Cell competition is a short-range communication originally observed in Drosophila. Relatively little is known about cell competition in mammals or in non-epithelial cells. Hippo signaling and its downstream transcription factors of the Tead family, control cell proliferation and apoptosis. Here, we established an in vitro model system that shows cell competition in mouse NIH3T3 embryo fibroblast cells. Co-culture of Tead-activity-manipulated cells with normal (wild-type) cells caused cell competition. Cells with reduced Tead activity became losers, whereas cells with increased Tead activity became super-competitors. Tead directly regulated Myc RNA expression, and cells with increased Myc expression also became super-competitors. At low cell density, cell proliferation required both Tead activity and Myc. At high cell density, however, reduction of either Tead activity or Myc was compensated for by an increase in the other, and this increase was sufficient to confer ‘winner’ activity. Collectively, NIH3T3 cells have cell competition mechanisms similar to those regulated by Yki and Myc in Drosophila. Establishment of this in vitro model system should be useful for analyses of the mechanisms of cell competition in mammals and in fibroblasts.
Communication among cells plays important roles in the regulation of embryonic development and maintenance of tissue homeostasis. Among the various types of communication, cell competition is a unique type of short-range communication originally identified in Drosophila. Cells compare their fitness level with those of neighboring cells, and cells with relatively higher fitness become winners, and cells with relatively lower fitness are eliminated as losers (for reviews, see Baker, 2011; Levayer and Moreno, 2013; Vincent et al., 2013; Amoyel and Bach, 2014).
In Drosophila imaginal discs, several genes affecting cell proliferation cause cell competition. Minute mutants, which have mutations in ribosomal genes, were originally reported to show cell competition (Morata and Ripoll, 1975; Simpson, 1979). Subsequently, Myc involvement was identified (de la Cova et al., 2004; Moreno and Basler, 2004). Cells with additional copies of Myc become ‘super-competitors’ and can eliminate neighboring wild-type cells. More recently, another growth regulator, Hippo, was found to be involved in cell competition (Neto-Silva et al., 2010; Ziosi et al., 2010). Hippo signaling suppresses cell proliferation by repressing nuclear localization of the coactivator protein Yorkie (Yki) (for reviews, see Pan, 2010; Halder and Johnson, 2011; Yu and Guan, 2013). Yki-overexpressing cells become super-competitors and Yki mutant cells become losers. Together, Yki and the transcription factor Scalloped (Sd) induce Myc, and Yki–Sd and Myc cooperate in growth regulation (Neto-Silva et al., 2010).
Regulators of apico-basal cell polarity also play important roles in cell competition-like selection of cells. Cells mutant for the tumor suppressor genes scribble (scrib), lethal giant larvae (lgl), and discs large (dlg) lose cell polarity, and these cells are eliminated by surrounding normal epithelial cells (Brumby and Richardson, 2003; Grzeschik et al., 2007; Ohsawa et al., 2011). This differs slightly from cell competition as described above, and is sometimes referred to as intrinsic tumor suppression (Igaki, 2009). Defects in cell polarity regulate the Hippo pathway (Chen et al., 2010; Grzeschik et al., 2010; Ling et al., 2010; Robinson et al., 2010), which is a major inducer of selection. An Lgl-interacting protein, Mahjong (Mahj; also known as VprBP in mammals), is also a downstream effector of selection (Tamori et al., 2010).
Although Myc-regulated cell competition has been most studied, other mechanisms exist. Steep differences in Wingless (Wg) signaling trigger cell competition independently of Myc (Vincent et al., 2011). Differences in the activity of the transcription factor STAT also induce cell competition independently of Myc, Yki and Wg (Rodrigues et al., 2012).
Cell competition is also present in mammals. Cells from the mouse Minute mutant Belly spot and tail (Bst), which contains a mutation in the ribosomal protein gene Rpl24 (Oliver et al., 2004), became losers when chimeric mice were created with wild-type embryonic stem cells. In MDCK epithelial cells, cells knocked down for the polarity regulators Mahj (Tamori et al., 2010) or Scribble (Norman et al., 2012), or expressing active Ras (Hogan et al., 2009) or Src (Kajita et al., 2010) are selectively eliminated through intrinsic tumor suppression. In mouse hematopoietic stem and progenitor cells, p53-mediated cell competition selects the least damaged cells (Bondar and Medzhitov, 2010). More recently, an involvement of Myc in cell competition has been identified. In the epiblast of early post-implantation stage mouse embryos and in embryonic stem cells, Myc-driven cell competition eliminates unfit cells at the onset of differentiation (Clavería et al., 2013; Sancho et al., 2013). In spite of recent progress, data on cell competition in mammals remains limited. Additionally, because cell competition has primarily been studied in epithelial cells, it is unknown whether cell competition also takes place in fibroblasts.
To improve our understanding of mammalian cell competition, a simple in vitro assay system is required. We developed an in vitro assay system for cell competition using the mouse embryo fibroblast cell line NIH3T3. We focused on Tead family transcription factors (mouse homologues of Sd) because the roles of Hippo signaling, which controls transcriptional activity of Tead, in mammalian cell competition is unknown. In mammals, Hippo signaling regulates subcellular localization of two coactivator proteins, Yap1 and Wwtr1/Taz (Yki homologues, all referred to as Yap hereafter) (Zhao et al., 2007; Lei et al., 2008; Ota and Sasaki, 2008). Yap binds to Tead proteins to control cellular responses. Manipulation of Tead activity in NIH3T3 cells alters cell proliferation, apoptosis and confluent cell density (Ota and Sasaki, 2008).
