Abstract
Developmental studies revealed fundamental principles on how organ size and function is achieved, but less is known about organ adaptation to new physiological demands. In fruit flies, juvenile hormone induces intestinal stem cell (ISC) driven absorptive epithelial expansion balancing energy uptake with increased energy demands of pregnancy.
Here, we show 20-Hydroxy-Ecdysone (20HE)-signaling controlling organ homeostasis with physiological and pathological implications. Upon mating, ovarian 20HE acts on nearby midgut progenitors inducing Ecdysone-induced-protein-75B (E75B). Strikingly, the PPARγ-homologue E75B restricts ISC daughter cell differentiation towards absorptive enterocyte lineage ensuring epithelial growth even in the absence of local fate determining Notch. To our knowledge, this is the first time a systemic hormone is shown to direct local stem cell fate decisions. Given the protective, but mechanistically unclear role of steroid hormones in female colorectal cancer patients, our findings suggest a tumor-suppressive role for steroidal signalling by promoting postmitotic fate independent of deteriorated local signalling.
Introduction
Reproduction is an energetically costly process triggering multiple physiological adaptations of organs such as liver, pancreas and gastrointestinal tract upon pregnancy in various species (Hammond 1997, Roa & Tena-Sempere 2014). As a part of the hormonal response to mating and increased metabolic energy consumption, the female Drosophila melanogaster gut is remodeled in size and physiology by stimulating intestinal stem cell (ISC) driven epithelial expansion to achieve an even energy balance (Cognigni et al 2011, Klepsatel et al 2013, Reiff et al 2015).
Since the discovery of adult intestinal stem cells (Micchelli & Perrimon 2006, Ohlstein & Spradling 2006), local signaling pathways such as Notch (N), Jak/Stat, EGFR, Wnt/wingless, Insulin-receptor, Hippo/Warts and Dpp-signaling were shown to contribute to intestinal homeostasis under physiological and challenged conditions like bacterial infections (Miguel-Aliaga et al 2018)(and references therein). The midgut epithelium is maintained by ISC giving rise to only two types of differentiated cells: enteroendocrine cells (EE) and absorptive enterocytes (EC). Pluripotent ISC are able to self-renew or divide asymmetrically to give rise to committed EC precursor cells called enteroblasts (EB) and enteroendocrine precursor (EEP) cells, which upon timely activation of scute in ISC divide once more prior to terminal differentiation yielding a pair of EE (Chen et al 2018, Micchelli & Perrimon 2006, Ohlstein & Spradling 2006, Ohlstein & Spradling 2007). Nine out of ten ISC mitosis give rise to EB specified by N-activation in EB daughters (Micchelli & Perrimon 2006, Ohlstein & Spradling 2006, Ohlstein & Spradling 2007). However, post-mitotic EB retain a certain degree of plasticity by: (1) delaying their terminal differentiation through mesenchymal-to-epithelial transition (MET) (Antonello et al 2015a), (2) changing their fate to EE upon loss of lineage marker klu (klumpfuss) and (3) undergoing apoptosis as an additional homeostatic mechanism (Korzelius et al 2019, Reiff et al 2019).
Apart from local signaling pathways, systemically released JH is able to control ISC proliferation through heterodimers of Met (methoprene-tolerant) and gce (germ cells expressed) nuclear hormone receptors upon mating (Reiff et al 2015). JH coordinates Drosophila larval development in concert with the steroid hormone 20-hydroxy-ecdysone (20HE) (Bownes et al 1984, Gilbert et al 2002). In adult female flies, 20HE is released by the ovary and its titer increases upon mating (Ameku & Niwa 2016, Gilbert & Warren 2005, Harshman et al 1999), prompting us to investigate the Ecdysone-receptor (EcR) signaling cascade in organ plasticity during reproduction in the anatomically proximal posterior midgut (PMG). Downstream of EcR activation, we found upregulation of Ecdysone-induced protein 75B (E75B) splice variants ensuring absorptive EC production upon mating. Using the established Notch tumor paradigm, we found that 20HE remote controls EB differentiation independently of local fate determining Notch. The mechanism identified in this study not only plays a role in the physiology of mating, but also contributes to our understanding of the protective effects of steroid hormone signaling in the pathophysiology of human colorectal cancer.
Results
EcR controls intestinal stem cell proliferation and progenitor differentiation
The female fly intestine undergoes various physiological post-mating adaptations including a size increase of the absorptive epithelium (Cognigni et al 2011, Klepsatel et al 2013, Reiff et al 2015). Intrigued by postmating increases of 20HE titers (Ameku & Niwa 2016, Harshman et al 1999), we explored a role for EcR-signaling in mating adaptations of the adult Drosophila melanogaster intestine.
