A feed-forward loop between Toll/NF-κB and Rac1 promotes epithelial to mesenchymal transition of Ras-oncogenic hindgut enterocytes in Drosophila Open Access

Immune responses play a dual role in tumour development: on the one hand, immune cells are recruited at the tumour site and induce apoptosis and elimination of cancer cells. On the other hand, the continuous release of pro-inflammatory factors establishes a microenvironment permissive to tumour growth. For example, constitutive NF-κB activity is reported in 40% of colorectal cancers and is associated with disease progression and poor outcomes (Lind et al., 2001; Sakamoto et al., 2009; Wu et al., 2015; Yu et al., 2004). NF-κB-mediated upregulation of chemoattractants at the pre-metastatic niche may influence the metastasis site by creating an environment that supports the seeding of tumour cells (Collins et al., 1995; Xia et al., 2014). Moreover, cytokines and growth factors produced by immune and epithelial cells alike elicit regenerative signalling that may feed intestinal dysplasia and tumorigenesis (Panayidou and Apidianakis, 2013; Kux and Pitsouli, 2014; Karin and Clevers, 2016). Chronic exposure to inflammatory responses enhances the accumulation of genetic mutations leading to genomic instability. As a result, cancer cells escape from immune-surveillance and acquire hallmarks of cancer cells (Hanahan and Weinberg, 2011). However, NF-κB signalling within pre-cancerous cells is pleiotropic, inducing cell proliferation, apoptosis, or metastasis depending on additional signalling pathways that are concomitantly activated.

RAS oncogene expression is the second most common genetic event in human colorectal cancers, found in more than a third of the metastatic stage cancers (Patelli et al., 2021). Constitutive activation of RAS promotes aberrant proliferation and growth, apoptosis, loss of differentiation, and invasiveness (Wan et al., 2019). However, this genetic pleiotropy points to context dependent modes of action of RAS, which may hinder anti-cancer drug specificity in the clinic, even if the RAS pathway is the drug target itself (Takashima and Faller, 2013). We have previously found that sustained infection facilitates the Ras^V12 signalling to induce basal invasion and dissemination of Drosophila hindgut cells to distant sites. Infection promotes dissemination by inducing part of the Imd innate immune pathway, upstream of Relish/NF-κB, which converges with Ras^V12 signalling on JNK pathway activation and concomitant extracellular matrix degradation (Bangi et al., 2012).

Taking advantage of a Drosophila hindgut cell invasion and dissemination model, we pinpoint transcriptional, post-transcriptional, and cellular changes which explain the transition of Ras^V12 oncogene-expressing hindgut enterocytes regardless of infection from the normal-healthy to an abnormal-invasive state through a crosstalk between the archetypical Toll/NF-κB and Rac1 signalling.

To assess the cytological priming of hindgut enterocytes prior to their delamination upon Ras^V12 oncogene expression, we used two markers: (i) F-actin reporter in w;UAS-Ras^V12/+;byn-Gal4,tub-Gal80^ts,UAS-gfp/UAS-lifeact-Ruby and control flies, and (ii) Fasciclin III (FasIII), a septate junction protein via antibody staining of w;UAS-Ras^V12/+;byn-Gal4,tub-Gal80^ts,UAS-gfp/+ and control flies (Fig. 1D-D′). We find a higher density of F-actin (Fig. 1E,E′,F,F′), and a widespread loss of FasIII in response to Ras^V12 expression (Fig. 1I,I′G,G′).

To assess the invasive characteristics of the oncogenic hindgut enterocytes at the delamination foci, we stained for non-muscle myosin II heavy chain (MHC-II) (Fig. 1C,D), and matrix metalloproteinase 1 (Mmp1) (Fig. 1G,G′,H,H′). We find high MHC II and MMP1 expression in the delamination areas as indicated with arrows in Fig. 1D and H′, respectively.

Noticeably, Ras^V12 expressing hindguts are wider and longer than wild type (Fig. 1K-N), which can be partly attributed to increased cell growth via endoreplication, indicated by larger nuclei and higher DNA content per nucleus (Fig. 1O-Q). This is corroborated by the biggest and irregularly shaped enterocytes bearing bigger nuclei, as shown in Fig. 1A,B,G,H,I-J′,O,P. To emphasize, we circumscribe cell borders in Fig. 1G,H with dotted lines. To assess whether the increased width of the ileum is accompanied by more cells, we quantified the number of cells per ileum length and find that Ras^V12 expressing hindguts exhibit tentatively higher, that is, 91.3±10.7 s.d. (n=7) versus 81.8±23.6 s.d. (n=5) of the control per 200 μm of middle ileum length. However, the difference is not significant (P=0.52, Mann–Whitney U-test). We conclude that the bigger dimensions of the hindgut ileum upon Ras^V12 expression are primarily due to enterocyte growth.