In this study, we found that co-culture of Tead-activity-manipulated and normal NIH3T3 cells caused cell competition. The Tead proteins regulated Myc expression, and Tead activity and Myc cooperatively controlled cell proliferation and cell competition. The observed cell competition resembled cell competition controlled by Yki–Sd and Myc in Drosophila imaginal discs. The results suggest that cell competition is also present in embryonic fibroblasts, and that this simple in vitro system will facilitate analysis of the molecular mechanisms of mammalian cell competition.
Co-culture of cells with different Tead activities causes cell-competition-like behavior
In NIH3T3 cells, modulation of Tead activity alters cell proliferation (Ota and Sasaki, 2008). Reproducing previous results, when cells with increased Tead activity (TeadVP cells), which express a fusion protein of the DNA-binding domain of Tead2 and the transcription activation domain of HSV VP16 protein (Tead2VP16), were cultured alone (designated ‘single cultures’), the cells continued to proliferate beyond the normal confluent density and showed reduced proliferation at a higher density (Fig. 1A). In contrast, cells with decreased Tead activity (TeadEnR cells), which express a fusion protein of the TEA domain of Tead2 and the repression domain of Drosophila Engrailed (Tead2EnR), showed a reduced proliferation rate and their proliferation ceased at a lower saturation density (Fig. 1A). We confirmed that expression of a Tead reporter, 8×GT-IIC-Luc (Ota and Sasaki, 2008), as well as the Tead–Yap target genes Ankrd1 and/or Cyr61 (Zhao et al., 2008; Dupont et al., 2011), in these cells were modified, consistent with altered Tead activity (Fig. 1B–D).
To examine whether cells with different Tead activities communicate and regulate each other's behavior, we co-cultured Tead-manipulated NIH3T3 cells, marked by GFP expression, with normal (wild-type) NIH3T3 cells at a ratio of 2∶3 (Fig. 1E). The behavior of Tead-manipulated cells in co-culture clearly differed from the behavior of cells in single culture. Proliferation of TeadEnR cells was strongly suppressed, and the number of TeadEnR cells barely increased (Fig. 1G). In contrast, proliferation of normal cells increased (Fig. 1H).
To clarify the effects of co-culture on the growth of TeadEnR cells, chronological changes in relative cell numbers (cell numbers relative to the cell number on day 0), or the ‘normalized growth curves’, were compared between single cultures and co-cultures. In this paper, the word ‘growth’ is used to indicate chronological changes in cell numbers, which is the sum of cell proliferation and cell death (or elimination by other means). For control GFP-expressing cells, the normalized growth curves of single cultures and co-cultures were identical (Fig. 1I). Co-cultured TeadEnR cells showed a slight reduction in growth at day 4 and strong downregulation thereafter (Fig. 1K). In contrast, co-cultured normal cells showed an increased growth from day 4 (Fig. 1L). Calculation of the growth rates (Sancho et al., 2013) on days 4 and 12 showed that, in co-culture, the growth rates of TeadEnR and normal cells significantly decreased and increased, respectively (supplementary material Fig. S1C,D). Thus, co-culture suppressed the growth of TeadEnR cells but increased the growth of normal cells from day 4.
In the co-culture of TeadVP cells, the number of TeadVP cells increased at a constant rate up to day 14 (Fig. 1G), whereas the number of normal cells decreased after day 6, when the total cell number exceeded the normal confluent density (Fig. 1F,H). The growth of TeadVP cells and normal cells were unchanged between single cultures and co-cultures up to day 6. Thereafter, in co-cultures, the growth rates of TeadVP cells and normal cells significantly increased and decreased, respectively (Fig. 1J,L; supplementary material Fig. S1A,B).
The behavior of co-cultured cells resembles the cell competition observed in Drosophila. Cells with lower Tead activity behaved as losers, whereas cells with higher Tead activity behaved as winners. Cell densities at the onset of cell competition differed; cell competition in TeadEnR and TeadVP cells started at low and normal confluent densities, respectively. Although cell competition has been reported mostly in epithelial cells, our results suggest that similar intercellular interactions also operate in embryonic fibroblasts.
In Drosophila, cell competition is not generally caused by cells with different cell proliferation properties. For example, cells that express a constitutively active form of phosphoinositide 3-phosphate kinase (PI3K) have enhanced cell proliferation but do not cause cell competition (de la Cova et al., 2004; Senoo-Matsuda and Johnston, 2007). Therefore, we examined whether cells that express a myristoylated active form of the catalytic p110α subunit of PI3K (Myr-p110 cells, Fig. 1M) cause cell competition. In single culture, Myr-p110 cells continued cell proliferation beyond the normal confluent density, indicating enhanced cell proliferation (Fig. 1N). In co-cultures with normal cells, proliferation of Myr-p110 cells continued after confluence, and the total cell number increased (Fig. 1O,P). In contrast, the number of co-cultured normal cells remained constant once the total cell number reached confluence (Fig. 1P). Therefore, a reduction in the number of normal cells after confluence is not generally caused by cells with enhanced cell proliferation, but is specific to TeadVP cells.