EcR encodes for three different splice variants: EcR.A, EcR.B1 and EcR.B2 (Cherbas et al 2003, Talbot et al 1993), which we detected by PCR in intestinal tissue with highest expression for EcR.B2 (Fig.S1B). Using pan-EcR antibodies, we found EcR in ISC (positive for the Notch ligand Delta+), EB (Notch responsive element, NRE-GFP+) and EC (Discs-large-1, Dlg-1+, Fig.S1C-F’”). To investigate a role for the EcR in intestinal tissue homeostasis, we first manipulated EcR function in ISC and EB using the ‘ReDDM’ (Repressible Dual Differential Marker, Fig.S1A) tracing method to observe its overall impact on tissue renewal (Antonello et al 2015a). Briefly, ReDDM differentially marks cells having active or inactive Gal4 expression with fluoro-phores of different stability. Combined with the enhancer trap esg-Gal4, active in progenitors (ISC and EB), esgReDDM double marks ISC and EB driving the expression of UAS-CD8::GFP (>CD8::GFP) with short half-life and >H2B::RFP with long half-life. Upon epithelial replenishment, newly differentiated EC and EE stemming from ISC divisions retain a RFP+-nuclear stain due to fluorophore stability (Antonello et al 2015a). Crosses are grown at 18°C in which transgene expression is repressed by ubiquitous tubulin-driven temperature sensitive Gal80ts. By shifting adult females to 29°C, Gal80ts is destabilized, in turn enabling spatiotemporal control of esgReDDM-tracing and additional UAS-driven transgenes in progenitors (Fig.S1A).
After seven days of tracing control virgin females (VF, Fig.1B) and mated females (MF, Fig.1C) using esgReDDM show increases progenitor numbers (Fig.1G) and newly generated EC (Fig.1H) in the R5 region of the PMG confirming previous observations (Reiff et al 2015). Reducing EcR-levels with two different >EcR-RNAi stocks in MF resulted in a reduction of ISC/EB numbers (Fig.1D,G,S1K) and newly generated EC (Fig.1D,H,S1K) to levels comparable to VF controls (Fig.1B,G,H). We addressed knockdown efficiency of >EcR-RNAi in esgReDDM by staining with EcR-antibodies and measuring fluorescence intensity of progenitor cells in situ and found EcR-signal significantly decreased in both RNAi lines (Fig.S1G).
Independent of EcR-abundance, we found that expressing dominant-negative EcR.B2 isoforms or EcR-heterozygosity using EcRM554fs, a well described loss-of-function (LOF) allele, phenocopy >EcR-RNAi (Fig.S1H,I,J). Generally, EcR LOF leads to a similar phenotype as JH-receptor knockdown in MF (Reiff et al 2015). To investigate whether EcR-levels affect progenitor behavior, we overexpressed wildtype EcR.B2 and pan-EcR using esgReDDM in VF. We found neither induction of progenitor numbers nor an increase in new EC (Fig.1E,F,G,H), suggesting that EcR-dependent proliferation and differentiation of progenitors might be limited by 20HE availability (Fig.1I).
Ecdysone is increased upon mating and actively transported into progenitors to adapt intestinal physiology to pregnancy
20HE is a polar steroid dispersing through hemolymph binding to EcR to activate target gene transcription (Bownes et al 1984, Gilbert et al 2002). As orally administered 20HE is metabolized and cleared of rapidly, we used the potent non-steroidal EcR agonist RH5849, developed and used as pest control due to its stability, specificity and a 30-60 times higher efficacy compared to 20HE (Robinson et al 2008, Robinson et al 1987, Wing et al 1988). Testing concentrations from 1-100μg/ml, we observed expected larval molting defects for concentrations from 50μg/ml upwards in timed egg-layings (Fig. S2A). Feeding control esgReDDM VF 50μg/ml RH5849, we traced intestinal progenitors with pharmacologically activated EcR-signaling for seven days. RH5849 strongly induces ISC mitosis, reflected by an tenfold increase in newly generated progeny over mating induction (Fig.2A-C,K). Interestingly, RH5849 mediated EcR-activation leads to no accumulation of progenitors (Fig.2J) as observed in oncogenic manipulations or disruptions of intestinal homeostasis, suggesting a role for EcR in both, proliferation and differentiation (Antonello et al 2015a, Chen et al 2016, Patel et al 2015, Reiff et al 2019). Given the role of 20HE induced programmed cell death (PCD) in larval metamorphosis and the recently discovered role of EB PCD in adult midgut homeostasis (Jiang et al 1997, Jiang et al 2000, Reiff et al 2019), we addressed PCD by activated caspase-3 staining, but found no increase of PCD by RH5849 (data not shown).
In close anatomical proximity to the PMG, the ovaries are an established source for 20HE and ecdyster-oidogenesis increases upon mating (Ameku & Niwa 2016, Harshman et al 1999). Using an established reporter for 20HE signaling activation, we found similar increases in the PMG and ovaries (Fig.2H). Cellular 20HE uptake was recently shown to depend on Ecdysone Importer (EcI) belonging to the evolutionary conserved SLCO superfamily of solute carrier transporters. EcI LOF causes phenotypes indistinguishable from 20HE and EcR deficiencies in vivo (Okamoto et al 2018). To investigate a role of EcI, we confirmed membrane localization of >EcI tagged with HA by immunostaining in progenitors using esgReDDM (Fig.S2B). After seven days, forced expression of >EcI using esgReDDM in VF lead to no increase in progenitor numbers and new EC, underlining that VF 20HE levels are low (Fig.2D,J,K). Further supporting this hypothesis, >EcI in MF lead to an increase of newly generated EC exceeding typical mating induction of MF controls by 4.8 fold (Fig.2E,J,K). Blocking 20HE uptake by EcI-RNAi abolished ecdysone induced tissue expansion to VF control levels (Fig.2F,G,J,K). Both, pharmacological and genetic experiments, suggest that 20HE levels and import control EcR-activity upon mating (Fig.2I).
