Local Ecdysone synthesis in a wounded epithelium sustains developmental delay and promotes regeneration in Drosophila

Regeneration of damaged tissue often involves cells around the injury reentering the cell cycle to repopulate the damaged area. In some cases, local stem cells are the main source of new cells, whereas in others, adjacent uninjured cells partially dedifferentiate and divide. Regenerative capacity varies widely between species and is often lost at post-embryonic or juvenile stages, suggesting that it depends on redirection of developmental pathways toward regenerative growth. The associated pause in developmental growth ensures synchrony by allowing injured tissue to repair before rejoining a normal developmental trajectory.

In Drosophila melanogaster larvae, genetically induced death of epithelial cells in the wing pouch results in proliferation of surviving cells (the ‘blastema’) that restores adult wing size. This process is accompanied by concentration of Wingless (Wg) morphogen within the blastema and by blastema-specific gene expression, including the transcription factor Ets21C and secreted factors upd3 and Ilp8 (Insulin-like peptide-8), which enhance blastema proliferation and suppress production of the steroid hormone ecdysone (Ec) in the prothoracic gland (PG) (Smith-Bolton et al., 2009; Worley et al., 2022; Romão et al., 2021; Colombani et al., 2012, 2015; Garelli et al., 2012, 2015). This decline in circulating Ec elicits a pause in development by reducing activity of its cognate receptor (EcR). Mutations that impair key elements of this pause mechanism or that reduce injury-induced proliferation each reduce wing regeneration efficiency.

Loss of regeneration competence in a commonly used Drosophila wing injury system parallels a surge of Ec at the late larval-to-pupal transition (Smith-Bolton et al., 2009; Harris et al., 2016). This temporal correlation has led to the hypothesis that high Ec inhibits regeneration (Narbonne-Reveau and Maurange, 2019). However, Ec can promote regeneration in other species and is present during regeneration-competent stages of the Drosophila life cycle at lower levels that promote growth in the absence of injury (Das and Durica, 2013; Dye et al., 2017; Herboso et al., 2015; Karanja et al., 2022). Moreover, Ec/EcR are required for activity of the Dpp and Wg pathways (Parker and Struhl, 2020), which promote normal and regenerative wing disc growth.

The paradox that injury depletes circulating Ec at a time when the wing blastema is undergoing regeneration led us to assess Ec roles within injured discs. We find that, as EcR activity drops in uninjured disc regions, it rises in the blastema region. Local depletion of Ec biosynthesis enzymes has a limited effect on growth of uninjured wings but strongly impairs regrowth of injured wings; depleting the Ec catabolic enzyme Cyp18a1 reciprocally enhances regeneration. Overall, our data indicate that Ec/EcR signaling is required for injury-induced pupariation delay and regeneration, and that mRNAs encoding the regeneration regulators Upd3 and Ets21c are responsive to Ec. These data support a model in which the blastema is a specialized signaling environment that generates a local Ec source necessary for efficient tissue repair.

To assay ecdysone (Ec) roles during wing disc regeneration, we used an established wounding system utilizing rotund (rn)-GAL4 and tubulin-GAL80^ts transgenes to express a pulse of the apoptotic gene reaper in the larval pouch (hereafter rn>rpr^ts) (Fig. 1A) (Smith-Bolton et al., 2009). Flies lacking UAS-rpr were used as uninjured (‘mock’; rn>+^ts) controls. Combining the rn^ts injury system with RNAi lines is an effective approach to identify proteins within the pouch that promote regeneration (Skinner et al., 2015; Khan et al., 2017; Tian and Smith-Bolton, 2021; Brock et al., 2017; Narbonne-Reveau and Maurange, 2019). We adapted this local knockdown (KD) approach to test enzymes that synthesize Ec or its bioactive form 20-hydroxyecdysone (20E) (collectively termed Halloween genes) (Petryk et al., 2003; Warren et al., 2004; Chávez et al., 2000; Gilbert, 2004) (Fig. 1B).

