Portal de Periódicos da CAPES

Notch and Mef2 synergize to promote proliferation and metastasis through JNK signal activation in Drosophila

2012; Springer Nature; Volume: 31; Issue: 13 Linguagem: Inglês

10.1038/emboj.2012.129

1460-2075

S. K. Pallavi, Diana M. Ho, Chindo Hicks, Lucio Miele, Spyros Artavanis‐Tsakonas,

MicroRNA in disease regulation

Article11 May 2012free access Notch and Mef2 synergize to promote proliferation and metastasis through JNK signal activation in Drosophila S K Pallavi S K Pallavi Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Diana M Ho Diana M Ho Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Chindo Hicks Chindo Hicks Cancer Institute at University of Mississippi Medical Center, Jackson, MS, USA Search for more papers by this author Lucio Miele Lucio Miele Cancer Institute at University of Mississippi Medical Center, Jackson, MS, USA Search for more papers by this author Spyros Artavanis-Tsakonas Corresponding Author Spyros Artavanis-Tsakonas Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author S K Pallavi S K Pallavi Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Diana M Ho Diana M Ho Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Chindo Hicks Chindo Hicks Cancer Institute at University of Mississippi Medical Center, Jackson, MS, USA Search for more papers by this author Lucio Miele Lucio Miele Cancer Institute at University of Mississippi Medical Center, Jackson, MS, USA Search for more papers by this author Spyros Artavanis-Tsakonas Corresponding Author Spyros Artavanis-Tsakonas Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Author Information S K Pallavi1, Diana M Ho1, Chindo Hicks2, Lucio Miele2 and Spyros Artavanis-Tsakonas 1 1Department of Cell Biology, Harvard Medical School, Boston, MA, USA 2Cancer Institute at University of Mississippi Medical Center, Jackson, MS, USA *Corresponding author. Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA. Tel.:+1 617 432 7048; Fax:+1 617 432 7050; E-mail: [email protected] The EMBO Journal (2012)31:2895-2907https://doi.org/10.1038/emboj.2012.129 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Genetic analyses in Drosophila revealed a synergy between Notch and the pleiotropic transcription factor Mef2 (myocyte enhancer factor 2), which profoundly influences proliferation and metastasis. We show that these hyperproliferative and invasive Drosophila phenotypes are attributed to upregulation of eiger, a member of the tumour necrosis factor superfamily of ligands, and the consequent activation of Jun N-terminal kinase signalling, which in turn triggers the expression of the invasive marker MMP1. Expression studies in human breast tumour samples demonstrate correlation between Notch and Mef2 paralogues and support the notion that Notch–MEF2 synergy may be significant for modulating human mammary oncogenesis. Introduction The Notch signalling pathway defines one of the fundamental cell signalling mechanisms in metazoans, and is broadly responsible for controlling cell fates by profoundly affecting proliferation, differentiation and apoptosis throughout development (Artavanis-Tsakonas et al, 1999; Bray, 2006; Kopan and Ilagan, 2009; Artavanis-Tsakonas and Muskavitch, 2010). The Notch surface receptor interacts with membrane-bound ligands present on adjacent cells to trigger a proteolytic cascade that eventually cleaves the receptor and releases the intracellular domain (Nact) that subsequently translocates into the nucleus and participates in a transcriptional complex directing Notch-dependent transcription (Borggrefe and Oswald, 2009). Development of both vertebrate and invertebrate animal systems has been shown to be very sensitive to the dosage of Notch signalling. Dysregulation of Notch signalling almost invariably leads to mutant phenotypes and other human diseases including cancer (Mohr, 1919; Aster and Pear, 2001; Gridley, 2003). The genetic circuitry capable of modulating Notch activity has been reported to be very complex (Kankel et al, 2007; Krejcí and Bray, 2007), as is common with ancient conserved pathways, where regulatory mechanisms are numerous and diverse (Bray, 2006). Abnormal activation of the pathway has been