Loss of Rab5 drives non-autonomous cell proliferation through TNF and Ras signaling in Drosophila

Developmental Biology 395 (2014) 19–28

Contents lists available at ScienceDirect

Developmental Biology journal homepage: www.elsevier.com/locate/developmentalbiology

Loss of Rab5 drives non-autonomous cell proliferation through TNF and Ras signaling in Drosophila Kyoko Takino a,b,1, Shizue Ohsawa a,1, Tatsushi Igaki a,c,n a

Laboratory of Genetics, Graduate School of Biostudies, Kyoto University, Yoshida-Konoecho-cho, Sakyo-ku, Kyoto 606-8501, Japan Division of Genetics, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan c PRESTO, Japan Science and Technology Agency (JST), 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan b

ar t ic l e i nf o

a b s t r a c t

Article history: Received 12 March 2014 Received in revised form 26 August 2014 Accepted 5 September 2014 Available online 16 September 2014

Deregulation of the endocytic machinery has been implicated in human cancers. However, the mechanism by which endocytic defects drive cancer development remains to be clarified. Here, we find through a genetic screen in Drosophila that loss of Rab5, a protein required for early endocytic trafficking, drives non-autonomous cell proliferation in imaginal epithelium. Our genetic data indicate that dysfunction of Rab5 leads to cell-autonomous accumulation of Eiger (a TNF homolog) and EGF receptor (EGFR), which causes activation of downstream JNK and Ras signaling, respectively. JNK signaling and its downstream component Cdc42 cooperate with Ras signaling to induce upregulation of a secreted growth factor Upd (an IL-6 homolog) through inactivation of the Hippo pathway. Such nonautonomous tissue growth triggered by Rab5 defect could contribute to epithelial homeostasis as well as cancer development within heterogeneous tumor microenvironment. & 2014 Published by Elsevier Inc.

Keywords: Tissue growth regulation Cell–cell interaction Tumorigenesis Rab5 Drosophila

Introduction Cancer development is achieved by the acquisition of multiple genetic alterations such as activation of oncogenes and inactivation of tumor suppressor genes (Kinzler and Vogelstein, 1996). Accumulating evidence suggests that cancer progression is also driven by cell–cell interactions within tumor microenvironment (Bissell and Hines, 2011; Jacks and Weinberg, 2002). The genetic mosaic techniques available in Drosophila allow us to analyze the behavior of cell clones with distinct oncogenic mutations (Xu and Rubin, 1993), making the Drosophila an ideal model organism for studying the role of cell–cell interactions in tumor development and progression in vivo (Brumby and Richardson, 2005; PastorPareja and Xu, 2013). Endocytosis is an evolutionarily conserved cellular process in which molecules on the cell surface such as various types of ligand–receptor proteins are internalized and transported to early endosomes, where they are either sorted into the recycling pathway back to the cell surface or trafficked to the lysosome for degradation. Thus, the endocytic machinery can regulate cellular

n Corresponding author at: Laboratory of Genetics, Graduate School of Biostudies, Kyoto University, Yoshida-Konoecho-cho, Sakyo-ku, Kyoto 606-8501, Japan. Fax: þ81 75 753 7686. E-mail address: [email protected] (T. Igaki). 1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.ydbio.2014.09.003 0012-1606/& 2014 Published by Elsevier Inc.

signaling triggered by the cell surface proteins including oncogenic ligand–receptor systems (Chia and Tang, 2009; Tomas et al., 2014). Indeed, it has been shown that deregulation of the endocytic machinery is frequently associated with human cancers (Chia and Tang, 2009; Floyd and De Camilli, 1998; Mosesson et al., 2008). However, the mechanism by which deregulated endocytosis drives tumor development and progression remains elusive. Rab5 is an evolutionarily conserved small GTPase required for the import of endocytic materials from the cell surface and early endocytic trafficking. Studies in Drosophila have shown that Rab5 acts as a tumor-suppressor protein in imaginal epithelium. For example, the imaginal tissue entirely mutant for rab5 gene results in tumorous overgrowth with multi-layered epithelial structure, an inability to terminally differentiate, and acquired invasive characteristics; therefore, rab5 gene is termed as a “neoplastic” tumor suppressor (Lu and Bilder, 2005). Similarly, loss-of-function mutations in rabenosyn (rbsn), an evolutionarily conserved Rab5 effector, also exhibit “neoplastic” phenotype (Morrison et al., 2008). Furthermore, imaginal tissues mutant for other endocytic genes such as vps22, vps25, vps36, vps45, erupted (ept)/tsg101, and avalanche (avl)/syntaxin-7 also result in tumorous overgrowth (Herz et al., 2009; Morrison et al., 2008; Vaccari et al., 2009). Interestingly, these endocytic mutant clones are eliminated from the tissue when surrounded by wild-type cells (Ballesteros-Arias et al., 2013), suggesting that normal epithelial tissues exert an anti-tumor effect against oncogenic endocytic mutant cells.

