Stem cell-specific ecdysone signaling regulates the development of dorsal fan-shaped body neurons and sleep homeostasis

Abstract

Complex behaviors arise from neural circuits that assemble from diverse cell types. Sleep is a conserved behavior essential for survival, yet little is known about how the nervous system generates neuron types of a sleep-wake circuit. Here, we focus on the specification of Drosophila 23E10-labeled dorsal fan-shaped body (dFB) long-field tangential input neurons that project to the dorsal layers of the fan-shaped body neuropil in the central complex. We use lineage analysis and genetic birth dating to identify two bilateral type II neural stem cells (NSCs) that generate 23E10 dFB neurons. We show that adult 23E10 dFB neurons express ecdysone-induced protein 93 (E93) and that loss of ecdysone signaling or E93 in type II NSCs results in their misspecification. Finally, we show that E93 knockdown in type II NSCs impairs adult sleep behavior. Our results provide insight into how extrinsic hormonal signaling acts on NSCs to generate the neuronal diversity required for adult sleep behavior. These findings suggest that some adult sleep disorders might derive from defects in stem cell-specific temporal neurodevelopmental programs.

Keywords: central complex; dorsal fan-shaped body neurons; neural cell fate; neural identity; neural stem cells; sleep fragmentation; steroid hormones; temporal patterning; type II lineages.

Figure 1. Type II NSCs produce 23E10 dFB neurons

A) Schematics of larval Type II NSCs (8 per lobe: DM1–6, DL1–2), which divide asymmetrically over 120 hours ALH to generate INPs and express early and late temporal factors. The temporally expressed EcR mediates the switch from early to late gene transition. The Type II NSC and INP temporal factors are thought to contribute to the formation and diversification of neural lineages of the Drosophila central complex. We are investigating the role of ecdysone signaling in the specification and function of 23E10 dFB neurons, which are part of the Drosophila sleep-wake circuit. B) Schematics showing an intersectional genetic approach for Type II NSC lineage analysis utilizing Type II NSC specific flip-out approach. The expression of Asense-GAL80 in Type I NSCs ensures that Worniu-GAL4 is only expressed in Type II NSCs. Worniu-GAL4 induces Type II NSC lineage-specific expression of flippase (FLP) in all Type II NSCs, which excises a STOP cassette, activating LexAopGFP expression in a class-specific manner when crossed to a LexA driver. These flip-out events allow 23E10 dFB neurons to be labeled in green if produced from Type II NSCs. C) The 23E10 dFB neurons are labeled by reporter GFP in Type II NSC specific manner. 23E10 dFB neurons are shown in green (max projection), and nc82 labels neuropil (magenta) (projections showing only FB). The expression of GFP reporter in 23E10 dFB neurons confirms that they are derived from Type II NSCs. D) Quantification of the number of 23E10 dFB neuron cell bodies per hemibrain flipped by Type II NSC filtering. Error bars represent ± SD. Scale bars represent 40μm. n= 20 adult hemibrains.