shown to be oncogenic based on the observations that Notch behaves as an oncogene and activating mutations have been detected in almost half the cases analysed of T-cell acute lymphoblastic leukaemias (T-ALLs; Ellisen et al, 1991; Weng et al, 2004). Notch mutations in other tumours remain scarce or non-existent, yet there is a growing list of solid cancers, including breast, prostate, skin, brain, lung, colon and pancreas (Ranganathan et al, 2011), where Notch activity has been positively correlated with oncogenic events. We and others have shown that the activation of the Notch receptor can induce proliferation in a context-specific manner both in invertebrates and in vertebrates (Berry et al, 1997; Go et al, 1998; Baonza and Garcia-Bellido, 2000; Fre et al, 2005). Our own studies in the mouse intestine and mammary glands (Kiaris et al, 2004; Klinakis et al, 2006; Fre et al, 2009), however, lead to the conclusion that the activation of the Notch receptor per se is not necessarily oncogenic, but it is the synergy between Notch signals and other factors that drives proliferation, eventually leading to oncogenesis (Artavanis-Tsakonas and Muskavitch, 2010). These types of synergies may result in either the de-novo activation of new pathways or the enhancement of existing Notch target pathways, in both cases resulting in the acquisition of dramatic downstream functional changes. We sought to identify genes that can synergize with activated forms of Notch to trigger proliferative events using genetic screens in Drosophila as our tool. Such synergies have been uncovered before in Drosophila (Moberg et al, 2005; Ferres-Marco et al, 2006; Vallejo et al, 2011), suggesting the existence of several factors that can influence proliferation through synergistic or additive effects with Notch signals. The screen we carried out resulted in the identification of Mef2 (myocyte enhancer factor 2) as a crucial synergistic partner of Notch in triggering massive proliferation and an invasive metastatic phenotype. Mef2, a transcription factor, was first identified as a regulator of muscle gene expression (Gossett et al, 1989) that is required for myoblast differentiation (Bour et al, 1995; Lilly et al, 1995; Ranganayakulu et al, 1995) but also has broad and pleiotropic effects in development and disease (Potthoff and Olson, 2007; Homminga et al, 2011). Our mechanistic studies indicate that the synergy between Notch and Mef2 can be attributed to the activation of the Jun N-terminal kinase (JNK) signalling pathway. We furthermore show that this results in upregulation of the secreted ligand eiger, the orthologue of mammalian tumour necrosis factor (TNF) family cytokines. We examine the relevance of these results in human breast cancer samples and uncover a correlation between Notch and Mef2 human paralogues, suggesting that this synergy is likely to be relevant to human malignancies. Results Notch and Mef2 synergy affects proliferation The activation of the Notch receptor has been shown to induce proliferation in Drosophila (Go et al, 1998; Baonza and Garcia-Bellido, 2000) and in vertebrates (Fre et al, 2005; van Es et al, 2005). Constitutively activated Notch is oncogenic in several distinct contexts, often in combination with other factors (Radtke and Raj, 2003; Ranganathan et al, 2011). In order to identify these factors, we sought to carry out a genetic screen for modifiers of a 'large eye' phenotype induced by the overexpression of a constitutively active, ligand-independent form of the Notch receptor (Nact), that has been shown to be oncogenic in mammals (Kiaris et al, 2004). Ectopic expression of Nact in the Drosophila eye, using the eye-specific driver E1Gal4 (eyGal4), results in a large eye phenotype (Figure 1B) caused by elevated levels of cell proliferation as shown by an increase in EdU incorporation levels accompanied by slightly larger and distorted eye discs (Figure 1E). This proliferation increase is heterogeneous, resulting in spots of localized EdU incorporation within the undifferentiated (Elav-negative) cell compartment, both because of the non-homogeneous expression pattern of the E1Gal4 driver and because of context specificity. The phenotype