Binning adult wings by size revealed that local KD of the Halloween genes phantom (phtm), disembodied (dib) or shade (shd) in rn>rpr^ts animals significantly impaired wing regeneration relative to rn>rpr^ts without KD, whereas KD of the Ec catabolic enzyme Cyp18a1 (Guittard et al., 2011) enhanced regeneration (Fig. 1C). KD of the torso (tor) receptor, which promotes Halloween gene expression in the PG (Rewitz et al., 2009; Yamanaka et al., 2013) did not affect wing regeneration, whereas depletion of the peptide hormone Ilp8 impaired regeneration, as described previously (Boone et al., 2016; Boulan et al., 2019; Colombani et al., 2015, 2012; Garelli et al., 2012, 2015; Gontijo and Garelli, 2018; Skinner et al., 2015). Quantification of wing area ratios (each genotype relative to its corresponding uninjured control; Fig. 1D) in control (−), phtm^KD and Cyp18a1^KD confirmed that phtm^KD prevents adult wings from regrowing to normal size, whereas Cypa18a1^KD enhances regrowth. phtm^KD and Cyp18a1^KD each slightly reduced the absolute size of uninjured wings, perhaps due to non-specific effects of RNAi (Fig. S1).

Given the requirement for Ec biosynthesis enzymes in the regenerating wing disc, we assessed EcR activity in injured discs using 7xEcRE-GFP, a reporter line with seven EcR response elements (EcREs) upstream of GFP

(Hackney et al., 2007). Without injury, 7xEcRE-GFP activity increased between early and mid-third instar (L3) in the hinge and pouch, and in sensory organ precursors along the dorsoventral boundary (Fig. 2A,B). 7xEcRE-GFP patterns were unaffected by local phtm^KD in the pouch (Fig. 2C-E), consistent with the PG as the major Ec source in larvae.

During the first 24 h of recovery after rpr injury (R0-R24), 7xEcRE-GFP expression dropped to almost undetectable levels in control and phtm^KD injured discs (Fig. 2F-I). By R32 (mid-regeneration), 7xEcRE-GFP expression had risen in scattered dorsal/ventral hinge cells and within the pouch proper (Fig. 2J) and by R48 it strengthened along the ventral hinge and the inner dorsal hinge cells (Fig. 2K), which contribute to the wing regenerate (Smith-Bolton et al., 2009; Herrera et al., 2013; Worley et al., 2022). By late regeneration at R72, 7xEcRE-GFP activity began to resemble its pattern in uninjured mid-L3 wing discs (e.g. Fig. 2B). Importantly, local phtm^KD in the blastema suppressed 7xEcRE-GFP in R32 and R48 discs (Fig. 2J-M,P), but had little effect on 7xEcRE-GFP in R72 discs (Fig. 2N,O). Quantification of 7xEcR-GFP within the pouch and inner hinge region confirmed a reliance on phtm specifically during the R32-R48 period (graphs in Fig. 2P and Fig. S2O; see Fig. S2A-N for an R0-R72 time course), implying a transient reliance on local Ec midway through regeneration.

Use of the TNF receptor ligand eiger (rn>egr^ts) as an alternative to rpr strongly induced 7xEcRE-GFP and EcRE-lacZ in the pouch and suppressed activity in the hinge and notum, indicating that elevated EcR activity at the site of injury and a drop elsewhere are also features of the egr injury system (Fig. S2P-S). Notably, 7xEcRE-GFP expressing cells in rn>egr^ts injured discs overlapped the Wg-positive blastema (Fig. S2T) and a region of elevated EcR protein (Fig. S2U), and were adjacent to Dcp-1 (caspase)-positive cells (Fig. S2V). rn>egr^ts regeneration was also impaired by shd^KD or the shd² null allele (Fig. S2W). For subsequent experiments, we used the rn>rpr^ts system due to its direct apoptotic mechanism and less crumpled disc morphology (Smith-Bolton et al., 2009; Harris et al., 2016; Igaki et al., 2002, 2009; Yoo et al., 2002).

