From Cell Structure to Transcription: Hippo Forges a New Path

Leading Edge

Review From Cell Structure to Transcription: Hippo Forges a New Path Bruce A. Edgar1,*

Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle WA 98109, USA *Contact: [email protected] DOI 10.1016/j.cell.2006.01.005 1

The control of cell number during animal development is a longstanding puzzle. Recent studies in the fruit fly Drosophila melanogaster have defined a new signaling pathway that restricts cell proliferation in differentiating epithelia. The cytoskeletal proteins Merlin and Expanded, which play a role in cell adhesion and structure, control the activation of the Hippo/Salvador kinase complex, which in turn activates the Warts/Mats kinase complex. Warts/Mats kinase phosphorylates and inhibits Yorkie, a transcriptional coactivator that positively regulates cell growth, survival, and proliferation. This conserved signaling pathway contains several tumor-suppressor genes and regulates the contact inhibition of proliferation in cultured cells. Developmental and cancer biologists have long pondered how cell proliferation is restricted, first during embryogenesis to allow morphogenesis and later in adults to maintain homeostasis. I use the term “restricted,” rather than “promoted” because, although it is often assumed that cells must be stimulated to proliferate, the basal state of the single-celled eukaryotes that gave rise to metazoa was almost certainly one of proliferation, limited only by nutrients. With the arrival of multicellularity came the requirement that cells acquire characteristics akin to cooperation and altruism, such as the ability to restrain their growth and division in response to signals from other cells and to die when no longer needed. These properties are essential for proper development and for maintaining homeostasis in adults. Acquiring them necessitated the establishment of new mechanisms to allow cells to sense their density or numbers and to apply this information to regulate intrinsic cellular processes including growth, division, and survival. Although it is clear that genetic systems controlling cell numbers are vulnerable (consider cancer as an example), we are still far from understanding how these systems operate during normal development. In this review, I highlight some recent advances relevant to this puzzle. A New Signaling Complex The advent of modern methods for generating genetic mosaics in the fruit fly Drosophila melanogaster (Xu and Rubin, 1993) prompted screens in which clones of cells with mutations in random genes were produced, and the adults were scored for abnormal tumor-like growths. These screens identified many genes that, when mutated, caused the inappropriate proliferation of cells. Many of these genes turned out to be orthologs of known or suspected human tumor-suppressor genes, such as PTEN or TSC2. Studies of these genes in Drosophila have proven

useful for assigning functions to the proteins they encode and mapping their signaling pathways. Some new growth suppressors were also discovered. The first of these was warts/large tumor suppressor (wts/lats), a kinase of the nuclear Dbf-2-related (NDR) family (Justice et al., 1995; Xu et al., 1995). Cells with mutations in the wts gene exhibit dramatic outgrowths in a variety of fly epithelial tissues but do not display overt changes in cell identity or the ability to differentiate. Further screens isolated three other genes— hippo (hpo), salvador (sav), and mob as tumor suppressor (mats)—that gave strikingly similar tumorous phenotypes when clonally deleted in developing eyes, wings, or legs of the fly (Harvey et al., 2003; Kango-Singh et al., 2002; Pantalacci et al., 2003; Tapon et al., 2002; Udan et al., 2003; Wu et al., 2003, Lai et al., 2005). The similarities among the phenotypic effects caused by mutations in these four genes suggested that they might function in a common pathway. When deleted in wing epithelial cells, each of them causes a cell shape change termed apical hypertrophy, in which the apical cell surface expands away from the nucleus, taking on a domed shape, and areas of apical-lateral adhesion are reduced (Justice et al., 1995). These apical-lateral regions are where adherens junctions reside in insect epithelia, suggesting that this set of genes affects the function of adherens junctions (Bilder, 2004; Perez-Moreno et al., 2003). Clones of mutant cells produced in the wing or eye primordia (called “discs”) also exhibit smooth borders and a rounded shape, indicative of altered cell adhesion, although cell polarity and organization of the epithelial monolayer are maintained. Prior to differentiation, the mutant cells were also observed to grow and divide more rapidly than normal cells, without any change in cell size. A distinct effect was noted in the eye, where mutations in hpo, sav, or wts each cause inappropriate proliferation and survival of nonneural interommatidial cells just after Cell 124, January 27, 2006 ©2006 Elsevier Inc.  267