is dosage sensitive to Notch signal modulation and, as expected, could be enhanced or suppressed using gain- and loss-of-function mutations in Notch pathway elements and could therefore be used as the experimental parameter in a genetic screen for Notch modifiers. Using the Exelixis collection of insertional mutations, which covers ∼50% of the Drosophila genome (Kankel et al, 2007), we identified two separate Gal4-driven mutations, d06622 and d03191, that strongly enhanced the large eye phenotype caused by Nact (Figure 1C). Figure 1.Synergy between Notch and Mef2 in the eye. (A) Wild-type fly eye. (B) Activated Notch (Nact) in the eye results in a large eye when driven by the eye-specific E1Gal4 driver. (C) Exelixis Mef2 fly stock (d06622) synergizes with Nact in the eye to cause massively overgrown eyes. (D–G) EdU incorporation assay in eye-antennal discs expressing various constructs under E1Gal4. Discs are labelled with EdU (red) to mark dividing cells and stained for anti Elav (green) to mark differentiated cells. (D) Wild-type control; # marks antennal disc (in which E1Gal4 is not active). (E) Nact alone causes an increase in EdU levels (*) and (F) Mef2 (d06622) alone does not cause any change in EdU levels. On the other hand, (G) Nact together with Mef2 (d06622) causes an enormous increase in both EdU labelling and disc size. Note that EdU-positive cells appear to be restricted to the Elav-negative undifferentiated cells. (H–J) Notch and Mef2 synergize to induce expression of the Notch target gene wingless in the eye disc. Clones expressing Nact and/or Mef2 were generated (see Materials and methods). Eye-antennal discs containing GFP-labelled clones were stained with wingless (wg, red). Note that neither Nact (H) nor Mef2 (I) alone could induce wg in the eye, but together they turned on ectopic wg in all clones (J). Download figure Download PowerPoint Mapping of the insertion (Thibault et al, 2004) revealed that these mutations represent two separate gain-of-function alleles of Mef2, a pleiotropic, essential gene that belongs to the evolutionarily conserved MADS family of transcription factors implicated in the development of diverse tissues (Potthoff and Olson, 2007). Drosophila possesses a single Mef2 gene that codes for an ∼57 kilodalton protein (Nguyen et al, 2002), whereas vertebrates harbour four genes (Mef2 A, B, C, and D) (Potthoff and Olson, 2007). Western blot and immunofluorescence analyses of both alleles revealed elevated expression levels of the Mef2 protein compared with wild type, proving the nature of the mutations (Supplementary Figure S1). Neither of the Exelixis Mef2 alleles on their own affected the adult eye morphology and the corresponding eye discs appeared wild type (Figure 1F). In contrast, coexpression of a single copy of either of these two Mef2 alleles with Nact resulted in massively overgrown discs, showing excessive EdU incorporation (Figure 1G). Expressing an Mef2 transgene caused an even stronger synergy with Nact (Supplementary Figure S1). Costaining with the neuronal differentiation marker Elav revealed that the hyperproliferative cell compartment was restricted to the anterior of the morphogenetic furrow harbouring the undifferentiated cells of the eye disc (Figure 1G). To examine whether the observed synergy is confined to the eye or is a more general phenomenon, we extended our analysis to the developing wing and coexpressed Nact and Mef2 under two different wing-specific drivers, vgGal4 and MS1096Gal4. Compared with discs expressing the two genes individually, coexpression resulted, just as in the eye, in massively overgrown discs (Supplementary Figure S1), indicating that the synergy is not tissue specific. Emphasizing the synergistic nature of the phenotype, only the coexpression of Nact and Mef2, not either gene alone, in somatic clones of the eye discs induced the expression of the Notch target Wingless outside of its normal expression realm (Figure 1H–J), that is, the lateral margins of the eye discs (Treisman and Rubin, 1995), releasing Notch signalling from its normal context-dependent restrictions. As in the eye, coexpression of Nact and Mef2 resulted in the synergistic activation