To explore the signalling network involved in the transformation of Ras^V12 oncogene-expressing hindgut enterocytes we performed RNA-Seq. Comparison between hindguts expressing Ras^V12 in their enterocytes (w;UAS-Ras^V12/+;byn-Gal4,tub-Gal80^ts,UAS-gfp/+) and control hindguts (w;;byn-Gal4,tub-Gal80^ts,UAS-gfp/+) reveals differentially expressed genes (DEGs) at a P-value threshold of 0.05, all off which, but three, exhibited a log2 fold change >1 or <−1. Functional annotation reveals 47 deregulated genes within a ‘cytoskeleton’ functional cluster, many of which are small Rho-GTPase (Rho, Rac, Cdc42) signalling genes, including twinstar (tsr), actin-related protein 2/3 complex subunit 1, actin-related protein 2/3 complex subunit 3A and actin-related protein 2/3 complex subunit 5, Rho kinase (rok), non-muscle myosin II heavy chain zipper (zip), and non-muscle myosin II regulatory light chain spaghetti squash (sqh), as well as FasII and Cortactin (Table S1 and Fig. 2). Tsr, the cofilin homologue, promotes F-actin turnover and actin depolymerization from the pointed end, Rok is the effector kinase of Rho1, while the Arp2/3 complex is necessary for branched actin nucleation and formation of filopodia and lamellipodia. The relative increase in mRNA levels of sqh, zip, Arp2, Arp3, tsr, and Rho1 is further confirmed in RT-qPCR experiments (Fig. 3A,B). Innate immunity and stress DEGs include: (i) Toll signalling pathway and target genes, Gram-positive Specific Serine protease (grass), spirit, spheroide (sphe), Peptidoglycan Recognition Protein SA (PGRP-SA), 18 wheeler/Toll-2 (18w), Persephone (psh), l(2)34Fc, IM3 (BomS3), IM4 (Dso1) and Baramicin A1 and A2 (BaraA1, BaraA2), (ii) immune deficiency (Imd) pathway and target genes, Peptidoglycan Recognition Protein SD (PGRP-SD), kenny (key), Attacins A, B, D (AttA, AttB, AttD), Drosocin (Dro) and Neuron navigator (Nav or sick), (iii) JNK pathway and target genes, Jun-related antigen (Jra), Ets at 21C (Ets21C), and Rab30 (Table S2 and Fig. 2). Cell polarity and adhesion DEGS include: the collagen IV basement membrane genes, Cg25C (Col4a1) and viking (vkg), the septate junction genes, Tetraspanin2A (Tsp2A), big bang (bbg), mesh, Snakeskin (Ssk) and lethal (2) giant larvae (l(2)gl), and the matrix metalloproteinases, Mmp1 and Invadolysin (Table S3 and Fig. 2).

To assess their involvement in hindgut enterocyte fate upon Ras^V12 expression, we downregulated via RNAi the small Rho-GTPases, sqh and Arp2 for 14 days. In all cases, RNAi caused a significant reduction in hindgut enterocyte dissemination (Fig. 4A,B). Moreover, activated Rac1 (Rac1^V12) alone sufficed to induce strong cell dissemination at 7 days (Fig. 4B) and at 14 days (Fig. 4C), while activated Rok (Rok^CAT) expression induced mild cell dissemination at 14 days (Fig. 4C). We conclude that Rac1 is necessary and sufficient for cell dissemination, while genes described to act downstream, sqh and arp2, may facilitate Rac1 action.

Considering potential targets of Rac1 signalling, we assessed the expression of snail, a master regulator of epithelial to mesenchymal transition (EMT) implicated in tumour aggressiveness. We noticed a prominent increase in snail expression when Rac1 activity was induced for 10 and 14 days (Fig. 5A,E). Moreover, Rac1 was sufficient to induce Arp2, Arp3 and sqh (Fig. 5D,H), while Arp2 and Arp3 induction by Ras^V12 required sqh and Arp2 (Fig. 5B,C,F,G), pointing to a positive regulation among Arp2, Arp3 and sqh. However, the Arp2 and sqh genes were not required for snail induction in Ras^V12 enterocytes (Fig. 5A,E). Thus, Ras^V12 induces snail, Apr2, Apr3 and sqh, but also the regulation of Arp2 and Arp3 by sqh and Arp2.