Since we observed EcR in EC (Fig. S1C-F), we investigated its function in this cell type as it has previously been shown that mating through JH-signaling induces not only size adaptations through ISC proliferation, but also metabolic changes in absorptive EC by activating sterol regulatory element-binding protein (Srebp) upregulating lipid uptake (Reiff et al 2015). We addressed Srebp-activity using a GFP-reporter that reflects upregulation of lipid uptake gene expression (Athippozhy et al 2011, Reiff et al 2015) and confirmed previous observations showing a 2.2 fold increase in Srebp-activity (mean fluorescence intensity, Fig.S2C,D) (Reiff et al 2015). MF-flies with EcR-knockdown fail to induce Srebp-activity over VF controls (Fig.S2E-F,H), whereas feeding VF with RH5849 induced Srebp-activity 2.4 fold (Fig.S2G,H).
These data support the idea that mating related synergistic effects of JH- and 20HE-signalling drive changes of intestinal homeostasis and physiology. The JH-signal is transduced by the transcription factor Krüppel-homolog 1 (Kr-h1) in adult MF intestines, which prompted us to explore a function for the ‘classical’ Ecdysone target genes (Jindra et al 2013, Reiff et al 2015).
Ecdysone-induced protein 75B splice variants control ISC proliferation and enteroblast differentiation
During Drosophila development, 20HE pulses lead to binding of EcR directly to regulatory regions of early ecdysone response genes Ecdysone-induced protein 74A (E74A) and E75B (Bernardo et al 2014, Karim & Thummel 1991, Segraves & Hogness 1990) (Fig.3A). First, we performed conventional PCR to analyse JH- and Ecdysone target gene expression. Signal for Kr-h1-A (Kr-h1 in the following), but not Kr-h1-B, and for all splice variants of E74A and E75B was found (Fig.3B). Using quantitative real time PCR (qPCR) analysis, we confirmed an increase of JH-pathway activity upon mating using Kr-h1 as control (Fig.3C) (Reiff et al 2015). To our surprise, we found E74A expression unchanged, in contrast to its prominent role in the control of germline stem cell (GSC) proliferation in oogenesis (Ables & Drummond-Barbosa 2010). Instead, we found induction of E75B-A and -C splice variants (Fig.3C). E75B encodes for three protein isoforms (E75B-A, -B and –C) that differ in their N-terminal domain structure (Segraves & Hogness 1990) and E75B-B lacks one of the two zinc-finger DNA binding domains rendering it incapable of direct DNA binding (White et al 1997).
Intrigued by mating increases of E75B-A/-C levels, we manipulated E75B function using esgReDDM. Reducing E75B levels with RNAi, we found a significant increase in ISC and EB numbers (Fig.3D,E,I), but not in EC generation compared to VF controls (Fig.3J). In MF, E75B knockdown reduced newly generated EC compared to MF controls, pointing to a role of E75B in EB differentiation downstream of EcR-activation (Fig.3E,J). Additionally, we investigated homozygous E75B-A mutants using MARCM clonal analysis with a known LOF allele (Rabinovich et al 2016). MARCM clones of E75B-A (E75A81) are significantly larger compared to controls (Fig.S4A-D) and contain few clonal EC (GFP+/Dlg-1+, Fig.S4C) reflecting disturbed differentiation (Lee & Luo 1999). Notably, we did not detect EE in E75B-A clones by co-staining for the EEmarker prospero (GFP+/pros+, arrowheads Fig.S4C). These LOF experiments hint to a role for mating induced E75B-A/-C splice variants in differentiation of progenitors downstream of EcR.
Forced expression of E75B splice variants with esgReDDM resulted in three prominent phenotypes: (1) E75B-A and E75B-C strongly drive differentiation into EC (Dlg-1+, Fig.3F,H,J) depleting the entire progenitor pool (Fig.3F,H,I). (2) The E75B-B isoform induces ISC proliferation in VF (Fig 3G) and raised progenitor numbers suggesting slowed down but not inhibited terminal differentiation (Fig.3G,I,J). (3) All E75B manipulations block EE differentiation (Fig.S4A) possibly reflecting a general block for E75B-RNAi and E75B-B isoform (Fig.3E,G), whereas E75B-A and E75B-C instantaneously drive progenitors into EC differentiation preventing further ISC division and thus EE differentiation (Fig.3I,J). Manipulating tissue homeostasis alters midgut length (Hudry et al 2016) and in line with this, E75B manipulations also reduce midgut length by either slowing down terminal differentiation (E75B-B and E75B-RNAi) or depleting the progenitor pool (E75B-A/–C, Fig.3K). These findings strongly suggest a role for E75B in EB differentiation.
After initial N specification, EB lineage is maintained by the N target klumpfuss (klu). klu+-EB retain some plasticity to change fate to EE upon reduction of klu-levels (Korzelius et al 2019, Reiff et al 2019). We used kluReDDM to explore the function of EcR and E75B in EB lineage identity (Fig.S4B). No changes in EB fate were observed upon knockdown of EcR and E75B (Fig.S4A,C-E). E75B-A and E75B-C expression with kluReDDM phenocopied EB differentiation effects observed in esgReDDM (Fig.S4H,J,K,L) without non-autonomously inducing proliferation in wild-type ISC of VF and MF (Fig.S4K,L). Expressing E75B-B in kluReDDM lead to delayed EB differentiation similar to esgReDDM and non-autonomous induction of ISC mitosis (Fig.3G, Fig.S4I, Fig.S4M) (Patel et al 2015, Reiff et al 2019). Having identified E75B-A/-C as mating induced differentiation effector of 20HE signaling, we hypothesized and investigated a synergism of 20HE and JH hormonal signaling pathways controlling epithelial expansion upon mating.