Cells with lower Tead activities are preferentially eliminated by apoptosis
We next asked whether the observed cell competition involved apoptosis. In co-cultures of GFP-expressing and normal cells (control co-cultures), the ratio of apoptotic cells, marked by cleaved caspase-3, increased after day 5 (Fig. 2A,D). This result suggests that proliferation of NIH3T3 cells did not completely cease even after confluence and that a constant cell number was maintained by eliminating the same number of cells through apoptosis. In co-cultures of TeadVP and normal cells, normal cells showed increased apoptosis compared with TeadVP cells at day 12 (Fig. 2B,E). In co-cultures of TeadEnR and normal cells, TeadEnR cells showed clearly higher apoptosis from day 2 onward, and the frequency increased over time (Fig. 2C,F). Therefore, the observed cell competition involved increased apoptosis of cells with relatively lower Tead activities. Live imaging of low-cell-density co-cultures revealed that some TeadEnR cells collapsed and detached from the culture dishes after contact with normal cells (Fig. 3A; supplementary material Movies S1, S2). TeadEnR cells tended to collapse after having either a prolonged or a large area of contact with normal cells (Fig. 3B,C), raising the possibility that cell–cell contact plays a role in apoptosis of TeadEnR cells at low densities.
To further examine the role of apoptosis in our co-culture system, we treated cells with the pan-caspase inhibitor N-Benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethyl ketone [Z-VAD(OMe)FMK]. Consistent with the hypothesis that normal NIH3T3 cells are slowly eliminated through apoptosis at confluence, treatment of control co-cultures with the caspase inhibitor after confluence (day 6 to day 9) significantly increased the total cell number (Fig. 2G). Treatment of co-cultured TeadVP and normal cells with a caspase inhibitor significantly suppressed the reduction in the relative number of normal cells and also slightly increased TeadVP cells (Fig. 2H). Similarly, treatment of co-cultured TeadEnR and normal cells with caspase inhibitor between day 4 and day 7 significantly suppressed the reduction of TeadEnR cells and increased the relative number of normal cells (Fig. 2I). These results suggest that both loser and winner cells undergo apoptosis at high and low frequencies, respectively, and that reduction of loser cells depends on apoptosis. Because inhibitor treatment suppressed apoptosis of both loser and winner cells, it remains unknown whether or not apoptosis of loser cells has any effect on the increase in winner cell.
Tead2-overexpressing cells become winners without causing overgrowth
During the analyses, we noticed that Tead2-overexpressing cells (Tead cells) also became weak winners. In single cultures, Tead and normal cells showed essentially identical growth (Fig. 4A). In co-cultures, the total cell number stayed constant once cells reached confluence (Fig. 4B). However, Tead cells continued to grow past confluence at a slower rate and normal cells decreased accordingly (Fig. 4C,D). Comparison of normalized growth curves and growth rates between single and co-cultures showed that a significant difference appeared after cells reached confluence, and Tead cells replaced normal cells without causing overgrowth (Fig. 4E,F; supplementary material Fig. S1E,F).
The reduction in the relative number of normal cells was accompanied by increased apoptosis, indicated by the presence of cleaved caspase-3 in cells (Fig. 4G,H), and caspase inhibitor treatment suppressed reduction (Fig. 4I). Normal cells adjacent to clusters of Tead cells tended to show apoptosis (Fig. 4G). Several lines of evidences suggest that Tead cells have increased Tead activity. Expression of a Tead reporter gene, 8×GT-IIC-Luc, and the Tead–Yap target gene Cyr61, were moderately increased in Tead cells (Fig. 1B,D). In Tead cells, Yap showed stronger nuclear localization than in control GFP-expressing cells in single cultures (Fig. 4J–M), as has been shown for epithelial cells (Diepenbruck et al., 2014). In co-cultures, Tead cells also showed significantly stronger nuclear Yap signals than co-cultured normal cells (Fig. 4K,N). Therefore, Tead cells have higher Tead activity than normal cells and higher Tead activity makes cells winners.
Tead activity determines competitive activity of cells
For consistency, we primarily used modified Tead2 at a 2∶3 ratio (Tead∶normal cells) for co-cultures throughout the study. However, these conditions were not important, as cells expressing Tead1 (Tead1VP16, Tead1EnR or Tead1) showed essentially the same behaviors as Tead2-expressing cells in co-cultures (Fig. 5A–C; supplementary material Fig. S1G–J). Furthermore, Tead cells behaved as winners irrespective of co-culture ratio (1∶9, 1∶1 and 3∶2) (Fig. 5D–F). The involvement of Tead was also shown by knocking down Tead1 with short hairpin RNAs (shRNAs). Tead1-knockdown cells (shTead1 cells) showed reduced expression of Tead1 mRNA and Tead1 proteins (Fig. 5G, supplementary material Fig. S2A) and reduced proliferation in single cultures (Fig. 5H). When co-cultured with normal cells, shTead1 cells became losers and showed a reduced growth rate from low cell density (Figs 5I–L; supplementary material Fig. S2B,C) that was reminiscent of TeadEnR cells. Similar to TeadEnR cells, shTead1 cells showed a reduced expression of the target gene Ankrd1 (Fig. 1C). Nuclear accumulation of Yap was also reduced in shTead1 cells (supplementary material Fig. S2D). These results suggest that shTead1 cells have reduced Tead activity, which makes cells losers. Thus, in this co-culture system, activity of Tead proteins, but not the co-culture ratio or type of Tead proteins, determines the competitive activity of the cells.