To validate the induction of Toll/Toll-like pathway genes by Ras^V12 (Table S2), we assessed by RT-qPCR the activation of these genes in hindguts expressing Ras^V12 in their enterocytes. We observed significant induction of the expression of 18w, grass, spz5 and snail, pointing to their contribution in oncogenic cell formation (Fig. 6A,B). In line with these results, cytoplasm to nucleus fractionation experiments show that Ras^V12 expression increases the protein level of nuclear Dorsal/NFκ-B, a transcription factor integral to Toll signalling (Fig. 6C,D). While upstream and downstream genes, spz5 and snail, respectively, are transcriptionally induced by Ras^V12, the lack of induction of spz, Toll, tube, pelle, dif and dorsal Toll at the mRNA level (Fig. 6B) points to a post-transcriptional activation of the core Toll pathway genes.

Next, we examined the necessity and sufficiency of signalling through the Toll pathway for hindgut enterocyte dissemination, with and without Ras^V12 oncogene expression. We find that silencing the extracellular Toll pathway components grass, spz, and spz5, the receptors Toll and 18w, and the Toll pathway NF-κΒ transcription factors Dif and dorsal, causes a decrease in cell dissemination. The decrease is dramatic upon expression of Toll^RNAi, highlighting the involvement of Toll innate immune signalling in the process (Fig. 7A). Accordingly, constitutive activation of Toll signalling by expressing an activated form of spz (spz^Act) or cactus downregulation via RNAi (cactus^RNAi) induces low but significant cell dissemination even in the absence of Ras^V12 oncogenic signalling (Fig. 7B). Moreover, co-expression of Ras^V12 with spz^Act or cactus^RNAi enhances oncogenic cell dissemination (Fig. 7C).

To investigate whether there is a crosstalk between Toll and the small Rho-GTPase responses, we genetically modified the Toll pathway and assessed Arp2, Arp3 and sqh expression. We find that the expression of the three cytoskeletal genes in the hindgut is significantly reduced upon spz, Toll or dorsal downregulation in oncogenic hindgut enterocytes, while Arp2 can be induced by constitutively activated spz (UAS-spz^Act) alone (Fig. 8A,B,C). Moreover, (i) spz, spz5, Toll or dorsal downregulation impedes Sqh phosphorylation induced by Ras^V12 (Fig. 8D), (ii) spz, Toll and dorsal downregulation reduces snail expression induced by Ras^V12, and (iii) spz^Act alone can induce snail, pointing to a mechanism through which Toll signalling promotes oncogenic cell dissemination (Fig. 8E). Interestingly, although spz expression does not increase upon Ras^V12 expression (Fig. 8F), spz5 expression increases upon either Ras^V12 or spz^Act expression, and decreases to baseline upon spz, Toll or dorsal downregulation in Ras^V12 expressing flies (Fig. 8G). Spz5 was first identified as the ligand of Toll-6, but it also binds and activates Toll (Mishra-Gorur et al., 2019). Thus, Toll signalling regulates itself, snail and cytoskeletal genes downstream of Rac1 in hindgut enterocytes.

To test whether Rac1 signalling genes can affect the Toll pathway, we measured the expression of spz and spz5 upon cytoskeletal gene downregulation and induction. Although spz expression did not change in Ras^V12 expressing hindguts upon cytoskeletal gene downregulation or upon sqh induction (Fig. 9A), spz5 expression was decreased by 3-fold in Ras^V12 expressing hindguts upon sqh^RNAi or Arp2^RNAi expression (Fig. 9B). Nonetheless, both spz and spz5 contributed to the repression of DE-cadherin (Fig. 9C,D). DE-cadherin protein levels are strikingly reduced in Ras^V12 expressing hindguts, while downregulation of any of the two ligands, spz or spz5, results in significant de-repression of DE-cadherin (Fig. 9C,D). However, while Rac1 downregulation tentatively represses DE cadherin, similarly to Ras^V12, inhibition via sqh^RNAi or Arp2^RNAi in the presence of Ras^V12 does not derepress DE-cadherin (Fig. 9E,F), suggesting that Toll signalling is repressing DE-cadherin directly.

Invasion of primary tumour cells in neighbouring tissues requires the degradation of basement membrane. Accordingly, MMP1 protein is induced in delaminating Ras^V12 hindgut cells (Fig. 1G,H), and Mmp1 expression is higher in Ras^V12 hindguts (Fig. 10A). Interestingly, downregulation of sqh or Arp2 in Ras^V12 enterocytes reduces Mmp1 expression to baseline levels, whereas activated Rac1 alone mimics the effect of Ras^V12. In line with our hypothesis that Rac1 and Toll signalling crosstalk, downregulation of spz, Toll and dorsal inhibits the induction of Mmp1 by Ras^V12 (Fig. 10A).

To validate the role of the Toll pathway in JNK activation, a positive regulator of Mmp1 (Bangi et al., 2012), we performed western blotting with antibodies against phosphorylated JNK (p-JNK). We find an increase in the levels of p-JNK upon Ras^V12 expression and a reduction of this activation upon downregulation of spz, spz5 and Toll, but not dorsal, suggesting that JNK activation by Ras^V12 requires upstream Toll signalling (Fig. 10B-D). Notably, we normalized p-JNK levels to those of the ‘housekeeping’ protein Syntaxin to indicate the relative amount of phosphorylated JNK within the cells, attributable to the total amount of JNK protein and the fraction of it being phosphorylated.

