Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 May 6.
Published in final edited form as: Dev Cell. 2019 Apr 11;49(3):461–472.e5. doi: 10.1016/j.devcel.2019.03.015

Oncogenic Notch triggers neoplastic-tumorigenesis in a transition-zone like tissue microenvironment

Sheng-An Yang 1, Juan Martin Portilla 1, Sonja Mihailovic 1,a, Yi-Chun Huang 1, Wu-Min Deng 1,2,*
PMCID: PMC6504601  NIHMSID: NIHMS1524561  PMID: 30982664

Summary

During the initial stages of tumorigenesis the tissue microenvironment where the pro-tumor cells reside plays a crucial role in determining the fate of these cells. Transition zones, where two types of epithelial cells meet, are high-risk sites for carcinogenesis, but the underlying mechanism remains largely unclear. Here we show that persistent upregulation of Notch signaling induces neoplastic tumorigenesis in a transition zone between the salivary gland imaginal ring cells and the giant cells in Drosophila larvae. In this region, local endogenous JAK-STAT and JNK signaling creates a tissue microenvironment that is susceptible to oncogenic Notch induced tumorigenesis, whereas the rest of the salivary gland imaginal ring is refractory to Notch-induced tumor transformation. JNK signaling activates a matrix metalloprotease (MMP1) to promote Notch-induced tumorigenesis at the transition zone. These findings illustrate the significance of local endogenous inflammatory signaling in primary tumor formation.

Graphical Abstract

graphic file with name nihms-1524561-f0001.jpg

Blurb

The transition zone between two types of epithelial cells is vulnerable to cancer transformation. Using Drosophila salivary gland imaginal rings, Yang et al. show that local endogenous JAK-STAT and JNK signaling and MMP1 expression, make the transition zone a tumor hotspot, where Notch hyperactivation is sufficient to induce malignant tumorigenesis.

Introduction

The “seed and soil” hypothesis proposed by Stephen Paget (1889) highlights the importance of tissue microenvironment for secondary tumor formation (Fidler and Poste, 2008; Langley and Fidler, 2011). In this theory, depending on the cancer type, cancer cells (seeds) metastasize to preferential organs (soil) that bear the microenvironment favoring the colonization and growth of the metastatic cancer cells. For example, breast cancers commonly metastasize to bone, since the bone-derived chemokines osteopontin and osteonectin trigger the metastasis of chemokine receptor-expressing breast cancer cells (Coniglio, 2018; Rahim et al., 2014). In contrast, how tissue microenvironment contributes to primary tumor formation is less understood, although it is well known that specific tissue regions appear to be more susceptible to tumor transformation. One such microenvironment is the folded hinge region in Drosophila wing imaginal discs, a tumor hotspot, where tumors always form following knockdown of neoplastic tumor suppressors such as scribble (scrib) and lethal giant larvae (lgl). In this microenvironment, the endogenous inflammatory JAK-STAT signaling pathway and cytoarchitectural structures help to form the tumor hotspot (Tamori et al., 2016).

In mammals, a location where primary tumors are prone to grow is the so-called transition zone, where two types of epithelial cells meet. Transition zones are known to be associated with metaplasia, a premalignant phase of carcinogenesis (McNairn and Guasch, 2011). For example, Barrett’s esophagus (a human pre-neoplastic metaplasia), specifically arises from the transitional cells at the squamous-columnar junction when CDX2 is overexpressed (Jiang et al., 2017). Another example is the anal canal tumors in mouse models, which always originate in the transition zone between the anterior mucosal epithelium of the large intestine and the posterior stratified squamous epithelium of anal skin, when TGF-β signaling is lost (Guasch et al., 2007). Although transition zones are established tissue hotspots for carcinogenesis, precisely how the tissue microenvironments facilitate tumor origination remains unclear. Here we report a Drosophila transition-zone model, the posterior boundary of the salivary gland imaginal ring, which contains progenitor cells for adult salivary glands. Previously, we have shown that Notch signaling regulates the size of this progenitor cell pool throughout larval development (Yang and Deng, 2018).

The evolutionarily conserved Notch pathway plays crucial roles in proliferation, survival, and differentiation in metazoans. Mutations in Notch have been shown associated with many types of cancer such as T-cell acute lymphoblastic leukemia (T-ALL), breast cancer, prostate cancer, and glioma (Grabher et al., 2006; Reedijk, 2012; Stockhausen et al., 2010). Interestingly, Notch can act as a tumor suppressor or an oncogene in different cancer types. In T-ALL, Notch1 plays an oncogenic role in driving cell proliferation and promotes cancer progression via activation of the Myc gene (Weng et al., 2006). In contrast, Notch promotes cell differentiation and cell cycle arrest in the skin (Lowell et al., 2000; Rangarajan et al., 2001). Loss of Notch in this tissue results in the increase of the basal epidermal layer and carcinomas (Nicolas et al., 2003; Rangarajan et al., 2001). In Drosophila imaginal discs, overexpression of Notch disrupts expression of cell cycle regulation genes and leads to overgrowth (Baonza and Freeman, 2005; Baonza and Garcia-Bellido, 2000).

Here, we report that persistent upregulation of Notch signaling induces neoplastic transformation at the posterior end of salivary gland imaginal ring, a transition zone where diploid imaginal ring cells and polyploid salivary gland cells meet (Fig. 1). Different from the neighboring tissues, this transitional region shows high levels of inflammatory signaling, JAK-STAT and JNK, that are necessary for neoplastic tumor growth. JNK induces the expression of matrix metalloprotease, MMP1, specifically in this tissue hotspot to promote tumorigenesis.

Figure 1. Schematic representation of wildtype Drosophila salivary gland imaginal ring and imaginal ring tumor.

Figure 1.

Open in a new tab

(A) The 3D view of wildtype salivary gland imaginal ring (light blue) and larval salivary gland (orange). (B) Cross section through the anterior-posterior axis from the white panel in (A) to show the lateral view of imaginal ring. All images were traced from tissues with Phalloidin staining. SG, salivary gland; ImR, imaginal ring; A, anterior (left); P, posterior (right).

Results:

The anatomy of salivary gland imaginal ring

The salivary gland imaginal ring, located between the duct cells and salivary gland giant cells, is a well-organized tubular structure that contains a single layer of epithelial cells with their apical side facing the lumen (Fig. 1). The imaginal ring cells are precursors for the adult salivary gland cells. These precursor cells undergo rapid proliferation from late second instar larval stage, and grow from 8–10 cells to about 150 by the end of the third instar larvae (Yang and Deng, 2018) The imaginal ring cells are much smaller and are more densely organized compared with the neighboring salivary gland giant cells (Fig. 1).

Overexpression of an active form of Notch results in neoplastic tumor formation in salivary gland imaginal rings

Using the temporal and regional gene expression targeting (TARGET) technique (McGuire et al., 2004), we induced conditional expression of an active form of Notch, the Notch intracellular domain (NICD), driven by Act-Gal4ts (Actts= Act-Gal4, Gal80ts) during the third instar larval stage. When NICD was overexpressed, the larvae showed prolonged larval development and a giant-larvae phenotype (Fig. 2A, the final sizes of these larvae were about twice the size of the wildtype controls), in addition to imaginal ring overproliferation as reported previously (Yang and Deng, 2018). After seven days of continued Notch activation, abnormal cell masses (tumors) were detected in the salivary gland imaginal rings (100%, n=1225, Fig 2B). Staining with the apical-basal polarity markers atypical protein kinase C (aPKC, an apical marker) (Morais-de-Sá et al., 2010) and dystroglycan (DG, a basal marker) (Deng et al., 2003), and Phalloidin, which labels F-actin specifically, we found that these tumors had disrupted cell arrangement and tissue organization. The cells stacked upon each other and became multilayered, in stark contrast to the well-organized, monolayered epithelium of the wildtype imaginal ring (Fig. 2C).

Figure 2. Over-active Notch signaling leads to neoplastic tumor formation in the salivary gland imaginal ring.

Figure 2.

