Introduction

In the past decade, a keen interest has been shown to explore the oncogenic cooperation with loss of cell polarity in tumor progression and malignancy. Studies in Drosophila have revealed that the oncogenic form of Ras cooperates with loss of tumor suppressors, namely scrib, lgl and dlg to cause tumor cell invasion [1, 2]. The oncogenic form of Notch has also shown to cooperate with scrib−/− to induce neoplastic overgrowth [2]. The loss-of-function mutation of Scribble complex genes (scrib, lgl and dlg) results in disruption of epithelial integrity followed by neoplastic tissue hyperproliferation [3,4,5]. However, the tumor formation caused by loss of scrib, lgl and dlg has been found to be restricted by the compensatory JNK mediated apoptosis [2, 6,7,8]. Among the Scrib complex genes, lgl was the first neoplastic tumor suppressor gene described in Drosophila [9]. The phenotypes of lgl mutant tissues show close similarity with that of the human epithelial cancers [10,11,12]. Although it has been shown that Notch cooperates with scrib−/− to induce neoplastic growth, it is still unknown whether Notch works in the same way with loss-of-function of lgl also. Recently, Lgl has been shown to regulate Notch signaling via endocytosis [13]. However, it gives no substantial evidence on coupling of lgl-Notch effect on tumorigenesis. In the present study, we checked the effect of a tumor suppressor gene mutation, lgl, in activated Notch background, and found that lgl downregulation synergizes with activated Notch to induce overgrowth and migratory behavior. Here, we show that Nact/lgl-IR tissues display the hallmarks of tumor overgrowth. Moreover, our study revealed that the effect of Nact/lgl-IR tumor is mediated by the activation of JNK signaling through the upregulation of its receptor, Grindelwald.

Main text

Methods

Detailed description of methods used in this study is provided in Additional file 1.

Results

Oncogenic Notch synergizes with RNAi mediated downregulation of lgl to promote tissue overgrowth

Coexpression of both lgl-IR and Notchact in the Drosophila eye discs using ey-GAL4 dramatically induced overgrowth (Fig. 1d, d″) as compared to that of only Nact overexpressed (Fig. 1b, b″) or only lgl-IR overexpressed (Fig. 1c, c″) eye discs. To further describe the phenotype of Nact/lgl-IR tumor, expression of Matrix metalloproteinase 1 (MMP1) was monitored. MMPs are enzymes with clear association to tumor cell invasion and cancer progression [14, 15]. Coexpression of Notchact and lgl-IR resulted in massive upregulation of MMP1 expression throughout the entire eye disc (Fig. 1d′) as compared to that of only Nact or only lgl-IR (Fig. 1b′, c′). Further, we extended our observation into the brain since ey-GAL is mildly expressed in the brain also. Except endogenous expression, no MMP1 activation was observed in the ey-GAL4 driven lgl-IR (Fig. 1g′) and Nact larval brain (Fig. 1f′). In case of Notchact/lgl-IR larval brain, excessive amount of GFP marked cells with enhanced MMP1 expression was observed in the optic lobes (Fig. 1h, h′). The increment in GFP and MMP1 expression was also found in the ventral nerve cord (VNC) of Notchact/lgl-IR larval brain (Fig. 1h, h′ marked with arrows). This indicates that the weak expression of ey-GAL4 in VNC is also inducing MMP1 expression in Notchact/lgl-IR tissue. When we quantified the amount of GFP in upper region of VNC, a significant increment in the amount of GFP in Notchact/lgl-IR was found as compared to that of the controls (Additional file 2: Figure S1a). We also quantified the presence of MMP1 in the VNC of Notchact/lgl-IR (Additional file 2: Figure S1b), which clearly shows a significant increase as compared to that of the controls. Moreover, transcript levels of mmp1 in the cephalic complex were also found to be upregulated in Notchact/lgl-IR tumor as compared to that of the controls (Additional file 2: Figure S1c).