differentiation. This sugof apoptosis gene product, gested a common function DIAP-1, and the cell cycle for these genes during eye regulator Cyclin E (Harvey development (see Figure et al., 2003; Pantalacci 1). Normally, interommaet al., 2003; Tapon et al., tidial cells are produced in 2002; Udan et al., 2003; excess, and most of them Wu et al., 2003). DIAP-1 are culled in a wave of and Cyclin E are important apoptosis during eye difregulators of cell survival ferentiation. In hpo, sav, or and cell cycle progression, wts mutants, these interomrespectively, and thus they matidial cells fail to die but probably account, at least instead divide several extra in part, for the survival times, causing expansion and continued proliferation and some disorganization of cells with mutations in of the eye. Mutations in the these genes. Other potenmats gene were reported tial clues to function have to cause a somewhat more come from studies of the severe phenotype, in which mammalian Warts orthodifferentiation was abrologs, LATS1 and LATS2/ gated and, perhaps as a Kpm, which implicate both result, substantial apoptosis proteins in the regulation of occurred at late stages (Lai mitosis. LATS1 and LATS2/ et al., 2005). When very large Kpm are phosphorylated Figure 1. Scanning Electron Micrographs Demonstrating cell clones with mutations in and activated in mitosis, the Overgrowth Phenotypes Conferred by Mutations in the Merlin/Hippo/Yorkie Pathway in Drosophila melanogaster wts or hpo were produced, and when overexpressed (A) A normal fly thorax. however, they too showed they can bind to and inhibit (B) A tumorous outgrowth of the thorax caused by clonal deletion of pronounced defects in cell the mitotic kinase Cdk1 merlin and expanded. (C—E) The crystalline arrangement of cells in the pupal retina (C) is differentiation (Wu et al., (Tao et al., 1999; Yang et similarly disrupted by overproliferation and inappropriate survival of in2003; Xu et al., 1995), sugal., 2001; Xia et al., 2002; terommatidial cells in hippo (D) and merlin expanded (E) mutant retinas. gesting that the severe pheKamikubo et al., 2003; Iida Cell boundaries revealed using anti-Dlg antibodies. Images courtesy of notype conferred by mats et al., 2004). This suggests Georg Halder. mutations may not actually a possible explanation for reflect a distinct function. LATS-dependent growth The cause of these differentiation defects is unclear, but suppression. Perhaps more importantly, loss-of-function given that they vary according to clone size, they could be studies revealed that cells lacking LATS1 or LATS2 exhibit due to the failure of the mutant cells to effectively transduce defects in exiting mitosis (Bothos et al., 2005) and in cytokilong-range differentiation signals. nesis (Yang et al., 2004; McPherson et al., 2004) that lead to Given these similarities in mutant phenotypes, perhaps multinucleate cells, centrosome amplification, and genomic it is not surprising that the Hippo and Sav proteins were instability. Functions in mitosis or cytokinesis might contribfound to physically interact with each other, as were the ute to the tumor-suppressor activity of the LATS genes in Warts and Mats proteins (see Figure 2). Biochemical assays mammals (Takahashi et al., 2005; St. John et al., 1999), showed that Hippo, a Ste-20/MST2 type kinase, phosbut such defects do not provide a ready explanation for the phorylates and activates Warts and that Sav, a WW-repeat increased cell growth and survival seen in wts mutant cells protein, facilitates this reaction potentially by acting as a in Drosophila. Moreover, as mitotic defects have not been scaffold. Furthermore, Mats was found to bind to Warts, observed in fly cells with mutations in wts, hpo, sav, or mats, but not to Hippo, and to stimulate Warts kinase activity. these functions may be specific to the mammalian LATS1 Genetic tests performed in vivo supported these interacand LATS2 genes. tions. Mats has also been shown to physically and genetiThus, despite the intriguing overgrowth phenotypes cally interact with another NDR-family kinase, tricornered conferred by loss of genes encoding Hippo complex com(trc), which is required for the morphogenesis of epidermal ponents, Hippo has until recently remained an orphan hairs, bristles, and dendrites (He et al., 2005). This might signaling entity, lacking known upstream inputs and, for explain the relative severity of the mats mutant phenotype. the most part, critical downstream effectors. Given that the Hippo and Warts complexes could not easily be slotDownstream of Hippo and Warts ted into any known signaling pathway, it was assumed Targeted searches for downstream effectors revealed that that they must define a new signaling pathway importhe loss of hpo, sav, wts, or mats caused a cell-autonotant for containing proliferation, at least in epithelial cells. mous increase in the amounts of the Drosophila inhibitor Although the increase in Cyclin E and DIAP-1 may explain 268  Cell 124, January 27, 2006 ©2006 Elsevier Inc.