Open in a new tab

(A) Giant larvae phenotype resulting from Actts> NICD. (B) Mitotic activity was determined by PH3 staining (red) in wildtype salivary gland imaginal rings, and Actts>NICD at 2 days and 7 days after temperature shift. The graph shows the mitotic index (the ratio of PH3 positive cells to total cells counted) for tumor hotspot and coldspot on D2 (2 days of NICD induction; coldspot, 7.7%; hotspot, 8%) and D7 (coldspot, 0.7%; hotspot, 6.7%). n=25 for each. Error bars, mean±S.E.M. (C) Staining with aPKC and DG (white) indicates cell polarity. Phalloidin (PHA, white) displays the cell organization in wildtype salivary gland imaginal rings, coldspot hyperplasia and hotspot neoplasms in the salivary gland imaginal ring Actts>NICD 7 days after temperature shift. (D) The adult hosts after transplantation with Actts>GFP salivary gland imaginal rings (control), or Actts>NICD, GFP tumors. GFP signal was absent in controls even on D21 after transplantation. ATP, tumors after transplantation. White dashed lines outline the host adult flies. (E) Transplantation of the Notch-induced tumors with GFP marker. BTP, tumors before transplantation (n=12); ATP (n=15). Error bars, mean±S.E.M. (F) GFP (green) shows the invasive tumor cells in secondary sites. PHA (red) outlines tissue boundaries. H, tumor hotspot; C, tumor coldspot; ImR, imaginal ring. Orange dashed lines circle the posterior end tumors (B and C). Nuclei labeled with DAPI (green in B, blue in C, E and F). Scale bar, 20μm. See also Figure S1

Because the larvae with NICD misexpression died 7–9 days (D) after NICD induction, we transplanted individual primary tumors and wildtype salivary gland imaginal rings into the abdomens of wildtype adult flies. The transplanted tumors grew steadily in the hosts (n=89, Fig. 2D). The average size of tumors on D21 after transplantation was about 30-fold the original size (Fig. 2E). To confirm the long-term proliferative ability, pieces of the transplanted tumors were reinjected into new hosts for several generations (140 days after the initial NICD induction, six re-transplantations). These tumors continued uncontrolled growth in the hosts, suggesting tumor immortality. In contrast, the wildtype control imaginal rings, though survived in the host, were unable to grow after transplantation (n=30). Among the hosts with transplanted tumors, we noticed that GFP-positive cells disseminated occasionally from the transplanted solid tumors and migrated to host abdominal tissues, which include the ovarian muscle sheath (4 in 60), the oviduct (4 in 60) and the Malpighian tubule (3 in 60) (Fig. 2F). Unlike the initial spherical transplanted tumors (Fig. 2E), the secondary invasive tumor cells were spread out and associated with the host tissues (Fig. 2F). The GFP-positive tumor cells attached to host tissues at secondary sites only showed up 21 days after transplantation, and did not appear in hosts on D3, D7 and D14 after transplantation (50 hosts examined in each group). These results suggest that the disseminated GFP-positive tumor cells are derived from the transplanted tumor, and not from dissociated cells during the process of transplantation. These findings further confirm the neoplastic characteristics of these tumor cells, which can invade and colonize secondary sites.

NICD-induced neoplastic tumors specifically grow at the posterior end of the salivary gland imaginal ring

Upon examination of the location of NICD-induced primary tumors in salivary glands, we found that they were always located at the posterior end of the salivary gland imaginal ring next to the giant polyploid cells (Fig. 2B). The rest of the imaginal ring, although showing over-proliferation, still maintained the monolayer-epithelial organization and an intact apical-basal epithelial polarity (Fig. 2C). An M-phase marker, Phospho-histone 3 (PH3) (Gurley et al., 1978), was predominantly detected in the tumorous posterior end (Fig. 2B). The mitotic rate was much lower in the anterior region of the imaginal ring (Fig. 2B). These findings suggest that various regions of the salivary gland imaginal ring respond to Notch activation differently in tumor transformation. To determine whether the anterior NICD+ hyperplasia would eventually undergo neoplastic transformation, we followed them in transplants and found that they survived on D21 after transplantation. However, they maintained a similar size and an organized tubular structure, easily distinguishable from the posterior neoplasm (Fig. S1A), suggesting that the anterior cells, although over-proliferated initially, fail to undergo neoplastic transformation even after long-term exposure to Notch signaling.

To determine precisely which cells in salivary gland imaginal rings undergo tumorigenesis upon NICD overexpression, we screened for Gal4 lines showing specific expression patterns in the imaginal ring, and identified retained (retn)-Gal4 (entire salivary gland imaginal ring), shotgun (shg)-Gal4 (posterior-most row of cells in the salivary gland imaginal ring and giant salivary gland cells), string (stg)-Gal4 (anterior salivary gland imaginal ring), and Posterior sex comb (Psc)-Gal4 (anterior imaginal ring and giant salivary gland cells) (Fig. 3A). Unlike Actts, NICD overexpression using these Gal4 lines during L3 did not prolong larval development enough to allow sufficient time for tumor development. Nonetheless, we examined the larvae on D3 after NICD induction, by which time dysplasia with multilayering and loss-of-cell-polarity phenotypes had begun to show at the posterior end of the imaginal ring when Actts was used as a driver (88%, n=210) (Fig. 3B). Consistently, multilayering of cells at the posterior was also found when NICD was overexpressed by retnts (64.3%, n=241) or shgts (4.7%, n=236), both of whose expressions included the row of posterior-most salivary gland imaginal ring cells (Fig. 3A, B). Using the two anterior-Gal4 lines to drive NICD overexpression, although overproliferation was detected (Fig. S2A), no multilayering was detected in the entire salivary gland imaginal ring (stgts: n=264; Pscts: n=283) (Fig. 3B). We further performed transplantation analyses to determine the long-term effect of NICD overexpression on these imaginal rings. When NICD was driven by retnts or shgts, the posterior dysplasia grew continuously and underwent neoplastic transformation in the host (retnts: 15-fold change; shgts: 3-fold change on D21; Fig. S1B). However, Pscts>NICD imaginal rings still showed an organized single-layer tubular structure 21 days after transplantation, and the tissue size did not change significantly (Fig. S2B). Taken together, these data indicate that the posterior-most imaginal ring cells, but not the anterior cells, are susceptible to oncogenic Notch-induced neoplastic transformation. This posterior boundary region fits the definition of a “transition zone,” as it is where the diploid imaginal ring cells and polyploid salivary gland cells meet. The cells within the transition zone are therefore referred as “transition cells” hereafter.

Figure 3. Tumors arise from the posterior-most cells of the salivary gland imaginal ring.

Figure 3.

Open in a new tab

(A) Gal4-driven GFP expression patterns (green) in the salivary gland imaginal ring with indicated genotypes. White dashed lines outline the posterior end salivary gland imaginal ring cells. H, tumor hotspot. (B) Cross-section for salivary gland imaginal ring with indicated genotypes. Salivary gland imaginal ring cells were outlined by PHA (red) and Dlg (green), which is enriched in a narrow basal-lateral domain, was used for cell polarity analyses. Second panel: line drawings trace PHA staining and magenta areas show the multilayer structures at posterior ends. The bottom panel magnifies the regions indicated by white boxes. Arrows indicate the mislocalized Dlg staining at posterior ends. Nuclei labeled with DAPI (blue). Scale bar, 20μm. See also Figure S1 and S2.

JAK-STAT signaling promotes Notch-induced tumorigenesis in the transition zone

To determine why the transition cells but not the anterior cells are susceptible to NICD-induced tumorigenesis, we examined the activation pattern of JAK-STAT signaling in the salivary gland. This is because endogenous JAK-STAT activity is key to tumor hotspot formation in the hinge region of the wing imaginal disc (Tamori et al., 2016). Indeed, STAT92E-GFP, a reporter for JAK-STAT activation (Bach et al., 2007), showed strong expression in the transition zone (Fig 4A). STAT92E-GFP was also highly expressed in NICD-induced imaginal ring tumors (Fig 4C). JAK-STAT signaling in these transition cells is probably activated by neighboring salivary gland giant cells, as Uupaired 1 (Upd1), a ligand for JAK-STAT signaling, is highly expressed in the giant cells adjacent to the transition cells (Fig. 4A and 4B).

Figure 4. JAK-STAT signaling enhances Notch-induced tumor progression in the salivary gland imaginal ring.

Figure 4.

Open in a new tab

(A) STAT92E-GFP expression (green) and expression of RFP driven by upd1-Gal4 (red) in the wildtype salivary gland imaginal ring. (B) Upd1-YFP expression in wildtype and loss-of-upd1 salivary gland imaginal rings. Upd1-YFP expression is completely suppressed by upd1-RNAi, suggesting the YFP expression is not background signal. (C) STAT92E-GFP expression (white) in NICD-induced imaginal ring tumor. (D) Salivary gland imaginal ring tumors with indicated genotypes. Orange dashed lines circle the tumor regions. Graph shows the average sizes of tumors. Numbers in the graph indicate sample sizes. ***, P<0.001. Error bars, mean±S.E.M. Nuclei labeled with DAPI [blue in (A, B and C), white in (D)]. Scale bar, 20μm. See also Figure S3.

