Neuroendocrine Control of Intestinal Regeneration Through the Vascular Niche in Drosophila .

Robust and controlled intestinal regeneration is essential for the preservation of organismal health and wellbeing and involves reciprocal interactions between the intestinal epithelium and its microenvironment. While knowledge of regulatory roles of the microenvironment on the intestine is vast, how distinct perturbations within the intestinal epithelium may influence tailored responses from the microenvironment, remains understudied. Here, we present previously unknown signaling between enteroendocrine cells (EE), vasculature-like trachea (TTCs), and neurons, which drives regional and global stem cell proliferation during adult intestinal regeneration in Drosophila . Injury-induced ROS from midgut epithelial cells promotes the production and secretion of Dh31, the homolog of mammalian Calcitonin Gene-Related Peptide (CGRP), from anterior midgut EE cells. Dh31 from EE cells and neurons signal to Dh31 receptor within TTCs leading to cell autonomous production of the vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF)-like Pvf1. Tracheal derived Pvf1 induces remodeling of the tracheal stem cell niche and regenerative ISC proliferation through autocrine and paracrine Pvr/MAPK signalling, respectively. Interestingly, while EE Dh31 exerts broad control of ISC proliferation throughout the midgut, functions of the neuronal source of the ligand appear restricted to the posterior midgut. Altogether, our work has led to the discovery of a novel enteroendocrine/neuronal/vascular signaling network controlling global and domain specific ISC proliferation during adult intestinal regeneration.


Introduction
The high regenerative nature of the intestinal epithelium, which is essential for organismal wellbeing, relies on tight coordination of interdependent stem cell intrinsic processes and multicellular paracrine signals (Hageman et al., 2020, Zhou andBoutros, 2023).Consistently, research on intestinal regeneration requires in vivo model systems that account for the complex inter-cellular and inter-tissue signaling between the gut epithelium and its natural microenvironment.Due its physiological sophistication, vast single cell molecular information and richness of tools for spatially and temporally restricted genetic manipulations, the fruit fly Drosophila melanogaster has been an excellent in vivo model system for the study of multiorgan communication including the gut (Medina et al., 2022).
Akin to the mammalian intestine, the adult Drosophila midgut regenerates through the action of stem cells (Micchelli andPerrimon, 2005, Ohlstein andSpradling, 2005) operating under the control of signals emerging from a rich microenvironment, collectively referred to as the intestinal stem cell (ISC) niche (Jiang et al., 2011, Jiang et al., 2009a, Amcheslavsky et al., 2014).In the mammalian gastrointestinal tract, diverse cellular types, including intestinal epithelium resident cells and cells from the associated mesenchyme provide physical support, contractility, and multiple stem cell niche factors such as Wnts, BMP, EGFs and R-Spondin, which instruct ISC self-renewal in homeostatic and challenging conditions (Kim et al., 2020, McCarthy et al., 2020, Stzepourginski et al., 2017, Niec et al., 2022, Palikuqi et al., 2022).
In addition to ISCs, the fly midgut epithelium consists of two subtypes of undifferentiated stem cell progeny, namely, enteroblasts (EBs) and preenteroendocrine cells, which are precursors of absorptive enterocytes (ECs) and secretory enteroendocrine cells (EEs), respectively (Chen et al., 2018).As in mammals, cells within the fly midgut epithelium and gut associated tissues, such as the visceral muscle (VM), terminal tracheal cells (TTCs) and enteric neurons (ENs) are components of the intestinal microenvironment that contribute to robust ISC proliferation and differentiation as part of a complete intestinal regeneration program (Medina et al., 2022, Petsakou et al., 2023).
A large body of research and cutting-edge technology has led to the discovery of the multiple components of the intestinal microenvironment and the myriad of stem cell niche factors they produce to instruct ISC function during tissue regeneration (Hageman et al., 2020, Kim et al., 2020, Harnack et al., 2019).However, less is known about how local and/or challenge-specific changes within the intestinal epithelium might influence the intestinal microenvironment to achieve robust and regenerative responses of the intestine, tailored to the environmental demands.
Secretory EE cells within the Drosophila and mouse intestine are a highly diverse, specialized and compartmentalized progeny of ISCs with the ability to sense a wide range of cues from within the intestine and the external environment (Guo et al., 2019, Gehart et al., 2019, Beumer et al., 2020, Lebrun et al., 2017).Signals received by EE cells are translated into the production of peptide hormones acting at short range, in an autocrine or paracrine fashion, or in a more classical endocrine or systemic mode, via their release into the circulation (Gribble andReimann, 2016, Guo et al., 2022).The molecular and functional versatility of EE cells make them likely candidates to endow the gut with a highly refined regenerative response to the diversity of challenges faced by the intestinal epithelium.While nutrient regulated responses of EE cells and their constitutive function in the maintenance of intestinal homeostasis and growth have been recognized for decades (Scopelliti et al., 2014, Amcheslavsky et al., 2014, Martin et al., 2005, Drucker et al., 1996, Sasaki et al., 2001, Lin et al., 2022, Scopelliti et al., 2019), little is understood about the regulation and function of EE cells during intestinal regeneration.Recent datasets and publicly available toolkits for in situ gene expression analysis and independent genetic manipulation of EE peptide hormones and their cognate receptors, open a door for new and fundamental discoveries on the communication between EE cells and the intestinal microenvironment (Hung et al., 2020, Li et al., 2022, Liu et al., 2022b).
Here, we uncover multi-tissue signalling crosstalk between intestinal EE cells, gut associated vascular-like trachea and neurons, involving Calcitonin Gene-Related Peptide and PDGF/VEGF like factors and controlling global and domain specific intestinal regeneration in Drosophila.
We analysed the proportion of EE cells in adult Drosophila midguts from mated females by staining with an antibody against the transcription factor, Prospero (Pros), and observed that damage to the midgut epithelium caused by oral exposure with the entomopathogen Pseudomonas entomophila (Pe), the DNA-damage agent bleomycin or the basal membrane disruptor dextran sodium sulphate (DSS), caused a significant increase in Pros positive EE cells (Figure 1A, 1B).Mouse intestinal samples from animals subject to DNA damage by whole body irradiation, or with inflammatory colitis induced by treatment with DSS also showed a noticeable increase in the number of Chromogranin A (ChgA) positive EE cells in the colon (Figure 1C,1C',1D).No significant change in EE cell numbers was observed in the small intestine (Figure 1C'' and 1D).Taken together, our results suggest that increased EE cell proportion is a conserved feature of Drosophila and mouse intestines undergoing acute damage, which recapitulate observations made in the context of related human intestinal perturbations (Sciola et al., 2009, Zissimopoulos et al., 2014, Zhang et al., 2019).Next, we used Drosophila to investigate the nature of such EE phenotype and its functional significance to the intestinal response to damage.