of wingless in the wing, outside of its normal expression domain of the wing margin (Couso et al, 1994; Supplementary Figure S1). The activation of ectopic wingless by Nact and Mef2 demonstrates that these two genes together but not alone can synergistically affect downstream gene expression. We thus conclude that it is the synergy between Nact and Mef2 that is responsible for inducing hyperproliferation and the accompanying phenotypes. Notch and Mef2 synergy induces metastatic behaviour and MMP1 expression In addition to hyperproliferation, coexpression of Nact and Mef2 in the eye consistently, albeit at low penetrance (∼5%), causes ectopic eyes in the thorax or abdomen (Figure 2A). Similar phenotypes in Drosophila have been previously reported and associated with invasive and metastatic behaviour (Ferres-Marco et al, 2006; Palomero et al, 2007). Figure 2.Notch and Mef2 synergize to induce MMP1 expression and invasiveness. (A) An adult fly coexpressing Nact and the Exelixis Mef2 stock d06622 under E1Gal4 displays an ectopic eye in the abdomen (arrow). (B–G) Wing discs expressing various UAS constructs (as indicated) under MS1096Gal4 were stained for MMP1 (green). Note that Mef2 alone (D) induces MMP1 expression only in the D/V boundary, which is reduced by RNAi-mediated downregulation of endogenous Notch (G). In contrast, Nact and Mef2 together (E) strongly induce MMP1 throughout the entire wing disc as compared with wild type (B) or Nact alone (C). (H–J) GFP-positive clones (green) expressing Nact and/or Mef2 were generated (see Materials and methods) and stained for MMP1 (red). (H, H′) Nact does not induce MMP1 in clones, (I, I′) Mef2 induces some MMP1 only in clones near the D/V boundary (arrowheads; D/V boundary is marked by a dotted line) but not in clones that are further away (arrows, clone boundaries marked by dashed yellow line), and (J, J′) Nact and Mef2 together induce MMP1 in all clones regardless of their location. (K–N) Wing discs expressing various UAS constructs (as indicated) along with UASGFP under dppGal4 were stained for engrailed (en, red), which is mutually exclusive of the dppGal4 domain. The A/P boundary is marked with a dotted line. Low magnification images showing the topology of the wing disc are shown in K, L, M, and N. (K–K′″, L–L′″) Both control discs and discs expressing Nact alone show restricted expression of dppGAL/UASGFP along the A/P boundary, although in the Nact discs we see overgrowth of the dppGAL4 region, but all growth is into the anterior compartment of the discs and a distinct A/P boundary is maintained. (M–M′″) Mef2 expression alone results in single cells marked by GFP (arrow in M′″) invading the posterior compartment of the wing disc counterstained by en. (N–N′″) Coexpression of Nact and Mef2 not only causes very strong dppGal4/UASGFP staining along the A/P boundary but also results in groups of GFP-positive cells (arrows) invading the posterior compartment, characteristic of invasive metastatic tissue. These groups of cells do not express en (circle in N″), indicative of their anterior nature. All wing discs are oriented with dorsal to the top and posterior to the right. Download figure Download PowerPoint Matrix metalloproteinases (MMPs) have been shown to degrade basement membrane and are thus required for the metastatic behaviour of cells in both mammals (McDonnell et al, 1991; Wang et al, 2010) and Drosophila (Uhlirova and Bohmann, 2006). Therefore, we used MMP1 as a molecular marker to further characterize the Nact and Mef2 invasive phenotype. Wild-type wing discs show endogenous MMP1 in the trachea and a small region of the notum, consistent with previously published data (Figure 2B) (Page-McCaw et al, 2003). Expression of Nact using the wing pouch-specific driver MS1096Gal4 only occasionally causes a detectable upregulation of MMP1; however, MMP1 expression is markedly and consistently increased in Mef2 expressing wing discs (Figure 2C and D). We note that, even though Mef2 is ectopically expressed in the entire pouch region, MMP1 upregulation is seen only near the dorsal/ventral (D/V) boundary (as marked by the vg-LacZ reporter; Supplementary Figure S2A), which is also the defined