The interplay between JH and 20HE in intestinal progenitors
Our data suggests that mating induced JH and 20HE signaling may sequentially affect ISC proliferation and EB differentiation through their effectors Kr-h1 and E75B (Fig.4A) (Reiff et al 2015). In VF lacking mating induction of both hormones (Ameku & Niwa 2016, Harshman et al 1999, Reiff et al 2015), forced >Kr-h1 expression crossed to esgReDDM doubles progenitor reproducing previous results (Fig.4C,J) (Reiff et al 2015). Initially, we aimed to perform a full genetic epistatis analysis for E75B and Kr-h1, but this analysis was hampered as we failed to recombine Kr-h1-RNAi with E75B splice variant expression stocks.
However, we were able to analyze midguts of MF with simultaneous >Kr-h1-RNAi/E75B-RNAi (Fig.4D), which did not show any mating induction as could be expected from single >Kr-h1-RNAi (Reiff et al 2015). Forced Kr-h1-expression and simultaneous E75B-RNAi increase progenitor numbers with only few new EC (Fig.4E,F), supporting the necessity of E75B in EB differentiation (Fig.3E) and Kr-h1 in ISC proliferation (Fig.4C). Surprisingly, we observed a strong increase of progenitor numbers forcing Kr-h1 and E75B-B expression (Fig.4H,J). These results indicate an additive role of Ecdysone and JH-signaling on ISC proliferation forcing E75B-B and Kr-h1-expression, whereas the latter transduces the mating related proliferation response (Fig.3C).
Most interestingly, E75B-A and –C, despite of raised Kr-h1-levels, drive progenitors into EC differentiation largely phenocopying sole E75B-A and -C expression (compare Fig.3F,H and Fig.4G,I). Our data on E75B-A and -C corroborates a potent role for Ecdysone-induced EB differentiation. ISC fate decisions in the adult midgut rely on N-signaling (Micchelli & Perrimon 2006, Ohlstein & Spradling 2006, Ohlstein & Spradling 2007). During Drosophila follicle cell development N opposes EcR-function (Sun et al 2008), which prompted us to investigate E75B-A/-C as effectors of EcR-signaling in the context of N-dependent EB differentiation.
Ecdysone signaling overcomes Notch deficiency through E75B-A and E75B-C remote controlling enteroblast differentiation
In the Drosophila and mammalian intestinal stem cell niche, N-signaling specifies EE and EC fate (Jensen et al 2000, Micchelli & Perrimon 2006, Ohlstein & Spradling 2006, Ohlstein & Spradling 2007, VanDussen et al 2012). In Drosophila, N mutant ISC produce ISC-like progenitor cells and EE entirely lacking EC generation, resulting in an intestinal epithelium accumulating a lack of newly produced EC and compensatory proliferation (Fig.5A,B) (Chen et al 2018, Guo & Ohlstein 2015, Micchelli & Perrimon 2006, Ohlstein & Spradling 2006, Ohlstein & Spradling 2007, Patel et al 2015).
Using N-LOF tumors as an experimental paradigm lacking EC production, we sought to investigate the differentiation inducing properties of E75B-A/-C. Reduction of N by RNAi or a dominant-negative N-receptor using esgReDDM lead to tumors of different sizes, which we quantified and classified according to their number (Fig.5B). We predominantly found tumors of 5-10 cells (Fig.5B, S5A), but also few tumors containing more than 100 ISC/EE-like cells with indistinguishable single or multiple ISC origin (class IV, Fig.5B,S5A).
Co-expression of E75B-A/-C with N-LOF strongly reduced total tumor number (Fig.5C,D,F) and smaller class I, II and III tumors (Fig.S5A), whereas E75B-RNAi results in confluent class IV tumors all along the PMG (Fig.5E). Although tumor number is strongly reduced (Fig.5F), occasional class IV tumors (<1/PMG) are able to escape suppressive E75B-A/-C showing no reduction in tumor size (Fig.S5A), suggesting that further tumorigenic mechanisms such as Upd- or EGF-ligands counteract the tumor suppressive function of 20HE signaling in class IV tumors (Patel et al 2015).
Since E75B-A and E75B-C re-enable EC differentiation despite the lack of N (Fig.5C,D Dlg-1+/RFP+), we investigated whether EcR-pathway activation is able to trigger EC fate in a N-LOF context. We activated EcR-signalling with RH5849 (Fig.6B) and raised intracellular levels of 20HE expressing >EcI in esgReDDM(Fig.6D). Either manipulation resulted in numerous confluent tumors reflecting the previously observed role of 20HE in proliferation (Fig.6F). In line with this, >EcI- and >EcR-RNAi significantly reduced tumor burden (Fig.6C,E,F).