20

K. Takino et al. / Developmental Biology 395 (2014) 19–28

Intriguingly, some of these endocytic mutant clones stimulate growth of surrounding wild-type tissue through autonomous activation of Notch signaling (Giebel and Wodarz, 2006; Herz et al., 2006; Herz et al., 2009; Moberg et al., 2005; Thompson et al., 2005; Vaccari and Bilder, 2005), while others do not (Herz et al., 2009; Vaccari et al., 2009). Thus, the effects and their underlying mechanisms of deregulated endocytosis in tumorigenesis still remain elusive. We have recently conducted a genetic screen in Drosophila imaginal epithelium to identify mutations that drive nonautonomous tumor progression through cell–cell interactions. In the screen, we found that mutations that cause mitochondrial dysfunction cooperate with oncogenic Ras (RasV12) to induce nonautonomous overgrowth of surrounding tissue. In this case, mitochondrial dysfunction and Ras activation (mito-/-/RasV12) cooperatively induce production of reactive oxygen species (ROS), which leads to activation of JNK signaling. JNK signaling further cooperates with Ras signaling to inactivate the Hippo pathway, leading to upregulation of its target Unpaired (Upd, an IL-6 homolog), thereby causes overgrowth of surrounding tissue (Ohsawa et al., 2012). Here, we find through the screen that lossof-function mutations for rab5 in RasV12-expressing cells (rab5-/-/ RasV12) induce non-autonomous overgrowth of surrounding tissue. Unlike other endocytic mutations, rab5-/-/RasV12 clones did not cause non-autonomous overgrowth through Notch signaling. Instead, the non-autonomous overgrowth triggered by rab5-/-/ RasV12 clones is caused by the Hippo pathway-mediated Upd induction. Interestingly, distinct from mitochondrial dysfunction, loss of Rab5 is sufficient for the Hippo pathway-mediated induction of Upd, as Rab5 defect causes activation of both JNK and Ras signaling. Mechanistically, loss of Rab5 leads to accumulation of Eiger (a tumor necrosis factor (TNF) homolog) and epidermal growth factor receptor (EGFR) at the plasma membrane of mutant cells, which results in activation of downstream JNK and Ras signaling, respectively. Our data suggest that dysfunction of Rab5 could contribute to epithelial homeostasis and cancer development by non-autonomous mechanisms within a heterogeneous cell populations.

Materials and methods Fly strains and generation of clones Fluorescently-labeled mitotic clones (Lee and Luo, 1999; Xu and Rubin, 1993) were produced in larval imaginal discs using the following strains: eyFLP5, Act 4y þ 4 Gal4, UAS-GFP; FRT82B, TubGal80 (82B tester), eyFLP5, Act 4 y þ 4Gal4, UAS-GFP; diap1-lacZ, FRT82B, Tub-Gal80 (82B diap1-lacZ tester), Tub-Gal80, FRT40A, UASRasV12; eyFLP6, Act 4 y þ 4Gal4, UAS-GFP (40 A Ras tester), TubGal80, FRT40A; eyFLP6, Act 4y þ 4Gal4, UAS-GFP (40 A tester), UASp35; Tub-Gal80, FRT40A; eyFLP6, Act 4 y þ 4Gal4, UAS-GFP (40 A p35 tester), egr1 and UAS-Eigerregg1 (kind gifts from Masayuki Miura), UAS-BskDN (a kind gift from Takashi Adachi-Yamada), UAS-RafDN (a kind gift from Willis Li), UAS-Wts (a kind gift from Tian Xu), P{w þ mc ¼ lacW}Rab5LL00467 (Drosophila Genetic Resource Center, Kyoto), Stat92E06346, vps25K08904, ex-lacZ (ex697), diap1-lacZ (thj5C8), UAS-p35, and UAS-RasV12 (Bloomington Stock Center), UAS-upd1-RNAi (National Institute of Genetics Stock Center), and UAS-notch-RNAi (Vienna Drosophila RNAi Center). Histology Larval tissues were stained with standard immunohistochemical procedures using rabbit anti-Rab5 antibody (1:50) (a kind gift from Marcos Gonzalez-Gaitan), rabbit anti-Upd polyclonal

antibody (1:500) (a kind gift from Douglas Harrison), mouse Mmp1 monoclonal antibody (1:100 from 1:1:1 cocktail of 3A6B4, 3B8D12, and 5H7B11; DSHB), rabbit anti-Eiger abtibody R1 (1:500) (a kind gift from Masayuki Miura), rabbit anti-Capicua polyclonal antibody (1:300) (a kind gift from Iswar Hariharan), mouse anti-βgalactosidase antibody (1:500; Sigma), goat anti-EGFR antibody dl-20 (1:500; Santa Cruz Biotechnology), and were mounted with DAPI-containing SlowFade Gold Antifade Reagent (Molecular Probes). Images were taken with a Leica SP5 or Zeiss LSM700 META confocal microscope.

Results Dysfunction of Rab5 causes non-autonomous overgrowth through Upd To study the genetic basis of tumor development through cell– cell communications, we performed a genetic screen in Drosophila eye imaginal disc to identify mutations that cause oncogenic RasV12-expressing cells to induce non-autonomous growth of surrounding tissue. Clones of cells expressing RasV12 overgrow and develop into benign tumors in the eye disc (Karim and Rubin, 1998; Pagliarini and Xu, 2003). Using the MARCM technique (Lee and Luo, 1999), we introduced additional homozygous mutations by P-element insertions in GFP-labeled RasV12-expressing clones and screened for those that triggered non-autonomous growth (nag) (Fig. 1A), which exhibits overgrown ‘folded-eyesphenotype in the adults (Ohsawa et al., 2012). We isolated a nag mutant (nag26) that bears a P-element insertion within the rab5 gene locus (Fig. 1B–D). Immunostaining analysis revealed that nag-26 homozygous clones show no detectable Rab5 protein expression in the eye disc (Fig. 1E). In addition, the non-autonomous overgrowth phenotype was cancelled by overexpression of wild-type Rab5 in nag-26 homozygous clones (data not shown). We further confirmed using a dominant-negative form of Rab5 (Rab5DN) or previously reported rab5 null allele rab52 that dysfunction of Rab5 in RasV12-expressing clones indeed triggers non-autonomous tissue overgrowth (Fig. S1A, data not shown). Interestingly, loss of Rab5 seemed to be sufficient to stimulate non-autonomous overgrowth, as clones of cells mutant for rab5 or expressing Rab5DN, with simultaneous expression of the cell death inhibitor p35 (which allows rab5 deficient clones to survive), also caused nonautonomous overgrowth (Fig. 1F, Fig. S1B). We found that Rab5DN þ p35-expressing clones also induced non-autonomous overgrowth in the wing disc (data not shown). The nonautonomous overgrowth could be mediated by secreted growth factors such as Upd (Giebel and Wodarz, 2006; Herz et al., 2006; Moberg et al., 2005; Ohsawa et al., 2012; Thompson et al., 2005; Vaccari and Bilder, 2005). Indeed, heterozygosity for Stat92E, a component of the Drosophila JAK/STAT pathway (which is activated by Upd), in animals bearing rab5-/-/RasV12 clones notably suppressed non-autonomous overgrowth (Fig. S1C). Furthermore, reducing upd expression in rab5-/-/p35 clones suppressed the nag phenotype (Fig. 1G). Consistent with these results, clones of cells mutant for rab5 or expressing Rab5DN simultaneously expressing p35 upregulated Upd (Fig. 1H, Fig. S2A). Importantly, upregulation of Upd was observed in rab5-/- clones in the absence of p35 expression as well (Fig. S2B), indicating that suppression of apoptosis by p35 is not required for the induction of Upd. These data show a sharp contrast between Rab5 dysfunction and mitochondrial dysfunction, as neither mitochondrial dysfunction on its own nor the combination of mitochondrial dysfunction with p35 expression induces Upd expression and non-autonomous overgrowth (Ohsawa et al., 2012). Thus, loss of Rab5 seems to