Fig. 1
figure 1

Oncogenic Notch synergizes with lgl-IR to induce tissue overgrowth. Fluorescent micrographs of eye imaginal discs and larval brains are shown. ey-GAL4/+ (control) eye imaginal disc shows expression of GFP (a) and MMP1 (a′). ey-GAL4 driven UAS-Nact results in enlarged disc size (b) and shows slight enhancement in MMP1 expression (b′). UAS-lgl-IR eye disc driven by ey-GAL4 (c) induces MMP1 expression (c′). UAS-Nact coexpressed with UAS-lgl-IR results in massively overgrown eye discs (d) and significant enhancement in MMP1 expression throughout the tissue (d′) compared to that of only lgl-IR (c′) or only Notchact (b′) overexpressed eye discs. Images a″–d″ are merges of those in a–a′, b–b′, c–c′ and d–d′ with DAPI, respectively. Expression of GFP and MMP1 in brains of ey-GAL4 driven UAS-Nact (f, f′) and UAS-lgl-IR (g, g′) remain similar as of wild-type brain (e, e′). h UAS-Nact and UAS-lgl-IR coexpressed brain shows massive expression of GFP in the optic lobes and in ventral nerve cord (arrowhead marks the expression of GFP in VNC). h′ The optic lobes and VNC of Nact/lgl-IR brain shows extensive expression of MMP1 (arrow marks the expression of MMP1 in VNC). Images e″–h″ are merges of those in e–e′, f–f′, g–g′ and h–h′ with DAPI, respectively. All eye discs are oriented with dorsal to the left and anterior to the top position. Dorsal view of the brains is shown. Scale bars, 50 µm (a–d, a′–d′, a″–d″) and 100 µm (e–h, e′–h′, e″–h″)

Full size image

In order to examine the cytoskeleton network and cell–cell adhesion, we marked the tissues with phalloidin and adherens junction marker proteins, Armadillo (Arm) and Cadherin (DE-Cad). The F-actin network marked by phalloidin revealed a defective actin cytoskeleton network in Nact/lgl-IR tumor tissues compared to that of controls (Additional file 3: Figure S2). In the same way, the localization of DE-Cad and Arm were also deregulated in Nact/lgl-IR tumorous eye discs (Additional file 4: Figure S3a–d, e–h). We, next, determined if neuronal differentiation was defective in Nact/lgl-IR tumor using a neuronal marker, Elav that marks the differentiated neurons in eye disc and brain. Remarkably, coexpression of Nact and lgl-IR led to severe loss of Elav positive cells in the eye disc and abnormal expression of Elav in the optic lobes indicating an impaired neuronal differentiation (Additional file 4: Figure S3i–l, m–p).

In parallel, we also used dominant-negative version of Notch to see the effect of depletion of Notch signaling on lgl-IR tumors. Previously, expression of mamDN in lgl− tissues partially rescued the lgl− mosaic adult eye phenotype [13]. Our analysis also found that reduction of Notch signaling partially rescued the phenotypes of lgl loss-of-function flies (Additional file 5: Figure S4). Thus, our analysis support the notion that the lgl loss-of-function wing phenotype is dependent on elevated Notch signaling, consistent with the previous study [13].

Involvement of JNK pathway in N act /lgl-IR tumor

Previous studies in Drosophila have revealed that oncogenic Ras along with loss of lgl or scrib or dlg induces JNK signaling, which is crucial for tumor invasion [7, 16]. This prompted us to check the expression of Puckered (puc), a transcriptional target of JNK signaling and widely used to check the activation of JNK signaling. An enhancer trap allele, puc-LacZ [17] was used to monitor the activation of JNK signaling. Coexpression of both Nact and lgl-IR resulted in intense upregulation of puc throughout the wing disc (Fig. 2d), indicating the activation of JNK signaling in Nact/lgl-IR tumor. We also observed a significant increase in size of the wing disc in Nact and lgl-IR coexpressed condition compared to that of the wild-type, only Nact, and only lgl-IR wing discs (Fig. 2i).

Fig. 2
figure 2

Activation of JNK signaling in Nact/lgl-IR tumor. Fluorescent micrographs of wing imaginal discs are shown. a Endogenous expression of puc in Drosophila wing imaginal disc is shown. b Only Nact overexpression shows slight increment in puc expression. c Downregulation of lgl in wing disc induces puc expression. d A significant increase in puc expression was found, when Nact is coexpressed with lgl-IR. a′–d′ are the merge images of DAPI along with a–d, respectively. Images (e–h) show the expression of Egr in wild-type, Nact, lgl-IR, and Nact/lgl-IR wing discs, respectively. The expression of Egr remains unchanged in all genotypes. e′–h′ are the merge images of DAPI along with e–h. i Analyses of differences in wing disc size show that there is a significant increase in Nact/lgl-IR wing imaginal disc size as compared to the wild-type, Nact, and lgl-IR wing disc sizes. j Real-time PCR analysis was performed to estimate the transcript levels of JNK pathway ligand egr and its receptors wgn and grnd in Nact/lgl-IR tumor. The transcript levels of egr and wgn in Nact/lgl-IR were significantly reduced, when compared to that of wild-type; however no significant change in transcript levels were seen, when compared to that of only Nact and lgl-IR. Interestingly, grnd transcript levels were significantly upregulated in Nact/lgl-IR tumor as compared to that of the wild-type, Nact and lgl-IR wing discs. qPCR was normalized with rps17 and repeated for three times. Analysis of data was done using two-way ANOVA with Tukey’s multiple comparison test; data represents mean ± SEM (***p < 0.001; **p < 0.01; *p < 0.05 and ns p > 0.05). All wing discs are oriented with dorsal to the top and posterior to the right. Scale bars: 100 µm (a–d, a′–d′, e–h, e′–h′)