More importantly, ecdysone-pathway activation led to significantly higher numbers of newly generated EC (arrowheads in Fig.6B,D,G), suggesting an upregulation of E75B-A and E75B-C in Notch tumors upon EcR-activation through RH5849 and >EcI. To exclude remaining N activity, we generated N mutant MARCM clones for N (N55e11) (Guo & Ohlstein 2015). In agreement with our previous observations activating 20HE signaling in N clones led to a tenfold increase in EC numbers over controls (Fig.S6A-C) and forced expression of >E75A/-C to instant EC formation (Fig.S6D-E).
Taken together, our results implicate EcR-signaling in control of ISC proliferation, EB differentiation and metabolic adaptation of EC. Importantly, our data establishes 20HE as a systemic signal controlling reproductive adaptations in synergism with JH. Mating induced E75B-A and E75B-C act as differentiation effectors of EcR activity in EB inducing remote fate decisions. In neoplasms marked by a lack of fate specification (e.g. Notch-deficiency), our data suggests a new mechanism how steroid hormone signaling suppresses tumor growth by promoting post-mitotic cell fate.
Discussion
Systemic hormones and intestinal remodeling upon mating
Growing offspring involves significant metabolic adaptations to raised energy demands in mothers. Drosophila melanogaster uses a classical r-selected reproductive strategy with high number of offspring and minimal parental care (Pianka 1970). Upon mating, egg production is increased tenfold supported by metabolic and behavioral adaptations such as food intake and transit, nutrient preference and a net increase of absorptive tissue in the PMG depending on JH release (Carvalho et al 2006, Cognigni et al 2011, Reiff et al 2015, Ribeiro & Dickson 2010).
Here, we describe a role for the steroid-hormone ecdysone controlling intestinal tissue remodeling upon mating through EcR-signaling. The unliganded heterodimer of EcR and ultraspiracle (USP) binds its DNA consensus sequences serving as a repressor that turns into an activator upon ligand binding inducing early response genes E74A, E75B and broad (br) (Schwedes et al 2011, Uyehara & McKay 2019). Initially, we investigated knockdown of USP, br and E74A in intestinal progenitors and observed effects on progenitor survival independent of mating status (Fig.3C, Zipper and Reiff, unpublished data).
Our data supports the idea that 20HE and JH synergistically concert intestinal adaptations balancing nutrient uptake to increased energy demands upon mating (Fig.1-4). Mating induces ovarian ecdysteroid biosynthesis resulting in increased 20HE titers inducing egg production by ovarian GSC (Ables & Drummond-Barbosa 2010, Ameku & Niwa 2016, Ameku et al 2017, Harshman et al 1999, König et al 2011, Morris & Spradling 2012, Uyehara & McKay 2019). In the adjacent PMG, we describe the impact of ovarian 20HE on E75B-A/-C expression (Fig.6H) and show that 20HE import triggers EcR-activation: i) in VF with low ovarian 20HE levels, overexpression of wild-type EcR and EcI are unable to elicit neither ISC proliferation nor differentiation effects. ii) pharmacological activation using RH5849 as well as (iii) forced EcI expression strongly induce proliferation and EB differentiation in MF mediated by E75B-A/-C.
The interplay between E75B and Kr-h1 controls intestinal size adaptation
Downstream of EcR, we found differential intestinal regulation of E75B splice variants upon mating. E75B activates egg production in ovarian GSC (Ables & Drummond-Barbosa 2010, König et al 2011, Morris & Spradling 2012, Uyehara & McKay 2019) and its knockdown reduces germarial size and number. In accordance with differentiation defects found in EB in our work, the number of cystoblasts, the immediate daughters of GSC, is reduced as well (Ables & Drummond-Barbosa 2010, Morris & Spradling 2012).
In our study, we dissected splice variant specific roles for E75B isoforms. E75B-B expression is at the lower detection limit and unchanged upon mating (Fig.3C). However, forced E75B-B expression raised ISC mitosis in VF (Fig.S4M). E75B-B mutants are viable and fertile and E75B-B is interacting with DNA only upon forming heterodimers with Hormone receptor 3 (Hr3) temporarily repressing gene expression (Bialecki et al 2002, Sullivan & Thummel 2003, White et al 1997). Further investigation of E75B-B, especially transcriptional regulation, is necessary to elucidate in which physiological context, else than mating, E75B-B controls ISC proliferation.
Mating induction of E75B-A/-C, but not E75B-B, suggests different transcriptional regulation. This idea is supported by the finding that larval epidermis from Manduca sexta exposed to heightened JH-levels is sensitized to 20HE, which results in a 10-fold increase of E75B-A, but not of E75B-B expression (Dubrovskaya et al 2004, Zhou et al 1998). An in silico analysis of 4kb regulatory regions upstream of Kr-h1 using JASPAR revealed several EcR/USP and Met consensus sequences (Khan et al 2018), suggesting that both hormonal inputs may also converge on Kr-h1 driving ISC proliferation (Fig.4D) (Reiff et al 2015). We propose a synergistic interplay between JH- and 20HE-signaling upon mating, in which the JH pathway effector Kr-h1 increases proliferation in ISC and the EcR effectors E75B-A/-C drive EB into differentiation thereby ensuring mating induced organ size growth. The experimental investigation of E75B and Kr-h1 regulatory regions incorporating endocrine signals from at least two different organs and leading to differential activity and function in ISC and EB will be a fascinating topic for future studies.