area of active endogenous Notch signalling (Neumann and Cohen, 1998), suggesting that Mef2-dependent MMP1 expression may require Notch signals. This notion was corroborated by the fact that when we inhibit endogenous Notch activity using an RNAi against Notch in the D/V boundary of discs expressing Mef2, we detected a dramatic downregulation of the ectopic MMP1 expression (Figure 2G). When Nact and Mef2 are coexpressed, MMP1 is highly upregulated throughout the entire wing disc (Figure 2E; Supplementary Figure S2B). To ensure that these observations are not specific to the MS1096Gal4 driver, we repeated this analysis using another wing driver, vgGal4, and obtained similar results. To extend this analysis we generated somatic clones overexpressing Nact and Mef2 and, consistent with the above observations, ectopic expression of Nact alone did not induce MMP1 in clones (Figure 2H and H′). Mef2 alone induced some punctate expression of MMP1 only in clones that were located near the D/V boundary but not in clones located away from the boundary (Figure 2I and I′). On the other hand, coexpression of Nact and Mef2 resulted in high levels of MMP1 in every single clone regardless of its position relative to the D–V boundary (Figure 2J and J′). We furthermore find that the basement membrane immediately surrounding Nact+Mef2 clones is disrupted as indicated by gaps in the laminin expression pattern (Supplementary Figure S2D and D′). Thus, the synergy between Nact and Mef2 not only causes hyperproliferation but also induces MMP1 expression and basement membrane degradation. The ectopic eye phenotype and MMP1 upregulation seen in cells coexpressing Nact and Mef2 all point to the possibility that these cells have acquired invasive characteristics. To test this hypothesis directly, we asked whether cells confined to a specific cellular compartment could be induced to cross the boundary to another compartment upon coexpression of both Nact and Mef2. For this, we used the decapentaplegic-Gal4 (dppGal4) driver, which is expressed in the anterior compartment of the A/P boundary of the wing imaginal disc (Raftery et al, 1991); since Dpp represseses engrailed (en), cells expressing dppGal4 and En are mutually exclusive, creating, in normal wing discs, a clearly defined and contiguous A/P boundary (Sanicola et al, 1995). We used dppGal4 to express Nact and/or Mef2, along with GFP to mark dppGal4-expressing cells, and asked whether or not GFP-labelled cells could migrate away from the boundary into the En expression domain. In clones expressing Nact alone we never saw GFP-positive cells mixing with the En cells (Figure 2L and L′″). In contrast, single GFP-positive cells were seen some distance away from the A/P boundary among the En-positive cells upon overexpression of Mef2 alone (Figure 2M and M′″). These single cells may be apoptotic given their morphology, especially given that Mef2 causes an increase in caspase-3 cleavage (see data below in Figure 6). Most strikingly, when Nact and Mef2 were coexpressed, we observed larger clusters of cells invading the En domain (Figure 2N and N′″). Notably, the GFP-positive cells in these clusters did not themselves express en, indicating that they are indeed cells that have migrated out of the dppGal4 domain. This invasive behaviour of Nact and Mef2 coexpressing cells was further explored using a larval metastasis assay established by Pagliarini and Xu (2003), in which gene expression is driven in eye discs by eyFlp; clusters of cells (which may be as small as single cells) marked with GFP are then monitored for migration into the larval body. Consistent with the dppGal4 results, Nact alone did not induce any migration of GFP-labelled cells while Mef2 expression did result in a few migrating cells (average no. of clusters=4). These primarily appeared to be single cells rather than larger clusters. This Mef2 phenotype was significantly enhanced when Nact was coexpressed, such that both the size and number of ectopic GFP-positive cell clusters was markedly increased (average no. of clusters=6) (Supplementary Figure S3). Live imaging of GFP-labelled somatic clones coexpressing Nact and Mef2 also revealed that individual