K. Takino et al. / Developmental Biology 395 (2014) 19–28

21

Fig. 1. Loss of Rab5 drives non-autonomous overgrowth. (A) A genetic screen for identifying mutations, in conjunction with oncogenic RasV12, that induce non-autonomous growth of surrounding tissue. Using the MARCM system, homozygous mutations (P-element insertions) were introduced in clones of cells expressing RasV12 and GFP in Drosophila eye imaginal disc. Non-autonomous overgrowth was assessed by eye phenotype and the size of GFP-expressing clones in the adult. (B and C) Eye of adult fly bearing GFP-labeled wild-type (B) or rab5 LL00467-/-/RasV12 clones in otherwise wild-type tissue. Light level (left) and fluorescent (right) images are shown. (D) The piggyBac line nag-26LL00467-/- has a P-element insertion in the rab5 gene locus. Exons are indicated by boxes. The coding region is marked in blue. (E) GFP-labeled rab5LL00467-/-/p35 clones were induced in eye-antennal disc and was stained with anti-Rab5 antibody. The nuclei were visualized by DAPI staining. (F) Eye of adult fly bearing GFP-labeled rab5 LL00467-//p35 clones. Light level (F) or fluorescent (F’) image is shown. (G) Eye of adult fly bearing GFP-labeled rab5LL00467-/-/p35þ upd-RNAi clones. Light level (left) and fluorescent (right) images are shown. (H) GFP-labeled rab5LL00467-/-/p35 clones were induced in eye-antennal disc and was stained with anti-Upd antibody. The nuclei were visualized by DAPI staining. Genotypes are as follows: Tub-Gal80, FRT40A/FRT40A; eyFLP6, Act 4y þ 4Gal4, UAS-GFP/ þ (B), Tub-Gal80, FRT40A, UAS-RasV12/PBac{SAstopDsRed} LL00467, FRT40A; eyFLP6, Act 4yþ 4Gal4, UAS-GFP/ þ (C), UAS-p35/ þ or Y; Tub-Gal80, FRT40A/PBac{SAstopDsRed}LL00467, FRT40A; eyFLP6, Act4 yþ 4Gal4, UAS-GFP/ þ (E, F, H), and UAS-p35/ þ or Y; Tub-Gal80, FRT40A, UAS-RasV12/PBac{SAstopDsRed}LL00467, FRT40A; eyFLP6, Act 4yþ 4Gal4, UAS-GFP/UAS-upd1-RNAi (G).

22

K. Takino et al. / Developmental Biology 395 (2014) 19–28

cause non-autonomous overgrowth through Upd induction by a distinct mechanism from mito-/-/RasV12 cells.

degradation. Indeed, Eiger protein level was upregulated in Rab5DN þp35 clones (Fig. 3F).

Loss of Rab5 induces Upd expression via inactivation of the hippo pathway

Eiger-JNK signaling promotes non-autonomous growth through Cdc42

It has been reported that clones of cells mutant for vps25 or ept, the components of the ESCRT (endosomal sorting complex required for transport) complexes, cause non-autonomous overgrowth in Drosophila imaginal epithelium (Giebel and Wodarz, 2006; Herz et al., 2006; Moberg et al., 2005; Thompson et al., 2005; Vaccari and Bilder, 2005). The ESCRT complexes are required for sorting ubiquitinated transmenbrane proteins at endosomes into multi-vesicular bodies (MVBs), which fuse with lysosomes for degradation of the internalized proteins. It has been shown that loss of vps25 or ept results in elevated Notch signaling, which leads to upregulation of its target Upd (Moberg et al., 2005; Thompson et al., 2005; Vaccari and Bilder, 2005). Indeed, reduction in notch expression in vps25-/- clones blocked the non-autonomous overgrowth phenotype (Fig. S3A and B). However, notch-RNAi had no effect on the nag phenotype caused by rab5-/-/RasV12 clones (Fig. S3C). Furthermore, we found that Rab5DN þp35 clones do not activate Notch signaling (Fig. S3D and E). Thus, loss of an early endosomal component Rab5 seems to cause non-autonomous overgrowth via distinct mechanism from loss of ESCRT proteins. Another possible mechanism by which loss of Rab5 causes Upd induction is through inactivation of the Hippo pathway (Karpowicz et al., 2010; Ohsawa et al., 2012; Staley and Irvine, 2010). Hippo signaling is an evolutionarily conserved kinase cascade that phosphorylates and inactivates the transcriptional co-activator Yokie (Yki) (Halder and Johnson, 2011; Pan, 2010; Zhao et al., 2010). We found that clones of Rab5DN þp35 cells upregulate Yki target genes such as expanded (ex) and diap1 (Fig. 2A and B, higher magnification images are shown in Fig. S4). Furthermore, forced activation of the Hippo kinase cascade by overexpressing Yki kinase Warts (Wts) in rab5-/-/p35 clones significantly suppressed Upd expression and non-autonomous overgrowth (Fig. 2C and D). These data suggest that loss of Rab5 causes upd expression through inactivation of the Hippo pathway.