Regulation of the EE-cell derived peptide hormone Dh31 is required to
fulfill the regenerative capacity of the midgut epithelium upon damage EE cells secrete peptide hormones that are constitutively required for the maintenance of ISC homeostasis and tissue growth in both fruit flies (Scopelliti et al., 2014, Amcheslavsky et al., 2014) and mammals (Drucker et al., 1996, Sasaki et al., 2001).However, the wide diversity and exquisite regionalization of EE cells in the gastrointestinal track (Beumer et al., 2020, Guo et al., 2019, Gehart et al., 2019) strongly suggest the potential for specialized and regulated responses of these cells to a wide range of stimuli.We therefore asked whether the global increase in EE cell numbers observed in damaged intestines may reflect on an EE cell sub-type specific response and/or function in tissue regeneration.We performed an RNA interference (RNAi) mini-screen to knockdown 9 different peptide hormones produced by EE cells using the temperature sensitive driver voilá-Gal4 ts , which labels all EEs in the midgut (Scopelliti et al., 2014).We measured intestinal regeneration by counting the number of proliferative ISCs (PH3 positive) within the posterior midgut of mated females fed overnight with the pathogen Pe.We observed that knocking down the expression of Related to the neuroepithelial characteristics of EE cells, peptide hormones secreted by those gut cells are also produced by neuronal cells in the brain.
Therefore, genetic drivers broadly targeting EE cells, including voilà-Gal4 and Dh31-Gal4 are expressed in both, the midgut and neurons (Scopelliti et al., 2014, Lin et al., 2022).To address EE specific functions of Dh31, we screened for a combination of genetic tools that would reduce Gal4 expression and therefore impair gene targeting within neurons, without compromising gene manipulations in EE cells (Figure S1C).
We found that the cholinergic line, ChAT-Gal80 (Kitamoto, 2002) strongly decreased GFP expression driven by voilá-Gal4 in central and peripheric neurons, with no effect on the midgut (Dicer-2;voilá ts ;Chat-gal80>GFP; Figure S1C).Using such genetic combination we observed that, while a significant reduction of Dh31 expression was observed in the heads of voilá ts driven Dh31-RNAi animals, Dh31 knockdown using voilá ts ,ChAT-Gal80 did not affect brain gene expression levels (Figure 1J).Consistently, we observed no reduction in Dh31 immunostaining in either the brain or ventral nerve cord (VNC) of voilà ts ,ChAT-Gal80>Dh31-IR animals, while midgut protein signal was undetectable (Figure 1K, S1D).Importantly, gut specific Dh31 knockdown was sufficient to impair regeneration in damaged midguts as revealed by reduced ISC proliferation in the anterior and posterior midgut (Figure 1L).However, overexpression of Dh31 or activation of Dh31 + EE cells through expression of the mammalian capsaicin receptor (VR1) was not sufficient to promote ISC proliferation in the absence of damage (Figure S1E, S1F).We noted that, in addition to EE cells, Dh31 was expressed in a subset of enteric neurons, most predominantly, in hindgut innervating neurons (Cognigni et al., 2011) (Figure 2A).Pan-neuronal knockdown of Dh31 under the control of nSyb-gal4 (nSyb ts ) decreased Dh31 levels in hindgut enteric neurons (Figure 2A') and gene expression in the head (Figure 2B) but did not affect Dh31 immunostaining in EE cells (Figure 2A').Interestingly, neuronal Dh31 knockdown resulted in reduced regenerative ISC proliferation restricted to the posterior midgut (Figure 2C).Altogether, these results indicate that EE cell derived Dh31 is broadly needed for damage induced regeneration of the adult Drosophila midgut.Furthermore, an additional source of the neuropeptide, produced by neurons, contributes locally to posterior midgut regeneration.