Previous studies have shown that Upd3 and Ilp8 secreted from growth-perturbed wing discs trigger developmental delay by suppressing systemic production of Ec in the PG; this drop in Ec delays the larval-to-pupal transition and allows injured wing discs to regrow and equalize with uninjured discs before pupation (Smith-Bolton et al., 2009; Worley et al., 2022; Romão et al., 2021; Colombani et al., 2015, 2012; Garelli et al., 2012, 2015). Removing Ilp8 impairs this delay and prevents efficient regeneration of adult wings (Hackney et al., 2012; Halme et al., 2010; Jaszczak et al., 2016; Colombani et al., 2012; Garelli et al., 2012; Skinner et al., 2015). Significantly, phtm^KD suppressed injury-induced pupariation delay to a similar extent as an Ilp8^KD control, whereas Cypa18a1^KD had the opposite effect of enhancing the delay; local torso^KD did not affect pupation timing (Fig. 3A,B). All RNAi lines tested slightly accelerated pupariation (Fig. S3A,B) and slightly reduced wing size (Fig. S1), which could suggest non-specific effects of the RNAi processing machinery.

Given the pro-growth role of Ec in discs (Dye et al., 2017; Herboso et al., 2015; Karanja et al., 2022; Gokhale et al., 2016; Nogueira Alves et al., 2022), we next assessed effects of phtm^KD and Cyp18a1^KD on cell division within the blastema, focusing on R48 when 7xEcRE-GFP activity is most dependent on phtm (see Fig. 2). In injured discs, phtm^KD did not significantly alter the absolute level of pouch+hinge EdU relative to injured control (−) (Fig. S3C) but did reduce the fraction of EdU located with the pouch+hinge relative to overall total EdU (Fig. 3F-I). This effect appears to be attributable to a partial reduction of EdU incorporation in areas outside the pouch+hinge of rn>rpr^ts+phtm^KD injured discs (Fig. 3F-H; see dotted white outline of whole disc area). Slowed cell division and reduced EdU in regions outside the blastema is a well-documented feature of injury-induced growth perturbation and results from inhibition of Ec production in the PG, which systemically slows growth of uninjured tissue (Boulan et al., 2019; Hackney et al., 2012; Halme et al., 2010; Jaszczak et al., 2016; Skinner et al., 2015; Worley et al., 2022; Smith-Bolton et al., 2009; Kiehle and Schubiger, 1985). Consistent with a local role for Ec and EcR in this intra-organ feedback mechanism, phtm^KD had a regional effect on growth; the blastema of rn>rpr^ts+phtm^KD discs comprised a smaller percentage of the total disc area relative to injured controls, whereas the blastema of Cyp18a1^KD discs was enlarged relative to total disc area (Fig. 3J) and exhibited an upward trend in the fraction of EdU signal within the pouch+hinge region (Fig. 3H,I). phtm^KD discs were larger than injured controls, and Cyp18a1^KD discs were smaller than injured controls at the same time point in recovery (Fig. 3K). In mock ablated discs, EdU fluorescence within the pouch+hinge was not significantly changed by phtm^KD and was slightly decreased by Cyp18a1^KD (Fig. S3E). The EdU pouch+hinge fraction (pouch+hinge EdU fluorescence/total EdU fluorescence) was unchanged in uninjured phtm^KD and Cyp18a1^KD discs relative to control (−) (Fig. 3C-E; Fig. S3F), which mirrored the very mild effect of phtm^KD or Cyp18a1^KD on wing size (see Fig. 1D and Fig. S1). Parallel analysis of mock-injured rn>+^ts discs did not detect significant differences in overall disc size or pouch+hinge areas (Fig. S3G,H). Analysis of the mitosis (M-phase) marker phospho-Histone H3 (pH3) revealed a drop in mitotic index of injured pouch+hinge cells depleted of phtm or dib relative to injury alone (rn>rpr^ts) (Fig. S3I-T), suggesting that local Ec is also required for maximal mitotic activity within the blastema. Overall, these data are consistent with phtm^KD impairing both mitotic progression within the blastema and intra-organ signaling required to inhibit growth of disc regions outside the blastema.