To assess whether JNK facilitates cell dissemination mediated by Toll or Rac1, we induced Toll signalling via cactus^RNAi and Rac1 signalling via Rac1^V12 or sqhEE, in combination with JNK signalling activation via hep^Act in hindgut enterocytes and assessed cell dissemination at 7 days of transgene induction (Fig. 11). No delamination was observed by JNK pathway activation alone, and no synergy was found by simultaneous induction of sph. However, JNK activation enhanced the effect of Rac1 or Toll signalling on dissemination (Fig. 11). This agrees with previous findings showing that JNK activation enhances Ras^V12 induced dissemination (Bangi et al., 2012). Thus, JNK activity facilitates but does not suffice for cell dissemination.

Constitutively active mutant forms of Ras are found in more than a third of human cancers and in as many metastatic colorectal cancers (Murugan et al., 2019; Patelli et al., 2021). Thus, targeting Ras signalling is an attractive therapeutic strategy. However, pharmacological targeting of Ras signalling is hampered by the alternative ways it is activated (Samatar and Poulikakos, 2014). Moreover, constitutively active Ras tumours induce cell death in surrounding cells, promoting Ras clonal expansion (Moreno et al., 2019). A more sophisticated targeting of Ras signalling may be achieved through combinatorial strategies that are directed at multiple nodes of the signalling network.

Here, we used the Drosophila hindgut epithelium model to reveal adjunct nodes of the pro-metastatic Ras^V12 signalling as additional targets of therapeutic intervention. Drosophila Ras^V12 triggers cell entry into the S phase and promotes tissue overgrowth and tissue invasion, recapitulating properties of mammalian cancer development (Prober and Edgar, 2002; Bangi et al., 2016, 2012). Our transcriptomic analysis was performed on day 7 post Ras^V12 induction, when enterocyte dissemination is still low per previous findings (Bangi et al., 2012), and suggested genes in at least three functional categories: (1) cytoskeleton, (2) immunity and stress and (3) cell polarity and adhesion as mediators of Ras oncogene signalling on cell motility. These were statistically significantly altered in expression (most of them orders of magnitude below the P<0.05 limit) and increased between 2.2- and 6.9-fold, 2.3- and 273-fold, and 2.1- and 27-fold (Tables S1-S3), respectively. However, the signalling pathways uncovered by this, and further analysis involve genes that are not necessarily activated by Ras^V12 at the transcription level. Many receptors and kinases are generally not at all activated transcriptionally. For example, the transcription factor of the Toll pathway, Dorsal, is activated via nuclear translocation. Thus, transcriptional activation for many of the genes studied is suggestive rather than a prerequisite for their involvement in the process of tissue transformation.

Previous findings indicated that the Ras oncogene signalling effect on cell motility is mediated by Rac1 signalling and downstream cytoskeletal rearrangements via the polymerization of actin branches and cellular contraction (Ridley, 2001; Brumby et al., 2011). To validate the role of Rac1 signalling in our model, we performed experiments showing that Rac1 is necessary and sufficient for cell dissemination. Rac1 acts as a master regulator of all aspects of cell dissemination downstream of Ras^V12 activation via a Rac1-Rho-Rok-sqh/zipper/tsr/Arp2/3 pathway, which orchestrates actin and stress fibre formation, consistent with previous studies on Drosophila hemocytes and development (Williams et al., 2007; Verdier et al., 2006).

Of note, Ras^V12 and Ras^V12 Apc^null Drosophila midgut progenitors do not produce metastatic tumours unless they also express snail (Christofi and Apidianakis, 2013; Campbell et al., 2019). Accordingly, we show that Rac1 promotes cell dissemination by inducing snail, in addition to actin cytoskeleton signalling genes, Rok, sqh, Apr2, and Apr3, and a positive feedback loop with the grass-spz/spz5-Toll-cactus/dif/dorsal pathway. Downregulation of core Toll signalling genes significantly decreases cell dissemination, with Toll receptor silencing almost abolishing it. Toll sufficiency was assessed by the expression of activated spz and cactus^RNAi, each of which induced significant cell dissemination on its own. We used two spz^RNAi transgenes, one of which did not inhibit sqh phosphorylation by Ras^V12, while it inhibited suppression of DE-cadherin and induction of JNK phosphorylation. Likely, Toll signalling suppresses DE-cadherin and induces JNK phosphorylation directly or strongly, while it may induce sqh phosphorylation indirectly or less strongly.