the boost in cell numbers seen in mutant tissues, these two target molecules cannot explain the stimulation of cell growth (increased cell mass) or the intriguing cell shape and adhesion defects displayed by these cells. Moreover, how the Warts kinase complex controls even its known targets (cyclin E, diap-1) is not obvious, with some researchers suggesting transcriptional control and others favoring regulation at the protein level. Two recent papers have redrawn this lonely scene, adding a downstream effector (Huang et al., 2005), some upstream inputs (Hamaratoglu et al., 2005), and an interesting feedback loop. The downstream effector, Yorkie (Yki), was discovered by Duojia “DJ” Pan’s group not by phenotypic screens in flies but by virtue of its ability to bind to Warts in a yeast twohybrid screen (Huang et al., 2005). Yorkie is the Drosophila ortholog of YAP (Yes-associated protein), which has been characterized as a non-DNA binding transcriptional coactivator in mammalian cells (Yagi et al., 1999). Yorkie fulfills all of the criteria required of a Warts effector: it not only binds to Warts but appears to recapitulate all of its tumorsuppressor functions when overexpressed in cell clones in developing wings and eyes of the fly. Moreover, loss-offunction mutations in yki arrest cell growth and are epistatic for hpo and wts, meaning that hpo, yki, or wts, yki doublemutant cells grow just as poorly as yki single-mutant cells. This suggested that the Hippo/Sav and Warts/Mats complexes might act upstream of Yorkie as negative regulators, and indeed, they were shown experimentally to suppress transcriptional activation by Yorkie. Biochemical analysis of cultured Drosophila cells showed that Yorkie is phosphorylated by Warts in a Hippo/Sav-dependent fashion, suggesting that Warts normally suppresses Yorkie activity by phosphorylating it. The specific role, if any, that phosphorylation plays in regulating the activity of Yorkie remains to be determined. The discovery of Yorkie is a substantial advance because it will allow the identification of the binding partners that bring it to the promoters of target genes. This will facilitate the comprehensive cataloging of Hippo/Warts targets that are regulated transcriptionally. Given the near-complete phenocopy of wts, hpo, and sav mutant phenotypes by overproduced Yorkie, this may cover most of the relevant targets and thus explain the cell adhesion, cell shape, and growth effects of the pathway, which are presently mysterious. The study from Huang et al. (2005) may provide the means to answer other pressing questions about the Hippo pathway. For instance, it has been suspected that Hippo/ Warts activity may increase as cells differentiate, thereby suppressing proliferation and promoting apoptosis, but so far there has been little evidence that the pathway is subject to spatiotemporal regulation at the onset of differentiation. Despite the observation that it has more profound effects in differentiating cells than proliferating ones, and in some cell types (interommatidial) than others (photoreceptors), all Hippo pathway components appear to be uniformly expressed during the stages of eye development when the mutant phenotypes unfold. Assays of Yorkie activity, most likely using phosphospecific antibodies but potentially

based on Yorkie nuclear localization or yki reporter gene expression, may reveal how the Hippo pathway is regulated in space and time during development. Such studies are likely to uncover considerable complexity, as evidenced by a striking recent report showing that Warts, Hippo, and Sav play a second role as essential regulators of photoreceptor cell type choice in the postmitotic Drosophila eye (Mikeladze-Dvali et al., 2005). In this case, the transcriptional targets, which might again be regulated by Yorkie, are rhodopsin genes essential for color vision and have no apparent relevance to growth suppression. This suggests that Yorkie probably activates distinct sets of genes using different DNA binding partners and that its spectrum of targets will vary greatly according to developmental stage and cellular context. Interestingly, MikeladzeDvali et al. (2005) also found that wts transcript expression correlated with photoreceptor cell type and was regulated by a conserved pleckstrin-homology (PH) domain-containing protein, Melted, in a bistable loop in which Melted and Warts can mutually repress each other at the transcriptional level. This loop ensures that photoreceptors express either Melted and one rhodopsin isoform (rh5, used for shorter wavelength colors) or Warts and a different rhodopsin isoform (rh6, used for longer wavelength colors), to guarantee that no cells express both rhodopsins simultaneously. Interactions between melted and wts, or cell type-specific transcription of wts, have not yet been noted during the growth phase in the eye or other organs. Up the Nile This leads us to another mystery that has dogged researchers in this field for several years: what are the upstream inputs that regulate the Hippo/Warts pathway, if it is regulated? Again, to the chagrin of the fly geneticists, forward genetic screens gave few clues. A candidate gene approach taken by Georg Halder’s group, however, recently provided a lead that should help to focus the search (Hamaratoglu et al., 2005). This study followed up on earlier work (McCartney et al., 2000) that showed that cells doubly mutant for two genes that encode band 4.1 superfamily proteins, merlin and expanded, exhibited epidermal outgrowths similar to those caused by loss of hpo, wts, or sav. merlin and expanded encode members of the ezrin/radixin/moesin (ERM) family of FERM domain-containing proteins. They are partially redundant for function and can form heterodimeric complexes. Both proteins localize adjacent to adherens junctions and are thought to link transmembrane proteins to the actin cytoskeleton (Bretscher et al., 2002; Sun et al., 2002; McClatchey and Giovanni, 2005). expanded mutants exhibit mild overgrowth phenotypes in flies (Boedigheimer et al., 1997; Boedigheimer and Laughon, 1993), and mammalian fibroblasts and endothelial cells lacking Merlin, encoded by the neurofibromatosis 2 (NF2) tumorsuppressor gene, are resistant to contact inhibition of proliferation (Lallemand et al., 2003; Okada et al., 2005). Interestingly, Lats2 mutant mouse embryonic fibroblasts also exhibit this defect (McPherson et al., 2004), suggesting that LATS2 may act downstream of Merlin in mammals as Cell 124, January 27, 2006 ©2006 Elsevier Inc.  269

likely involved, but the general implication is that the Hippo and Warts kinases communicate cues reflecting cell structure, emanating from Merlin and Expanded, to the nucleus via Yorkie (see Figure 2).