Figure 2. 23E10 dFB neurons are generated by late DL1 and DM1 Type II NSCs

A) The schematic of CLIn intersectional genetics illustrates the genetic components enabling lineage analysis and genetic birth dating of Type II NSC lineages. The CLIn flies use a Type II NSC-specific promotor, stg14 (magenta), to express KD recombinase specifically in Type II NSCs. KD recombinase removes a stop sequence, bringing FLP recombinase (grey) in frame with a heat shock promotor, but only in Type II NSCs. Upon heat shock, FLP is stochastically expressed, removing another stop sequence and activating Cre recombinase (purple), specifically in Type II NSCs. The active Cre recombinase removes GAL80 and makes LexA::p65 active, thus enabling the lineage-specific expression of reporter mCherry (cyan) and the expression of mCD8GFP in a class-specific manner possible when crossed to GAL4. (Adopted from Ren et al. 2016) B) The schematic shows how CLIn allows lineage analysis of Type II NSCs. The stochastic heat shock mediated FLP event in a DL1 Type II NSC labels all neurons and glia with mCherry (cyan) born from this NSC and, when crossed to a cell class-specific GAL4 co-labels the neurons with GFP. C) Composite confocal image of a single DL1 NSC clone induced at 0h ALH labels most 23E10 dFB neurons (green). All lineages from DL1 NSC are labeled in cyan (mCherry), and the neuropil of the adult fly brain is stained with nc82 (magenta). D) Confocal image of single DM1 NSC clone induced at 0h ALH labels 1–2 23E10 dFB neurons. The DM1 lineages are labeled in cyan (mCherry), and the neuropil of the adult fly brain is stained with nc82 (magenta). E) The schematic illustrates heat shock administered at various time points during larval development. The red lightning bolt symbol indicates heat shock given at three specific time points: 0h, 48h, and 76h after larval hatching (ALH). F-H) Clones induced at 0h and 48h ALH label the 23E10 dFB neurons (F, G), whereas clones induced at 76h ALH do not label any 23E10 dFB neurons. H) Quantification of 23E10 dFB neuron cell bodies labeled per hemibrain when clones are induced at 0h, 48h, and 76h ALH. Error bars represent ± SD. Scale bars represent 20μm, n=16 adult hemibrains for each time point.

Figure 3. Ecdysone signaling regulates 23E10 dFB neuron specification

A) A schematic illustrating EcR-FlpStop2.0 conditional knock-out strategy. In this strategy, the expression of FLP recombinase in Type II NSCs flips the tdTomato sequence into the correct frame with a UAS promoter, allowing it to label mutant cells specifically under the control of Type II specific GAL4 in red color. Additionally, the FLP event also inverts the STOP sequence - transcription-based disruption (Tubα1terminator and 10x Ribozyme sequence) and translation disruption (MHC splice acceptor paired with STOP codons) - to generate a premature stop, disrupting EcR expression and function. B) A schematic depicting normal Type II NSC expressing EcR at 55h ALH, which results in the activation of EcR-induced downstream genes. When EcR function is lost, the expression of downstream genes is disrupted. C, C’) Shows a control brain with dFB neurons labeled by 23E10-LexA driving GFP expression, highlighting their projection pattern in the FB. The nc82 (magenta) labels the neuropil in the adult brain in all subsequent figures. D, D’) Upon EcR loss of function in Type II NSCs, 23E10 dFB neurons are not specified E, E’) Blocking ecdysone signaling in Type II NSCs using EcR-DN results in significant loss of 23E10 dFB neurons. F-H’) In control brains, the 23E10 dFB neurons project to layer 6 of the FB (F, F’). In animals with EcR loss of function (G, G’), or expressing EcR-DN (H, H’), the surviving dFB neurons misproject to the ventral FB layers indicated by yellow arrows. I) One-way ANOVA test (followed by Dunnett’s Multiple Comparison Test) quantification of 23E10 dFB neuron cell bodies. Error bars represent ± SD. Asterisks indicate the level of statistical significance: *p<0.05, **p<0.01, ***p<0.001, **** p<0.0001, NS, non-significant. Cell bodies are indicated by white arrows. The dashed line outlines the FB. Scale bars represent 20μm, n=12 adult hemibrains. See also Figure S1.

Figure 4. Ecdysone-induced gene E93 is necessary for 23E10 dFB fate

A) 23E10 dFB neurons, labeled in green (A), express E93 in cyan (A’, A”). B) Control dFB neurons labeled by 23E10-LexA project to the dorsal FB. C) E93 knock-down in Type-II NSCs using Pointed-GAL4 results in the complete loss of 23E10 dFB neurons (C, C’). D) E93 overexpression in Type-II NSCs using Pointed-GAL4 does not alter the number or morphology of 23E10 dFB neurons (D, D’). E) Expression of E93 under Pointed-GAL4 fails to rescue 23E10 dFB neurons in EcR loss- of-function background (E, E’). F-G’) Compared to control brains (F, F’), the experimental brains with UAS-E93 expressed in EcR loss-of-function background show defects in axonal targeting (G, G’). The axonal projections of 23E10 dFB neurons ectopically innervate ventral layers of FB, as indicated by yellow arrows. H) Quantification of 23E10 dFB neuron cell bodies per hemibrain using one-way ANOVA test followed by Šidák’s Multiple Comparison Test. Error bars represent mean ± SD; Asterisks indicate the level of statistical significance: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, NS, non-significant. Cell bodies are indicated by white arrows. The dashed line outlines the FB. Scale bars represent 20μm, n=14 adult hemibrains for each genotype. See also Figures S2 and S3.