E75B/PPARγ and progenitor fate commitment in physiology and pathology
A key discovery of our study is the role of E75B-A/-C in progenitor differentiation downstream of EcR activation. To our knowledge, this is the first time that N independent EB fate determination is reported. Notch signaling specifies EB as EC committed progenitors by the activation klumpfuss (Korzelius et al 2019, Reiff et al 2019). When tissue demand arises, EB are thought to physically separate from ISC and become independent of N. Detached EB acquire motility and are able to postpone their terminal epithelial integration after initial differentiation (Antonello et al 2015a, Antonello et al 2015b, Martin et al 2018, Siudeja et al 2015).
The ecdysone pathway acting through E75B-A/-C may provide such a signal for terminal EB differentiation, which is supported by several lines of evidence: i) EB-specific knockdown of E75B stalls their differentiation using kluReDDM (Fig.S4E), suggesting E75B activity after initial N-input. ii) klu+-EB do not change fate upon E75B knockdown, suggesting a more mature EB differentiation status that already lost plasticity (Fig.S4A) (Korzelius et al 2019, Reiff et al 2019). iii) Forced E75B-A/-C expression in EB leads to their immediate differentiation independent of N-input (Fig.5C,D,S6D,E). Supporting an EB specific role, E75B was found to be highly expressed in progenitors in a recent study from the Perrimon lab (Hung et al 2018).
E75B is a long term predicted PPARγ-homologue (peroxisome proliferator-activated receptor gamma) in Drosophila, which got recently confirmed in pharmacogenetic approaches demonstrating that E75B-mutant flies are irresponsive to PPARγ activating drugs (Joardar et al 2015, King-Jones & Thummel 2005). This homology is of particular interest, as human PPARγ plays a role in i) pregnancy related adaptations of lipid metabolism (Waite et al 2000) and ii) as target in colorectal cancer (CRC) (Sarraf et al 1999).
Decreased PPARγ levels are associated with development of insulin resistance and failure of lipolysis in obese pregnant women (Rodriguez-Cuenca et al 2012, Vivas et al 2016). Here, we found an EcR-dependent upregulation of Srebp and PPARγ/E75B in adult female intestines that activate genes with known function in fatty acid synthesis and activation (Horton et al 2003, Reiff et al 2015, Seegmiller et al 2002). This interplay between SREBP, PPARγ and steroid hormone signaling needs to be dissected in detail to reveal to which extent adaptation of intestinal size and lipid metabolism contribute to pathophysiological conditions like diabetes.
In CRC, PPARγ loss-of-function mutations are observed in around 8% of patients (Sarraf et al 1999). PPARγ expression correlates with good prognosis and PPARγ is epigenetically silenced during CRC progression (Ogino et al 2009, Pancione et al 2010). In agreement with our in vivo experiments on E75B-A/-C in EB differentiation (Figs.3,5,6), in vitro studies identified PPARγ promoting differentiation in various colon cancer cell lines (Cesario et al 2006, Shimizu & Moriwaki 2008, Yamazaki et al 2007, Yoshizumi et al 2004).
In a CRC mouse model, biallelic loss of PPARγ leads to a 4-fold increase tumor incidence and reduced survival in female mice over males (ApcMin/+/PPARγ-/-) (McAlpine et al 2006), whereas males develop around three times more colon tumors with wild type PPARγ (Cooper et al 2005). In line with this, ovariectomized ApcMin/+ mice develop significantly more tumors and have decreased PPARγ expression, highlighting a possible link between PPARγ and estrogen signaling. Meta-analyses of hormone replacement studies have shown that estrogen confers a lower risk to develop CRC and better survival rates for female CRC patients (Chen et al 1998, Hendifar et al 2009, Lin et al 2012).
It is tempting to speculate that this mechanism of steroid hormones signaling through EcR/ER (estrogen receptor) to activate E75B/PPARγ is conserved from flies to humans. Indeed, human estradiol and 20HE activate the EcR with similar affinity and elicit E75B transcription by binding to EcRE (Ecdysone-Responsive Elements). EcRE can be converted to functional ERE (estrogen responsive elements) by a simple change of nucleotide spacing suggesting high conservation (Martinez et al 1991, Schwedes et al 2011).
Taken together, our findings reveal that mating induced steroid hormone release from the ovaries signal to adjacent intestinal stem cells controlling their proliferation and, more importantly, differentiation of committed precursor cells through E75B/PPARγ. Remote control of cell fate ensures absorptive epithelial cell production during mating related intestinal adaptations and in a Drosophila intestinal tumor model marked by loss of enterocyte fate specification. E75B/PPARγ acts independent of Notch-signaling, the only currently known precursor fate specifier in the intestinal stem cell niche. Mechanistically, we propose that E75B/PPARγ exerts an anti-neoplastic role by promoting progenitor differentiation into postmitotic enterocyte fate, thereby reducing the pool of mitotically active cells. Several studies describe pleiotropic roles for PPARγ. Combining knowledge about PPARγ protein structure and a distinct function of PPARγ in intestinal homeotasis may facilitate future studies using already available agonists and antagonists in the search for better treatment of colorectal cancer.