To determine whether JAK-STAT signaling is necessary for NICD-induced transition-zone tumorigenesis, we knocked down JAK-STAT pathway components in Actts>NICD tumors. The average tumor sizes were reduced by 79% (hop-RNAi) or 41% (Stat92E-RNAi) on D7 after NICD induction when compared with the NICD-alone tumors (Fig 4D), suggesting that JAK-STAT signaling promotes Notch-induced transition-zone tumorigenesis in salivary glands. Next, we tested whether elevated JAK-STAT signaling can lead to tumor transformation when NICD is induced in the anterior end of the imaginal ring. As Actts driven STATΔNΔC (active STAT, Ekas et al., 2010), or hop resulted in early larval lethality, we used the retnts driver instead. Overexpression of STATΔNΔC or hop alone did not cause any obvious changes in imaginal ring size or morphology (Fig S3A). When retnts was used to co-overexpress NICD and STATΔNΔC, no tumorigenesis or multilayering of cells was detected in the anterior; the multilayer structures and disruption of cell polarity again only occurred in the posterior transition zone (100%; n=144, Fig S3B). Taken together, these findings suggest that JAK-STAT signaling is required but insufficient for NICD-induced tumorigenesis in the salivary gland imaginal ring.

To determine whether JAK-STAT promotes tumor growth through regulating cell proliferation, we knocked down components of the JAK-STAT pathway by Actts during the third larval instar and examined cell proliferation and imaginal ring morphology. These imaginal rings showed normal mitotic activity and morphology when compared with wildtype controls (Fig. S3D). In addition, loss of JAK-STAT did not change the overall size of the anterior hyperplasia caused by NICD overexpression (Fig 4D). These results indicate that JAK-STAT is not involved in cell proliferation regulation in salivary gland imaginal rings.

JNK signaling is required for NICD induced tumorigenesis in the transition zone

We further searched for other signaling pathways that might be also active in the transition cells of the salivary gland imaginal ring. From the screen of a collection of signaling reporter lines, we found that two reporters of the JNK pathway, Puckered-LacZ (Puc-LacZ) (Ring and Martinez Arias, 1993) and TPA response element-RFP (TRE-RFP) (Chatterjee and Bohmann, 2012), were both expressed in the transition cells (Fig 5A), suggesting JNK signaling is active at the transition zone. Eiger (Egr), a TNF superfamily member, is a ligand for JNK signaling activation in Drosophila (Igaki et al., 2002). Using an antibody against Egr to stain the imaginal rings, strong signal was detected at the border between the salivary gland giant cells and the imaginal ring cells, but not at other parts of the imaginal ring (Fig. 5D). Knockdown of egr in the imaginal ring by Actts or retnts eliminated almost all Egr staining and resulted in lower TRE-RFP intensity (Fig. 5C and 5D), suggesting Egr is likely produced by the imaginal ring cells. In contrast, knockdown of egr in salivary gland giant cells (driven by RhoGEF3-Gal4) showed no change in TRE-RFP or Egr expression levels (Fig. 5C and 5D).

Figure 5. JNK signaling promotes Notch-induced tumorigenesis in the salivary gland imaginal ring.

Figure 5.

Open in a new tab

(A) Puc-LacZ and TRE-RFP (white) are used for JNK activation in wildtype salivary gland imaginal ring. Egr antibody staining (white) in wildtype salivary gland imaginal ring. (B) Tumors in salivary gland imaginal ring of indicated genotypes. Graph shows the average sizes of tumors. Numbers indicate sample sizes. Orange dashed lines circle the tumor regions. (C) TRE-RFP expressions in the salivary gland imaginal rings with the indicated genotypes. The graph shows the relative RFP intensities. (D) Egr antibody staining for the indicated genotypes. ***, P<0.001; ns, not significant. Error bars, mean±S.E.M. Nuclei labeled with DAPI [blue in (A and C), white in (B)]. Scale bar, 20μm. See also Figure S3.

To determine whether JNK signaling is required for NICD-induced salivary gland imaginal ring tumorigenesis, we made use of the RNAi lines against egr, basket (bsk), the Drosophila homologue of mammalian JNK, and TGF-β activated kinase 1 (Tak1), one of the MAPK kinase kinase (MAPKKK) that activates bsk. Compromised JNK signaling significantly reduced the size of NICD-induced posterior end tumors (bsk-RNAi: 52% reduction, Tak1-RNAi: 61% reduction, egr-RNAi: 28% reduction) (Fig. 5B), indicating the involvement of JNK signaling in NICD-induced transition-zone tumor formation. To determine whether the effect of JNK signaling is due to regulation of cell proliferation, we knocked down bsk or Tak1 in the imaginal ring, but could not detect any obvious changes in imaginal ring size or mitotic activity (Fig. S3D). Additionally, loss of JNK did not change the size of anterior hyperplasia induced by NICD overexpression (Fig 5B). These findings suggest that JNK signaling is not required for cell proliferation in imaginal ring cells.

Next, we asked whether ectopic activation of JNK in the anterior imaginal ring could facilitate NICD-induced tumorigenesis. Unfortunately, flies with co-overexpression of NICD and Bsk (or Tak1) driven by Actts died one day after Gal4 induction. Using the retnts to co-overexpress Tak1 and NICD, salivary gland imaginal ring cells almost vanished because of pervasive cell death (Fig. S3C), suggesting that high levels of JNK activity have an apoptosis-promoting role in the salivary gland imaginal ring, as reported in other tissues (Chen et al., 1996; Dhanasekaran and Reddy, 2008; Yang and Su, 2011). Indeed, overexpression of Bsk or Tak1 alone also caused loss of imaginal ring cells (Fig. S3C).

MMP1 is downstream of JNK signaling to promote NICD-induced tumor formation

JNK signaling has a complicated relationship with tumorigenesis. On one hand, it can cause cell death and thus reduces the tumor size. On the other hand, JNK can promote tumor progression by activating downstream targets such as matrix metalloproteases (MMPs) (Ghiso et al., 1999; Westermarck and KÄHÄRI, 1999). In fact, MMP1 is a marker for tumor transformation in Drosophila (Srivastava et al., 2007; Uhlirova and Bohmann, 2006), and we did detect high levels of MMP1 in NICD-induced imaginal ring tumors (Fig. 6A). Surprisingly, we also found that MMP1 is normally expressed at the transition cells in salivary gland imaginal rings (Fig. 6B). This expression pattern is a result of JNK activation in the transition zone, as MMP1 expression decreased significantly following disruption of JNK signaling by bsk or Tak1 knockdown (Fig. 6B and S4A). In contrast, disruption of JAK-STAT signaling by hop or Stat92E RNAi was unable to decrease MMP1 expression (Fig. S4B), suggesting that MMP1 expression is induced by JNK activation, but not by JAK-STAT signaling.

Figure 6. JNK-induced MMP1 is sufficient to induce tumor formation in coldspot.

Figure 6.

Open in a new tab

(A) MMP1 staining in Notch-induced tumor. (B) MMP1 staining in salivary gland imaginal rings with indicated genotypes. (C) Tumors in salivary gland imaginal ring of indicated genotypes. Graph shows the average sizes of tumors. Numbers indicate sample sizes. **, P<0.005; ns, not significant. Error bars, mean±S.E.M. (D) Salivary gland imaginal ring cells were outlined by PHA (red). Dlg (green) was used for cell polarity analyses. The line drawings trace the PHA staining and magenta areas showing the multilayer structures in both coldspot and hotspot regions. Bottom panel magnifies the regions indicated by white boxes. Arrows indicate the ectopic Dlg staining. Nuclei labeled with DAPI [blue in (A, B and D), white in (C)]. Orange dashed lines circle the tumor regions in (A and C). Scale bar, 20μm. See also Figure S1, S2, S3, S4 and S5.

To determine whether MMP1 is involved in NICD-induced tumorigenesis in the salivary gland imaginal ring, we knocked down MMP1 and found a reduction of NICD-induced tumor growth (Fig. 6C), suggesting that MMP1 promotes Notch-induced tumorigenesis in the transition zone. In contrast, loss of MMP2 was not sufficient to suppress Notch-induced tumor growth (Fig. 6C). As a control, knockdown of MMP1 alone did not cause smaller imaginal rings or lower proliferation rates (Fig S3D).