We next sought to identify downstream component of JNK signaling that mediates inactivation of the Hippo pathway in Rab5 deficient cells. In our candidate approach, we found that Cdc42, which has been shown to be essential for nuclear localization of YAP (a mammalian homolog of Yki) (Reginensi et al., 2013), acts downstream of JNK signaling to upregulate Upd. We first observed that blocking Cdc42 activity by a dominant-negative form of Cdc42 (Cdc42DN) significantly suppressed Eiger/JNK-induced cell death phenotype in the eye (Fig. 4A and B). In addition, expression of Cdc42DN blocked nuclear translocation of Yki in Rab5DN þ p35expressing cells (Fig. 4D, compare to Fig. 4C), suggesting the role of Cdc42 in inactivating the Hippo pathway. Furthermore, expression of Cdc42DN in Rab5DN þp35 clones blocked Upd induction as well as non-autonomous overgrowth (Fig. 4E and F, higher magnification images are shown in Fig. S5A and B). In contrast, expression of Cdc42DN did not affect JNK activation in Rab5DN þp35 clones (Fig. 4G). We also found that Cdc42 protein level was upregulated in Rab5DN þp35 clones (Fig. S5C). These results suggest that, in rab5 deficient clones, Eiger-JNK signaling causes Upd induction and non-autonomous growth through Cdc42.

Loss of Rab5 induces Upd expression through Eiger-JNK signaling We next examined the mechanism by which loss of Rab5 leads to Yki activation. It has been reported that Yki activity is positively regulated by JNK signaling (Ohsawa et al., 2012; Staley and Irvine, 2010; Sun and Irvine, 2011). Indeed, Rab5DN þp35 clones upregulated the expression of mmp1 (Fig. 3A), a transcriptional target of JNK signaling (Uhlirova and Bohmann, 2006). Furthermore, blocking JNK signaling in Rab5DN þp35 clones by a dominant-negative form of the Drosophila JNK Bsk (BskDN) suppressed MMP1 expression (data not shown), Upd expression (Fig. 3B), as well as nonautonomous overgrowth (Fig. 3C). These data indicate that loss of Rab5 leads to activation of JNK signaling that contributes to Yki activation. We next investigated the mechanism by which loss of Rab5 causes JNK activation. In imaginal discs, JNK signaling can be activated by a TNF superfamily ligand Eiger (Igaki et al., 2002; Moreno et al., 2002). Significantly, removal of eiger gene from the entire imaginal tissue blocked non-autonomous overgrowth triggered by Rab5DN þ p35 clones as well as upregulation of MMP1 and Upd in these clones (Fig. 3D and E). These data indicate that loss of Rab5 induces Upd expression through activation of Eiger-JNK signaling. A possible mechanism for activation of Eiger signaling in Rab5 deficient clones could be accumulation of Eiger protein in mutant cells because of an impaired endocytosis-mediated

Loss of Rab5 causes activation of Ras signaling Our genetic data so far presented suggest that loss of Rab5 causes inactivation of the Hippo pathway through the Eiger–JNK– Cdc42 pathway. However, JNK activation is not sufficient, as clones of cells solely activating JNK signaling in the imaginal disc do not cause non-autonomous overgrowth (Igaki et al., 2006; Igaki et al., 2009; Ohsawa et al., 2011), indicating that loss of Rab5 must induce an additional downstream effect(s) essential for the inactivation of the Hippo pathway. Our recent study has shown that JNK signaling cooperates with Ras signaling to inactivate the Hippo pathway (Ohsawa et al., 2012). Interestingly, we found that Rab5DN þ p35 clones downregulated Capicua (Cic) expression (Fig. 5A and B), indicating that these clones possess elevated activity of Ras signaling (Tseng et al., 2007). In addition, blocking Ras–MAPK signaling by a dominant-negative form of Ras effector Raf (RafDN) in Rab5DN þp35 clones blocked Upd expression (Fig. 5C) as well as non-autonomous overgrowth (Fig. 5E). In contrast, RafDN had no effect on the expression of JNK target MMP1 (Fig. 5D). These data suggest that loss of Rab5 causes activation of Eiger–JNK–Cdc42 and Ras–MAPK signaling, which cooperate together to inactivate the Hippo pathway. Ras signaling is activated through EGFR in Rab5 deficient cells Finally, we investigated the mechanism by which loss of Rab5 causes activation of Ras signaling. It has been shown that the endocytic machinery negatively regulates EGFR signaling that triggers activation of Ras signaling (Downward, 2003). Significantly, protein level of EGFR was strongly increased in Rab5DN þ p35 clones (Fig. 5F and G). Furthermore, reduced egfr expression by egfr-RNAi in rab5-/-/p35 clones blocked Upd expression (Fig. 5H) as well as non-autonomous overgrowth (Fig. 5I). We also found that co-activation of Cdc42 and Ras was not sufficient for Upd induction (data not shown), thus placing JNK, Cdc42, and Ras signaling parallel in the induction of Upd. Together, these data indicate that loss of Rab5 leads to cellular accumulation of Eiger and EGFR, thereby triggers activation of downstream JNK, Cdc42,

K. Takino et al. / Developmental Biology 395 (2014) 19–28

23

Fig. 2. Non-autonomous overgrowth is caused through inactivation of the Hippo pathway. (A and B) GFP-labeled clones overexpressing Rab5DN and p35 were induced in eye-antennal disc of ex-LacZ/þ (A) or diap1-LacZ/ þ (B) fly and was stained with anti-β-galactosidase antibody. The nuclei were visualized by DAPI staining. (C) GFP-labeled Rab5DN þ p35þ Wts-expressing clones were induced in eye-antennal disc and stained with anti-Upd antibody. The nuclei were visualized by DAPI staining. (D) Eye of adult fly bearing Rab5DN þ p35þWts clones. Light level (left) and fluorescent (right) images are shown. Genotypes are as follows: UAS-p35/ þ or Y; eyFLP5, Act4y þ 4Gal4, UASGFP/ex-lacZ; FRT82B, Tub-Gal80/FRT82B, UAS-Rab5DN (A), eyFLP5, Act 4y þ 4Gal4, UAS-GFP/UAS-p35; diap1-lacZ, FRT82B, Tub-Gal80/FRT82B,UAS-Rab5 DN (B), and UAS-p35/ þ or Y; eyFLP5, Act 4y þ 4Gal4, UAS-GFP / UAS-Wts; FRT82B, Tub-Gal80/UAS-Rab5DN, FRT82B (C, D).

and Ras signaling, leading to the cooperation of these signaling to cause the Hippo pathway-mediated Upd induction (Fig. 6).