Figure 2. The Emerging Hippo Pathway

Parts of two epithelial cells are shown, with Merlin (Mer) and Expanded (Ex) interacting with plasma membrane proteins (identity unknown) and the actin cytoskeleton and transducing a signal that stimulates the phosphorylation of Warts (Wts), which is in a complex with Mats (Mob as tumor suppressor), by Hippo (Hpo), which is in a complex with Salvador (Sav). Activated Warts phosphorylates and inhibits the transcriptional coactivator Yorkie (Yki), thereby suppressing cell growth and proliferation and allowing programmed cell death. merlin and expanded are also targets of Hippo signaling, creating a growth-suppressive feedback loop. Not all reported physical interactions or phosphorylations are shown.

its ortholog, Warts, does in Drosophila (see below). In flies, the overgrowth phenotype of merlin, expanded double-mutant cells is much more severe than that caused by either single mutant. Notably, the double-mutant phenotype is strikingly similar to that conferred by loss of wts, hpo, or sav (see Figure 1). When generated in the Drosophila eye, merlin, expanded double-mutant cells grow into smooth-bordered clones that proliferate inappropriately during the differentiation phase and then evade the normal wave of apoptosis that culls excess interommatidial cells. As with hpo and wts complex mutants, merlin, expanded double-mutant cells also sustain increased transcription of cyclin E and diap-1. Hamaratoglu et al. (2005) went on to use a clever battery of epistasis tests to show that hpo and wts are required for and act genetically downstream of merlin and expanded in containing growth and culling excess interommatidial cells at differentiation. For instance, overexpressed expanded promoted apoptosis but was incapable of doing so in hpo mutant tissue. Biochemical tests indicated that Merlin and Expanded do not bind directly to Hippo, Sav, or Warts; however, they do stimulate the phosphorylation of Warts by Hippo and thereby probably activate Warts to signal downstream. Consistent with this notion, transcriptional activation by Yorkie was shown to be suppressed by high levels of Merlin and Expanded. How these ERM proteins stimulate the interaction between Hippo and Warts is still unclear. Intermediate factors are 270  Cell 124, January 27, 2006 ©2006 Elsevier Inc.

Searching for the Source Although the control of the Hippo complex by ERM family proteins extends the pathway upstream a step, the exact source of the input signals is still murky. Merlin can heterodimerize with Ezrin (in humans) or Expanded (in Drosophila), and it has been reported to bind numerous factors that have potential functional relevance. These include cytoskeletal actin, β-catenin, βII-spectrin, phosphati­dylinositol (4,5)-bis­ phosphate (PIP2), paxillin, and several transmembrane proteins such as (in humans) CD44, a receptor for the extracellular matrix component hyaluronic acid (Sun et al., 2002; McClatchey and Giovanni 2005). Most of these binding capabilities reside in the FERM domain of Merlin, which is highly conserved in metazoa and distinct from those of other ERM family members (Golovnina et al., 2005). ERM proteins including Merlin can switch between open (unfolded) and closed (self-binding) states dependent upon phosphorylation, protein-protein interactions, and mutations such as those in the human Merlin/NF2 gene that cause neurofibromatosis. This conformational change is believed to determine Merlin’s interactions with its binding partners, and it is the closed, hypophosphorylated conformation that appears to exert growth-suppressive activity. Levels of hypophosphorylated Merlin increase upon loss of adhesion in cultured (NIH 3T3) cells, and levels of both the hypo- and hyperphosphorylated protein increase dramatically when such cells become confluent, or are deprived of serum growth factors (Shaw et al., 1998). These changes are assumed to accompany increased Merlin activity since they are associated with growth arrest. Similar changes in the growth suppressive activity of Merlin have not yet been described in differentiating tissues in vivo but will be very interesting to look for. The aforementioned findings have suggested that Merlin might integrate growth stimulatory and inhibitory signals from transmembrane receptors like CD44 and the EGFR or adhesion molecules such as the cadherins (Sun et al., 2002; McClatchey and Giovanni 2005), but so far the functional data to support this attractive hypothesis is somewhat limited. Of note are studies performed in fibroblasts and endothelial cells showing that Merlin acts in a mutually antagonistic loop with p21-activated kinase (PAK) and its activator, the small GTPase Rac (Shaw et al., 2001; Okada et al., 2005). Merlin phosphorylation is Rac and PAK dependent, whereas the cortical localization (and presumed activity) of Rac can be suppressed by Merlin. Of particular significance is the observation that activated forms of PAK or Rac can release human endothelial cells from contact inhibition, just as does loss of Merlin activity (Lallemand et al., 2003), and that an activated variant of Merlin can block this effect. Since Rac activity can be stimulated by growth factors via integrins and receptor tyrosine kinases, several authors have suggested that Merlin activity is suppressed by such signals, at least at low cell density. At high cell den-