Next, we asked whether MMP1 overexpression could lead to transformation of anterior coldspot cells with NICD expression. MMP1 overexpression alone with Actts led to lethality, but with retnts, the larvae survived with no obvious abnormalities in the imaginal ring (Fig S3A). When Pscts, the anterior driver, was used to co-overexpress NICD and MMP1, basally delaminated cells were detected in the anterior region (Fig. 6D). Transplantation of Pscts >NICD+MMP1 imaginal rings resulted in more dysplasia-like basal delaminations. However, these delaminated cells did not propagate further to form neoplasms in the host (Fig. S2B). When retnts, the driver for the entire imaginal ring, was used, basal delamination and loss of epithelial organization were found at both the posterior and anterior ends of the imaginal ring (Fig. 6D). In addition, the retnts >MMP1+NICD combination hastened tumor progression as compared with NICD overexpression alone (Fig. S5A). When retnts>NICD+MMP1 tumors were transplanted in adult hosts, the tumors increased rapidly in size (21-fold change at D21, Fig. S1B). In some retnts>NICD+MMP1 primary tumors (13%, n=198, Fig. S5A), and all transplanted tumors (n=11, Fig. S1B), the anterior tubular structures were no longer identifiable, while neoplastic transformation appeared in the entire imaginal ring. The cell polarity and tissue organization were grossly disrupted in these tumors (Figure S5B). This phenotype is markedly different from the Actts>NICD tumors in which the anterior hyperplasia maintained its tubular structure even after transplantation (Fig. S1A). Also different from the spherical-shaped Actts>NICD tumor transplants (Fig. 2E), retnts>NICD+MMP1 tumor transplants showed irregular shapes with groups of cells forming protrusions trying to break away from the tumor mass (27%, n=11, Fig. S1B).

MMP1 regulates the extracellular matrix organization at the transition zone

The degradation of the extracellular matrix (ECM) is an essential process for tumor progression, by which tumor cells are able to extrude basally from the native epithelial layers and detach from the original tumor mass for further metastasis (Lu et al., 2011; Sekiguchi and Yamada, 2018). During this process, MMPs play important roles in the degradation of ECM components (Page-McCaw et al., 2007). In the salivary gland imaginal ring, NICD-induced tumors always arise through basal delamination at the transition zone, where MMP1 is highly expressed. We hypothesize that the endogenous MMP1 is involved in creating a weak spot in the basement membrane that allows basal delamination of pro-tumor cells. To test this hypothesis, we first examined the expression of viking (vkg)-GFP. vkg encodes a subunit of collagen IV and is a major component in the basal membrane. However, vkg-GFP showed a uniform pattern in the imaginal ring—its expression levels were similar between the anterior and posterior regions, and were not changed when MMP1 was overexpressed or knocked down (Fig. S6A). These results suggest that MMP1 has a minimal impact on the distribution of collagens in the imaginal ring.

We next examined the expression of Perlecan (Pcan), an ECM heparan sulfate proteoglycan (Farach-Carson and Carson, 2007; Sarrazin et al., 2011), which has been shown to be expressed in the salivary gland (Schneider et al., 2006). In wildtype controls, Pcan expression was robust in the anterior coldspot area, highlighting the basement membrane, whereas Pcan expression was significantly lower in the transition zone (Fig. 7A and 7B). In imaginal rings with sustained NICD-overexpression, the level of Pcan expression in the posterior neoplasm was also lower than the anterior hyperplastic region (Fig. S6E). To determine whether the reduced Pcan level is caused by increased MMP1 in the transition zone, we knocked down MMP1 and found that Pcan expression in the transition zone was upregulated (3.3-fold higher than control, Fig. 7A and 7C). Consistently, overexpression of MMP1 decreased the level of Pcan in the basement membrane covering the anterior end of the imaginal ring (35% lower than control, Fig. 7A and 7D). To test whether JNK signaling is involved in the regulation of Pcan, we also knocked down JNK components and found that loss of JNK resulted in increased Pcan expression in the transition zone (Fig. 7A, E and F). Taken together, these results indicate that MMP1, a JNK target, is the key enzyme to regulate ECM organization in the salivary gland imaginal ring transition zone.

Figure 7. MMP1 promotes degradation of Pcan in the salivary gland imaginal ring.

Figure 7.

Open in a new tab

(A) Upper panel: the illustrated wildtype Pcan expression (on the basal side of anterior coldspot cells). Lower panel: the Pcan antibody staining (white) in wildtype control, loss of MMP1 and overexpression of MMP1. Nuclei labeled with DAPI (Blue). Blue lines indicate the anterior coldspot spot areas and red lines indicate the posterior tumor hotspots. The red-green spectrum indicates the intensity of Pcan staining. Red to green: high intensity to low. (B-F) Quantification of the relative intensities of Pcan staining. (B): the relative intensities for Pcan staining in anterior coldspot and posterior hotspot of controls. The control coldspot was used for normalization. (C): the relative intensities for Pcan staining in posterior hotspots of control and loss of Mmp1. The control hotspot was used for normalization. (D): the relative intensities for Pcan staining in in anterior coldspot of control and overexpression of Mmp1. The control coldspot was used for normalization. (E): the relative intensities for Pcan staining in posterior hotspots of control and loss of bsk. The control hotspot was used for normalization. (F): the relative intensities for Pcan staining in posterior hotspots of control and loss of Tak1. The control hotspot was used for normalization. **, P<0.005; ***, P<0.001. Error bars, mean±S.E.M. n=24 for each. See also Figure S6.

Discussion:

We show in this study that the posterior boundary of the Drosophila salivary gland imaginal ring is a tissue microenvironment favorable for oncogenic Notch-induced neoplastic transformation. Upon receiving sustained Notch signaling, the anterior and posterior regions of the salivary gland imaginal ring are evidently different. In the anterior region, we can only detect hyperplastic overproliferation. In contrast, dysplasia and malignant neoplasms are always generated at the posterior boundary region, a transition zone that is rich in endogenous JAK-STAT and JNK signaling. Our findings suggest that pro-tumor cells exploit the local signaling to overcome the tumor-suppressive role that epithelial tissues normally possess. Interestingly, Notch activity in the posterior transition zone of wildtype salivary gland is relatively low (Fig. S7). Given that hyperactivation of Notch in this region induces tumorigenesis consistently, a normal low level of Notch activity in transition cells is probably needed to protect them from undergoing neoplastic transformation.

The inflammatory signaling in transition zones and tumor hotspots

Previously, we have proposed the tumor hotspot theory that highlights the tissue microenvironmental effect on the initial stages of tumorigenesis (Tamori and Deng, 2017; Tamori et al., 2016). The tumor hotspot in the wing disc hinge area possesses high levels of JAK-STAT activity that is necessary for tumorigenesis caused by the loss of neoplastic tumor suppressor genes (nTSG), such as scrib and lgl. The strong JAK-STAT activity at the salivary gland imaginal ring transitional region is similarly needed for oncogenic Notch-induced tumorigenesis. Although the initiating mutations are different between these two tissues, both microenvironments possess active JAK-STAT signaling that promotes neoplastic tumor transformation, suggesting that JAK-STAT signaling may have a general role in the initial stages of tumorigenesis in Drosophila epithelial tissues. Indeed, previous reports have indicated that the terminal regions of the follicular epithelium in the developing egg chambers are also prone to the multilayering and tumorigenesis phenotypes (Lu and Bilder, 2005), and not surprisingly, this region is high in JAK-STAT signaling (Poulton and Deng, 2007; Xi et al., 2003). The ligands for JAK-STAT signaling in Drosophila are Unpaired 1–3, homologs of the mammalian IL-6 (Harrison et al., 1998; Jiang et al., 2009), which is also involved in tumor-associated inflammation. During colitis-associated oncogenesis, inflammation-induced IL-6 and STAT3 are required for cell survival and proliferation (Bollrath et al., 2009; Grivennikov et al., 2009).

The precise role for JAK-STAT signaling in the transition-cell development and transformation is still unclear. Compromised JAK-STAT signaling has a minimal impact on the growth of the salivary gland imaginal ring. During transition-zone tumorigenesis, JAK-STAT signaling is probably involved in promoting dysplasia to neoplasia transition. As we have shown in this study, the anterior coldspot cells undergo only basal delamination to form dysplasia in Pscts> NICD +MMP1 imaginal rings. These anterior cells may need elevated JAK-STAT activity to complete a full transformation into a neoplasm. It is noticeable that when NICD and MMP1 are co-overexpressed in the entire imaginal ring (using the retnts driver), both the posterior and the anterior cells seem to be fully transformed. In these tumors, tissue organization and cell polarity are completely disrupted, and no anterior tubular structures are identifiable (Fig. S5B). Furthermore, the transplanted retnts> NICD +MMP1 tumors grow rapidly and are highly malignant (Fig. S1B). It is debatable how the anterior cells become fully transformed in these tumors. Possibly, the retnts> NICD +MMP1 imaginal ring has a much weakened basement membrane to allow posterior tumor cells migrate anteriorly and provide a ligand to activate JAK-STAT signaling in anterior cells, thus complete the neoplastic transformation. In many Drosophila tumors, the JAK-STAT ligands, Upd 1–3, have been shown upregulated significantly (Classen et al., 2009; Xie et al., 2017). A preliminary study with RNA-Seq analysis of imaginal ring tumors also indicated the upregulation of these ligands (data not shown).