Tead regulates Myc
The presence of cells with different Tead activities caused apoptosis of cells with lower Tead activity. Although Tead–Yap is important for resistance to apoptotic stresses and/or anoikis (Overholtzer et al., 2006; Zhao et al., 2007; Ota and Sasaki, 2008; Zhao et al., 2012), it is unknown whether reduction of Tead activity directly causes cell death. Therefore, we hypothesized that other factors might also be involved in the regulation of Tead-triggered cell competition. We focused on the proto-oncogene Myc as a candidate, because Myc has been shown to be an important factor for controlling cell competition in Drosophila (de la Cova et al., 2004; Moreno and Basler, 2004) and because expression of Myc is regulated by Yki–Sd (Neto-Silva et al., 2010; Ziosi et al., 2010).
We first asked whether Myc is regulated by Tead proteins in NIH3T3 cells. In normal cells, the expression of Myc mRNA depended on cell density: Myc expression was high at a low cell density, at day 2, but low at a high cell density, at days 5 and 10 (Fig. 6A, normal). In single cultures, TeadVP enhanced Myc expression irrespective of cell density, whereas TeadEnR suppressed Myc expression at a low cell density (Fig. 6A). Examination of the genomic sequence within and surrounding the Myc gene using the JASPAR CORE database identified a cluster of seven potential Tead-binding sites in the first intron of Myc (Fig. 6B; supplementary material Fig. S3A). To examine whether Tead–Yap binds to any of these sites, we performed a chromatin immunoprecipitation (ChIP) assay by using anti-Tead1 and anti-Yap1 antibodies. At low cell density, interactions with both Tead1 and Yap1 were observed with all seven sites, although these interactions were weaker than that of Cyr61 (Fig. 6C). At high cell density, the interaction with Tead1 was unchanged, but the interaction with Yap1 was reduced for all sites (Fig. 6C). These results are consistent with the hypothesis that cell-density-dependent Hippo signaling directly controls Myc mRNA expression by regulating the formation of the Tead–Yap complex.
At the protein level, TeadVP and Tead cells showed ∼20% higher levels of Myc protein, whereas the Myc protein level in TeadEnR and shTead1 cells was only 35% and 53% of that in control cells, respectively, at day 2 (Fig. 6D–G). However, such clear differences were not observed at day 5 (Fig. 6D–G). These results suggest that Tead controls Myc expression at low cell density in single cultures, but post-transcriptional regulation might also operate for Myc proteins at high cell density.
Winner cells commonly express higher levels of Myc proteins
To further examine the possible involvement of Myc in our experimental system, we next examined expression of Myc proteins under various competitive high-density co-culture conditions at day 5 by immunostaining of Myc proteins. In control (non-competitive) co-cultures, average Myc expression levels were identical in normal and control GFP-expressing cells (Fig. 6H,M). In co-cultures of normal cells with cells that had high Tead activity (TeadVP and Tead), Myc levels were significantly higher in the cells with the high Tead activity than in normal cells (Fig. 6I,K,M). In contrast, in co-cultures of normal cells with cells that had low Tead activity (TeadEnR and shTead1), Myc levels were significantly higher in the normal cells (Fig. 6J,L,M). Such clear differences in Myc levels at high density were not observed in single cultures (Fig. 6D–G), suggesting that interactions between cells with different Tead activities induced stronger expression of Myc in winner (i.e. relatively higher Tead activity) cells. In co-cultures with the cells with high Tead activity, some normal cells adjacent to the high Tead activity cells also showed strong Myc signals (Fig. 6I,K, arrowheads). At day 10, this phenomenon was also observed in co-cultures of Tead cells, and differences in relative Myc levels between Tead and normal cells were reduced (supplementary material Fig. S3C, arrowheads, Fig. S3D). In co-cultures of TeadVP cells, a strong Myc signal was restricted to TeadVP cells, and the difference in relative Myc levels between TeadVP and normal cells was enhanced (supplementary material Fig. S3B,D). Strong Myc signals in winner cells were mostly restricted to cells adjacent to loser (i.e. relatively lower Tead activity) cells (e.g. supplementary material Fig. S3E), which might indicate intercellular communication. Taken together, under conditions of Tead-triggered cell competition, interactions between cells with different Tead activities promoted higher levels of Myc proteins in winner cells, which raises the possibility that Tead-triggered competition involves Myc.
Cells with increased Myc expression become winners
To examine directly the role of Myc in cell competition in NIH3T3 cells, we used cells expressing exogenous Myc (Myc cells) (Fig. 7A). In single cultures, Myc cells showed an enhanced growth and proliferated beyond the normal confluent density (Fig. 7B). When co-cultured with normal cells, Myc cells became winners once they exceeded confluent density (Fig. 7C–E). Comparison of normalized growth curves between single cultures and co-cultures revealed that the onset of cell competition (i.e. reduction of normal cells) in Myc cell co-cultures was later, i.e. at a higher density, than that of TeadVP cell co-cultures (Fig. 7F–H). This difference likely reflects the smaller size of Myc cells (data not shown). These behaviors are similar to those of TeadVP cells, except in timing.