Based on a previous RNA-seq study that identified phtm mRNA among a group of injury-induced transcripts in sorted cells from rn>rpr^ts blastemas (Khan et al., 2017), we reanalyzed a public single-cell RNA-sequencing (scRNA-seq) dataset derived from injured (rn>egr^ts system) and uninjured larval wing discs (Floc'hlay et al., 2023). This data confirms phtm induction in wound cells and further refined this expression to ‘beta’ wound cells (Fig. S4A,B,H) (as defined in Floc'hlay et al., 2023). Expression of sro (shroud), a Halloween gene that acts upstream of phtm in the Ec synthesis pathway, was detected throughout the disc independent of injury (Fig. S4C). Following injury, the Ec-inducible genes ImpE1, ImpL2, ImpL3 (also known as Ldh) (Paine-Saunders et al., 1990; Moore et al., 1990; Osterbur et al., 1988; Natzle et al., 1988, 1986) and Cyp18a1, which is induced by EcR as part of a feedback inhibitory loop (Rewitz et al., 2010), were induced in blastema cells (Fig. S4D-H). Analysis of public EcR CUT&RUN data (Uyehara and McKay, 2019) indicate that all four of these loci are bound by EcR in larval wing cells (Fig. S4I).

Based on evidence of Ec synthesis and EcR activation in injured wings, we assessed effects of phtm^KD, dib^KD or Cyp18a1^KD on mRNAs encoding upd3, Ilp8, and the regeneration-specific transcription factor Ets21c (Fig. 4) (Worley et al., 2022; Santabárbara-Ruiz et al., 2015; Katsuyama et al., 2015; Pastor-Pareja et al., 2008; La Fortezza et al., 2016). Pouch-specific depletion of phtm or dib in the background of rn>rpr^ts significantly impaired injury-induction of upd3 mRNA relative to uninjured controls (as detected by qPCR) at R48 (48 h recovery), whereas depletion of Cyp81a1 increased upd3 expression at R48 (Fig. 4A). A upd3-lacZ enhancer trap reported elevated upd3 transcription in the ventral hinge at R24 of rn>rpr^ts discs (Fig. 4B, left panels), which overlapped the region of phtm-dependent EcR activity post-injury (see Fig. 2). By R48, upd3-lacZ appeared in large pouch cells that resemble the pattern of EcR protein detected in post-injury discs (Fig. S2U). Importantly, upd3-lacZ expression was reduced at R24 and R48 when tested in age-matched rn>rpr^ts+phtm^KD disc (Fig. 4B, optical slices in right panels;Fig. S5, z-stack). Quantification of fluorescence intensity confirmed dependence of upd3-lacZ on local phtm expression (Fig. 4C). We found no evidence that phtm, dib or Cyp18a1 was required for expression of Ilp8 mRNA or an Ilp8-GFP reporter in injured (Fig. 4A,D) or uninjured discs (Fig. S6A). phtm^KD and dib^KD also had no significant effect on Ets21 levels; however, Cyp18a1^KD strongly elevated Ets21c mRNA levels at both time points post-injury (Fig. 4A). This effect of Cyp18a1^KD suggested that excess Ec can promote Ets21c mRNA accumulation but is not normally required for it. Recent evidence indicates that wing injury alters epithelial permeability (DaCrema et al., 2021) and could thus affect the ability of Upd3 and Ilp8 to escape the disc lumen and signal to distant tissues. We measured this parameter using fluorescein conjugated dextran (FITC-dextran) but found no evidence that this parameter responds to local phtm^KD or Cyp18a1^KD (Fig. S6B).

Based on these collective findings, we propose that local Ec influences wing disc regeneration by enhancing local expression upd3 and potentially Ets21c (Fig. 4E). This enhanced dependence on a local Ec in genetically wounded discs contrasts with cells in developing wing discs that normally respond to Ec derived from the PG (Parker and Struhl, 2020). As EcR activity and nuclear accumulation of EcR protein itself remained restricted to the pouch and hinge of rn>rpr^ts or egr^ts damaged discs (e.g. see Fig. 2J-L and Fig. S2U), locally available Ec appears to signal over a relatively short range. It is worth noting that this paradigm of local Ec synthesis driving local EcR activity is similar to the initiation of border cell (BC) migration in the ovary (Domanitskaya et al., 2014), which relies on Ec generated by a few follicle cells to promote invasion of the BC cluster between nurse cells (Bai et al., 2000). How Ec is concentrated locally in these two scenarios is unclear, but could rely on secretion, uptake or local sequestration, perhaps including low levels of Ec produced by the PG during the developmental pause. EcR activity has also been shown to be induced within larval wing discs damaged by exposure to gamma irradiation (Karanja et al., 2022), which parallels the effect of rn>rpr^ts and egr^ts observed in this study. However, the source of Ec following irradiation was not examined. Intriguingly, activating EcR in irradiated discs by Ec feeding or blocking EcR using a dominant-negative transgene (EcR^DN) each affect expression of the Ilp8-GFP reporter and Wg protein, which plays a key role in regeneration (Karanja et al., 2022). Neither of these factors responds to knockdown of Cyp18a1 or Halloween genes in rn>rpr^ts damaged wing discs, implying potential differences in either in the effect of Rpr-killing versus ionizing radiation, or in methods used to identify EcR candidate targets. Broadly, the identity of gene targets that respond to the pulses of EcR activity detected in the pouch of rn>rpr^ts and egr^ts models would reveal much about how a damaged wing disc evades the systemic loss of Ec from the PG by upregulating Halloween gene activity in the pouch, and how this local Ec contributes to the unique transcriptional signature necessary for tissue repair.