Using the salivary gland imaginal ring model, we also identified the JNK pathway that is endogenously active and involved in tumor formation at the transition zone. JNK is known to be related to tumor-promoting inflammation in mammals, and has a prominent role in tumor initiation and progression (Wagner and Nebreda, 2009). Reduction of JNK is sufficient to inhibit inflammation in mouse liver tissue and further suppress hepatitis and hepatocellular carcinoma (Han et al., 2016). In Drosophila, loss of apicobasal polarity in imaginal discs activates JNK signaling to drive tumor growth, and this activated JNK further cooperates with oncogenic RasV12 to induce invasion and metastasis (Igaki et al., 2006). In oncogenic Src induced-tumorigenesis, JNK promotes non-autonomous tumor growth via regulating Yokie activation (Enomoto and Igaki, 2013). Interestingly, JNK signaling can induce apoptosis that effectively blocks tumorigenesis in certain circumstances (Tournier, 2013). Indeed, excessive JNK activity induces cell death in both the wildtype and NICD-overexpressed imaginal rings, which reduce the tumor size significantly (Fig. S3C).

In this study, we show that Egr is involved in JNK activation in the imaginal ring transition zone. However, JNK signaling might not be fully dependent on Egr in the imaginal ring, as moderate levels of JNK activation is still detectable after removal of Egr expression (Fig. 5C). In addition, the reduction of tumor size by egr-RNAi was not as significant as bsk or Tak1 knockdowns (Fig. 5B), suggesting some JNK activity is probably maintained when Egr is knocked down. Thus, in addition to Egr, JNK can probably be activated by another ligand(s) or in a ligand-dependent manner or is induced.

Our data reveal that the combination of Notch/JNK/JAK-STAT is sufficient to induce tumorigenesis in the transition cells, and the Notch/JNK induced-MMP1 combination is sufficient to induce dysplasia in the anterior imaginal ring. In other Drosophila epithelial tissues, such as the eye/antenna and wing discs, hyperactivated Notch can also cooperate with JNK signaling to induce tumorigenesis. Notch/Src activates Egr-independent JNK signaling, which causes high levels of JAK-STAT and MMP1 expression, hyperproliferation and apoptosis (Ho et al., 2015). Also, in Drosophila eye/antenna and wing discs, activated Notch cooperates with Mef2 to promote JNK signaling in an Egr-dependent manner, which also results in high MMP1 expression and hyperproliferation, and mild JAK-STAT activation phenotypes (Pallavi et al., 2012). Additionally, in scrib mutated discs, overexpression of Notch is sufficient to transform JNK signaling from a pro-apoptotic role into a tumor-promoting factor that enhances overgrowth (Brumby and Richardson, 2003). These studies suggest that Notch and inflammatory signaling are involved in triggering transformation in multiple epithelial cell types.

Inflammatory signaling is associated with tumorigenesis at other transition zones in mammals. For example, inflammatory macrophages are detectable in the mucosa underlying the wildtype anal epithelium close to the anorectal junction (Guasch et al., 2007). Esophageal adenocarcinoma has been confirmed as an inflammation-associated cancer partially because reflux-induced chronic inflammation contributes to the formation of Barret’s esophagus (Wild and Hardie, 2003). Although the link between inflammation and tumorigenesis is well established, whether there are endogenous inflammatory signaling activities at mammalian transition zones to promote tumorigenesis remains unclear. Considering that our study provides direct evidence showing local endogenous JNK and JAK-STAT activation facilitates neoplastic tumor formation at the transition zone, it would be of interest to further investigate the expression and function of local endogenous inflammatory signaling at other transition zones.

The regulation of extracellular matrix components shapes the transition zone as a tumor hotspot

In this study, we show that MMP1 is a downstream target of JNK signaling at the salivary gland imaginal ring. Additionally, MMP1 can induce dysplasia formation in the anterior region that is normally refractory to Notch-induced tumorigenesis. MMP1 expression has been used as a molecular marker for tumors in Drosophila (Uhlirova and Bohmann, 2006). MMP1 is one of two MMPs in Drosophila for ECM remodeling; it acts by cleaving the main components of ECM. We find that in wildtype imaginal ring the expression of an ECM component Pcan is different between the anterior and posterior regions, and this differential expression is regulated by MMP1. Pcan connects the components of ECM, including laminins and collagens (Farach-Carson and Carson, 2007), and interacts with cellular dystroglycan and integrins to link cells and the ECM (Sekiguchi and Yamada, 2018). In the salivary gland imaginal ring, MMP1 is probably involved in ECM remodeling and the cell-ECM interaction through controlling the degradation of Pcan. During Notch-induced tumorigenesis, the robust ECM in the anterior prevents the delamination of overproliferative cells, whereas these pro-tumor cells could extrude from the frail ECM structure in the posterior end. Elevated expression of MMPs has been shown to be related to metastasis (Deryugina and Quigley, 2006; Shay et al., 2015), and our findings suggest they also play crucial roles in creating the microenvironment that allows carcinogenesis to take the initial steps.

In the wing disc tumor hotspot, the ECM alignment is more robust compared to the wing pouch region. This organization is thought to form a strong basement membrane, which prevents nTSG-loss of function cells delaminating from the basal side. Instead, the cells delaminate from the apical side and become exposed to high levels of JAK-STAT signaling, and thus undergo tumor growth in the lumen. Basally extruded cells, in contrast, normally undergo apoptosis and are removed by macrophages (Tamori et al., 2016). In the salivary gland imaginal ring transition zone the role of the ECM plays seems to be opposite: a weaker ECM allows the pro-tumor cells to pile up and undergo tumor transformation at the basal side. It is possible macrophages are lacking or less accessible in this region.

Dysplasia is a key transition point from benign to pre-malignant condition. During human cancer progression, dysplasia often develops from metaplasia, which is a conversion of one epithelial cell type to another. For example, Barret’s esophagus is a metaplastic transformation, in which intestinal Goblet cells replace the normal squamous mucosa of the esophagus. However, in advanced Barret’s esophagus, dysplastic cells can be detected, which also indicates a precancerous condition (Haggitt, 1994). Delamination and loss of apical-basal polarity are two important features to confirm a dysplasia (Lisovsky et al., 2009, 2010; Marinowic et al., 2017; Tamori et al., 2016). Our studies show that gain of MMP1 can cause NICD induced hyperplastic cells to become dysplastic in the anterior imaginal ring. In the transition zone, because the MMP1 level is normally high, dysplasia is detected shortly after NICD induction. These findings suggest that the salivary gland imaginal ring transition zone can be used as a model to decipher the regulatory steps and mechanisms involved in progression from benign to malignant tumor.

The triggers for initial tumorigenesis at transition zones and tumor hotspots

Tumor hotspots are specific terroir for tumorigenesis (Tamori and Deng, 2017), providing the right “soil” for different pro-tumor-cell seeds to grow into a tumor. Depending on the initiating mutations, the microenvironmental soil suitable for the various possible pro-tumor-cell seeds will be different. In tumor hotspots, a single dysregulated oncogene or the loss of a single tumor suppressor can produce neoplastic tumors due to the existence of favorable endogenous cancer-promoting factors. In tumor coldspots, however, additional mutations are needed to modify the tissue microenvironment so that pro-tumor cells can further develop into a tumor. Cancer development has long been viewed as a complex multi-step process that requires accumulation of oncogenic and tumor suppressor mutations. Our findings reveal an alternative route, where in tumor hotspots such as transition zones, overexpression of a single proliferation-promoting signal is sufficient to induce malignant tumor transformation, suggesting the existence of vulnerable sites in animal bodies that are highly susceptible to tumorigenic mutations.

STAR Methods

Contact for Reagent and Resource Sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Wu-Min Deng (wumin@bio.fsu.edu).

Experimental Model and Subject Details

Drosophila strains and culture

Drosophila stocks were maintained and crossed at 21–22 °C, unless otherwise indicated. Unless other controls are mentioned, w1118 was used as wild-type control.

For the NICD-induced tumor analyses during larval stages, Gal80ts was applied to control the timing of NICD expression. Flies were reared at 18°C to inhibit Gal4 function until late-second instar and then were transferred to 29°C to induce NICD expression. When using the Act-Gal4 driver for tumorigenesis studies, a few larvae with NICD overexpression survived until D9 after transfer to 29°C, but most of them died about D7. Therefore, we studied the tumors in the D7 larvae. For the tumor studies driven by other Gal4 drivers, tumorigenesis in D3 larvae (ready for pupation) were used.