Dh31 production is regulated by intestinal damage
We next assessed the distribution of Dh31 + EE cells in the midgut using a Dh31-Gal4>mCD8-GFP gene expression reporter and an antiserum against Dh31 protein.Consistent with previous reports (Veenstra et al., 2008, Benguettat et al., 2018), we detected high levels of Dh31 immune-labelling and gene expression in posterior midgut regions R4c and R5 (Figure 3A, B', C Altogether, these results suggest domain specific upregulation of Dh31 mRNA and protein in response to pathogenic damage, which is restricted to the anterior midgut.EE cells store peptide hormones into vesicles and secrete them locally or systemically into the haemolymph following activation by specific stimuli in a process that involves increase in intracellular Calcium (Ca 2+ ) (Scopelliti et al., 2018, Lin et al., 2022).To further interrogate the nature of upregulated Dh31 upon midgut damage (Figure 3A-D), as to whether this may reflect increased peptide retention or higher peptide production and secretion, we measured Dh31 + EE cell activity ex vivo by assessing cytoplasmic Ca 2+ levels using GCaMP7, a genetically encoded calcium sensor.We expressed GCaMP7 in Dh31 + EE cells after different periods of 5% sucrose or Pe feeding.We detected increased Dh31 + cell activity in the anterior midgut as soon as 4h of infection, which remained consistently high after 8h and 16h of infection (Figure 3E, F).Co-expression of GCaMP7 and the nuclear reporter RedStinger, using a Dh31 knock-in Gal4 line Dh31-T2A-Gal4, showed that almost all Dh31 + EE cells were activated in the anterior midgut after 16h of Pe infection when compared to sucrose fed animals (Figure S2B).We did not detect changes in basal levels of Ca 2+ in Dh31 + EE cells within the posterior midgut, which remained constitutively high regardless of treatment (Figure 3G).Taken together, our results show a compartmentalized regulated response of EE cells to intestinal damage, resulting in increased production and secretion of Dh31 from the anterior compartment of the adult Drosophila midgut.

Gut epithelial derived reactive oxygen species induce Dh31 production in Drosophila.
Intestinal damage triggered by pathogenic bacteria infection generates high levels of reactive oxygen species (ROS) in the gut lumen, mainly via enterocytes (ECs), as a protective mechanism from the host against the pathogen (Morris and Jasper, 2021).Dh31 + EE cells are responsive to extracellular ROS generated by short-term bacterial infection (Benguettat et al., 2018).Blocking ROS in the adult Single cell and bulk RNAseq as well as loss of functional experiments, have previously demonstrated the presence of Dh31R in midgut ECs, EEs and the visceral muscle (VM) (Dutta et al., 2015, Benguettat et al., 2018, Hung et al., 2020).
We used a Dh31-R knock-in Gal4 reporter, which recapitulates the endogenous expression pattern of the Dh31-R gene (Deng et al., 2019), to express GFP (Dh31-R>GFP) and therefore visualize receptor expression in situ.Consistent with previous reports (Dutta et al., 2015, Benguettat et al., 2018, Hung et al., 2020), we observed GFP expression in VM and ECs in the anterior midgut (Figure 5A and S3A), we also observed expression in uncharacterised enteric neurons of the crop and anterior midgut, and PDF expressing enteric neurons (Cognigni et al., 2011)  We hypothesized that MAPK activating factors other than Bnl/FGF might be regulated by Dh31R signaling in the trachea.Analysis of our TTC Targeted DamID (TaDa) data (Perochon et al., 2021), revealed binding of RNA Pol II to genes encoding for Drosophila vascular endothelial growth factor (VEGF) and plateletderived growth factor (PDGF)-like Pvf1 and Pvf3 (Parsons and Foley, 2013) (Figure S4C).Autocrine Pvf/Pvr signalling drives homeostatic ISC self-renewal through MAPK activation (Bond and Foley, 2012).We hypothesized that tracheal Pvf may activate midgut regeneration by signalling paracrinally to Pvr expressed within ISCs.
Interestingly, RNAi knockdown of Pvf ligands from TTCs showed that Pvf1, but not