sities, however, the growth suppressive activity of Merlin is dominant to mitogenic signaling, and thus it is difficult to view Merlin as a simple signal transduction component. Alternative mechanisms, in line with the traditional view that ERM proteins form a structural linkage between transmembrane components and the actin cytoskeleton, are also interesting to consider (McClatchey and Giovanni, 2005). Such alternatives do not necessarily require specific ligands or receptors as regulators of Merlin/Expanded function. In one scenario, Merlin and Expanded may act as structural components or regulators of cell structure—possibly compromising plasma membrane-cytoskeletal crosslinking when absent—and this would be sensed by the Hippo complex. This possibility seems consistent with the rounded cell-clone morphologies observed in Drosophila (Hamaratoglu et al., 2005) and the observation that Merlin is essential for the maintenance of adherens junctions in mouse embryonic fibroblasts (Lallemand et al., 2003). If Merlin and Expanded act as structural components, then numerous other regulators of cell adhesion and the cytoskeleton might also be expected to affect Hippo/Warts activity. Among the more intriguing genes to consider are the Src protooncogene, which encodes for a non-receptor tyrosine kinase, and its negative regulator encoded by C-Src kinase (csk). The literature implicating Src as a regulator of cell adhesion and actin cytoskeletal function is vast (Frame, 2004), and both functions are consistent with genetic analysis in flies. Interestingly, loss of csk, which activates Src, causes overgrowth phenotypes in the fly eye that are not dissimilar to those observed in hpo, wts, or merlin, expanded mutants: interommatidial cells undergo ectopic proliferation and fail to be appropriately culled by apoptosis (Read et al., 2004). In this case, the effects downstream of Src were attributed in part to activation of Jun-kinase and STAT signaling, but the potential involvement of the Hippo complex and Yorkie were not assessed. Interestingly, Nf2 mutant cells, which lack Merlin, also have increased Jun-kinase activity (Shaw et al., 2001). Another plausible role for Merlin and Expanded would be to act as sensors of a mechanical property, such as tension, at the membrane-cytoskeleton interface. In this capacity, they might provide a means for the Hippo complex to monitor cell-cell contact or adhesion. Growth suppression by cell-cell contact is of course well known in cell culture, and, as mentioned above, Merlin, Rac, and LATS2 mediate this response in several cell types (Lallemand et al., 2003; McPherson et al., 2004; Okada et al., 2005). Though it is unclear whether contact inhibition is a useful concept for understanding how cell proliferation is contained during differentiation in vivo, the possibility is intriguing to consider. Developmental biologists sometimes assume that specific ligands and receptors must be used to arrest cell proliferation upon differentiation, but in fact there are few known signals that can be classified as dedicated differentiation signals. In Drosophila, signals that are used for patterning in early development (such as EGFR ligands, BMPs, Wnts, Notch ligands) are used again to drive differentiation (see for example, Firth and Baker [2005]). As part of their job

as promoters of differentiation and morphogenesis, these signals cause profound changes in cytoskeletal organization, cell adhesion, cell shape, and polarity. Hence, it may not be unreasonable to consider structural changes, such as increased adhesion or a stabilized cytoskeletal configuration, as the proximal trigger for cell cycle arrest upon differentiation. Merlin and Expanded might be used to sense, or orchestrate, these structural changes. A Homeostat for Cell Structure? In their studies of merlin and expanded, Hamaratoglu et al. (2005) discovered another intriguing phenomenon. They show that loss of hpo or wts causes a robust upregulation of Merlin and Expanded protein expression. This was attributed to increased levels of transcription, and thus (though it remains to be tested) the obvious expectation is that it is mediated via increased Yorkie activity. This negative feedback loop probably explains why loss of merlin upregulates expanded expression (McCartney et al., 2000) and so may explain how Merlin and Expanded complement each others’ function. The potential impact of this feedback loop on cytoarchitecture and growth control is especially interesting to consider. Although the Drosophila studies have not yet detailed the effects of Merlin or Expanded on cell structure, expanded mutant wings do show evidence of altered cell shape (Boedigheimer et al., 1993), and studies of human Merlin indicate that it can alter adhesion, motility, and actin cytoskeletal morphology (Sun et al., 2002; Lallemand et al., 2003; McClatchey and Giovanni 2005). If Merlin and Expanded were to act as both sensors of and regulators of cell structure, then feedback via Hippo signaling could generate a closed loop, or homeostat, that might maintain some critical cytoarchitectural parameter (such as plasma membrane-cytoskeletal tension) at an optimum. For example, decreased plasma membrane-cytoskeletal tension could suppress Hippo signaling and thereby enhance cell growth and survival, but also upregulate Merlin. The increased Merlin activity would then reestablish tension and activate Hippo signaling, rendering the growth response transient. Hence, the outlines of a mechanism of tissue homeostasis, as used in wound healing or contact inhibition, begin to emerge. Hippo and Us Drosophila researchers refer to hpo, sav, wts, and mats as “tumor suppressors” and have proposed that they will have a similar function in humans. Indeed, the components of this pathway are all well conserved, and some of the interactions observed in flies have been validated in mammalian cells. MST2 (Hippo) for instance, can activate LATS1 (Warts) (Chan et al., 2005), and human Mats homologs (the MOB proteins) have been found to associate with multiple NDR type kinases homologous to Warts (Tamaskovic et al., 2003). The evolutionary conservation is underscored by the demonstration that the human LATS1, MATS1, MST2, and YAP genes can all functionally rescue their respective Drosophila mutants in vivo. Therefore, it seems that the basic outlines of the pathway as deterCell 124, January 27, 2006 ©2006 Elsevier Inc.  271