The following stocks were used for the activation of signaling pathways: STAT-92EGFP (Bloomington #26198) for JAK-STAT activation; Puc-LacZ (gifts from D. Bohmann) and TRE-RFP (Bloomington #59012) for JNK activation. vkg-GFP (vkgG454) was used to mark collagen IV expression (Morin et al., 2001). NRE-eGFP was used for Notch activation (Housden et al., 2014). upd1-YFP (gift from Y. Yamashita), a fusion protein of Upd1 and YFP (yellow fluorescent protein), was used for Upd1expression. The following Gal4 lines were used for ubiquitous or site-specific studies: Act-Gal4 (Bloomington #4414), Retn-Gal4 (Bloomington #47433), shg-Gal4 (Bloomington #63856), Psc-Gal4 (Bloomington #49344), stg-Gal4 (Bloomington #62700), and RhoGEF3-Gal4 (Bloomington #65647). The following UAS stocks were used: UAS-NICD (Domanitskaya and Schüpbach, 2012), UAS-NIR (Bloomington #33611), UAS-STATΔNΔC, UAS-STAT92EIR (Inverted repeat, Bloomington #33637), UAS-hop (Harrison et al., 1995), UAS-hopIR (Bloomington #32966), UAS-egrIR (Bloomington #58993), UAS-bsk (Bloomington #9310), UAS-bskIR (Bloomington #53310), UAS-Tak1 (Bloomington #58810), UAS-Tak1IR (Bloomington #33404), UAS-Mmp1IR (Bloomington #31489), UAS-Mmp2IR (Bloomington #65935), and UAS-Mmp1.f2 (Bloomington #58703).

Method Details

Transplantation assays

The transplantation procedure was modified from previous studies (Beaucher et al., 2007; Caussinus and Gonzalez, 2005). The tumors were obtained from the larvae following NICD overexpression or NICD and MMP1 co-expression for 7 days. The posterior salivary gland cells and anterior duct cells were removed carefully and only the tumor cells were kept. Entire imaginal ring tumors were transferred to w1118 adult hosts. After injection, adult hosts were maintained at 25°C for recovery for 2 days and then transferred to 29°C for 21 days. Then, tumors were dissected from hosts for size measurement on the 21st day after transplantation. For re-transplantation, the tumors from 21 days after transplantation were dissected and reinjected to w1118 adult hosts.

Immunostaining

Larvae were dissected, fixed, and stained with antibodies as described previously (Tamori et al., 2016). The following primary antibodies were used: rabbit anti-PH3 (1:200, Millipore), rabbit anti-aPKC (1:1000, Santa Cruz Biotechnology), rabbit anti-DG (1:1000), mouse anti-Dlg [1:40, 4F3, Developmental Studies Hybridoma Bank (DSHB)], mouse anti-MMP1 (1:1:1 mixture of 3A6B4, 3B8D12, and 5H7B11 were diluted 1:15, DSHB), mouse anti-DCP1 (Cell signaling, 1:200), rabbit anti-Pcan (1:500) and mouse anti-β-gal (Promega, 1:500). Alexa Fluor 488- or 546-conjugated goat anti-mouse and anti-rabbit secondary antibodies (Invitrogen, 1:400) were used. Nuclei were labeled with DAPI (Invitrogen, 1:1000). F-actin were labeled with Phalloidin (1:50, Invitrogen).

Fluorescence microscopy and image analyses

Images were captured with Zeiss LSM 800 confocal microscopes. ImageJ software or Zen microscope software (Zeiss) were used for image analyses and processing.

For tumor size measurement, the Z-Stack function in Zen microscope software (Zeiss) was used to capture tumors from top to bottom. Measure Stack plugin in ImageJ was applied for the measurement of 3D tumor volume.

For the intensity analyses of TRE-RFP, STAT-92EGFP, vkg-GFP and the antibody staining for Pcan and MMP1, the single sections with brightest intensity were taken by confocal microscope. Mean intensities in posterior end cells were measured by ImageJ. The average intensities from the control group were used to normalize the average intensities from experimental groups. For the Red-Green spectrum for Pcan intensities, the LSM function in Zen (Zeiss) was used.

Information of detailed genotypes

All flies with Gal80ts were reared at 18°C for the first 7 days, and then transferred to 29°C for additional days as indicated. The detailed genotypes of each experiment are described below:

  • Figure 2.

  • (A) w1118 (third instar larval stage, L3)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)
  • (B) w1118 (L3)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/+ (2D, 7D at 29°C)
  • (C) w1118 (L3)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)
  • (D) Act-Gal4, UAS-GFP/+; tub-Gal80ts/+ (wild type imaginal rings dissected from L3)
    • Act-Gal4, UAS-GFP/UAS-NICD; tub-Gal80ts/+ (imaginal ring tumors dissected from 7D at 29°C)
    • w1118 (Adult hosts)

  • (E) Act-Gal4, UAS-GFP/UAS-NICD; tub-Gal80ts/+ [before transplantation (3D at 29°C), 21D after transplantation]

  • (F) Act-Gal4, UAS-GFP/UAS-NICD; tub-Gal80ts/+ (21D after transplantation)
    • w1118 (Adult hosts)

  • Figure 3.

  • (A) Act-Gal4/UAS-GFP (L3)
    • UAS-GFP/+; retn-Gal4/+ (L3)
    • shg-Gal4/UAS-GFP (L3)
    • UAS-GFP/+; Psc-Gal4/+ (L3)
    • UAS-GFP/+; stg-Gal4/+ (L3)
  • (B) w1118 (3D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/+ (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; retn-Gal4/+(3D at 29°C)
    • shg-Gal4/UAS-NICD; tub-Gal80ts/+ (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; Psc-Gal4/+(3D at 29°C)
    • tub-Gal80ts/UAS-NICD; stg-Gal4/+(3D at 29°C)

  • Figure 4.

  • (A) upd1-Gal4; STAT92E-GFP/+; UAS-RFP/+ (L3)

  • (B) upd1-YFP (L3)
    • Act-Gal4/tub-Gal80ts; upd1-YFP/UAS-upd1IR (L3)

  • (C) Act-Gal4/UAS-NICD; tub-Gal80ts/STAT92E-GFP (7D at 29°C)

  • (D) Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-Stat92EIR (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-hopIR (7D at 29°C)

  • Figure 5.

  • (A) Puc-LacZ (L3)
    • TRE-GFP (L3)
  • (B) Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-bskIR (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-Tak1IR (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-egrIR (7D at 29°C)
  • (C) tub- Gal80ts/TRE-RFP; retn-Gal4/+ (2D at 29°C)
    • tub-Gal80ts /TRE-RFP; retn-Gal4/UAS-egrIR (2D at 29°C)
    • tub-Gal80ts/TRE-RFP; RhoGEF3-Gal4/UAS-egrIR (2D at 29°C)
  • (D) w1118 (2D at 29°C)
    • Act-Gal4/+; tub-Gal80ts/UAS-egrIR
    • tub-Gal80ts /+; retn-Gal4/UAS-egrIR (2D at 29°C)
    • tub-Gal80ts/+; RhoGEF3-Gal4/UAS-egrIR (2D at 29°C)

  • Figure 6.

  • (A) Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)

  • (B) w1118 (2D at 29°C)
    • Act-Gal4/+; tub-Gal80ts/UAS-Mmp1IR (2D at 29°C)
    • Act-Gal4/+; tub-Gal80ts/UAS-Tak1IR (2D at 29°C)
    • Act-Gal4/+; tub-Gal80ts/UAS-bskIR (2D at 29°C)
  • (C) Act-Gal4/UAS-NICD; tub-Gal80ts/+ (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-Mmp1IR (7D at 29°C)
    • Act-Gal4/UAS-NICD; tub-Gal80ts/UAS-Mmp2IR (7D at 29°C)
  • (D) w1118 (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; retn-Gal4/UAS-Mmp1.f2 (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; Psc-Gal4/UAS-Mmp1.f2 (3D at 29°C)
  • Figure 7.
    • w1118 (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; s retn-Gal4/ UAS-Mmp1IR (3D at 29°C)
    • tub-Gal80ts /UAS-NICD; retn-Gal4 /UAS-Mmp1.f2 (3D at 29°C)
    • tub-Gal80ts/UAS-NICD; retn-Gal4 / UAS-bskIR (3D at 29°C)
    • tub-Gal80ts /UAS-NICD; retn-Gal4/ UAS-Tak1IR (3D at 29°C)

Quantification and Statistical Analysis

All quantification data were analyzed by two-tailed Student’s t-test, expect for Figure S5A, which was analyzed by Chi-Square test. P-values are provided by GraphPad Prism 8. The sample sizes and p-values are descripted in main text, figures or figure legends.