Interestingly, FlyAtlas anatomical gene expression data show that spz is expressed in high levels in the hindgut compared to the midgut, suggesting that hindgut enterocytes might be primed for cell dissemination via the Toll pathway. Moreover, spz5 participates in this network, as it can bind and induce Toll in addition to Toll-6, and both spz and spz5 drive Toll-6-mediated JNK activation and cell migration in the context of organotropic metastasis (Mishra-Gorur et al., 2019; Chowdhury et al., 2019; Nonaka et al., 2018). Whether Spz and Spz5 ligands are released only by hindgut enterocytes or also by neighbouring or remote cells is yet to be determined.

In marked similarity with the Ras^V12 signalling network in hindgut enterocytes, spz-Toll-dorsal-twist/snail signalling and small Rho-GTPases operate during embryonic gastrulation (Rich and Glotzer, 2021) and culminate in the repression of DE-cadherin, a core component of adherens junctions (Bauer et al., 2006). Moreover, we find that 18w interacts with Ras^V12 to induce cell dissemination, similarly to its synergism with Rho-GTPases in inducing apical constriction of salivary gland cells (Kolesnikov and Beckendorf, 2007). Of note, there are nine Toll receptors in Drosophila, and each may signal differently in different contexts, such as in border cell migration (Kleve et al., 2006), neuronal survival (Foldi et al., 2017), competition-induced cell death (Wu et al., 2015) and carcinogenesis (Mishra-Gorur et al., 2019; Ding et al., 2022). Since 18w/Toll2 affects cell dissemination without being part of the Toll pathway, additional Toll pathways may contribute to the Ras^V12 signalling network (Ding et al., 2022). Given the high expression of TLRs in multiple cancers (O'Neill, 2008; Zhang et al., 2009), and the conservation between Toll and TLR pathways (Gay and Gangloff, 2007; Gay et al., 2014), mammalian TLRs may have a similar role in sustaining oncogenic cell migration.

Rho GTPase signalling (Brumby et al., 2011) and Toll signalling (Mishra-Gorur et al., 2019) are necessary in other contexts to induce JNK and invasive cell behaviour downstream of the Ras oncogene. Notwithstanding the possibility of JNK-independent activation of Mmp1 by Toll or Rac1, the simplest explanation to our findings is that Toll and Rac1 crosstalk and induce JNK and in turn Mmp1, because: (i) Toll and Rac1 signalling can activate each other, (ii) Toll or Rac1 signal can induce JNK, (iii) JNK can induce Mmp1, and (iv) Toll or Rac1 can induce Mmp1. In previous studies, bacterial infection and Imd signalling was shown to stimulate JNK-mediated Mmp1 expression in the hindgut facilitating cell dissemination by Ras^V12 (Bangi et al., 2012). Here, western blotting analysis indicates that activation of JNK by Ras^V12 is necessary but not sufficient for cell dissemination and is under the control of Rac1 and Toll signalling, which, in addition, oversee cytoskeletal rearrangements and DE-cadherin suppression (Fig. 12). In light of their conserved functions, inhibition of Toll and Rac1 may restrict Ras-induced cell overgrowth and tissue invasion in mammalian cells. It will, thus, be of interest to test their combinatorial inhibition against various types of Ras-activated tumours.

All fly stocks and crosses were maintained at 18°C on a 12:12 h light: dark cycle on standard fly food (agar, cornmeal, sucrose and Brewer's yeast, Tegosept and propionic acid). Adult female flies were collected shortly after eclosion and aged to 3-5 days old into fresh vials prior to each experiment. For temperature controlled hindgut enterocyte expression the w;;byn-Gal4,tubGal80^ts,UAS-gfp/TM6b line was crossed to UAS lines: UAS-grassRNAi (Bl#51920), UAS-18wRNAi (Bl#30498), UAS-TollRNAi (VDRC_KK100078), UAS-spz5RNAi (Bl#67229), UAS-spzRNAi (Bl#34699), UAS-spzRNAi (VDRC_KK105017), UAS-SpzAct (gift by Petros Ligoxygakis), UAS-DorsalRNAi (VDRC_KK105491), UAS-DifRNAi (VDRC_KK100537), UAS-CactusRNAi (Bl#34784), UAS-CactusRNAi (Bl#34775), UAS-sqhRNAi (Bl#31542), UAS-sqhRNAi (Bl#33892), UAS-Rac1 (Bl#6291), UAS-RokCAT (Bl#6669), UAS-Rac1N17 (gift by Norbert Perrimon), UAS-sqhE20E21 (Bl#64411), UAS-Arp3RNAi (Bl#32921), UAS-RhoGEF (gift by Norbert Perrimon), UAS-Arp2RNAi/TM3,sb (Bl#27705), UAS-Arp3RNAi (VDRC_GD35260), UAS-HepAct (Bl#9306), yw; UAS-Lifeact-Ruby (Bl#35545), w; Byn-Gal4, tubGal80ts, UAS-gfp/TM6B (Bangi et al., 2012), w;UAS-RasV12 (Bl#64196). Crosses were reared at 18°C and adult flies were transferred to 29°C to induce the transgenes before experiments. We crossed females of the Gal4 line to UAS lines of interest using the reference line, w¹¹¹⁸, as a basic control. Thus, half of the nuclear DNA, all the mitochondrial DNA, and the egg nutrients were supplied by mothers of the same genotype and were identical among the progeny compared. To alleviate concerns regarding remaining genetic background differences among the UAS lines and the reference line, we used alternative UAS lines per gene and many UAS lines per pathway in multiple instances.