Full size image

To check the mode of activation of JNK signaling, we examined the transcript level expression of ligand eiger (egr), and its receptor wengen (wgn), in Nact/lgl-IR tumor. egr and wgn transcript levels were found to be depleted in case of Nact/lgl-IR tumor as compared to that of the controls (Fig. 2j). Recently, another member in tumor necrosis factor receptor superfamily, Grindelwald (Grnd), found to be associated with loss of cell polarity and neoplastic growth [18]. Interestingly, a significant upregulation of grnd transcripts in Nact/lgl-IR tumor was found, when compared to that of the wild-type, only Nact and only lgl-IR tissues (Fig. 2j). We went on to check the protein level expression of Egr in Nact/lgl-IR tumors. Immunostaining with anti-Egr antibody [19] revealed that there is no change in the level of Egr protein expression in Nact/lgl-IR tumor (Fig. 2h) as compared to that of the wild-type, only Nact and only lgl-IR tissues (Fig. 2e–g). As Egr is known to be also expressed by the tumor-associated hemocytes, leading to signaling activation [20], these immune cells may be in this case responsible for Grnd activation, but their poor adhesion to the tumor tissue may make them escape Immunofluorescence detection.

To further confirm the involvement of JNK signaling as a downstream event of Nact/lgl-IR cooperation, we blocked JNK signaling in Nact/lgl-IR tumor, and checked whether blocking JNK could affect the Nact/lgl-IR tumor. The massive upregulation of MMP1 in Nact/lgl-IR tumor (Additional file 6: Figure S5a) was drastically suppressed, when bsk-DN (a dominant negative allele of Drosophila JNK gene, basket) was expressed in the background (Additional file 6: Figure S5b). In addition, coexpression of bsk-DN with Nact; lgl-IR resulted in a reduced wing disc size as compared to Nact/lgl-IR overexpressed wing disc (Additional file 6: Figure S5c). These results indicate that JNK signaling may be involved in the tumorous overgrowth of Nact/lgl-IR tissues.

N act /lgl-IR tumor induces cell death

Eluding apoptosis is considered as one of the acquired capabilities of many types of cancer; however, studies also explain that elevated oncogenic signaling induces apoptosis or senescence [21]. When we checked the status of cell death in Nact/lgl-IR tumor, we observed a significant amount of acridine orange (Compare Fig. 3d with a–c) and caspase positive cells (Compare Fig. 3i with f–h) indicating severe cell death. Since loss of lgl in a tissue known to induce cell competition to remove the unfit cells [22], dying cells in Nact/lgl-IR tissue could be an indication of cell competition. To check the effect of cell death on overgrowth and MMP1 expression, we blocked cell death by expressing a caspase inhibitor, p35 (Fig. 3e, j). It was found that blocking cell death in Nact/lgl-IR overexpressed condition did not obstruct MMP1 expression (Fig. 3o). Coexpression of p35 with Nact/lgl-IR resulted in an increased wing disc size as compared to Nact/lgl-IR overexpressed wing disc (Fig. 3r). As the caspase inhibitor, p35 is known to block cell death [23], the increase in the tissue size is expected since blocking cell death in Nact/lgl-IR tumor allowed more cells to overgrow that, in turn, increased the disc size.