The Drosophila lines rn-GAL4, tub-GAL80^ts; rn-GAL4, tub-GAL80^ts, UAS-rpr; and rn-GAL4, tub-GAL80^ts, UAS-egr have been previously described (Smith-Bolton et al., 2009; gifts from R. Smith-Bolton, University of Illinois Urbana-Champaign, USA), as has 7xEcRE-GFP (Hackney et al., 2007; gift from V. Henrich, UNC Greensboro, USA) and upd3-lacZ (Bunker et al., 2015; gift from I. Hariharan, UC Berkeley, USA). The following lines were obtained from the Bloomington Drosophila Stock Center (BDSC): Ilp8-GFP (Mi[MIC]Ilp8^MI00727) (BDSC #33079); UAS-torso-RNAi (BDSC #58312); UAS-phm-RNAi (BDSC #55392); UAS-shd-RNAi (BDSC #67356) (Jugder et al., 2021); shd² (BDSC #4219); UAS-cyp18a1-RNAi (BDSC #64923) (Idda et al., 2020); and 7xEcRE-lacZ (BDSC #4517). The following lines were obtained from the Vienna Drosophila Stock Center (VDSC): UAS-dilp8-RNAi (v102604) (Kashio et al., 2016); UAS-phm-RNAi (v104028) (Beebe et al., 2015); UAS-dib-RNAi (v101117) (Deady et al., 2015); UAS-shd-RNAi (v108911) (Knapp and Sun, 2017).

Ablation experiments were performed as previously described (Smith-Bolton et al., 2009). We used the (1) w¹¹¹⁸;;rn-GAL4, tub-GAL80ts,UAS-reaper and (2) w¹¹¹⁸;;rn-GAL4, tub-GAL80ts, UAS-eiger genetic ablation systems to study regeneration and (3) w¹¹¹⁸;;rn-GAL4, tub-GAL80ts as our mock ablation control. These are abbreviated as (1) rn>rpr^ts, (2) rn>egr^ts and (3) rn>+^ts. Development was synchronized by collecting eggs on grape plates. Eggs were collected for 4 h at 25°C then placed at 18°C. After 2 days at 18°C, 50 L1 larvae were picked and placed into churned fly food vials. For rn>rpr^ts experiments, on day 7 after egg lay (AEL) temperature shifts were performed to induce ablation. Vials were moved from 18°C to a 29°C circulating water bath for 20 h. For rn^ts>egr, temperature shifts were performed at day 8 AEL for 40 h. For wing disc experiments, dissections were performed at several time points during regeneration: immediately following ablation (R0), and after 24, 48 and 72 h of regeneration. Mock ablation dissections were performed at R24 or R48 after mock ablation and developmental stage (early L3-late L3) was further confirmed by colorimetric assay (Hailstock et al., 2023).

Adult wings were plucked and mounted in Gary's Magic Mountant (Canada balsam dissolved in methyl salicylate). Images were taken on a Nikon SMZ800N microscope at 25× magnification (2.5× objective×10× eyepiece) using a Leica MC170HD camera and a Nii-LED High intensity LED illuminator. Area of wing was measured using FIJI.