GFP-positive cells could be seen migrating away from the main cluster of cells within a 45-min timeframe (Supplementary Movie S1). In contrast, the cells in clones expressing only Nact remained as a static cluster (Supplementary Movie S2). Clones expressing Mef2 alone do not show any movement, although they do send out transient processes (Supplementary Movie S3). We note that Mef2 clones are smaller than Nact+Mef2 clones because the synergy between Nact and Mef2 results in increased proliferation and hence a larger clone; therefore, we cannot exclude the possibility that the lack of migrating cells seen in Mef2 clones could be due to decreased cell numbers, a notion that is supported by the fact that we do indeed see a few migrating Mef2 cells in the dppGal4 and larval metastasis assays. Notch–Mef2 synergy affects epithelial integrity A noteworthy aspect of clones coexpressing Nact and Mef2 is that they display a distinctive morphology. These clones are rounded, sometimes even forming hollow ball-shaped structures, and often appear to be protruding from the flat epithelial cell layer of the disc (Figure 3C and C′), consistent with the notion that Nact and Mef2 coexpression causes disruption of epithelial integrity, leading to metastatic cellular behaviour. We were thus prompted to examine key molecules previously linked with loss of epithelial integrity (Rodahl et al, 2009), such as actin and β-integrin. Figure 3.Notch and Mef2 synergize to disrupt epithelial integrity. (A–F′) Clones expressing Nact and/or Mef2 were generated using hsflp; Actin5C>y+>Gal4;UASGFP (10′, 37°C heat-shock 24–48 h (A–C′) or 48–72 h (D–F′) AEL) and stained with phalloidin (red, A–C′), or anti-β-integrin (red, D–F′). Coexpression of Nact and Mef2 upregulates the expression of the epithelial markers actin (C–C′) and β-integrin (F–F′) and appears to cause changes in cell adhesion properties. Note the characteristic nature of the Nact and Mef2 clones, which protrude out of the wing disc membrane as a rounded and well-defined group of cells (C, arrow). All wing discs are oriented with dorsal to the top and posterior to the right. Download figure Download PowerPoint Clones expressing Nact alone showed no detectable change in either actin or β-integrin expression (Figure 3A–A′ and D–D′). Mef2 expression caused an upregulation of actin and β-integrin levels, which was clearly further enhanced by coexpressing Nact (Figure 3B–B′). Discs large (Dlg), which normally localizes to basolateral junctions, is also upregulated and mislocalized throughout Nact+Mef2 clones, including at the apical surface (Supplementary Figure S2E and E′), raising the possibility of a loss of normal apicobasal polarity. Notably, the highest levels of actin and β-integrin were observed at the clone borders raising the possibility that these cells may have altered cell adhesion properties, which could promote formation of cellular ensembles that can thus escape the epithelial integrity of the surrounding wild-type cells. Mef2 is linked to the JNK signalling pathway It has been previously demonstrated that the JNK signal activation due to mutations in scribbled, a Drosophila tumour suppressor gene, causes an upregulation of MMP1, resulting in invasive cellular behaviour (Uhlirova and Bohmann, 2006). Furthermore, upregulation of JNK signalling has been linked to the control of both epithelial integrity and proliferation. We were thus led to examine whether the mechanism underlying the invasive, hyperproliferative behaviour of Nact and Mef2 cells is related to JNK signalling. To monitor JNK signal activation, we used the transgenic reporter, puckered-lacZ (puc-LacZ) (Martín-Blanco et al, 1998). Puckered is a transcriptional target as well as a feedback inhibitor of JNK. Activating Nact alone in the D/V boundary cells using vgGal4 caused a slight upregulation of puc-LacZ (Figure 4B). In contrast, Mef2 alone led to a clear upregulation of puc-LacZ, which was significantly amplified upon coexpression of Nact (Figure 4C and D; Supplementary Figure S2C). Clonal analysis also showed the same result (Supplementary Figure S4). Figure 4.Blocking the JNK signalling pathway with BskDN results in rescue of Notch+Mef2 