Figure 5. EcR and E93 are required in DL1 Type II NSCs to specify 2310 dFB neurons

A-G) Control (A), EcR, and E93 loss-of-function (B, C) show loss of 23E10 dFB neurons, labeled by 23E10-lexA driving GFP expression. In control brains (D, D’), 23E10 dFB neurons exhibit normal axonal projection to the dorsal layers of FB. However, in E93 (E, E’) and EcR (F, F’) loss-of-function conditions, the axonal projections of the 23E10 dFB neurons are impaired and expand into ventral layers of FB, as indicated by yellow arrows. G) Quantification of 23E10 dFB neuron cell bodies per hemibrain upon DL1/DL2 specific knockdown of E93 and EcR, analyzed using one-way ANOVA followed by Dunnett’s Multiple Comparison Test. Error bars represent mean ± SD. Asterisks indicate the level of statistical significance: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, NS, non-significant. Cell bodies are indicated by white arrows. The dashed line outlines the FB. Scale bars represent 20μm. n= 12 adult hemibrains for each genotype.

Figure 6. E93 expression in a restricted time window regulates 23E10 dFB neuronal fate

A) Schematic of TARGET system illustrating GAL80ts-mediated restricted knockdown of E93. At 18C, GAL80ts is active, preventing the E93 RNAi expression by inhibiting Pointed-GAL4 activity. At 29C, GAL80ts is inactive, allowing E93 RNAi expression temporally. B) Schematic of the experimental setup showing E93 RNAi flies reared at different temperatures throughout the larval life cycle, from 0 to 120h ALH. Flies with E93 RNAi and GAL80ts were initially grown at 18°C and then shifted to 29°C around 40 hours ALH to enable E93 RNAi expression in late Type II NSCs. C, C’) Shows a loss of 23E10 dFB neurons, labeled with GFP, at 29C upon E93 knockdown (E93 RNAi without dicer). D, D’) Displays loss of 23E10 dFB neurons upon continuous E93 RNAi expression in animals grown at 29C throughout development (GAL80ts is inactive at 29C). E, E’) Shows significant loss of 23E10 dFB neurons when UAS-E93 RNAi is restricted to late Type II NSCs using GAL80ts. Flies were grown at 18C until 40 hours ALH and then shifted to 29C to inactivate GAL80ts. F, F’) The 23E10 dFB neuron number remains normal when flies expressing UAS-E93 RNAi combined with GAL80ts are grown continuously at 18C. G) Quantification of 23E10 dFB neuron cell bodies per hemibrain using one-way ANOVA followed by Dunnett’s Multiple Comparison Test. Error bars represent ±SD. Asterisks indicate statistical significance: *p<0.05, **p<0.01, ***p<0.001, **** p<0.0001, NS, non-significant. Cell bodies are indicated by white arrows. The dashed line outlines the FB. Scale bars represent 20μm. n= 12 adult hemibrains for each genotype.

Figure 7. Knockdown of E93 in larval Type II NSCs impairs adult sleep

Sleep traces (A) and quantification of day and night sleep duration (B), sleep bout number (C), and sleep bout length (D) in flies expressing E93-RNAi under control of pointed-GAL4 (red) compared to genetic controls (black, gray) using the standard 5-minute quiescence threshold for sleep. (E-H) shows the same data for long bout sleep (≥60-minute quiescence threshold for sleep). N=60,55,59 from left to right (A-H). Quantification of sleep duration (I), sleep bout number (J), and sleep bout length (K) in E93-RNAi flies and controls during the first 3 hours of a baseline day or following a night (12 hours) of sleep deprivation. N=51,47,57 from left to right. Error bars represent SEM; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, NS, non-significant by Kruskal-Wallis test with Dunn’s multiple comparison corrections. See also Figure S4.