Discussion Drosophila imaginal tissue entirely mutant for rab5 results in autonomous overgrowth and invasive behavior (Ballesteros-Arias

et al., 2013; Lu and Bilder, 2005), which makes the rab5 gene to be a “neoplastic” tumor suppressor (Bilder, 2004). In this study, we provide evidence that impaired Rab5 function can also promote tumor development through cell–cell communication. Clones of cells mutant for rab5 stimulate growth of surrounding wild-type cells through upregulation of Upd. Given that cancers arise from a small clone of oncogenic cells within a normal tissue, the nonautonomous tumor progression mechanism triggered by rab5

24

K. Takino et al. / Developmental Biology 395 (2014) 19–28

Fig. 3. Loss of Rab5 induces Upd expression through Eiger-JNK signaling. (A) GFP-labeled Rab5DN þp35 clones were induced in the eye-antennal disc and stained with antiMmp1 antibody. The nuclei were visualized by DAPI staining. (B) GFP-labeled Rab5DN þ p35þ BskDN clones were induced in the eye-antennal disc and stained with anti-Upd antibody. The nuclei were visualized by DAPI staining. (C) Eye of adult fly bearing Rab5DN þ p35þ BskDN clones. Light level (left) and fluorescent (right) images are shown. (D) Eye of adult fly bearing Rab5DN þp35 clones in eiger mutant background. Light level (left) and fluorescent (right) images are shown. (E) GFP-labeled Rab5DN þ p35 clones were induced in eiger mutant eye-antennal disc and stained with anti-Upd and anti-Mmp1 antibodies. The nuclei were visualized by DAPI staining. (F) GFP-labeled Rab5DN þp35 clones were induced in the eye-antennal disc and stained with anti-Eiger antibody. The nuclei were visualized by DAPI staining. Arrowheads (yellow) indicate representative Rab5DN þp35 clones upregulating Eiger expression. Arrow (white) indicates endogenous expression of Eiger near the morphogenetic furrow. Genotypes are as follows: UAS-p35/ þ or Y; eyFLP5, Act4 y þ 4Gal4, UAS-GFP/ex-lacZ; FRT82B, Tub-Gal80/FRT82B, UAS-Rab5DN (A, F), BskDN/UAS-p35; eyFLP5, Act 4y þ 4Gal4, UAS-GFP/ þ; FRT82B, Tub-Gal80/FRT82B, UAS-Rab5DN (B, C), and UAS-p35/eyFLP1; Act 4y þ 4Gal4, UAS-GFP, egr1/egr1; FRT82B, Tub-Gal80/UAS-Rab5DN, FRT82B (D, E).

K. Takino et al. / Developmental Biology 395 (2014) 19–28

25

Fig. 4. Eiger-JNK signaling promotes non-autonomous growth through Cdc42. (A and B) Eiger (A) or EigerþCdc42DN (B) was expressed under the control of the eye-specific promoter GMR. (C and D) GFP-labeled Rab5DN þ p35 (C) or Rab5DN þp35þ Cdc42DN (D) cells were induced in the posterior compartment of the wing disc and stained with anti-Yki antibody. Wing discs were subjected to heat-shock (37 1C) for 12 h to induce gene expression. The nuclei were visualized by DAPI staining. (E) GFP-labeled Rab5DN þ p35þ Cdc42DN clones were induced in the eye-antennal disc and stained with anti-Upd antibody. The nuclei were visualized by DAPI staining. (F) Eye of adult fly bearing Rab5DN þ p35þ Cdc42DN clones. Light level (left) and fluorescent (right) images are shown. (G) GFP-labeled Rab5DN þ p35þ Cdc42DN clones were induced in the eyeantennal disc and stained with anti-Mmp1 antibody. The nuclei were visualized by DAPI staining. Genotypes are as follows: UAS-Eigerregg1/ þ; GMR-gal4/ þ(A), UAS-Eigerregg1/ UAS-Cdc42DN; GMR-gal4/ þ (B), UAS-p35/ þ or Y; en-Gal4, UAS-GFP/ þ ; Tub-Gal80ts/UAS-Rab5DN, FRT82B (C), UAS-p35/ þ or Y; en-Gal4, UAS-GFP/UAS-Cdc42DN; Tub-Gal80ts/UASRab5DN, FRT82B (D), and UAS-p35/ þ or Y; eyFLP5, Act4y þ 4Gal4, UAS-GFP/UAS-Cdc42DN; FRT82B, Tub-Gal80/UAS-Rab5DN, FRT82B (E, F, G).