To examine further the roles of Myc in cell competition, we next suppressed endogenous Myc by expressing shRNAs. We only obtained weak Myc knockdown cells (low-Myc or shMyc cells), which showed ∼30–40% reduction of Myc protein expression (Fig. 7I,J), likely reflecting the essential role of Myc in proliferation. In single cultures, the growth of low-Myc cells was slower than that of normal cells, and low-Myc cells stopped proliferating at a lower density (Fig. 7K). When co-cultured with normal cells, low-Myc cells showed slow growth, but maintained cell numbers after confluence (Fig. 7L–N). Therefore, although overexpression of Myc made cells winners or super-competitors as in Drosophila, moderate reduction of Myc did not make cells losers.
Tead activity and Myc cooperate in cell proliferation and cell competition
Because both Tead activity and Myc play important roles in regulation of cell proliferation and cell competition, we next evaluated whether they play distinct roles. If Tead and Myc play the same roles, it would be expected that a reduction in the levels of Myc would be rescued by increased Tead activity. For this purpose, we expressed TeadVP in shMyc cells (TeadVP+shMyc cells). Although TeadVP cells showed increased Myc protein levels, TeadVP+shMyc cells maintained reduced Myc proteins similar to shMyc cells (Fig. 8A,B). In single cultures, TeadVP+shMyc cells initially showed slow growth, similar to shMyc cells, indicating that increased Tead activity did not rescue slow growth resulting from low Myc levels at a low cell density (Fig. 8C). Nevertheless, TeadVP+shMyc cells continued proliferating beyond normal confluent density, similar to TeadVP cells (Fig. 8C). When co-cultured with normal cells, TeadVP+shMyc cells showed slow growth, resembling shMyc cells at low cell density. However, at high cell density, TeadVP+shMyc cells continued proliferating, and the number of co-cultured normal cells decreased (Fig. 8D–F). In co-culture, the growth rates of TeadVP+shMyc cells and normal cells significantly increased and decreased, respectively, between days 4 and 12 (supplementary material Fig. S1K–N). Therefore, the effects of TeadVP in shMyc cells differ between low and high cell densities.
As an opposing experiment, we next evaluated whether reduced Tead activity could be compensated by increased Myc expression. We expressed Myc in TeadEnR cells (TeadEnR+Myc cells). At day 2, TeadEnR cells showed lower Myc expression, and TeadEnR+Myc cells showed Myc levels similar to those in normal low-density cells (Fig. 8G,H). In single cultures, TeadEnR+Myc cells initially showed slow growth, similar to TeadEnR cells (Fig. 8I), indicating that recovery of Myc expression did not rescue slow growth resulting from reduced Tead activity at low cell density. Nevertheless, at high cell density, TeadEnR+Myc cells continued to proliferate beyond normal confluent density, partially mimicking Myc cells (Fig. 8I). When co-cultured with normal cells, similar to TeadEnR cells, the growth rate of TeadEnR+Myc cells was significantly reduced at low density (Fig. 8K; supplementary material Fig. S1Q–R). However, unlike TeadEnR cells, TeadEnR+Myc cells slowly recovered proliferation once the total cell number reached confluent density; normal cells were then eliminated (Fig. 8J–L; supplementary material Fig. S1U,V). Elimination of normal cells at confluent density was also observed in co-culture with Myc-expressing cells (DsRed+Myc cells) (supplementary material Fig. S1T). Thus, in co-culture, although Myc did not increase the growth rates of TeadEnR cells at low cell density, Myc did prevent elimination of TeadEnR cells and rendered the cells winners at a high cell density.
To examine the role of apoptosis in Myc-triggered cell competition, we treated cells with Z-VAD(OMe)FMK. Treatment of co-cultured DsRed+Myc (essentially the same as Myc cells) and normal cells with caspase inhibitor between day 9 and day 12 significantly suppressed the reduction of normal cells and increased DsRed+Myc cells (supplementary material Fig. S4A), suggesting that the reduction of normal cells in Myc-triggered cell competition involves apoptosis. Treatment of co-cultured TeadEnR+Myc and normal cells with caspase inhibitor between day 9 and day 12 also significantly suppressed the reduction of normal cells, but TeadEnR+Myc cells were not affected (supplementary material Fig. S4B). This result suggests that, at high cell densities, expression of Myc not only prevents apoptosis of TeadEnR cells but also confers winner activity, which induces apoptosis of normal cells.
Collectively, these results suggest that both Tead activity and Myc cooperatively control cell proliferation and cell competition, but their requirements differ depending on cell densities. At low density, both Tead activity and Myc are required and reduction of either one restricts proliferation. At high cell density, an increase of either one is sufficient to promote cell proliferation and confer winner activity.
In vitro model of cell competition in the mouse embryonic fibroblast cell line NIH3T3
In this study, we established an in vitro model system for cell competition using the mouse embryonic fibroblast cell line NIH3T3. Most previous studies on cell competition have been performed in the context of epithelial cells. More recently, cell competition has also been identified in non-epithelial cell types, including Drosophila macrophage-like S2 cells, mouse embryonic stem cells, and hematopoietic stem or progenitor cells. Our identification of cell competition in mouse embryonic fibroblasts suggests that cell competition is a widely used type of intercellular communication that is also present in fibroblasts.