Larval discs were dissected in 1×PBS, fixed for 20 min in 4% paraformaldehyde at room temperature, rinsed 3× in 1×PBS, then permeabilized for 30 min at room temperature (RT) in 1×PBS+0.3% Triton X-100 (1×PBST 0.3%). Samples were then blocked for 1 h in 1×PBS+10% normal goat serum (NGS), and then incubated in primary antibody in 1×PBS+0.1% Triton X-100 (1×PBST 0.1%)+10% NGS overnight at 4°C. The next day, samples were washed for 5 min 5× in 1×PBST 0.1%, incubated for 45 min at RT in secondary antibody in 1×PBST 0.1%+10% NGS, and washed for 5 min 5× in 1× PBST 0.1%. Nuclei were then stained with DAPI/Hoechst for 10 min and washed 2× in 1× PBS, before mounting discs in n-propyl gallate (4% w/v in glycerol). Wing discs were imaged on a Nikon A1R HD25 confocal system using 20× and 40× objectives. Primary antibodies used were: anti-Wg antibody [Developmental Studies Hybridoma Bank (DSHB), 4D4; 1:50], anti-β-galactosidase (DSHB, 40-1a; 1:250), anti-EcR common (DSHB, DDA2.7; 1:100), anti-cleaved DCP-1 (Cell Signaling Technology, 9578S; 1:100) and anti-Phospho-histone H3 (Cell Signaling Technology, 9701S; 1:100). Fluorescently labeled secondary antibodies were from Jackson ImmunoResearch and used at 1:50 (115-166-003). Nucleic acid staining was performed by incubating discs for 10 min with Hoechst 3342 (Thermo Fisher Scientific) or DAPI (Invitrogen).

EdU staining was performed according to the Click-iT EdU Cell Proliferation Kit following the manufacturer's instructions (Thermo Fisher Scientific; C10340). Briefly, live discs were incubated in Schneider's medium (Thermo Fisher Scientific; 662249) with EdU at 100 µM concentration for 20 min at RT in the dark. After incubation, discs were fixed in 4% paraformaldehyde for 20 min before proceeding with standard antibody staining as described above. Click-iT EdU detection was then performed according to the Click-iT EdU Proliferation Kit instructions.

Pupariation rates were quantified by counting newly formed pupae every 8 h. Timing was reported as h AEL, starting at the beginning of the 4 h egg lay period. Pupariation was determined as when animals stopped moving and darkened in color.

We dissected 50 discs from equivalently staged larvae of each genotype and them added to 100 µl of Trizol. RNA was extracted using the Qiagen RNeasy kit (Qiagen; 74106) according to the manufacturer's instructions. cDNA conversion was performed using SuperScript-III RT kit (Thermo Fisher Scientific; 18080093) according to the manufacturer's instructions, using 1 µg of extracted RNA. qPCR analysis was performed in triplicate using SYBR Green Master Mix (Qiagen, 204143) on a QuantStudio6 Real-Time PCR System. The entire assay was repeated three times using discs from separate ablation experiments. Refer to Table S1 for a full list of primers used. Quantitative qPCR analysis was performed using the ΔΔC_T method and expression levels were normalized to rp49. Statistical analysis was performed in GraphPad Prism9 using a two-tailed Student's t-test.

The dextran assay was performed as previously described (DaCrema et al., 2021). Larvae were dissected, inverted and cleaned in Schneider's medium at the time points noted. Carcasses were then placed in 25 mg/ml 10 kD fluorescein-conjugated dextran (Thermo Fisher, D1820) diluted 1:8 in Schneider's medium and incubated for 30 min at RT in the dark while rocking. Carcasses were then washed 1× with Schneider's medium and the fixed with 4% paraformaldehyde for 20 min, before proceeding with normal antibody staining as described above.