26

K. Takino et al. / Developmental Biology 395 (2014) 19–28

Fig. 5. Loss of Rab5 causes activation of Ras signaling. (A and B) GFP-labeled Rab5DN þp35 clones were induced in the eye-antennal disc and stained with anti-Capicua (Cic) antibody. The nuclei were visualized by DAPI staining. (B) is high-magnification image of the boxed area in (A). (C and D) GFP-labeled Rab5DN þp35þRafDN clones were induced in the eye-antennal disc and stained with anti-Upd (C) or anti-Mmp1 (D) antibody. The nuclei were visualized by DAPI staining. (E) Eye of adult fly bearing Rab5DN þp35þ RafDN clones. Light level (left) and fluorescent (right) images are shown. (F and G) GFP-labeled Rab5DN þp35 clones were induced in the eye-antennal disc and stained with anti-EGFR antibody. The nuclei were visualized by DAPI staining. (G) is high-magnification image of the boxed area in (F). (H) GFP-labeled rab5-/-/p35þegfr-RNAi clones were induced in the eye-antennal disc and stained with anti-Upd antibody. The nuclei were visualized by DAPI staining. (I) Eye of adult fly bearing rab5-/-/p35þ egfrRNAi clones. Light level (left) and fluorescent (right) images are shown. Genotypes are as follows: UAS-p35/ þ or Y; eyFLP5, Act4 y þ 4Gal4, UAS-GFP/ þ ; FRT82B, Tub-Gal80/ FRT82B, UAS-Rab5DN (A, B, F, G), UAS-p35/ þ or Y; eyFLP5, Act 4y þ 4Gal4, UAS-GFP/UAS-RafDN; FRT82B, Tub-Gal80/FRT82B, UAS-Rab5 DN (C, D, E), and UAS-p35/ þ or Y; TubGal80, FRT40A/PBac{SAstopDsRed}LL00467, FRT40A; eyFLP6, Act4y þ 4 Gal4, UAS-GFP/UAS-egfr-RNAi (H, I)

K. Takino et al. / Developmental Biology 395 (2014) 19–28

Fig. 6. A model for non-autonomous growth regulation by loss of Rab5. Loss of Rab5 upregulates levels of Eiger and EGFR proteins at the cell surface, which cause downstream JNK and Ras activation, respectively. While Cdc42 acts downstream of JNK signaling to promote cell death, it cooperates with Ras and JNK signaling to induce Upd expression through inactivation of the Hippo pathway, resulting in non-autonomous overgrowth of surrounding tissue.

defect could be crucial for cancer development compared to its autonomous oncogenic role. Similar non-autonomous overgrowths have been previously observed in mosaic tissues bearing clones of cells mutant for ESCRT genes such as vps22, vps25, and ept (Herz et al., 2006; Herz et al., 2009; Moberg et al., 2005; Thompson et al., 2005; Vaccari and Bilder, 2005). Unlike these ESCRT mutant clones, the non-autonomous overgrowth caused by Rab5 defect is not mediated through Notch signaling but through inactivation of the Hippo pathway, a frequent change observed in many types of human cancers(Bao et al., 2011; Chan et al., 2011; Fernandez and Kenney, 2010; Pan, 2010; Zeng and Hong, 2008; Zhao et al., 2010). Our data also indicate that dysfunction of endocytic machinery either at the early or late membrane trafficking results in upregulation of Upd through distinct mechanisms. Thus, our findings show that deregulation of signaling at the level of endosomal regulation causes multiple parallel signaling pathway defects that affect each other in a non-linear manner to promote tumorigenesis (Fig. 6). Intriguingly, clones of cells lacking rab5 function (rab5-/-/p35 or Rab5DN þp35 cells) exhibited enlarged nuclei (Figs. 1E and 5B), suggesting that rab5 mutant cells undergo endoreplication. We also noticed that inhibition of JNK signaling, Ras-Raf signaling, or Yki activity in rab5 mutant clones not only blocked nonautonomous overgrowth but also suppressed the polyploid induction (data not shown). Thus, it would be interesting to investigate in the future studies the molecular link between polyploid induction and non-autonomous growth regulation Endocytosis has long been recognized as a way to terminate signal transduction by degrading activated receptor complexes after internalization from the cell surface. In this study, we found that Eiger/TNF and EGFR are accumulated at the cell surface of rab5 mutant cells, which leads to activation of their downstream JNK-Cdc42 and Ras–MAPK signaling. Importantly, it has been shown in a variety of cellular and developmental systems that endocytic machinery can also promote activation of ligand–receptor signaling such as EGF-EGFR signaling in endosomes. Indeed, in Drosophila imaginal disc, elevated endocytosis in clones of cells mutant for apico-basal polarity gene scribble promotes translocation of Eiger from the plasma membrane to endocytic vesicles, which results in activation of downstream JNK signaling in endosomes (Igaki et al., 2009). These observations suggest that

27

deregulated endocytosis, either elevated or attenuated intracellular trafficking, can trigger activation of tumor-promoting ligand– receptor signaling such as TNF and EGF signaling. It has been shown that clones of “undead cells” (cells exposed to apoptotic stimuli but are rescued from cell death by a caspase inhibitor p35) cause overgrowth of surrounding tissue by producing secreted growth factors such as Wg and Dpp, a phenomenon called ‘apoptosis-induced proliferation’ (Mollereau et al., 2013). The non-autonomous stimulation of neighboring cell proliferation was found in Drosophila wing disc when massive cell death was induced within the tissue by radiation (Haynie and Bryant, 1977). In this case, unaffected viable cells surrounding dying cells proliferate to compensate for cell loss, which is called ‘compensatory proliferation’ (Mollereau et al., 2013). Thus, the non-autonomous overgrowth triggered by “undead cells” is considered to be an abnormally enhanced outcome of the compensatory proliferation. In the physiological condition, such compensatory cell proliferation could have a role in fine-tuning of normal developmental processes and tissue homeostasis, as well as cancer development (Bergmann and Steller, 2010; Morata et al., 2011). A similar multicellular event with a combination of cell death and compensatory cell proliferation can be observed in ‘cell competition’, a phenomenon whereby cells with higher fitness actively eliminate neighboring cells with lower fitness by inducing cell death (de Beco et al., 2012; Morata and Ripoll, 1975; Tamori and Deng, 2011; Vincent et al., 2013). Interestingly, it has recently been shown that clones of cells deficient for rab5 are eliminated from the imaginal disc by cell competition (Ballesteros-Arias et al., 2013). Thus, Eiger–JNK–Cdc42 and EGFR-Ras signaling activated downstream of Rab5 defect could have a role in compensatory cell proliferation during cell competition. Given that the pathways examined are evolutionarily conserved, it would be interesting to investigate whether the non-autonomous effect of rab5 deficiency plays a role in tissue homeostasis and cancer development in higher organisms.