Multiple mechanisms regulate cell competition. We identified cell competition in NIH3T3, similar to that induced by Yki (Hippo) and Myc in other experimental systems in the following ways: (1) elimination of loser cells involves apoptosis (Moreno et al., 2002; Senoo-Matsuda and Johnston, 2007; Tyler et al., 2007; Clavería et al., 2013; Sancho et al., 2013), (2) differences in Tead activity induces cell competition (Neto-Silva et al., 2010; Ziosi et al., 2010), (3) expression of exogenous Myc makes cells super-competitors (de la Cova et al., 2004; Moreno and Basler, 2004; Senoo-Matsuda and Johnston, 2007; Clavería et al., 2013; Sancho et al., 2013), and (4) Tead transcriptionally regulates Myc, and Tead activity and Myc cooperatively control cell proliferation and cell competition (Neto-Silva et al., 2010).
Cell competition in NIH3T3 cells also differed from other systems. As is common in established cell lines, NIH3T3 cells do not completely cease proliferation at confluence. As a result, apoptosis occurs in both winner and loser cells, and the differing frequencies of apoptosis determine which cells are winners and which are losers. In Drosophila imaginal discs, apoptosis takes place only in loser cells, and the enhanced proliferation of winner cells, known as compensatory proliferation, depends on apoptosis of loser cells (reviewed in Morata et al., 2011; Ryoo and Bergmann, 2012). The roles of apoptosis in cell competition can vary with the experimental system. Indeed, cell competition in mouse hematopoietic stem and progenitor cells does not involve apoptosis (Bondar and Medzhitov, 2010). In spite of such differences, elimination of normal cells in co-cultures was specifically induced by cells with increased Tead activities and Myc levels, and not by cells with enhanced cell proliferation (Myr-p110 cells). This resembles cell competition in Drosophila (de la Cova et al., 2004; Senoo-Matsuda and Johnston, 2007).
Although an in vitro mammalian cell competition system has been established using the epithelial MDCK cell line (Hogan et al., 2009; Kajita et al., 2010), competition is induced by the interaction between transformed and normal cells, and it remains unknown whether differences in Tead activity and/or Myc also trigger cell competition in this system. Therefore, NIH3T3 cells should be useful for studying molecular mechanisms of cell competition controlled by Tead (the Hippo signal) and Myc in mammalian fibroblasts. To the best of our knowledge, this is the first demonstration of cell competition in fibroblasts and of Tead-triggered cell competition in mammals. In Drosophila, differences in Wingless signaling (Vincent et al., 2011) and STAT activity levels (Rodrigues et al., 2012) induce cell competition independently of Myc and Yki. Examining whether these mechanisms are also present in NIH3T3 cells is of future interest.
Tead and Myc cooperatively control cell proliferation and cell competition
In NIH3T3 cells, Tead–Yap bound to Myc and controlled expression of Myc RNA, indicating that Tead controls Myc at the transcriptional level. Drosophila Myc is also transcriptionally regulated by Sd–Yki, and this transcriptional regulation establishes the feedback mechanism between Myc and Yki (Neto-Silva et al., 2010; Ziosi et al., 2010). In NIH3T3 cells, regulation of Myc proteins in competitive co-cultures appeared to be more complex. In high-density single cultures, cells showed similar low levels of Myc proteins, irrespective of Tead activity. However, in competitive co-cultures, winner cells (with relatively higher Tead activity) flanked by loser cells (with relatively lower Tead activity) showed higher levels of Myc proteins. Such local increase in Myc proteins at the interface cannot be explained by simple transcriptional regulation by Tead. It is likely that local communication between cells with different Tead activity also controlled the amount of Myc proteins. Indeed, similar upregulation of Myc proteins in winner cells has also been reported in co-cultures of normal and mutant embryonic stem cells (Sancho et al., 2013). Because Myc protein stability is controlled by multiple signaling pathways (Hann, 2006; Vervoorts et al., 2006), it is likely that such post-transcriptional regulation is also involved in Myc regulation.
In NIH3T3 cells, cells with increased Myc expression became winners and cells with reduced Myc expression were not eliminated by co-cultured normal cells, although the low-Myc cells showed reduced cell proliferation. This differs from the roles of Myc in Drosophila imaginal discs and mouse epiblasts, in which increased and decreased Myc make cells winners and losers, respectively (Neto-Silva et al., 2010; Clavería et al., 2013). Although our results were reproducible, we could not obtain cells with strong knockdown of Myc, probably because of the crucial role of Myc in cell proliferation. Therefore, we cannot completely exclude the possibility that further reduction of Myc would make cells losers.
In NIH3T3 cells, the roles of Tead activity and Myc in regulation of cell proliferation and cell competition depend on cell density. At low cell density, both Tead activity and Myc are required, and the reduction in either protein could not be compensated for by an increase in the other. Similar interdependence is also present in Yki and Myc in Drosophila (Neto-Silva et al., 2010). In contrast, at high cell density, an increase of either protein was sufficient to support cell proliferation and to make cells winners. Although the mechanisms underlying such density-dependent responses remain unknown, the mechanisms responsible for density-dependent control of Tead activity and Myc proteins (Ota and Sasaki, 2008; Zhao et al., 2008; Wada et al., 2011; Aragona et al., 2013; Mori et al., 2014) are likely to be involved.