Local and global regulation of ISC proliferation by neuroendocrine Dh31 during intestinal regeneration
Our study provides new evidence of complex fine-tuning of intestinal regeneration trough multi-tissue signaling and the importance of damage dependent regulation of EE cell function and specific peptide hormone secretion in this process.
Even though EEs have been previously shown to contribute to intestinal homeostasis in flies and mammals (Amcheslavsky et al., 2014, Scopelliti et al., 2014, Drucker et al., 1996, Sasaki et al., 2001), our work represents the first report of EE and neuronal derived peptide hormone and PDGF/VEGF as a vascular stem cell niche regulating intestinal regeneration.
Interestingly, although damage regulated production and secretion of EE Dh31 is specific to the peptide produced in the anterior midgut, the impact of Dh31 on ISC proliferation in the regenerating midgut is global.A similar observation was described in another study, where Dh31 from the anterior midgut is necessary to promote peristaltic muscle contraction throughout the midgut via activation of Dh31-R in the visceral muscle (Benguettat et al., 2018).This could be explained by Dh31 acting not only at short range/paracrinally in the midgut, but also via its release into the extracellular space (Lin et al., 2022) and, therefore, targeting its receptor in regions of the midgut far away from where its production site.Additionally, the 3D conformation of the midgut, which brings the anterior and posterior regions into proximity (Buchon et al., 2013), could facilitate the communication between cells across gut sections.
The area of biggest need for high levels of Dh31 appears to be within the posterior region of the regenerating midgut, where we detected expression and exclusive functionality of an additional source of Dh31, provided by neurons (Figure 2).This may be related to the increased ISC numbers and regenerative activity of the posterior midgut, which is likely to require higher thresholds of proliferative factors and downstream activated signals than other, perhaps more quiescent, sections of the tissue.
Apart from its short-range action within the midgut and midgut associated tissues, Dh31 secretion into the hamolymph has clear systemic consequences (Lin et al., 2022), which we have yet to explore in the context of intestinal damage.EEderived Dh31 was recently shown to signal to neurons in the brain altering feeding and courtship behaviour (Lin et al., 2022).Dh31-R is highly expressed in neurons in the brain, ventral nerve cord and corpora allata (Deng et al., 2019, Kurogi et al., 2023).Additionally its isoform Dh31-R RC is widely expressed in the glia (Deng et al., 2019).Therefore, damage induced Dh31 from the midgut may regulate systemic traits, which could in turn be potentially instructive to the process of intestinal regeneration and beyond.
We observed a conserved phenomenon of increased EEs proportion during intestinal injury in both flies and mammals.Given the regenerative role of EEs described here and additional protective and immune roles ascribed to this cells (Benguettat et al., 2018, Kamareddine et al., 2018), inducible changes in EE cell numbers might be associated to broad biological functions like worthy of further investigation in flies and mammals.

Dh31 and its communication with the vascular intestinal stem cell niche
We show here that Dh31 participates in intestinal regeneration via communication with its receptors in the vascular-like tracheal component of the ISC niche.Interestingly, signal relay to associated tissues is a common mechanism used by peptide hormones in both flies as humans (Scopelliti et al., 2014, Amcheslavsky et al., 2014, Estall and Drucker, 2006, Baldassano and Amato, 2014), as perhaps the most efficient strategy to translate local EE cues into global tissue responses.
While our results suggest stronger impairment of midgut regeneration when knocking down Dh31 receptor in terminal tracheal cells compared to other niche cells, we cannot rule out a role for Dh31 signaling to receptor in other midgut associated tissues and epithelial cells, as we observed small but significant role in ISC proliferation when affecting Dh31R expression in such cellular compartments.
Dh31 has similarities with the human Calcitonin gene related peptide (CGRP), which is a splicing variant from the Calcitonin gene (Furuya et al., 2000).Similarly, CGRP have very diverse functions in organismal homeostasis, acting locally or systemically in multiple organs.However, CGRP is not found expressed in EE cells but mainly in immune cells and central and peripheral neurons (Russell et al., 2014).
Our evidence on the expression and function of neuronal derived Dh31 is a clear parallel to the latter.Indeed, as we observed here, CGRP communicates with the cardiovascular system and is known to be a potent vasodilator by activating cAMP pathway through its receptor in endothelial cells and the smooth muscle that surrounds them (Russell et al., 2014).Additionally, CGRP has also been reported to have anti and pro-inflammatory roles and facilitates wound healing possibly by local upregulation of growth factors such as VEGF, FGF, TGF-β (Toda et al., 2008, Mishima et al., 2011).In the intestine, new studies are beginning to show the importance of CGRP in inflammatory contexts (Xu et al., 2019, Manion et al., 2023).
However, its CGRP functions in tissue regeneration are largely unexplored and our study provides the first evidence of such a role in the fly midgut, prompting similar investigations in mammals.