Acknowledgments We thank Aya Betsumiya, Toshiharu Sawada, Mayumi Akatsuka, Yoshiaki Kanemoto, and Yoshitaka Sato for technical support; AHM Alam for comments on the manuscript; Masato Enomoto and Mikio Furuse for discussions; Marcos Gonzalez-Gaitan, Douglas Harrison, Masayuki Miura, Iswar Hariharan for antibodies; Takashi Adachi-Yamada, Masayuki Miura, Willis Li, Henry Sun, Tian Xu, the Bloomington Stock Center, the Vienna Drosophila RNAi Center, the National Institute of Genetics Stock Center (NIG-FLY), and the Drosophila Genetic Resource Center (DGRC, Kyoto Institute of Technology) for fly stocks. This work was supported in part by grants from the Japanese Ministry of Education, Science, Sports, Culture and Technology to S.O and T.I, Grant-in-Aid for Scientific Research on Innovative Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan to S.O and T.I, the Japan Society for the Promotion of Science to T.I, the Japan Science and Technology Agency to T.I, the G-COE program for Global Center for Education and Research in Integrative Membrane Biology to K.T, S.O and T.I, R1 (1:250), mouse anti-phospho-or fly stocks., D. Bilder, Pernille Rorth abtibody 1 antibody R1 (1:250), mouse anti-phospho-, Tomizawa Jun-ichi & Keiko Fund of Molecular Biology Society of Japan for Young Scientist to S.O, the Novartis Foundation for the Promotion of Science to S.O., the Nakajima Foundation to S.O., the Inoue Science Research Award to S.O., the Takeda Science Foundation to S.O and T.I, the Senri Life Science Foundation to S.O. and T.I, the Kao Function for Art and Science to S.O, the Shiseido Female Researcher Science grant to S.O., Uehara Memorial Foundation to S.O. and T.I., Suzuken

28

K. Takino et al. / Developmental Biology 395 (2014) 19–28

Memorial Foundation to T.I., Inamori Foundation to T.I., and the Human Frontier Science Program Career Development Award to T. I. K.T is a JSPS Fellow.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ydbio.2014.09.003. References L. Ballesteros-Arias, V. Saavedra, G. Morata, 2013. Cell competition may function either as tumour-suppressing or as tumour-stimulating factor in Drosophila. Oncogene. Bao, Y., Hata, Y., Ikeda, M., Withanage, K., 2011. Mammalian Hippo pathway: from development to cancer and beyond. J. Biochem. 149, 361–379. Bergmann, A., Steller, H., 2010. Apoptosis, stem cells, and tissue regeneration. Sci. Signal 3, re8. Bilder, D., 2004. Epithelial polarity and proliferation control: links from the Drosophila neoplastic tumor suppressors. Genes Dev. 18, 1909–1925. Bissell, M.J., Hines, W.C., 2011. Why don’t we get more cancer? A proposed role of the microenvironment in restraining cancer progression. Nat. Med. 17, 320–329. Brumby, A.M., Richardson, H.E., 2005. Using Drosophila melanogaster to map human cancer pathways. Nat. Rev. Cancer 5, 626–639. Chan, S.W., Lim, C.J., Chen, L., Chong, Y.F., Huang, C., Song, H., Hong, W., 2011. The Hippo pathway in biological control and cancer development. J. Cell Physiol. 226, 928–939. Chia, W.J., Tang, B.L., 2009. Emerging roles for Rab family GTPases in human cancer. Biochim. Biophys. Acta 1795, 110–116. Downward, J., 2003. Targeting RAS signalling pathways in cancer therapy. Nat. Rev. Cancer 3, 11–22. Fernandez, L.A., Kenney, A.M., 2010. The Hippo in the room: a new look at a key pathway in cell growth and transformation. Cell Cycle 9, 2292–2299. Floyd, S., De Camilli, P., 1998. Endocytosis proteins and cancer: a potential link? Trends Cell Biol. 8, 299–301. Giebel, B., Wodarz, A., 2006. Tumor suppressors: control of signaling by endocytosis. Curr. Biol. 16, R91–92. Halder, G., Johnson, R.L., 2011. Hippo signaling: growth control and beyond. Development 138, 9–22. Haynie, J.L., Bryant, P., 1977. The effects of X-rays on the proliferation dynamics of cells in the imaginal wing disc of Drosophila melanogaster. Rouxs Arch. Dev. Biol. 183, 85–100. Herz, H.M., Chen, Z., Scherr, H., Lackey, M., Bolduc, C., Bergmann, A., 2006. vps25 mosaics display non-autonomous cell survival and overgrowth, and autonomous apoptosis. Development 133, 1871–1880. Herz, H.M., Woodfield, S.E., Chen, Z., Bolduc, C., Bergmann, A., 2009. Common and distinct genetic properties of ESCRT-II components in Drosophila. PLoS One 4, e4165. Igaki, T., Kanda, H., Yamamoto-Goto, Y., Kanuka, H., Kuranaga, E., Aigaki, T., Miura, M., 2002. Eiger, a TNF superfamily ligand that triggers the Drosophila JNK pathway. EMBO J. 21, 3009–3018. Igaki, T., Pagliarini, R.A., Xu, T., 2006. Loss of cell polarity drives tumor growth and invasion through JNK activation in Drosophila. Curr. Biol. 16, 1139–1146. Igaki, T., Pastor-Pareja, J.C., Aonuma, H., Miura, M., Xu, T., 2009. Intrinsic tumor suppression and epithelial maintenance by endocytic activation of Eiger/TNF signaling in Drosophila. Dev Cell. 16, 458–465. Jacks, T., Weinberg, R.A., 2002. Taking the study of cancer cell survival to a new dimension. Cell 111, 923–925. Karim, F.D., Rubin, G.M., 1998. Ectopic expression of activated Ras1 induces hyperplastic growth and increased cell death in Drosophila imaginal tissues. Development 125, 1–9. Karpowicz, P., Perez, J., Perrimon, N., 2010. The Hippo tumor suppressor pathway regulates intestinal stem cell regeneration. Development 137, 4135–4145. Kinzler, K.W., Vogelstein, B., 1996. Lessons from hereditary colorectal cancer. Cell 87, 159–170. Lee, T., Luo, L., 1999. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22, 451–461.