An unexpected observation from the current experiments was that the expression of Myc in some normal cells adjacent to cells with high Tead activity in co-cultures. Although the mechanism that induces Myc expression in these cells is unknown, expression of Myc probably suppressed apoptosis of normal cells (or cells with lower Tead activity), as shown for TeadEnR+Myc cells. Such mechanisms are likely to prevent unnecessary cell competition when differences in Tead activity are small.
In conclusion, we established an in vitro model system for cell competition using the mouse embryonic fibroblast cell line NIH3T3. Analysis revealed that NIH3T3 cells possess cell competition mechanisms similar to those regulated by Yki (the Hippo signal) and Myc in Drosophila. Establishment of such a simple model system should facilitate analyses of the mechanisms of cell competition in mammals and non-epithelial cells.
MATERIALS AND METHODS
NIH3T3 cells (Jainchill et al., 1969) were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Tead-manipulated cells were prepared as described previously (Ota and Sasaki, 2008). Cells expressing Myc or Myr-p110 were prepared in a similar manner by generating retroviral vectors containing the respective cDNAs introduced into pMYs-IRES-EGFP (Kitamura et al., 2003) or pMYs-IRES-DsRedT4 (newly constructed). Double-infected cells were isolated using a FACSAria cell sorter (BD Biosciences) as EGFP and DsRedT4 double-positive cells. pLKO.1-GFP lentiviral vector expressing shRNA was infected to generate knockdown cells. For all assays, 5×104 or 2×104 cells were seeded onto 35-mm dishes or LAB-TEK II chamber slides coated with 0.1% gelatin. The medium was changed every other day at low density (up to day 4) and every day at high density (after day 5).
Counting of cell numbers for growth curve analyses was performed using either a hemocytometer or a Tali image cytometer (Life Technologies). Growth curves of Myr-p110 cells (Fig. 1M–P), TeadVP+shMyc cells (Fig. 7C–F), TeadEnR+Myc cells (Fig. 7I–L), and cells after caspase inhibitor treatment (Fig. 2G–I, Fig. 3I) were determined using a Tali image cytometer. Others were determined by using a hemocytometer.
All cell culture experiments described in this paper were performed independently at least twice to confirm reproducibility, and the results of one representative set of experiments are shown. For each experiment, the values from two or three independent samples were averaged; they are presented as mean±s.e.m.
Caspase inhibitor treatment
Cells were treated with 50 µM Z-VAD(OMe)FMK (Merk) between days 4 and 7 (co-culture of TeadEnR with normal cells), days 6 and 9 (co-culture of TeadVP or Tead with normal cells), or days 9 and 12 (co-culture of DsRed+Myc or TeadEnR+Myc with normal cells).
Luciferase assays of 8×GT-IIC-Luc transfected cells were performed as previously described (Ota and Sasaki, 2008).
Immunofluorescence staining of cultured cells
Immunostaining of cells was performed as previously described (Ota and Sasaki, 2008). The following primary antibodies were used at the indicated dilutions: mouse anti-Yap1 monoclonal antibody (Abnova; 1∶500); rabbit anti-cleaved caspase-3 polyclonal antibody (Asp 175) (Cell Signaling Technology; 1∶800); rat anti-GFP monoclonal antibody (Nacalai Tesque; 1∶500); and rabbit anti-c-Myc (Y69) monoclonal antibody (Abcam, 1∶500). Nuclei were counterstained with Hoechst 33258 (Dojindo).
Quantification of fluorescent signals
For quantification of cells positive for cleaved caspase-3, more than 200 cells were analyzed for each sample. Quantification of nuclear and cytoplasmic Yap signals and nuclear Myc signals was performed using ImageJ (NIH) and/or MetaMorph (Molecular Devices) software. More than 50 and 20 cells were analyzed for Myc and Yap signals, respectively.
Total RNA was isolated from the cells by using TRIzol (Invitrogen) and was used for cDNA synthesis by using PrimeScript RT Master Mix (Takara Bio, Japan), according to the manufacturer's instructions. Quantitative RT-PCR (qRT-PCR) analyses for the indicated genes were performed with SYBR Premix Ex TaqII (Takara Bio, Japan) and a PCR 7500 Fast Real-Time PCR system (Applied Biosystems) following the manufacturer's instructions. The following primer sets were used for qRT-PCR: TBP, 5′-GTGATGTGAAGTTCCCTATAAGG-3′ and 5′-CTACTGAACTGCTGGTGGGTCA-3′; Myc, 5′-AGTGCTGCATGAGGAGACAC-3′ and 5′-CTTCTCCACAGACACCACATC-3′; Ankrd1, 5′-TACTGAGAGTAGAGGAGCTG-3′ and 5′-TTGGCCGGAAGTGTCTTCAGGT-3′; Tead1, 5′-CAAGACGTCAAGCCCTTTGTG-3′ and 5′-GAGAATTCCACCAGGCGAAG-3′; Cyr61, 5′-AAGAGGCTTCCTGTCTTTGGC-3′ and 5′-AGACGTGGTCTGAACGATGC-3′.
The qRT-PCR experiments were performed in duplicate or triplicate. The Ct values for each gene amplification were normalized by subtracting the Ct value calculated for TBP. To quantify Myc expression in Myc-overexpressing cells, DNase-I-treated RNA was used.