Fly stocks and husbandry
A complete list of fly strains used in this paper is included in the Key Resources Table .All flies were kept in temperature-controlled incubators with a 12-12-hour light-dark cycle on cornmeal-based rearing medium.Fly stocks were kept at 18°C.
Crosses for experiments containing the temperature sensitive tubulin-Gal80 ts were set up at 18°C.Their progeny was kept for 3 days in fresh standard medium in the same temperature as parental crosses and then transferred to 29°C for 5-8 days, unless otherwise stated.Flies at 29°C were transferred to fresh medium every two days.Other crosses were set up and kept at 25°C.Only mated female flies were used for experiments.

Capsaicin treatment
Flies were fed a 5% sucrose solution (Mock) or sucrose containing capsaicin (50µM) for 16h applied on a glass microfiber filter.

Immunofluorescence
Guts from adult female flies were dissected in PBS and immediately fixed in 4% formaldehyde (FA, EM grade, Polysciences) diluted in PBS for 1h at room temperature (RT) and then washed three times in PBS-T (PBS + 0.2% Triton X100).
Samples were incubated with the primary antibody diluted in PBT (PBS-T + 2 % BSA) overnight at 4°C.Guts were washed three times in PBS-T and then incubated with secondary antibody in PBT for 2h at RT.After incubation, guts were washed three times for 20min in PBS-T and mounted on glass slides with 13mm x 0.12mm spacers (sigma) in Vectashield mounting medium with DAPI (Vector Laboratories).
Guts stained with anti-DH31, were blocked in 7% goat serum (in PBS-T) for 1h before primary antibody incubation.For anti-PVF1 and anti-pMAPK staining, fixation was performed in 4% formaldehyde for 1h, followed by 5min methanol fixation added to the sample gradually and then 10min in 100% methanol.Samples were blocked in 7% goat serum for 1h before primary antibody incubation.
For brains and ventral nerve cord (VNC) dissection, whole flies or fly heads were first washed in 70% ethanol for 30sec and fixed as described above.Heads were transferred to PBS-T and dissected.After, they were washed two times in PBS-T and blocked in 7% goat serum for at least 1h.Samples were incubated with primary antibodies in PBT for 48 hours at 4°C followed by the steps mentioned above.
A list of all antibodies and dilutions used in this study is included in the Key Resources Table.

Tissue imaging
Confocal microscope images were taken on a Zeiss LSM 710, Zeiss LSM 780 or Zeiss LSM 880 using identical acquisition conditions for all samples from a given experiment.Images were processed on ImageJ and Photoshop.
Alternatively, Images were taken with the automated confocal microscope Opera Phenix (Perkin Elmer) using a 5X objective for pre-scan and 20X water objective for re-scan.Images were processed and analysed with Harmony (Perkin Elmer) or ImageJ.

Quantifications from immunofluorescence images
Confocal images from the anterior and/or posterior midgut were used to quantify signal intensity using ImageJ.PVF1 signal was obtained from maximal projections of a limited number of Z-stacks for each terminal trachea cell.For Dh31 quantifications, whole gut images from the Opera Phenix were exported as Tiff files and stitched.
Regions of interest were cropped and quantified with ImageJ using a MACRO (Appendix 1).
Enteroendocrine cell proportion was obtained from whole midgut images acquired with Opera Phenix.Total number of Prospero+ objects were divided by the total number of DAPI+ objects using Harmony (Perkin Elmer).
Total number of pH3+ cells were scored manually on visual inspection using an epifluorescent microscope Olympus BX51 in the anterior and/or posterior midgut.

Terminal tracheal cells coverage
Whole posterior midgut images were dragged to ImageJ and only the green channel exhibiting TTCs labelled by GFP was analysed.Maximal projection of all stacks had their threshold adjusted without creating a noisy background.Images were then skeletonized and the exact area to be quantified was selected.Measurement of signal coverage in a specific area was analysed.

Analysis of calcium levels in Dh31+ cells
Crosses and rearing of UAS-GCaMP7s and Dh31-Gal4 were done at 25°C.5-8 days old mated female flies expressing GCaMP7s specifically in Dh31+ cells were treated with 5% sucrose solution (Mock) or sucrose containing Pe at an OD 600 = 50 for 16h on a glass microfiber filter.Whole guts were carefully dissected in Schneider's Drosophila-Medium (Gibco) and directly mounted on slides containing the same medium.Guts were immediately observed with a Zeiss Observer 7. The number of green fluorescent cells was scored manually by visual inspection in the anterior and posterior midgut.