synergy. (A–D) Wing discs expressing various UAS constructs (as indicated) under vgGal4 and stained for puc-LacZ (green), a marker for JNK activity. Note that Mef2 induces puc-LacZ expression in the D/V boundary (C), which is enhanced many-fold when Nact is coexpressed (D). (E–I) DIC images of vgGal4-driven UAS constructs (genotypes as indicated) showing the varied size differences. Note that blocking the JNK pathway using BskDN (I) partially rescues the Nact and Mef2-induced gigantic disc phenotype, and also reduces the expression of MMP1 (J, K; red) and puc-LacZ (L, M; green) to near normal levels. All wing discs are oriented with dorsal to the top and posterior to the right. Download figure Download PowerPoint Blocking the JNK pathway rescues both invasiveness and overgrowth If JNK signalling is the crucial event downstream of the Nact and Mef2 synergy, then one might expect that blocking JNK signals should rescue the synergistic phenotype. We thus blocked the JNK pathway using a dominant-negative allele of the Drosophila JNK gene Basket, BasketDN (BskDN) (Adachi-Yamada et al, 1999). Strikingly, expression of BskDN, along with Nact and Mef2, suppressed the overproliferation phenotype, MMP1 upregulation, and puc-LacZ expression in wing discs (Figure 4F–M). BasketDN alone does not have any appreciable effect on disc size or MMP1 expression (data not shown). It therefore appears that the Nact and Mef2 synergistic phenotype is dependent upon activation of the JNK pathway. Notch and Mef2 synergy controls the expression of a JNK pathway-activating ligand Given that the Nact and Mef2 synergistic effects depend on JNK signals, we sought a deeper mechanistic insight into this relationship. The only known ligand that activates the JNK pathway in Drosophila is eiger (egr), a member of the TNF superfamily (Igaki et al, 2002a, 2002b; Moreno et al, 2002). Overexpression of Egr along with Nact in wing discs resulted in a large disc phenotype very similar to that seen by coexpressing Nact and Mef2. MMP1 was also synergistically upregulated in these discs (Supplementary Figure S5), suggesting that the regulation of the JNK pathway could be at the level of the ligand. To investigate this, we performed immunofluorescence for Egr in clones overexpressing Nact and/or Mef2. Wing discs with control LacZ-expressing clones show only endogenous Egr expression in the notum region of the disc, consistent with previously published data (Figure 5A–A″; Igaki et al, 2002a, 2002b). Clones overexpressing Nact or Mef2 alone in the wing disc did not show any notable ectopic expression of Egr; however, every clone coexpressing Nact and Mef2 displayed upregulated levels of Egr (Figure 5B–D″). Figure 5.Eiger, a TNF ligand, is required downstream of Nact and Mef2 signals. (A–D″) Clones expressing Nact and/or Mef2 were generated (see Materials and methods) and stained for Egr (green). Clones are marked by lacZ (red: anti-β-gal). Normal expression of Egr is seen in the notum (*) in each genotype. Note the synergistic upregulation of Egr expression (green) by Nact and Mef2 together (D–D″) as compared with either alone (B–B″ and C–C″). Only along the D/V boundary do a few clones that express Mef2 alone display slight upregulation of Egr (arrows in C′). (E–I) Rescue of the Nact and Mef2-induced giant disc phenotype in egr1/egr1 background. Expression of Mef2 (F–G) or Nact and Mef2 together (H–I) in the egr1/egr1 background rescues MMP1 to near normal levels and partially rescues the Nact+Mef2-induced giant disc phenotype. All wing discs are oriented with dorsal to the bottom and posterior to the left. (J) Gel shifts performed using a 200-bp fragment of the egr promoter bearing Su(H) and MEF2-binding sites and a control fragment with the sites mutated (see Materials and methods). Whole protein extract from wild-type and Nact+Mef2 overexpressing wing discs show gel shifts using the wild-type fragment that could be competed using cold probe. A mutant version of the same fragment does not show significant gel shifts, and therefore no competition. (K–M) Supershifts using antibodies against Nact (K) and Mef2 (L, M), using 30 mer complementary oligonucleotides carrying the Su(H) or M