Fig. 3
figure 3

Nact/lgl-IR tumor induces cell death. Fluorescent micrographs of wing imaginal discs are shown. Acridine orange-staining in wild-type (a), Nact over-expressed (b), lgl-IR over-expressed (c) and both Nact/lgl-IR coexpressed (d) wing discs are shown. Expression of cleaved caspase 3 in wild-type (f), Nact over-expressed (g), lgl-IR over-expressed (h) are shown. i Nact/lgl-IR coexpressed wing disc shows upregulation in cleaved caspase 3 expression. MMP1 expression in wild-type (k), Nact overexpressed (l), lgl-IR overexpressed (m) are shown. n Nact/lgl-IR wing disc shows massive upregulation in MMP1 expression. Blocking cell death in Nact/lgl-IR tissues by expressing UAS-p35 leads to absence of acridine orange positive cells (e) and cleaved caspase-3 marked cells (j); whereas the expression of MMP1 is found to be unaltered (o). p Intensity per unit area for acridine orange shows that there is significant amount of upregulation of acridine orange-positive cells in Nact/lgl-IR wing disc. q Intensity per unit area for caspase staining shows that the Nact/lgl-IR wing disc contains significantly upregulated level of caspase activity than that of wild-type, lgl-IR and Nact wing discs. r Wing disc size analyses show that the disc size is increased when p35 is expressed in the background of Nact/lgl-IR, as compared to Nact/lgl-IR. Analysis of data for intensity profiling was done using two-way ANOVA with Tukey’s multiple comparison test; data represents mean ± SEM (***p < 0.001; **p < 0.01 and ns p > 0.05. Analysis of data for wing disc size was done using Unpaired t test with Welch’s correction; data represents mean ± SEM *p < 0.05). All wing discs are oriented with dorsal to the top and posterior to the right. Scale bars: 100 µm for (a–e), and 50 µm for (f–j, k–o)

Full size image

Discussion

In the present study, we unveil a cooperation of Notch with RNAi-mediated downregulation of a polarity cum tumor suppressor gene, lgl to promote tumor overgrowth. Our data, presented here, illustrate that coexpression of Nact and lgl-IR in Drosophila eye disc results in overgrowth, loss of positional clues and upregulation of MMP1 expression, which is less prevalent in only Nact overexpression or only lgl-IR overexpression. Earlier the loss of polarity gene scribble found to cooperate with Notch signaling to promote neoplastic overgrowth [2]. Another two independent studies of similar context show that oncogenic Ras cooperates with loss of cell polarity genes (lgl, scrib, dlg) to induce metastasis and secondary tumor formation at distant sites [7, 14]. Interestingly, we found that Notch synergizes with loss of lgl to promote tumorous overgrowth and elevated expression of MMP1, and inhibiting Notch signaling rescues the defects caused by loss of lgl. It indicates the potential function of Notch signaling during lgl mediated tumor development. Our data also show distorted epithelial integrity in Nact/lgl-IR tumor that point towards epithelial to mesenchymal transition, where tightly joined epithelial cells with regularly spaced cell–cell junctions convert to mesenchymal cells which are of irregular shape without tight intracellular adhesion [24].

Further, we found upregulation of JNK signaling and its receptor Grindelwald in Nact/lgl-IR tumor. Two previous studies have shown that Notch cooperates with two different proteins to induce proliferation and metastasis by the activation of JNK signaling in ligand-dependent and -independent manner [25, 26]. In case of Nact/lgl-IR tumor, we show that the transcript levels of egr (ligand) and wgn (receptor) were not upregulated, whereas a significant upregulation of grnd transcripts in the Nact/lgl-IR tumor was observed. Earlier the active form of Grnd has shown to activate JNK signaling in vivo [18]. Thus, in case of Nact/lgl-IR tumor, JNK signaling might get activated through Grindelwald. Previously, it has been shown that JNK signaling can initiate tumor initiation and growth in Eiger-independent manner also [27].

Another most important hallmark of almost all types of cancer is the ability to evade apoptosis that, in turn, helps tumor cell population to increase in number [21]. In other similar tumor models such as Rasv12/dlg−/−, dying cells of dlg−/− clones evade apoptosis in presence of oncogenic Ras, where JNK signaling switches its role from proapoptotic to progrowth [7]. In contrast, Ras/scrib−/− and Ras/lgl−/−tumors were reported to show apoptosis [22, 28]. However, Notch/scrib−/− tumor did not show the presence of apoptosis [29]. In our case, Nact/lgl-IR tumor resulted in severe apoptosis along with strong overgrowth and MMP1 expression. These dying cells in Nact/lgl-IR tumor might be the indication of cell competition as there is a strong proliferation and overgrowth. In case of Nact/scrib−/− tumor, Notch is giving growth advantage to scrib−/− tissues by preventing cell death. However, in case of Nact/lgl-IR wing discs, activation of Notch failed to restrict the cell death caused by loss of lgl; rather its activation induces further cell death. These differences indicate that although oncogenic cooperation with loss of cell polarity results in similar tumor cell migration but certain property like cell death occurs depending on the context.

Limitations

  • The present study is not the first one to show the cooperation between Notch and loss of cell polarity genes. Activated Notch is known to cooperate with another cell polarity gene, scribble, to induce neoplastic overgrowth.

  • In the present study, experiments were performed using RNAi line of lgl, but not with the lgl loss-of-function mutants.