Material and Methods
Genetics and fly husbandry/ Fly strains
OregonR and w1118 flies served as wild-type controls. The following transgenes and mutations were employed: esgReDDM (Antonello et al 2015), kluReDDM (Reiff et al 2019), UAS-EcI-Flag-HA (Okamoto et al 2018), UAS-EcI (Okamoto et al 2018), UAS-EcI-RNAi (Okamoto et al 2018), UAS-E75B-A-Flag (Rabinovich et al 2016), UAS-E75B-B-Flag (Rabinovich et al 2016), UAS-E75B-C-Flag (Rabinovich et al 2016), UAS-NDN (J. Treisman), Dl::GFP (F. Schweisguth), N55e11 FRT19A (Guo & Ohlstein, 2015), Srebp-GAL4 (Kunte et al 2006). From Bloomington (BDSC): UAS-EcR-RNAi (BL58286), UAS-EcR.B2 (BL4934), UAS-EcR.B2W650A (BL9449), UAS-EcR.B2F645A (BL9450), EcRM554fs (BL4894), EcRE-lacZ (BL4516), E75A81(BL23654), E7507041 (BL11712), E75Δ51 (BL23652), UAS-CD8::GFP (BL5137), ovoD1 (BL1309), NRE::GFP (BL30727), NRE::GFP (BL30728), From Kyoto: UAS-Kr-h1 (DGRC120052). From FlyORF: UAS-EcR-HAF000480, UAS-Kr-h1F000495, UAS-MetF000644, UAS-gceF001915, From Vienna (VDRC) UAS-EcR-RNAiGD37059, UAS-E75B-RNAiGD44851, UAS-E75B-RNAiKK108399, UAS-N-RNAiGD14477, UAS-Met-RNAiKK100638, UAS-gce-RNAiKK101814.
MARCM clones
Mosaic analysis with a repressible cell marker (MARCM; Lee & Luo, 1999) clones were induced in midguts by Flippase under control of a heat-shock promoter. Expression of the Flippase was activated by 45 min in a 37°C-water bath to induce positively marked clones. The guts were dissected 5 days after induction. Clones in experimental and control flies were induced in parallel.
Food composition and fly keeping
Fly food contained 1424 g corn meal, 900 g malt extract, 800 g sugar beet syrup, 336 g dried yeast, 190 g soy fluor, 100 g agarose, 90 ml propionic acid and 30 g NIPAGIN powder (antimycotic agent) in 20 l H2O. The food was cooked for about an hour to reduce bioburden, then filled into small plastic vials and cooled down at RT. Flies were kept at 25°C except for crosses with temperature-sensitive GAL80ts (GAL4 repressor) which were kept at 18°C (permissive temperature) until shifted to 29°C (restrictive temperature) to activate GAL4-mediated transgene expression. Crosses with esgReDDM and kluReDDM were carried out as described previously (Antonello et al., 2015; Reiff et al., 2015, 2018). Due to persisting problems with mucous formation on food surface in vials with virgin females, all experiments distinguishing between mated and virgin female flies were run on food with twice the amount of NIPAGIN. Mucous formation was avoided because of massive induction of tissue renewal by pathogenic stress.
Hormone analogue treatments
A vial of fly food was reheated in the microwave until it became liquid, the hormone analogues were added, thoroughly mixed and filled into a new vial. For each ml of food 5 μl of RH5849 (340 μM final concentration; 20 μg/μl stock solution, diluted in MeOH; DRE-C16813000, DrEhrenstorfer™) and/or 0.5 μl of Methoprene (1.49 mM final concentration; PESTANAL®, analytical standard; Sigma-Aldrich; Reiff et al., 2015) were added. As a control, an equivalent volume of carrier solution (MeOH) was added to the food. Hormone analogue treatments were performed for the period of the 7 days ReDDM (Antonello et al., 2015) shift.
Immunohistochemistry
Guts were dissected in PBS and were transferred into 4% PFA immediately after dissection. The staining was performed on an orbital shaker. After 45 min of fixation the guts were washed one time with PBS for 10 min. Antibodies were diluted in 0.5% PBT + 5% normal goat serum. The incubation with primary antibodies (1:250 anti-Dlg-1 [mouse; Developmental studies Hybridoma Bank (DSHB)]; 1:50 anti-Pros [mouse; DSHB]; 1:2000 anti-pH3 [rabbit; Merck Millipore, 06-570]; 1:50 anti-EcR common Ag10.2 [mouse; DSHB]; 1:500 anti-HA High Affinity 3F10 [rat; Merck, Sigma-Aldrich]) was performed at 4°C over night. After a second washing step with PBS the guts were incubated with secondary antibodies (1:500 Goat anti-MouseAlexa647 [Invitrogen]; 1:500 Goat anti-RatAlexa647 [Invitrogen]; 1:500 Goat anti-RabbitAlexa647 [Invitrogen]) and DAPI (1:1000; 100 μg/ml stock solution in 0.18 M Tris pH 7.4; DAPI No. 18860, Serva, Heidelberg) for at least 3 h at RT. Guts were washed a last time with PBS prior to mounting in Fluoromount-G™ Mounting Medium (Electron Microscopy Sciences, emsdiasum).
Image acquisition
Posterior midguts were imaged using an LSM 880 Airyscan confocal microscope (Carl Zeiss) using “Plan-Apochromat 10x/0.45 M27”, “Plan-Apochromat 20x/0.8 M27” and “C-Apochromat 40x/1.20 W Corr M27” objectives. Image resolution was set to at least 2048×2048 pixels. At least three focal planes (1 μm distance) were combined into a Z-stack to image one cell layer and to compensate for gut curvature.