Supplementary Material

1

4 highlights.

  • The Drosophila posterior salivary gland imaginal ring is a transition-zone model

  • Persistent Notch signaling induces neoplastic transformation in the transition zone

  • Endogenous inflammatory JAK-STAT and JNK signaling promotes tumorigenesis

  • JNK-induced MMP1 shapes a weaker ECM structure for advanced tumorigenesis

Acknowledgements:

We thank Bloomington Drosophila Stock Center, Vienna Drosophila Research Center and Developmental Studies Hybridoma Bank for fly stocks and antibodies. We also thank D. Corcoran, A. Jevitt, J. Kennedy, J.S. Poulton, S. Row and X.-F. Wang for critical reading of the manuscript.

Funding:

W.-M.D. is supported by Florida Department of Health 8BC12, National Science Foundation IOS-1052333, and National Institutes of Health R01GM072562 and R01CA224381.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Supplementary Materials:

Figure S1–S7

Information of Detailed Genotypes

Declaration of Interests:

The authors declare no competing interests.

References:

  1. Baonza A, and Freeman M (2005). Control of Cell Proliferation in the Drosophila Eye by Notch Signaling. Dev. Cell 8, 529–539. [DOI] [PubMed] [Google Scholar]
  2. Baonza A, and Garcia-Bellido A (2000). Notch signaling directly controls cell proliferation in the Drosophila wing disc. Proc. Natl. Acad. Sci 97, 2609–2614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Beaucher M, Hersperger E, Page-McCaw A, and Shearn A (2007). Metastatic ability of Drosophila tumors depends on MMP activity. Dev. Biol 303, 625–634. [DOI] [PubMed] [Google Scholar]
  4. Bollrath J, Phesse TJ, von Burstin VA, Putoczki T, Bennecke M, Bateman T, Nebelsiek T, Lundgren-May T, Canli Ö, Schwitalla S, et al. (2009). gp130-Mediated Stat3 Activation in Enterocytes Regulates Cell Survival and Cell-Cycle Progression during Colitis-Associated Tumorigenesis. Cancer Cell 15, 91–102. [DOI] [PubMed] [Google Scholar]
  5. Brumby AM, and Richardson HE (2003). scribble mutants cooperate with oncogenic Ras or Notch to cause neoplastic overgrowth in Drosophila. EMBO J. 22, 5769–5779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Caussinus E, and Gonzalez C (2005). Induction of tumor growth by altered stem-cell asymmetric division in Drosophila melanogaster. Nat. Genet 37, 1125. [DOI] [PubMed] [Google Scholar]
  7. Chatterjee N, and Bohmann D (2012). A Versatile ΦC31 Based Reporter System for Measuring AP-1 and Nrf2 Signaling in Drosophila and in Tissue Culture. PLoS One 7, e34063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen YR, Wang X, Templeton D, Davis RJ, and Tan TH (1996). The role of c-Jun N-terminal kinase (JNK) in apoptosis induced by ultraviolet C and gamma radiation. Duration of JNK activation may determine cell death and proliferation. J. Biol. Chem 271, 31929–31936. [DOI] [PubMed] [Google Scholar]
  9. Classen A-K, Bunker BD, Harvey KF, Vaccari T, and Bilder D (2009). A tumor suppressor activity of Drosophila Polycomb genes mediated by JAK-STAT signaling. Nat. Genet 41, 1150–1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Coniglio SJ (2018). Role of Tumor-Derived Chemokines in Osteolytic Bone Metastasis. Front. Endocrinol. (Lausanne). 9, 313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Deng WM, Schneider M, Frock R, Castillejo-Lopez C, Gaman EA, Baumgartner S, and Ruohola-Baker H (2003). Dystroglycan is required for polarizing the epithelial cells and the oocyte in Drosophila. Development 130, 173–184. [DOI] [PubMed] [Google Scholar]
  12. Deryugina EI, and Quigley JP (2006). Matrix metalloproteinases and tumor metastasis. Cancer Metastasis Rev. 25, 9–34. [DOI] [PubMed] [Google Scholar]
  13. Dhanasekaran DN, and Reddy EP (2008). JNK Signaling in Apoptosis. Oncogene 27, 6245–6251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Domanitskaya E, and Schüpbach T (2012). CoREST acts as a positive regulator of Notch signaling in the follicle cells of Drosophila melanogaster. J Cell Sci 125, 399–410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Enomoto M, and Igaki T (2013). Src controls tumorigenesis via JNK□dependent regulation of the Hippo pathway in Drosophila. EMBO Rep. 14, 65 LP–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Farach-Carson MC, and Carson DD (2007). Perlecan—a multifunctional extracellular proteoglycan scaffold. Glycobiology 17, 897–905. [DOI] [PubMed] [Google Scholar]
  17. Fidler IJ, and Poste G (2008). The seed and soil hypothesis revisited. Lancet Oncol. 9, 808. [DOI] [PubMed] [Google Scholar]
  18. Ghiso JAA, Kovalski K, and Ossowski L (1999). Tumor dormancy induced by downregulation of urokinase receptor in human carcinoma involves integrin and MAPK signaling. J. Cell Biol 147, 89–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Grabher C, von Boehmer H, and Look AT (2006). Notch 1 activation in the molecular pathogenesis of T-cell acute lymphoblastic leukaemia. Nat. Rev. Cancer 6, 347–359. [DOI] [PubMed] [Google Scholar]
  20. Grivennikov S, Karin E, Terzic J, Mucida D, Yu G-Y, Vallabhapurapu S, Scheller J, Rose-John S, Cheroutre H, Eckmann L, et al. (2009). IL-6 and Stat3 Are Required for Survival of Intestinal Epithelial Cells and Development of Colitis-Associated Cancer. Cancer Cell 15, 103–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Guasch G, Schober M, Pasolli HA, Conn EB, Polak L, and Fuchs E (2007). Loss of TGFβ; Signaling Destabilizes Homeostasis and Promotes Squamous Cell Carcinomas in Stratified Epithelia. Cancer Cell 12, 313–327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gurley LR, D’Anna JA, Barham SS, Deaven LL, and Tobey RA (1978). Histone phosphorylation and chromatin structure during mitosis in Chinese hamster cells. FEBS J. 84, 1–15. [DOI] [PubMed] [Google Scholar]
  23. Haggitt RC (1994). Barrett’s esophagus, dysplasia, and adenocarcinoma. Hum. Pathol 25, 982–993. [DOI] [PubMed] [Google Scholar]
  24. Han MS, Barrett T, Brehm MA, and Davis RJ (2016). Inflammation Mediated by JNK in Myeloid Cells Promotes the Development of Hepatitis and Hepatocellular Carcinoma. Cell Rep. 15, 19–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Harrison DA, Binari R, Nahreini TS, Gilman M, and Perrimon N (1995). Activation of a Drosophila Janus kinase (JAK) causes hematopoietic neoplasia and developmental defects. EMBO J. 14, 2857–2865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Harrison DA, McCoon PE, Binari R, Gilman M, and Perrimon N (1998). Drosophila unpaired encodes a secreted protein that activates the JAK signaling pathway. Genes Dev. 12, 3252–3263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ho DM, Pallavi SK, and Artavanis-Tsakonas S (2015). The notch-mediated hyperplasia circuitry in drosophila reveals a Src-JNK signaling axis. Elife 4, 1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Housden BE, Li J, and Bray SJ (2014). Visualizing Notch signaling in vivo in Drosophila tissues. Methods Mol Biol 1187, 101–113. [DOI] [PubMed] [Google Scholar]
  29. Igaki T, Kanda H, Yamamoto-Goto Y, Kanuka H, Kuranaga E, Aigaki T, and Miura M (2002). Eiger, a TNF superfamily ligand that triggers the Drosophila JNK pathway. EMBO J. 21, 3009–3018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Igaki T, Pagliarini RA, and Xu T (2006). Loss of cell polarity drives tumor growth and invasion through JNK activation in Drosophila. Curr. Biol 16, 1139–1146. [DOI] [PubMed] [Google Scholar]
  31. Jiang H, Patel PH, Kohlmaier A, Grenley MO, McEwen DG, and Edgar BA (2009). Cytokine/Jak/Stat signaling mediates regeneration and homeostasis in the Drosophila midgut. Cell 137, 1343–1355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Jiang M, Li H, Zhang Y, Yang Y, Lu R, Liu K, Lin S, Lan X, Wang H, Wu H, et al. (2017). Transitional basal cells at the squamous–columnar junction generate Barrett’s oesophagus. Nature 550, 529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Langley RR, and Fidler IJ (2011). The seed and soil hypothesis revisited - the role of tumor-stroma interactions in metastasis to different organs. Int. J. Cancer 128, 2527–2535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lisovsky M, Dresser K, Baker S, Fisher A, Woda B, Banner B, and Lauwers GY (2009). Cell polarity protein Lgl2 is lost or aberrantly localized in gastric dysplasia and adenocarcinoma: an immunohistochemical study. Mod. Pathol. an Off. J. United States Can. Acad. Pathol. Inc 22, 977–984. [DOI] [PubMed] [Google Scholar]