RNA extraction and RT-qPCR
15 midguts, 30 anterior/posterior midguts or 30 heads were transferred immediately to 800 µL TRIzol (Ambion by Life technologies), smashed with a pestle and stored at -80°C until RNA extraction.RNA was extracted by precipitation.cDNA was obtained using the high-capacity cDNA reverse transcription kit (Applied Biosystems) following manufacturers protocol.Quantitative PCR was done at least in biological triplicates for each genotype/condition and run in technical triplicates on an Applied Biosystems 7500 Fast Real-Time PCR machine using PerfeCta SYBR Green FastMix (Quanta).
Data was analysed using QuantStudio design and analysis desktop software v1.4.3.
Expression of target genes was normalized to the internal control Rpl32 using standard curves.A list of primer sequences can be found in the Key Resources Table.

Mouse intestinal regeneration and immunohistochemistry
Mice (Mus musculus) were crossed to a C57BL/6 background and were subjected to 10Gy-72h before the experiment.For colitis induction, mice were fed a 2% solution of DSS in water and weighed daily.They were fed 2% DSS for 5 days and allowed to rest for 2 days.Any mouse that lost >20% of its body weight or exhibiting clinical signs such as hunching or excessive diarrhoea was humanely culled.Mice were then humanely killed by a rising concentration of CO2.The small intestine and colon were isolated and flushed with tap water.10 1-cm portions of intestine or colon were bound together with surgical tape and fixed in 10% neutral buffered formalin between 20 to 48h and then transferred to 70% ethanol and paraffinized.Paraffinized tissue were cut in 4µm sections and placed onto slides and incubated at 60°C overnight.
Before staining, the sections were dewaxed for 5min in xylene, followed by rehydration through decreasing concentrations of alcohol and a final wash with H 2 O for 5min.The formalin-fixed paraffin-embedded sections underwent heat-induced epitope retrieval in a Dako pretreatment module.