scRNA-seq data were analyzed using the online portal https://scope.aertslab.org/#/WingAtlas/*/welcome. The session loom ‘multiome_scRNA.scenic1x.scope’, located under sequential ‘wing disc’, ‘atlas’ and ‘full_multiome’ tabs, was loaded and analyzed using the ‘Compare’ feature with setting ‘number of displays:2’. ‘REWE37570709B79’ (Rotund GAL4 GAL80ts Ctrl) and ‘REW51C2157Eeed7’ (Rotund GAL4 GAL80ts Eiger) were loaded into adjacent windows with the following display settings: ‘Log transform’ (off), ‘CPM normalize’ (on), ‘Expression based plotting’ (on), ‘Dissociate viewers’ (off). Pouch and wound annotations were added to UMAPs using built-in ‘broad annotations’ setting. Counts per million (CPMs) for individual genes were annotated using the dropdown tabs and downloaded as .csv files for importation into Prism to generate Bubble plots. For visualization of EcR CUT&RUN peaks, datasets from Uyehara and McKay (2019) was downloaded from the NCBI SRA and mapped to the Dm6 genome using Galaxy tools (usegalaxy.org) and visualized using IGV viewer (igv.org). SRA files analyzed: SRX15238797, SRX15238795, SRX5173551, SRX5173550, SRX5173549, SRX5173548.

Only female animals were used for these experiments, as previous research has indicated that female and male flies display statistically different regenerative capacity. Populations from separate ablation experiments cannot be meaningfully compared with one another as studies from other labs have indicated seasonal variation in regeneration.

For binning experiments, adult wing size was scored following injury and regeneration into five categories (0%, 25%, 50%, 75% and 100%). Results were calculated per population, with n representing data from at least three separate ablation experiments. Statistical analysis was performed in GraphPad Prism9 using the χ² test.

Results were calculated per population, with n representing data from at least two separate ablation experiments. For injured wing size measurements, wings were normalized to the mean wing size of mock ablation wings from the same genotype. This was done to control for the small but statistically significant difference between mock ablation controls. Statistical analysis was performed in GraphPad Prism9 using a two-tailed Student's t-test.

ImageJ/FIJI software was used to quantify fluorescence intensity. All quantified images used the same imaging parameters within a given experiment. EcRE-GFP intensity was quantified by measuring the fluorescence within the wing disc pouch+hinge as defined by the outer edge of hinge Wg expression. To control for background variations across discs, a 52×52 pixel non-regenerating region in the notum was used to determine background EcRE-GFP expression. Ilp8-GFP intensity was quantified by measuring the total GFP fluorescence from the Ilp8-GFP-positive region of the disc using summed z-stacks. Intensity of βGal expressed from upd3-lacZ was quantified by measuring the fluorescence within the wing disc pouch+hinge as defined by the outer edge of hinge Wg expression from summed z-stacks through the disc. Statistical analysis was performed in GraphPad Prism9 using a two-tailed Student's t-test. Total n represents three separate ablation experiments.

ImageJ/FIJI software was used to quantify fluorescence intensity. All quantified images used the same imaging parameters within a given experiment. EdU intensity was measured from summed z-stacks of discs. For both intensity and area measurements, the pouch+hinge region of the disc was determined by the outer edge of hinge Wg expression. Percentage of total EdU fluorescence in pouch+hinge was determined: pouch+hinge EdU intensity/total disc EdU intensity. This allowed for normalization of any variation in incorporation between discs. Area of pouch+hinge was determined: area of pouch+hinge/total disc area. Statistical analysis was performed in GraphPad Prism9 using a two-tailed Student's t-test.

Z-stacks of 20 planes were taken through wing discs from the indicated genotypes. ImageJ/FIJI software for automatic cell counting was used to count the number of pH3-positive cells. Regions of the disc were determined by morphology and Wg antibody staining.

Survival curve analysis for pupariation timing experiments was performed in GraphPad Prism9 using the Log-rank (Mantel-Cox) test. Results are calculated per population, with n representing data from at least three separate ablation experiments.

We thank the Developmental Studies Hybridoma Bank, the Bloomington Drosophila Stock Center, the Harvard Transgenic RNAi Project and the Vienna Drosophila Resource Collection for antibodies and stocks. Fig. 4E created with BioRender.com. 7xEcRE-GFP was a gift of V. Henrich. Ablation stocks (rn-GAL4, tub-GAL80ts; rn-GAL4, tub-GAL80ts, UAS-rpr; rn-GAL4, tub-GAL80ts) were gifts of Rachel Smith-Bolton. We thank members of the Moberg laboratory for their discussions and Emory Integrated Cellular Imaging Core Facility for technical support.