Per Bangi et al. (2012), anesthetized female flies were lined up and immobilized on glass slides coated with Vaseline with their ventral sides up. The glass slide was then transferred into a Petri dish and submerged in 1x PBS solution (10x PBS: 1.3 M NaCl, 0,07 M Na₂HPO₄, 0,03 M NaH₂PO₄), so that a complete coverage of the fly bodies was achieved. Using forceps, the abdomens were opened to make a ‘fillet’, intestines and ovaries removed and the number of GFP+ foci along the abdominal cavity was counted using a Leica M165 FC dissecting stereoscope with a GFP filter under a ×10 magnification. Each experiment was performed in duplicate (n=30 in each replicate). Error bars throughout represent standard deviation of the mean (s.d.). P-values were calculated using the Fisher's exact test with a 5×2 contingency table to assess the five phenotypic classes between two conditions at a time.

Fifteen hindguts per genotype were dissected onto a silicon plate in 1x PBS drop and fixed for 1 h with 9% formaldehyde (FA) at room temperature (RT). For FasIII immunostaining, hindguts were fixed in 4% FA for 30 min at RT. Following completion of the fixation period the samples were rinsed three times with 1x PBS to remove any residual FA. Hindguts were then incubated into PBT (1× PBS, 0.5% BSA, 0.1% Triton X-100) blocking solution for 1 h to achieve tissue permeabilization and blocking of non-specific antibody binding sites. Then, blocking solution was removed and primary antibodies diluted in PBT solution were added and incubated with tissue samples overnight in the dark at 4°C. Primary antibodies: anti-Mmp1 antibody cocktail 1:100 (mouse, 3B8D12, 5H7H11 and 3A6B4 from DSHB), anti-Fas III 1:100 (mouse, 7G10 from DSHB). The next day, hindguts were washed three times for 10 min in PT (1x PBS, 0.2% Triton-X) solution. Secondary antibodies against mouse, rabbit or guinea pig conjugated to Alexa fluor 555 (Invitrogen) were used at 1:1000. Samples were incubated in secondary antibody solution including DAPI (Sigma, 1:3000 of 10 mg/ml stock) for 2 h at room temperature (RT), in the dark, with mild shaking. Finally, samples were washed three times with PT to remove unbound antibody and mounted on glass microscope slides in 20μl of Vectashield (Vector), covered with glass coverslips and sealed with nail polish. Images of the immunostained hindguts were captured using the LEICA TCS SP2 AOBS confocal microscope.

Length and width were assessed from GFP pictures acquired with a fluorescent Leica M165 FC stereomicroscope (at 6.3x magnification) and their width and length were analysed using ImageJ (https://imagej.net/ij/). Clicking on the Analyze, Set Scale option, using the length of the picture as 2048 pixels corresponding to 1.65 mm. Using the segmented line option, a line was drawn manually, starting from hindgut proliferating zone (HPZ) and ending just before the rectum. For the width measurement, straight lines indicating the width of pylorus (HPZ), the middle ileum and posterior ileum were manually drawn vertically to the gut length. The length of the lines was measured and thereby the gut's dimensions, by clicking on the Analyze, Measure option.

Per previous studies (Tamamouna et al., 2020) nuclear DNA was calculated from non-saturated confocal images using ImageJ. Images of the middle hindgut were captured at 63× magnification, zoom 1× and 1024×1024 format and produced as a maximum projection of 10-15 sections serial imaging. Sum projections were used to measure the total DAPI fluorescence from individual nuclei using ImageJ in anterior and posterior images of the midguts. Specifically, each and every nucleus (excluding nuclei that overlapped) was selected manually using the circular selection tool of the software and surface and integrated density were acquired. Data points of hundreds of cells arising from ≥3 hindgut parts were added to generate BoxPlots.