Dh31
caused the strongest and most consistent reduction in PH3 positive cells (Figure S1A, S1B), suggesting a preferential role for Dh31 producing EE cells in intestinal regeneration.Whole body Dh31 loss of function and EE cell gene knockdown using independent RNAi lines (Dh31-IR1 and Dh31-IR2) confirmed our initial observations that Dh31 is required for intestinal regeneration (Figure 1E, 1F).Similarly, knocking down Dh31 in Dh31 expressing cells (Dh31-Gal4 ts ) in animals subject to intestinal damage by oral administration of Pe, bleomycin or DSS, resulted in significantly reduced regenerative ISC proliferation (Figure 1G-I).
', D and S2A), and a low number of cells with weaker protein and gene expression signal in the rest of the homeostatic midgut (Figure 3A-D and S2A).On the other hand, oral infectious challenge to the intestinal epithelium with Pe, resulted in strong increase in immunolabelling of Dh31 + EE cells and gene expression in the anterior midgut (Figure 3B-D), with no significant changes detected in the posterior midgut (Figure 3B',C',D).
midgut by feeding Pe infected animals with the antioxidant N-acetyl cysteine (NAC) or by genetically inhibiting ROS production in ECs via DUOX knockdown led to a significant reduction in EE cell Dh31 immunolabelling in the anterior midgut (Figure 4A-D).Furthermore, feeding animals with a sucrose solution containing H 2 O 2 , a form of ROS commonly produced during bacterial infection, increased EE cell Dh31 immunolabelling specifically within the anterior midgut (Figure 4E-H).Therefore, ROS produced by midgut epithelial cells is necessary and sufficient to induce Dh31 production in EE cells of the adult anterior midgut.Tracheal Dh31 receptor signaling is required for ISC proliferation during midgut regeneration To understand how damage-induced Dh31 signals to promote intestinal regeneration, we next investigated the expression of the Dh31 receptor (Dh31R).
in the posterior midgut (Figure 5A and S3A).Interestingly, we also observed Dh31-R>GFP expression in terminal tracheal cells (TTCs) though out the full length of the midgut (Figure 5A, B).We used inducible RNAi lines to knockdown Dh31R from various midgut compartments and midgut associated tissues, including PDF producing neurons (Pdf ts >Dh31R-IR1), visceral muscle (how ts >Dh31R-IR1), ECs (Mex ts >Dh31R-IR1) and in TTCs (dSRF ts >Dh31R-IR1) (Figure S3B, C).While mild reduction in intestinal regeneration was observed upon Dh31R knockdown in the VM and ECs (Figure S3D,E), receptor downregulation in TTCs resulted in the most robust and consistent impairment of regenerative ISC proliferation throughout the midgut of Pe infected animals (Figure 5C, D and S3F) or fed with bleomycin (Figure 5E).Therefore, we focused the rest of our work on studying how Dh31 signalling to its tracheal receptor regulates intestinal regeneration.Dh31R signaling affects TTCs remodeling and MAPK activation in the regenerating midgut.Adult terminal tracheal cell plasticity and the production of stem cell niche factors by the trachea are essential to support the regenerative function of ISCs following midgut injury (Perochon et al., 2021, Tamamouna et al., 2021).Downregulating Dh31R in TTCs (dSRF ts >Dh31R-IR) led to significantly diminished TTC remodeling in Pe infected midguts (Figure 6A, B).These results suggest that the activation of Dh31/Dh31R signaling within adult gut trachea and its contribution to midgut regeneration, involves the induction of TTC remodeling.ROS inducible production of the Drosophila FGF-like ligand Branchless (Bnl) by the trachea and intestinal epithelium, drives regenerative ISC proliferation through paracrine activation of the FGF receptor Breathless (Btl) and downstream MAPK signaling in ISCs (Perochon et al., 2021, Tamamouna et al., 2021).While we observed significant downregulation of MAPK activation in ISCs from injured midguts with impaired tracheal Dh31R signaling (dSRF ts >Dh31R-IR) (Figure 6C, D), there was no impact on bnl upregulation (Figure S4A,B).These data suggest that Dh31R dependent tracheal remodeling and MAPK activation within ISCs is unlikely to be mediated by Bnl/Btl signaling.Tracheal derived Pvf1 drives midgut regeneration through activation of dual paracrine and autocrine Pvr/MAPK signaling.
Pvf3, was necessary to induce tracheal remodeling and ISC proliferation during midgut regeneration.Therefore, the functional role of Pvf1 produced by TTCs appears distinct from that of tracheal Bnl, which is redundant for TTC remodeling(Perochon et al., 2021).These results suggest that autocrine Pvf1/Pvr signalling within the trachea may impact ISC proliferation indirectly, through the regulation of TTC remodeling.Consistently, tracheal specific knock down of Pvr impairs TTC remodeling and MAPK signaling activity and inhibits ISC proliferation in regenerating midguts (Figure6H-L).Furthermore, knocking down Pvr in ISCs/EBs using the escargot-gal4 driver (esg ts >GFP) led to almost complete impairment of ISC proliferation and MAPK signalling activity in regenerating midguts (Figure6M-O).Altogether, our results uncover a new cellular source and dual signaling mode for Drosophila VEGF/PDGF-like ligand, Pvf1, regulating midgut regeneration from the tracheal stem cell niche via ISC and TTC receptors.This distinctive versatile characteristic of tracheal derived Pvf1 makes it a likely candidate for regulation by Dh31/Dh31R signaling during midgut regeneration.Dh31R signaling regulates tracheal Pvf1 production in the regenerating midgut To investigate a potential connection between Dh31/Dh31R signaling and Pvf1, we stained Dh31R-T2A-Gal4>mCD8-GFP midguts carrying endogenously tagged Pvf1 (Pvf1-HA) with anti-HA antibody (Figure 7A), which revealed colocalization between the two proteins.Immunostaining to detect either HA-tagged or unmodified Pvf1 revealed strong upregulation of Pvf1 in ECs and TTCs of Pe treated midguts (Figure 7B and S5A).Importantly, knocking down Dh31R from trachea (Figure 7B) or Dh31 from neurons (Figure 7D) impaired Pvf1 upregulation in TTCs of the posterior midgut (Figure 7C, E).Tracheal remodeling and Dh31 expression are highly dependent on damageinduced ROS production from the midgut epithelium (Perochon et al., 2021) (Figure 4), we therefore asked whether Pvf1 was also regulated by ROS.Indeed, Pvf1 upregulation was significantly prevented in Pe treated animals fed with the antioxidant NAC (Figure 7F, G).Altogether, our results uncover a novel EE/Neuronal/tracheal signaling network controlling global and domain specific ISC proliferation during adult midgut regeneration.Mechanistically, this interorgan signaling system involves regulated production of Dh31 by EE cells in the anterior midgut and constitutive sources of the ligand from EE cells in the posterior midgut and neurons.Dh31 signaling to Dh31R within TTCs induces the production of a previously uncharacterized source of Pvf1 from the trachea.Reciprocally, tracheal Pvf1 induces tracheal remodeling and regenerative ISC proliferation through autocrine and paracrine Pvr/MAPK activation, respectively (Figure 7H).