mined in flies will be conserved in humans. It is still unclear, however, whether the mammalian Yorkie ortholog, YAP, will account for the majority of LATS function, as Yorkie appears to do for Warts. YAP has been reported to work as a cofactor for a variety of human transcriptional regulators, including P73, P53BP-2, Runx2, SMAD7, ErbB4, PEBP2α, and hnRNPU, but the relevance of these to LATS or MST2 function is unknown. It is also puzzling that some of the described functions of YAP are opposite to those of Yorkie; for instance, YAP can promote apoptosis after DNA damage as a cofactor for P73 (Basu et al., 2003; Strano et al., 2001), whereas Yorkie is clearly an apoptosis suppressor in the fly eye. Furthermore, the potential connection between human Merlin and the LATS/MST2 complex remains wholly unexplored, as does the possibility that Rac and PAK regulate Warts/Hippo and LATS/MST2 signaling in flies and humans, respectively. As suggested by the Drosophila mutant phenotypes, a number of the mammalian genes in this emerging pathway have already been associated with tumorigenesis, with varying degrees of confidence. These include Lats1 and Lats2 (St. John et al., 1999; Takahashi et al., 2005), hMats1, and Sav1/WW45 (Tapon et al., 2002). NF2/Merlin, however, is the only one of these genes that is clearly causal for a human disease, neurofibromatosis, characterized by benign tumors of the central nervous system (McClatchey and Giovanni 2005). Determining whether the others are important tumor suppressors, regulators of tissue organization that lack cancer-promoting capability, or regulators of a completely disparate process such as color vision (Mikeladze-Dvali et al., 2005) will require that more stringent in vivo tests be performed in mice, similar to the mosaic knockouts done in Drosophila. Even if these genes prove not to be tumor suppressors in humans, they are likely to play important roles in preventing benign dystrophic diseases and ensuring normal cell and tissue architecture in developing epithelial organs. References Basu, S., Totty, N.F., Irwin, M.S., Sudol, M., and Downward, J. (2003). Akt phosphorylates the Yes-associated protein, YAP, to induce interaction with 14-3-3 and attenuation of p73-mediated apoptosis. Mol. Cell 11, 11–23. Bilder, D. (2004). Epithelial polarity and proliferation control: links from the Drosophila neoplastic tumor suppressors. Genes Dev. 18, 1909–1925. Boedigheimer, M., and Laughon, A. (1993). Expanded: a gene involved in the control of cell proliferation in imaginal discs. Development 118, 1291–1301. Boedigheimer, M., Bryant, P., and Laughon, A. (1993). Expanded, a negative regulator of cell proliferation in Drosophila, shows homology to the NF2 tumor suppressor. Mech. Dev. 44, 83–84. Boedigheimer, M.J., Nguyen, K.P., and Bryant, P.J. (1997). Expanded functions in the apical cell domain to regulate the growth rate of imaginal discs. Dev. Genet. 20, 103–110. Bothos, J., Tuttle, R.L., Ottey, M., Luca, F.C., and Halazonetis, T.D. (2005). Human LATS1 is a mitotic exit network kinase. Cancer Res. 65, 6568–6575. Bretscher, A., Edwards, K., and Fehon, R.G. (2002). ERM proteins and