Lentiviruses used for knockdown
The pLKO.1-GFP lentiviral vector was generated by replacing the puromycin resistance gene of the pLKO.1 vector with EGFP. The pLKO.1-GFP lentiviral vectors harboring the following shRNAs were used to generate knockdown cells: shTead1 #1-sense, 5′-CCGGGATCAACTTCATCCACAAGCTCTCGAGAGCTTGTGGATGAAGTTGATCTTTTTC-3′; shTead1 #1-antisense, 5′-AATTGAAAAAATGATCAACTTCATCCACAAGCTCGAGCTTGTGGATGAAGTTGATCAT-3′; shTead1 #2-sense, 5′-CCGGGATCAACTTCATCCACAAGCTCTCGAGAGCTTGTGGATGAAGTTGATCTTTTTC-3′; shTead1 #2-antisense, 5′-AATTGAAAAAGATCAACTTCATCCACAAGCTCTCGAGAGCTTGTGGATGAAGTTGATC-3′ (Zhao et al., 2008); shMyc #1 TRCN0000234923 sequence, 5′- CCGGACTTCACCAACAGGAACTATGCTCGAGCATAGTTCCTGTTGGTGAAGTTTTTTG-3′; shMyc #2 TRCN0000234925 sequence, 5′- CCGGTGGAGATGATGACCGAGTTACCTCGAGGTAACTCGGTCATCATCTCCATTTTTG-3′ (Sigma-Aldrich).
Control shRNA (Non-Mammalian shRNA Control, SHC002), sequence, 5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTtgtttttg-3′ (Sigma-Aldrich).
Viral infection and selection of infected cells using a cell sorter using EGFP as a marker were performed in essentially the same manner described for the preparation of retrovirus-infected cells (Ota and Sasaki, 2008).
Preparation of cell lysates and western blotting were performed as described previously (Suico et al., 2014). Chemiluminescent signals were detected and quantified with Image a Quant LAS4000mini (GE Healthcare). The following primary antibodies were used at the indicated dilutions: rabbit anti-c-Myc (Y69) monoclonal antibody (Abcam, 1∶1000); anti-γ-Tubulin (Sigma; 1∶2000); rabbit anti-Tead1 (Ota and Sasaki, 2008) (1∶200); anti-p110α (Cell Signaling; 1∶1000).
ChIP assays were performed using a SimpleChIP Plus Enzymatic Chromatin IP kit (Magnetic Beads, Cell Signaling) according to the manufacturer's instructions. Briefly, NIH3T3 cells were grown to ∼50% (low density) or 100% (high density) confluence. Cross-linked chromatin (10 µg) was immunoprecipitated overnight at 4°C with 2 µg of anti-Tead1 mouse monoclonal antibody (H-4, Santa Cruz Biotechnology) or anti-Yap1 mouse monoclonal antibody (Abnova) or normal mouse IgG (Santa Cruz Biotechnology). Quantitative PCR was performed using the following primer sets: Myc site 1–2, 5′-GCGTTGGAAACCCCGGTAAG-3′ and 5′-AGTCGCTCTACCCCGACTCA-3′; Myc site 3, 5′-ATCTGAGTCGGGGTAGAGCGA-3′ and 5′-GGGTCAGCGTCAGCCCATAG-3′; Myc site 4, 5′-TTTTGAAGCGGGGTTCCCGA-3′ and 5′-GAAGCGACCTCCCGGTTTGA-3′; Myc site 5–6, 5′-ATGTTGGGCTAGCGCAGTGA-3′ and 5′-CGGAACCGCTCAGATCACGA-3′; Myc site 7, 5′-AGTCCGACGAGCGTCACTGA-3′ and 5′-ACCACTCCCCTTTCAGCGTG-3′; Cyr61, 5′-CTCTGATGGATCTGAGAAGAGG-3′ and 5′-GCCCTTTATAATGCCTGCCTA-3′ (Diepenbruck et al., 2014); and Intergenic region, 5′-GCTCCGGGTCCTATTCTTGT-3′ and 5′-TCTTGGTTTCCAGGAGATGC-3′ (Diepenbruck et al., 2014). The fold enrichment was calculated using the 2−ΔΔCt method: 2−[(Ct ChIP – Ct input) − (Ct IgG− Ct input)], where Ct is the threshold cycle.
Statistical analyses were performed with Prism 5 statistical software (GraphPad) by using an unpaired two-tailed Student's t-test, t-test with Welch correction, one-way ANOVA followed by Tukey's multiple comparison test, Dunnett's multiple comparison test, or Kruskal–Wallis test followed by Dunn's multiple comparison test, depending on the data.
We thank Dr Y. Minami for NIH3T3 cells, and Dr T. Kitamura for the pMY vector system.
H.S. conceived and designed the research. M.O., H.M. and T.S. also designed the research. M.O. identified cell competition in NIH3T3 cells and initiated this project. H.M. and T.S. performed most of the experiments and data analyses. H.S. and T.S. wrote the paper.
This work was supported by grants from RIKEN; the Uehara Memorial Foundation; and the Mitsubishi Foundation [grant number 26123]; and by Grants-in-Aid for Scientific Research (KAKENHI) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) [grant numbers 21116003, 26112715 to H.S.]; Japan Society for the Promotion of Science (JSPS) [grant number 23247036 to H.S.]; and MEXT [grant number 25111724 to T.S.]
The authors declare no competing or financial interests.