Figure 2 :
Figure 2: Neuronal-derived peptide hormone Dh31 is required to fulfill the regenerative capacity of the posterior midgut epithelium upon damage (A, A') Confocal images of nSyb>GFP expression (green) in the adult Drosophila hindgut and posterior midgut co-stained with and anti-Dh31 (red).Dashed squares indicate the magnified view presented below each full-size figure panel.Images corresponds to hindguts and midguts from control animals w 1118 (A) and animals with Dh31 KD (Dh31-IR2) under the neuronal driver nSyb-gal4 ts (A') and fed with Sucrose (A, A'; left panels) or Pe (A, A'; right panels).(B) RT-qPCR of head Dh31 mRNA levels in control animals or animals with Dh31 KD (Dh31-IR2) under the neuronal

Figure 3 :
Figure 3: Dh31 production is regulated by intestinal damage.(A) Schematic view of Dh31 expression pattern in the midgut.Scale from white (less expression) to red (more expression).(B) Expression levels of Dh31 detected by immunostaining upon treatment with Sucrose or Pe in the anterior (B) and posterior midgut (B').(C, C') Quantifications of Dh31 signal intensity levels and Dh31+ cell numbers in the anterior (C) and posterior midgut (C').(D) Dh31 mRNA expression value in anterior and posterior midguts from Sucrose and Pe treated animals.(E) Exvivo visualization of GCamp7c (green) expression in Dh31+ cells in the anterior midgut of flies treated with Sucrose or Pe for 0h, 4h, 8h and 16h.(F) Quantification of GCamp7 + cells as in E. (G) Ex-vivo visualization of GCamp7 expression in Dh31+ cells in the posterior midgut of flies untreated or treated with Sucrose or Pe for 16h.Dashed line delineate the edges of the midgut.(C, C', D) Two-tailed unpaired Student's t-test.(F) Two-way ANOVA followed by Tukey's multiple comparisons test.* (P<0.1)** (P<0.002)**** (P<0.0001).Scale bar = 100 µm.

Figure 4 :
Figure 4: Gut epithelial derived reactive oxygen species induce Dh31 production in Drosophila.(A) Expression levels of Dh31 (grey) detected by immunostaining upon treatment with Sucrose, Pe or Pe + NAC in the anterior midgut.(B) Quantifications of Dh31 signal intensity value and Dh31+ cell numbers in anterior midguts as in A. (C) Duox knockdown in ECs (Mex-gal4) and effect on Dh31 protein levels (grey) in anterior midguts of flies treated with Sucrose and Pe.(D) Quantifications of Dh31 signal intensity value and Dh31+ cell numbers in anterior midguts as in C. (E, G) Expression levels of Dh31 (grey) upon treatment with Sucrose or hydrogen peroxide (H 2 O 2 ) in anterior (E) and posterior (G) midguts.(F, H) Quantifications of Dh31 signal intensity value and Dh31+ cell numbers in anterior midguts as in E and G, respectively.Dashed lines delineate the edges of the midgut.(B)Ordinary one-way ANOVA and (D) Two-way ANOVA followed by Tukey's

Figure 6 :
Figure 6: Dh31R signaling affects TTCs remodeling and MAPK activation in the regenerating midgut through activation of Pvf1/Pvr signaling.(A) Confocal imaging of tracheal coverage (green) in posterior midguts following Dh31R knockdown (Dh31R-IR1) in TTCs in animals treated with Sucrose or Pe.(B) Quantification of tracheal coverage in midguts as in A. (C) Representative confocal images of p-ERK (red) immunostaining in ISCs/EBs (small cells stained with anti-Armadillo (Arm; white) from posterior midguts upon Dh31R KD (Dh31R-IR1) in TTCs of animals treated with Sucrose or Pe.p-ERK is indicated in a fire scale (blue = low, yellow = high) to facilitate signal intensity visualization.(D) Quantification of p-ERK signal intensity in ISCs/EBs of posterior midguts as in C. (E) Confocal imaging of tracheal coverage (green) in posterior midguts following Pvf1 knockdown (Pvf1-IR1) in TTCs and treated with Sucrose or Pe.(F, G) Quantification of tracheal coverage (F) and PH3+ ISCs (G) in posterior midguts a in E. (H) Confocal imaging of tracheal coverage (green) in posterior midguts following Pvr knockdown (Pvr-IR1) in TTCs and treated with Sucrose or Pe.(I, J) Quantification of tracheal coverage (I) and PH3+ ISCs (J) in posterior midgut a in H. (K) Representative confocal images of p-ERK immunostaining (red) in the posterior midgut upon Pvr KD (Pvr-IR1) in TTCs (green) after treatment with Pe. (L) Quantification of p-ERK signal in TTCs of The sections were heated in Target retrieval solution high pH (Dako, cat.no.K8004) for 20min at 97 °C before cooling to 65 °C.The slides were removed and washed in Tris-buffered saline with Tween