  35. Lisovsky M, Ogawa F, Dresser K, Woda B, and Lauwers GY (2010). Loss of cell polarity protein Lgl2 in foveolar-type gastric dysplasia: correlation with expression of the apical marker aPKC-zeta. Virchows Arch. 457, 635–642. [DOI] [PubMed] [Google Scholar]
  36. Lowell S, Jones P, Le Roux I, Dunne J, and Watt FM (2000). Stimulation of human epidermal differentiation by delta-notch signalling at the boundaries of stem-cell clusters. Curr. Biol 10, 491–500. [DOI] [PubMed] [Google Scholar]
  37. Lu H, and Bilder D (2005). Endocytic control of epithelial polarity and proliferation in Drosophila. Nat. Cell Biol 7, 1232. [DOI] [PubMed] [Google Scholar]
  38. Lu P, Takai K, Weaver VM, and Werb Z (2011). Extracellular Matrix Degradation and Remodeling in Development and Disease. Cold Spring Harb. Perspect. Biol 3, 10.1101/cshperspect.a005058a005058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Marinowic DR, Majolo F, Sebben AD, da Silva VD, Lopes TG, Paglioli E, Palmini A, Machado DC, and da Costa JC (2017). Induced pluripotent stem cells from patients with focal cortical dysplasia and refractory epilepsy. Mol. Med. Rep 15, 2049–2056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. McGuire SE, Mao Z, and Davis RL (2004). Spatiotemporal gene expression targeting with the TARGET and gene-switch systems in Drosophila. Sci STKE 2004, pl6. [DOI] [PubMed] [Google Scholar]
  41. McNairn AJ, and Guasch G (2011). Epithelial transition zones: merging microenvironments, niches, and cellular transformation. Eur. J. Dermatology 21, 21–28. [DOI] [PubMed] [Google Scholar]
  42. Morais-de-Sá E, Mirouse V, and St Johnston D (2010). aPKC phosphorylation of Bazooka defines the apical/lateral border in Drosophila epithelial cells. Cell 141, 509–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Morin X, Daneman R, Zavortink M, and Chia W (2001). A protein trap strategy to detect GFP-tagged proteins expressed from their endogenous loci in Drosophila. Proc. Natl. Acad. Sci 98, 15050 LP–15055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nicolas M, Wolfer A, Raj K, Kummer JA, Mill P, van Noort M, Hui C, Clevers H, Dotto GP, and Radtke F (2003). Notch1 functions as a tumor suppressor in mouse skin. Nat. Genet 33, 416–421. [DOI] [PubMed] [Google Scholar]
  45. Page-McCaw A, Ewald AJ, and Werb Z (2007). Matrix metalloproteinases and the regulation of tissue remodelling. Nat. Rev. Mol. Cell Biol 8, 221–233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Pallavi SK, Ho DM, Hicks C, Miele L, and Artavanis □ Tsakonas S (2012). Notch and Mef2 synergize to promote proliferation and metastasis through JNK signal activation in &lt;em&gt;Drosophila&lt;/em&gt; EMBO J. 31, 2895 LP–2907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Poulton JS, and Deng W-M (2007). Cell-cell communication and axis specification in the Drosophila oocyte. Dev. Biol 311, 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Rahim F, Hajizamani S, Mortaz E, Ahmadzadeh A, Shahjahani M, Shahrabi S, and Saki N (2014). Molecular Regulation of Bone Marrow Metastasis in Prostate and Breast Cancer. Bone Marrow Res. 2014, 405920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rangarajan A, Talora C, Okuyama R, Nicolas M, Mammucari C, Oh H, Aster JC, Krishna S, Metzger D, Chambon P, et al. (2001). Notch signaling is a direct determinant of keratinocyte growth arrest and entry into differentiation. EMBO J. 20, 3427–3436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Reedijk M (2012). Notch signaling and breast cancer. Adv. Exp. Med. Biol 727, 241–257. [DOI] [PubMed] [Google Scholar]
  51. Ring JM, and Martinez Arias A (1993). puckered, a gene involved in position-specific cell differentiation in the dorsal epidermis of the Drosophila larva. Development 119, 251 LP–259. [PubMed] [Google Scholar]
  52. Sarrazin S, Lamanna WC, and Esko JD (2011). Heparan Sulfate Proteoglycans. Cold Spring Harb. Perspect. Biol 3, a004952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Schneider M, Khalil AA, Poulton J, Castillejo-Lopez C, Egger-Adam D, Wodarz A, Deng W-M, and Baumgartner S (2006). Perlecan and Dystroglycan act at the basal side of the Drosophila follicular epithelium to maintain epithelial organization. Development 133, 3805–3815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sekiguchi R, and Yamada KM (2018). Chapter Four - Basement Membranes in Development and Disease In Extracellular Matrix and Egg Coats, Litscher ES, and P.M.B.T.-C.T. in Wassarman DB, eds. (Academic Press; ), pp. 143–191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Shay G, Lynch CC, and Fingleton B (2015). Moving targets: Emerging roles for MMPs in Cancer Progression and Metastasis. Matrix Biol. 0, 200–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Srivastava A, Pastor-Pareja JC, Igaki T, Pagliarini R, and Xu T (2007). Basement membrane remodeling is essential for Drosophila disc eversion and tumor invasion. Proc. Natl. Acad. Sci 104, 2721–2726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Stockhausen M-T, Kristoffersen K, and Poulsen HS (2010). The functional role of Notch signaling in human gliomas. Neuro. Oncol 12, 199–211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Tamori Y, and Deng W-M (2017). Tissue-Intrinsic Tumor Hotspots: Terroir for Tumorigenesis. Trends in Cancer 3, 259–268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Tamori Y, Suzuki E, and Deng W-M (2016). Epithelial Tumors Originate in Tumor Hotspots, a Tissue-Intrinsic Microenvironment. PLoS Biol 14, e1002537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Tournier C (2013). The 2 Faces of JNK Signaling in Cancer. Genes Cancer 4, 397–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Uhlirova M, and Bohmann D (2006). JNK- and Fos-regulated Mmp1 expression cooperates with Ras to induce invasive tumors in Drosophila. EMBO J. 25, 5294–5304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wagner EF, and Nebreda ÁR (2009). Signal integration by JNK and p38 MAPK pathways in cancer development. Nat. Rev. Cancer 9, 537. [DOI] [PubMed] [Google Scholar]
  63. Weng AP, Millholland JM, Yashiro-Ohtani Y, Arcangeli ML, Lau A, Wai C, del Bianco C, Rodriguez CG, Sai H, Tobias J, et al. (2006). c-Myc is an important direct target of Notch1 in T-cell acute lymphoblastic leukemia/lymphoma. Genes Dev. 20, 2096–2109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Westermarck J, and KÄHÄRI V-M (1999). Regulation of matrix metalloproteinase expression in tumor invasion. FASEB J. 13, 781–792. [PubMed] [Google Scholar]
  65. Wild CP, and Hardie LJ (2003). Reflux, Barrett’s oesophagus and adenocarcinoma: burning questions. Nat. Rev. Cancer 3, 676. [DOI] [PubMed] [Google Scholar]
  66. Xi R, McGregor JR, and Harrison DA (2003). A Gradient of JAK Pathway Activity Patterns the Anterior-Posterior Axis of the Follicular Epithelium. Dev. Cell 4, 167–177. [DOI] [PubMed] [Google Scholar]
  67. Xie G, Chen H, Jia D, Shu Z, Palmer WH, Huang Y-C, Zeng X, Hou SX, Jiao R, and Deng W-M (2017). The SWI/SNF Complex Protein Snr1 Is a Tumor Suppressor in Drosophila Imaginal Tissues. Cancer Res. 77, 862–873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Yang S-A, and Deng W-M (2018). Serrate/Notch Signaling Regulates the Size of the Progenitor Cell Pool in Drosophila Imaginal Rings. Genetics 209, 829–843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Yang S-A, and Su M-T (2011). Excessive Dpp signaling induces cardial apoptosis through dTAK1 and dJNK during late embryogenesis of Drosophila. J. Biomed. Sci 18, 85. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1524561-supplement-1.pdf (4MB, pdf)

RESOURCES