Lu, H., Bilder, D., 2005. Endocytic control of epithelial polarity and proliferation in Drosophila. Nat. Cell Biol. 7, 1232–1239. Moberg, K.H., Schelble, S., Burdick, S.K., Hariharan, I.K., 2005. Mutations in erupted, the Drosophila ortholog of mammalian tumor susceptibility gene 101, elicit non-cell-autonomous overgrowth. Dev. Cell 9, 699–710. Mollereau, B., Perez-Garijo, A., Bergmann, A., Miura, M., Gerlitz, O., Ryoo, H.D., Steller, H., Morata, G., 2013. Compensatory proliferation and apoptosis-induced proliferation: a need for clarification. Cell Death Differ. 20, 181. Morata, G., Ripoll, P., 1975. Minutes: mutants of drosophila autonomously affecting cell division rate. Dev. Biol. 42, 211–221. Morata, G., Shlevkov, E., Perez-Garijo, A., 2011. Mitogenic signaling from apoptotic cells in Drosophila. Dev. Growth Differ. 53, 168–176. Moreno, E., Yan, M., Basler, K., 2002. Evolution of TNF signaling mechanisms: JNKdependent apoptosis triggered by Eiger, the Drosophila homolog of the TNF superfamily. Curr. Biol. 12, 1263–1268. Morrison, H.A., Dionne, H., Rusten, T.E., Brech, A., Fisher, W.W., Pfeiffer, B.D., Celniker, S.E., Stenmark, H., Bilder, D., 2008. Regulation of early endosomal entry by the Drosophila tumor suppressors Rabenosyn and Vps45. Mol. Biol. Cell 19, 4167–4176. Mosesson, Y., Mills, G.B., Yarden, Y., 2008. Derailed endocytosis: an emerging feature of cancer. Nat. Rev. Cancer 8, 835–850. Ohsawa, S., Sato, Y., Enomoto, M., Nakamura, M., Betsumiya, A., Igaki, T., 2012. Mitochondrial defect drives non-autonomous tumour progression through Hippo signalling in Drosophila. Nature 490, 547–551. Ohsawa, S., Sugimura, K., Takino, K., Xu, T., Miyawaki, A., Igaki, T., 2011. Elimination of Oncogenic Neighbors by JNK-Mediated Engulfment in Drosophila. Dev. Cell 20, 315–328. Pagliarini, R.A., Xu, T., 2003. A genetic screen in Drosophila for metastatic behavior. Science 302, 1227–1231. Pan, D., 2010. The hippo signaling pathway in development and cancer. Dev. Cell 19, 491–505. Pastor-Pareja, J.C., Xu, T., 2013. Dissecting social cell biology and tumors using Drosophila genetics. Annu. Rev. Genet. 47, 51–74. Reginensi, A., Scott, R.P., Gregorieff, A., Bagherie-Lachidan, M., Chung, C., Lim, D.S., Pawson, T., Wrana, J., McNeill, H., 2013. Yap- and Cdc42-dependent nephrogenesis and morphogenesis during mouse kidney development. PLoS Genet. 9, e1003380. Staley, B.K., Irvine, K.D., 2010. Warts and Yorkie mediate intestinal regeneration by influencing stem cell proliferation. Curr. Biol. 20, 1580–1587. Sun, G., Irvine, K.D., 2011. Regulation of Hippo signaling by Jun kinase signaling during compensatory cell proliferation and regeneration, and in neoplastic tumors. Dev. Biol. 350, 139–151. Tamori, Y., Deng, W.M., 2011. Cell competition and its implications for development and cancer. J. Genet. Genom. 38, 483–495. Thompson, B.J., Mathieu, J., Sung, H.H., Loeser, E., Rorth, P., Cohen, S.M., 2005. Tumor suppressor properties of the ESCRT-II complex component Vps25 in Drosophila. Dev Cell. 9, 711–720. Tomas, A., Futter, C.E., Eden, E.R., 2014. EGF receptor trafficking: consequences for signaling and cancer. Trends Cell Biol. 24, 26–34. Tseng, A.S., Tapon, N., Kanda, H., Cigizoglu, S., Edelmann, L., Pellock, B., White, K., Hariharan, I.K., 2007. Capicua regulates cell proliferation downstream of the receptor tyrosine kinase/ras signaling pathway. Curr. Biol. 17, 728–733. Uhlirova, M., Bohmann, D., 2006. JNK- and Fos-regulated Mmp1 expression cooperates with Ras to induce invasive tumors in Drosophila. EMBO J. 25, 5294–5304. Vaccari, T., Bilder, D., 2005. The Drosophila tumor suppressor vps25 prevents nonautonomous overproliferation by regulating notch trafficking. Dev. Cell. 9, 687–698. Vaccari, T., Rusten, T.E., Menut, L., Nezis, I.P., Brech, A., Stenmark, H., Bilder, D., 2009. Comparative analysis of ESCRT-I, ESCRT-II and ESCRT-III function in Drosophila by efficient isolation of ESCRT mutants. J. Cell Sci. 122, 2413–2423. Vincent, J.P., Fletcher, A.G., Baena-Lopez, L.A., 2013. Mechanisms and mechanics of cell competition in epithelia. Nat. Rev. Mol. Cell Biol. 14, 581–591. Xu, T., Rubin, G.M., 1993. Analysis of genetic mosaics in developing and adult Drosophila tissues. Development 117, 1223–1237. Zeng, Q., Hong, W., 2008. The emerging role of the hippo pathway in cell contact inhibition, organ size control, and cancer development in mammals. Cancer Cell 13, 188–192. Zhao, B., Li, L., Lei, Q., Guan, K.L., 2010. The Hippo-YAP pathway in organ size control and tumorigenesis: an updated version. Genes Dev. 24, 862–874. de Beco, S., Ziosi, M., Johnston, L.A., 2012. New frontiers in cell competition. Dev. Dyn. 241, 831–841.