272  Cell 124, January 27, 2006 ©2006 Elsevier Inc.

merlin: integrators at the cell cortex. Nat. Rev. Mol. Cell Biol. 3, 586– 599. Chan, E.H., Nousiainen, M., Chalamalasetty, R.B., Schafer, A., Nigg, E.A., and Sillje, H.H. (2005). The Ste20-like kinase Mst2 activates the human large tumor suppressor kinase Lats1. Oncogene 24, 2076–2086. Firth, L.C., and Baker, N.E. (2005). Extracellular signals responsible for spatially regulated proliferation in the differentiating Drosophila eye. Dev. Cell 8, 541–551. Frame, M.C. (2004). Newest findings on the oldest oncogene; how activated src does it. J. Cell Sci. 117, 989–998. Golovnina, K., Blinov, A., Akhmametyeva, E.M., Omelyanchuk, L.V., and Chang, L.S. (2005). Evolution and origin of merlin, the product of the Neurofibromatosis type 2 (NF2) tumor-suppressor gene. BMC Evol. Biol. 5, 69. Hamaratoglu, F., Kango-Singh, M., Nolo, R., Hyun, E., and Tao, C. JafarNejad, H., and Halder, G. (2005). The tumor suppressor genes NF2/Merlin and Expanded act through Hippo signalling to regulate cell proliferation and apoptosis. Nat. Cell Biol. 8, 27–36. Published online December 11, 2005. 10.1038/ncb1339. Harvey, K.F., Pfleger, C.M., and Hariharan, I.K. (2003). The Drosophila Mst ortholog, hippo, restricts growth and cell proliferation and promotes apoptosis. Cell 114, 457–467. He, Y., Emoto, K., Fang, X., Ren, N., Tian, X., Jan, Y.N., and Adler, P.N. (2005). Drosophila mob family proteins interact with the related tricornered (trc) and warts (wts) kinases. Mol. Biol. Cell 16, 4139–4152. Huang, J., Wu, S., Barrera, J., Matthews, K., and Pan, D. (2005). The Hippo signaling pathway coordinately regulates cell proliferation and apoptosis by inactivating Yorkie, the Drosophila homolog of YAP. Cell 122, 421–434. Iida, S., Hirota, T., Morisaki, T., Marumoto, T., Hara, T., Kuninaka, S., Honda, S., Kosai, K., Kawasuji, M., Pallas, D.C., and Saya, H. (2004). Tumor suppressor WARTS ensures genomic integrity by regulating both mitotic progression and G1 tetraploidy checkpoint function. Oncogene 23, 5266–5274. Justice, R.W., Zilian, O., Woods, D.F., Noll, M., and Bryant, P.J. (1995). The Drosophila tumor suppressor gene warts encodes a homolog of human myotonic dystrophy kinase and is required for the control of cell shape and proliferation. Genes Dev. 9, 534–546. Kamikubo, Y., Takaori-Kondo, A., Uchiyama, T., and Hori, T. (2003). Inhibition of cell growth by conditional expression of kpm, a human homologue of Drosophila warts/lats tumor suppressor. J. Biol. Chem. 278, 17609–17614. Kango-Singh, M., Nolo, R., Tao, C., Verstreken, P., Hiesinger, P.R., Bellen, H.J., and Halder, G. (2002). Shar-pei mediates cell proliferation arrest during imaginal disc growth in Drosophila. Development 129, 5719–5730. Lai, Z.C., Wei, X., Shimizu, T., Ramos, E., Rohrbaugh, M., Nikolaidis, N., Ho, L.L., and Li, Y. (2005). Control of cell proliferation and apoptosis by mob as tumor suppressor, mats. Cell 120, 675–685. Lallemand, D., Curto, M., Saotome, I., Giovannini, M., and McClatchey, A.I. (2003). NF2 deficiency promotes tumorigenesis and metastasis by destabilizing adherens junctions. Genes Dev. 17, 1090–1100. McCartney, B.M., Kulikauskas, R.M., LaJeunesse, D.R., and Fehon, R.G. (2000). The neurofibromatosis-2 homologue, Merlin, and the tumor suppressor expanded function together in Drosophila to regulate cell proliferation and differentiation. Development 127, 1315–1324. McClatchey, A.I., and Giovanni, M. (2005). Membrane organization and tumorigenesis—the NF2 tumor suppressor, Merlin. Genes Dev. 19, 2265–2277. McPherson, J.P., Tamblyn, L., Elia, A., Migon, E., Shehabeldin, A., Matysiak-Zablocki, E., Lemmers, B., Salmena, L., Hakem, A., Fish, J., et al.