For RNA extraction 100 hindguts (without the rectum) were dissected for each genotype in three biological replicates. Tissue dissections were performed in 1x PBS drops on a silicon plate followed by an immediate freezing on dry ice, so that maximizing tissue integrity and eliminating potential RNA degradation. Sample tissues were homogenized at 50 Hz for 10 min in 500 μl Qiazol using the TyssueLyser LT (Qiagen) and a stainless-steel bead (5 mm diameter). Next, the homogenized tissues were incubated at RT for 5 min. Sample phase separation was mediated by the addition of 100 μl chloroform (CHCI₃) followed by a vigorous shaking for 15-20 s and centrifugation at 12000 g for 15 min at 4°C. The resulting upper phase, which contains the RNA was carefully transferred into a clean Eppendorf tube. Finally, 250 μl isopropanol were added and mixed in the RNA sample and incubated at RT for 15 min. A second step of centrifugation at 12,000 g for 15 min at 4°C was performed to pellet the RNA. Supernatant was then discarded, and RNA pellet was carefully rinsed with 200 μl 70% EtOH. A last centrifugation step at 8000 g for 5 min at 4°C occurs to completely remove any residual EtOH. The pellet was then dissolved in 20 μl RNase free water and the concentration of RNA in the sample was determined using the NanoDrop 200c Spectrophotometer (Thermo Fisher Scientific).

For cDNA synthesis 800 ng of total RNA were used to synthesize the cDNA using Promega RQ1 RNase-Free DNase Kit according to the manufacturer’s protocol. RNA was treated with DNAse I mixture (RQ1 10× buffer, 1 μl RQ1 DNase enzyme from Promega) for 30 min at 37°C to eliminate the genomic DNA contamination. Reaction was ceased by the addition of RQ1 DNAse stop solution and 10 min incubation at 65°C. Reverse transcription was performed using 145.4 ng of the total DNAse-treated RNA by using the TaKaRa Prime ScriptTM RT Master Mix Kit. Following completion of the reverse transcription reaction, the cDNA is diluted (1:6) in RNase/DNase free water. For each qPCR reaction 6 μl cDNA, 4 μl gene specific primers and 10 μl KAPA SYBR Supermix (Kapa Biosystems) were used. qPCR amplification was then performed based on the following amplification program on the CFX96 Real-Time System/C1000 Thermal Cycler (Bio-Rad): 95°C for 30 s (initial denaturation), 40 cycles of 95°C for 10 s (denaturation), 60°C for 30 s (annealing), 65°C for 30 s (extension) and 65°C for 1 min (final extension). The expression of the genes of interest was normalized to the expression levels of two reference genes, rpl32 and α-tubulin using the 2^−ΔΔCt method. Data were analysed using the Bio-Rad CFX Manager 3.1 program.

For sub-cellular protein fractionation hindgut tissues were collected on dry ice as described earlier. Tissue homogenization was performed at 50 Hz for 10 min into a 2 ml tube containing 150 μl Buffer A with 0.1% Triton (10 mM HEPES, 10 mM KCL, 1.5 mM MgCl2, 0.34 mM sucrose, 10% glycerol, 100 μl, 0,1% Triton-X 100, 1 mM DTT, 1 tablet of protease inhibitors, dissolved in DNAse-RNAse free water) and a stainless-steel bead using the TyssueLyser LT (Qiagen). Following homogenization, samples were incubated on ice for 30 min with vortexing every 10 min. 50 μl of the protein lysate were transferred into a clean tube and referred to as the whole tissue extract (WTE). The remaining 150 μl of protein extract were centrifuged at 21,000 g for 10 min at 4°C producing a supernatant (S1) and a pellet (P1). S1 sample was subjected to the centrifugation step as previously for four times until the S5 supernatant (cytoplasmic fraction-CF) was collected. Following completion of each centrifugation step, the supernatant was transferred into a clean tube while the pellet was discarded. P1 pellet was treated with 100 μl Buffer A (without Triton) and centrifuged at 6000 g for 10 min at 4°C. This step was performed twice to eliminate the cytoplasmic protein contamination of pellet and ensure purity. At the final centrifugation step supernatant is removed and P1 pellet is dissolved into 70 μl of Buffer B (10 mM HEPES, 3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, 1 tablet of protease inhibitors, dissolved in DNAse-RNAse free water). The sample was then incubated on ice for 30 min with vortexing every 10 min. Completion of this step P1 denoted the nuclear protein fraction (NF) of the sample.

For western blotting the samples that were not subjected to sub-cellular fractionation were homogenized into Buffer A with 0.1% Triton followed by 30 min incubation on ice with vortexing every 10 min, as described above.

