neurons project normally to the mb calyx (arrow, 2-week adult). (K,right) Axons project either normally through the mb pedunculus or mistarget (K,left). Single arrow marks premature axon bifurcation (K,left) and double arrow marks termination at the top of the α lobe. (L) model (see discussion). worniuGAL4 is used to induce UAS transgene expression (A–E,G–K). Scale bar (D,I–K) equals 10µm. see also FigureS4.