(2004). Lats2/Kpm is required for embryonic development, proliferation control and genomic integrity. EMBO J. 23, 3677–3688. Mikeladze-Dvali, T., Wernet, M.F., Pistillo, D., Mazzoni, E.O., Teleman, A.A., Chen, Y.W., Cohen, S., and Desplan, C. (2005). The growth regulators warts/lats and melted interact in a bistable loop to specify opposite fates in Drosophila R8 photoreceptors. Cell 122, 775–787. Okada, T., Lopez-Lago, M., and Giancotti, F.G. (2005). Merlin/NF-2 mediates contact inhibition of growth by suppressing recruitment of Rac to the plasma membrane. J. Cell Biol. 171, 361–371. Pantalacci, S., Tapon, N., and Leopold, P. (2003). The Salvador partner Hippo promotes apoptosis and cell-cycle exit in Drosophila. Nat. Cell Biol. 5, 921–927. Perez-Moreno, M., Jamora, C., and Fuchs, E. (2003). Sticky business: Orchestrating cellular signals at adherens junctions. Cell 112, 535–548. Read, R.D., Bach, E.A., and Cagan, R.L. (2004). Drosophila C-terminal Src kinase negatively regulates organ growth and cell proliferation through inhibition of the Src, Jun N-terminal kinase, and STAT pathways. Mol. Cell. Biol. 24, 6676–6689. Shaw, R.J., McClatchey, A.I., and Jacks, T. (1998). Regulation of the neurofibromatosis type 2 tumor suppressor protein, merlin, by adhesion and growth arrest stimuli. J. Biol. Chem. 273, 7757–7764. Shaw, R.J., Paez, J.G., Curto, M., Yaktine, A., Pruitt, W.M., Saotome, I., O’Bryan, J.P., Gupta, V., Ratner, N., Der, C.J., et al. (2001). The Nf2 tumor suppressor, merlin, functions in Rac-dependent signaling. Dev. Cell 1, 63–72. St. John, M.A., Tao, W., Fei, X., Fukumoto, R., Carcangiu, M.L., Brownstein, D.G., Parlow, A.F., McGrath, J., and Xu, T. (1999). Mice deficient of Lats1 develop soft-tissue sarcomas, ovarian tumours and pituitary dysfunction. Nat. Genet. 21, 182–186. Strano, S., Munarriz, E., Rossi, M., Castagnoli, L., Shaul, Y., Sacchi, A., Oren, M., Sudol, M., Cesareni, G., and Blandino, G. (2001). Physical interaction with Yes-associated protein enhances p73 transcriptional activity. J. Biol. Chem. 276, 15164–15173. Sun, C.X., Robb, V.A., and Gutmann, D.H. (2002). Protein 4.1 tumor suppressors: getting a FERM grip on growth regulation. J. Cell Sci. 115, 3991–4000. Takahashi, Y., Miyoshi, Y., Takahata, C., Irahara, N., Taguchi, T., Tamaki, Y., and Noguchi, S. (2005). Down-regulation of LATS1 and LATS2 mRNA

expression by promoter hypermethylation and its association with biologically aggressive phenotype in human breast cancers. Clin. Cancer Res. 11, 1380–1385. Tamaskovic, R., Bichsel, S.J., and Hemmings, B.A. (2003). NDR family of AGC kinases—essential regulators of the cell cycle and morphogenesis. FEBS Lett. 546, 73–80. Tao, W., Zhang, S., Turenchalk, G.S., Stewart, R.A., St. John, M.A., Chen, W., and Xu, T. (1999). Human homologue of the Drosophila melanogaster lats tumour suppressor modulates CDC2 activity. Nat. Genet. 21, 177–181. Tapon, N., Harvey, K.F., Bell, D.W., Wahrer, D.C., Schiripo, T.A., Haber, D.A., and Hariharan, I.K. (2002). salvador promotes both cell cycle exit and apoptosis in Drosophila and is mutated in human cancer cell lines. Cell 110, 467–478. Udan, R.S., Kango-Singh, M., Nolo, R., Tao, C., and Halder, G. (2003). Hippo promotes proliferation arrest and apoptosis in the Salvador/Warts pathway. Nat. Cell Biol. 5, 914–920. Wu, S., Huang, J., Dong, J., and Pan, D. (2003). hippo encodes a Ste-20 family protein kinase that restricts cell proliferation and promotes apoptosis in conjunction with salvador and warts. Cell 114, 445–456. Xia, H., Qi, H., Li, Y., Pei, J., Barton, J., Blackstad, M., Xu, T., and Tao, W. (2002). LATS1 tumor suppressor regulates G2/M transition and apoptosis. Oncogene 21, 1233–1241. Xu, T., and Rubin, G.M. (1993). Analysis of genetic mosaics in developing and adult Drosophila tissues. Development 117, 1223–1237. Xu, T.A., Wang, W.Y., Zhang, S., Stewart, R.A., and Yu, W. (1995). Identifying tumor suppressors in genetic mosaics—the Drosophila lats gene encodes a putative protein-kinase. Development 121, 1053–1063. Yagi, R., Chen, L.F., Shigesada, K., Murakami, Y., and Ito, Y. (1999). A WW domain-containing yes-associated protein (YAP) is a novel transcriptional co-activator. EMBO J. 18, 2551–2562. Yang, X., Li, D.M., Chen, W., and Xu, T. (2001). Human homologue of Drosophila lats, LATS1, negatively regulates growth by inducing G(2)/M arrest or apoptosis. Oncogene 20, 6516–6523. Yang, X., Yu, K., Hao, Y., Li, D.M., Stewart, R., Insogna, K.L., and Xu, T. (2004). LATS1 tumour suppressor affects cytokinesis by inhibiting LIMK1. Nat. Cell Biol. 6, 609–617.

Cell 124, January 27, 2006 ©2006 Elsevier Inc.  273