Protein lysates were boiled in 4× Laemmli Buffer for 10 min at 95°C and separated using standard immunoblotting protocol. Small proteins (20-70 kDa) were separated through a 12% SDS–PAGE while larger proteins through an 8% SDS–PAGE. The gel was placed into an electrophoresis tank and run for 1-1.5 h in 200 V until efficient protein separation as indicated by the 240 kDa prestained protein ladder (Nippon Genetics) was achieved. Proteins were then transferred to a PVDF membrane, which had been previously soaked in methanol for 1 min. Protein transfer was carried out at 100 V for 1.5 h at 4°C. Next, the membrane was incubated in Ponceau S solution to confirm transfer efficiency, rinsed and blocked in 5% skimmed milk dissolved in 1× TBS-Tween 20 (TBS-T) buffer for 1 h on a shaking platform at RT. Then, the membrane was incubated with the primary antibody overnight at 4°C. Primary antibodies were diluted in 5% skimmed milk, 1x TBS-T, unless otherwise recommended. Primary antibodies: anti-E-cadherin 1:200 (rat, DCAD2 from DSHB), anti-sqh 1:5000 (rabbit, gift from Karess R.), anti-p-sqh 1:1000 (rabbit, 3671S from CST), anti-Mmp1 antibody cocktail 1:1000 (mouse, 3B8D12, 5H7H11 and 3A6B4 from DSHB). Following primary antibody incubation, the membrane was washed for 20 min with 1× TBS-T and then for additional 10 min with fresh 1× TBS-T. Finally, the membrane was incubated with an HRP-conjugated secondary antibody of choice (1:10,000 dilution) at RT for 1 h on a shaking platform. Then, the membrane was washed as before, and protein signal detection was performed using the SuperSignal™ West Femto Maximum Sensitivity Substrate (Invitrogen) and Syngene G-Box gel documentation system. Protein expression was quantified compared to the syntaxin (mouse, 1:1000, DSHB #8C3) loading control using the ImageJ software. The mean intensity of respective protein bands from three different immunoblots was used for the quantification.

RNA was isolated from three replicates of 100 control hindguts (w;;byn-Gal4tubGal80^tsUAS-gfp/+) and hindguts expressing the Ras^V12 oncogene (w;UAS-Ras^V12/+;byn-Gal4tubGal80^tsUAS-gfp/+) for 7 days at 29°C. RNA quality was assessed on a Bioanalyzer (Agilent Technologies) with RNA 6000 Nano Kit reagents and protocol (Agilent Technologies), followed by use of the 3′ mRNA-Seq Library Prep Kit protocol for Ion Torrent (QuantSeq-LEXOGEN Vienna, Austria) according to the manufacturer's instructions. Briefly, up to 500 ng of RNA was used for first and second-strand synthesis, followed by amplification. Library quality and quantity were assessed on a Bioanalyzer with DNA High Sensitivity Kit reagents and protocol (Agilent Technologies). The libraries were pooled together and templated and enriched on an Ion Proton One Touch system. Templating was performed using the Ion PI Hi-Q OT-II 200 Kit (Thermo Fisher Scientific), followed by sequencing with the Ion PI Hi-Q Sequencing 200 Kit on Ion Proton PI V2 chips (Thermo Fisher Scientific) on an Ion Proton System, as per the manufacturer's instructions.

Raw reads were aligned to the Drosophila genomic build dm6 through in two steps. Firstly, reads were mapped with hisat2 (Kim et al., 2019) using the default parameters and then, the unmapped reads were mapped with bowtie2 (Langmead and Salzberg, 2012) using the local and very sensitive parameters. Prior to the statistical testing procedure, the gene read counts were filtered for possible artifacts that could affect the subsequent statistical testing procedures. Genes presenting any of the following were excluded from further analysis: (i) genes with zero reads (6531), (ii) genes with length less than 500 (817 genes), (iii) genes whose average reads per 100 bp was less than the 25th quantile of the total normalized distribution of average reads per 100 bp (824 genes with cutoff value 0.26135 average reads per 100 bp), (iv) genes with read counts below the median read counts of the total normalized count distribution (3931 genes with cutoff value) normalized read counts. The total (unified) number of genes excluded due to the application of all filters was 12055. The resulting gene counts were subjected to differential expression analysis for byn-G4 versus byn-G4 U-Ras^V12 using the Bioconductor packagemetaseqR2 with the DESeq algorithm and the utr parameter (Fanidis and Moulos, 2021). DEGs were identified based on an absolute log2 fold change (|log2(FC)|>1) and a P-value <0.05. The code is publicly available at https://github.com/alex-galaras/panagi_et.al.2025. DEGs were further analysed using the DAVID (Database for Annotation, Visualization and Integrated Discovery) database to discover enriched functional-related gene groups. However, we manually curated the genes listed in three functional categories: cytoskeleton, immunity and stress, and cell polarity and adhesion, because of multiple false positive and false negative genes listed as cytoskeletal and immunity related.

We thank Christina Michael for technical assistance in fly preparation and experiments.