For determining whole midgut length, an Axio Zoom.V16 (Carl Zeiss) was employed with DAPI filter and 1x/0.25-objective.
JASPAR
The open-access webtool JASPAR (Khan et al., 2018) was utilized to predict transcription factor binding sites within the 5’-UTR of EcI. It was specifically scanned for binding motifs related to Ecdysone and Juvenile Hormone signaling.
Quantification of proliferation and intensity measurements
Maximum intensity projections were calculated from Z-stack images of posterior midguts by Fiji (ImageJ 1.51n, Wayne Rasband, National Institutes of Health, USA). Total cell number and RFP-positive cell count of ReDDM (Antonello et al., 2015) guts were analyzed semi-automatically by a self-written macro for Fiji whereas GFP-positive cells were counted manually (available from the authors).
For fluorescence intensity measurements, intestines were scanned with fixed laser settings and measured in Fiji. The region of interest was selected manually and mean intensity of the area was determined. This way, relative EcR protein levels were measured in antibody stainings and relative SREBP activity was analyzed in PMG cells expressing mCD8::GFP under the control of Srebp-GAL4.
RNA isolation and cDNA synthesis
The R5 regions or ovaries of at least 3 female flies were dissected and transferred into a droplet of RNAlater™ Solution (Invitrogen by Thermo Fisher Scientific) on ice. The tissue was homogenized in 100 μl peqGOLD TriFast (VWR Life Science) and total RNA was isolated as specified by the manufacturer. The following cDNA synthesis was performed with 250 ng of total RNA and the SuperScript™ IV Reverse Transcriptase (Invitrogen by Thermo Fisher Scientific) using a 1:1 mixture of oligo-dT primers and random hexamers directly upon RNA isolation. Prior to Real-time qPCR, cDNA samples were diluted 1:2 in dH2O.
Verification of gene expression in the adult midgut by PCR
To verify gene expression in the adult Drosophila midgut, PCRs were performed with primer pairs specific for ovo or isoform-specific primer pairs for EcR spanning at least one exon-exon boundary. PCRs were performed with Q5® High-Fidelity DNA Polymerase (NEB) for at least 30 cycles. Reaction products were loaded on an agarose gel (1.5%) and separated by electrophoresis to verify lengths of PCR products.
20-HE isolation and titer determination
20-HE titers were determined in VF and MF of the same age. Eclosed OregonR or heterozygous ovoD1 mutant flies were aged for 3d before mating. After 24 h or 48 h at least 20 adult female flies were pierced in the thorax with a needle and put into a 150 μl-PCR tube that was punctured in its very bottom. Hemolymph was harvested by centrifugation (5.000x g 5min RT) and collected in a 1.5 ml-reaction tube. Total weight of flies was determined before and after centrifugation. Typical yields of hemolymph were around 0.6-1 mg. The isolated hemolymph was mixed with 500 μl MeOH, centrifuged (12.000x g 20 min 4°C) and the supernatant was transferred into a new 1.5 ml-reaction tube. MeOH was evaporated at 30°C in a vacuum centrifuge (Eppendorf concentrator plus) and 20HE was resuspended in 100 μl EIA buffer and stored at −20°C until usage.
Ecdysone levels were determined using the 20-Hydroxyecdysone Enzyme Immunoassay kit according to manufacturer’s instructions (#A05120.96 wells; Bertin Bioreagent). 20HE titer was normalized to hemolymph yield.
Real-time qPCR
Expression levels of Ecdysone signaling pathway-associated genes were determined in VF and MF. Eclosed OregonR or EcRE-LacZ flies were aged for 4d before mating. After 72 h of mating, RNA was isolated and cDNA synthetized before running qPCRs. After an enzyme activation step (20 sec 95°C), 40 cycles of denaturation (2 sec 95°C), primer annealing (20 sec 58°C) and elongation (20 sec 72°C) were run. Primers were designed to anneal at 59°C. Reaction was set up in technical triplicates with KAPA SYBR® FAST Universal (Roche) in a total volume of 10 μl. All qPCR results were normalized to the house-keeping gene rp49 and include at least three biological replicates.
Statistical analysis
Quantifications were done as described in Antonello et al., 2015 and statistics were run in GraphPad Prism 6.01. For single comparisons, data sets were analyzed by two-sided unpaired t-test. For multiple comparisons, data sets were analyzed by one-way ANOVA and Tukey’s post-hoc test. Significant differences are displayed as * for p ≤ 0.05, ** for p ≤ 0.01, *** for p ≤ 0.001 and **** for p ≤ 0.0001.
SUPPLEMENTAL FIGURE LEGENDS
Acknowledgements
We thank Thomas Klein, Oren Schuldiner, Francois Schweisguth, Naoki Yamanaka, the Bloomington Drosophila Stock Center (NIHP400D018537), the Transgenic RNAi Project (TRiP) at Harvard Medical School (NIK/NIGMS R01-GM084947) and the Vienna Drosophila Resource Center (VDRC, http://www.vdrc.at) for providing transgenic fly stocks and Thomas Klein for being a very supportive host. We also thank Zeus Antonello, Nahuel Villegas, Hendrik Pannen and Thomas Klein for comments on the manuscript. The project is funded by a Deutsche Forschungsgesellschaft (DFG-Sachbeihilfe RE 34532-1) grant.