Oncogenic mutations in adenomatous polyposis coli (Apc) activate mechanistic target of rapamycin complex 1 (mTORC1) in mice and zebrafish

Colorectal cancer (CRC) is responsible for over 600,000 deaths annually worldwide (USCSW Group, 2013). Some 80% of sporadic CRCs result from mutations in adenomatous polyposis coli (APC) (Kinzler and Vogelstein, 1996; Cancer Genome Atlas, 2012). Patients with germline APC mutations develop familial adenomatous polyposis (FAP), which is marked by hundreds to thousands of adenomatous colon polyps and progression to invasive carcinomas (Groden et al., 1991; Kinzler et al., 1991; Miyoshi et al., 1992). APC mutations occur within a mutation cluster region (MCR) and result in expression of a truncated protein that lacks the C-terminal half (Miyoshi et al., 1992). Intestinal cells in humans, mice and rats with these mutations undergo loss of heterozygosity, thus initiating tumor development (Amos-Landgraf et al., 2007; Haramis et al., 2006; Kinzler and Vogelstein, 1996; Su et al., 1992).

As a negative regulator of the Wnt signaling pathway, APC is a core component of the degradation complex that mediates the turnover of β-catenin. APC mutations therefore stabilize and constitutively activate Wnt/β-catenin signaling, a key step in the development of CRCs (Korinek et al., 1997; MacDonald et al., 2009; Morin et al., 1997; Munemitsu et al., 1995). Overexpression of β-catenin in the colon leads to adenoma formation (Romagnolo et al., 1999) whereas knocking down β-catenin reduces adenoma size and frequency (Foley et al., 2008; Scholer-Dahirel et al., 2011). In addition, CRCs with wild-type APC frequently have stabilizing mutations in β-catenin (Morin et al., 1997; Polakis, 2000), providing compelling evidence for the role of β-catenin in colorectal carcinogenesis.

However, several groups have reported that nuclear localization of β-catenin, which is required to activate Wnt target genes, is infrequently observed in early adenomas of patients with FAP, sporadic human polyps and microadenomas in a rat model of FAP, despite Apc loss of heterozygosity (LOH) and elevated cytosolic β-catenin (Amos-Landgraf et al., 2007; Anderson et al., 2002; Bläker et al., 2004; Kobayashi et al., 2000). In addition, defects in intestinal differentiation were observed in zebrafish apc^mcr mutants without detectable nuclear β-catenin or activation of a Wnt transcription reporter (Phelps et al., 2009b). Although the absence of detectable nuclear β-catenin in adenomas could also reflect the sensitivity of the detection methods, the observations have nevertheless led to the identification of additional steps required for nuclear translocation of β-catenin, including activation of Ras and Rac1 (Phelps et al., 2009b; Zhu et al., 2012).

These observations have also raised the possibility that additional effectors downstream of APC might contribute to Apc loss-of-function phenotypes (Phelps et al., 2009a; Phelps et al., 2009b). APC has Wnt/β-catenin-independent roles, including regulation of apoptosis, microtubule dynamics, regulation of retinoic acid biosynthesis and cell-cell adhesion (Hanson and Miller, 2005; Phelps et al., 2009a). We previously found that APC directly enhances the activity of glycogen synthase kinase-3 (GSK-3) (Valvezan et al., 2012). GSK-3 in turn negatively regulates mechanistic target of rapamycin complex 1 (mTORC1) (Inoki et al., 2006) and thus we found that APC, acting through GSK-3, suppresses mTORC1 activity in cultured cells (Valvezan et al., 2012). Because mTORC1 promotes cell growth and proliferation and is aberrantly active in many cancers (Laplante and Sabatini, 2012), we hypothesized that oncogenic Apc mutations might activate mTORC1 independently of β-catenin, and that this activation is important for Apc mutant phenotypes.

To investigate a role for mTORC1 downstream of APC, we have examined oncogenic Apc mutations in zebrafish and mice. Homozygous Apc mutant mice die before gastrulation (Fodde et al., 1994), whereas homozygous mutant zebrafish (apc^mcr/mcr zebrafish) survive until early larval stages (Hurlstone et al., 2003). The Apc mutant zebrafish have multiple developmental defects including impaired circulation, hyperproliferation of the cardiac cushions, enlarged livers and defects in morphogenesis of the anterior-posterior axis, resulting in severe body curvature (Goessling et al., 2008; Hurlstone et al., 2003). These observations indicate a crucial role for APC in the development of many organs and tissues.

Here, we show that mTORC1 is aberrantly activated in apc^mcr/mcr zebrafish and that mTORC1 inhibition attenuates multiple phenotypes in Apc mutants, indicating that mTORC1 activation is important for these phenotypes. We also find that mTORC1 is robustly activated in intestinal adenomas in Apc^min mice. Taken together, these data suggest that mTORC1 is a crucial, β-catenin-independent effector that is activated by Apc mutation.

We previously found that APC directly enhances the kinase activity of GSK-3 (Valvezan et al., 2012). GSK-3 inhibits mTORC1 activity (Inoki et al., 2006) and thus knocking down APC reduces GSK-3 activity and activates mTORC1 in cultured cells (Valvezan et al., 2012). We therefore hypothesized that oncogenic APC mutations should activate mTORC1 in vivo. For these studies we used apc^mcr zebrafish, which contain a previously characterized mutation that causes a premature stop codon within the mutation cluster region (mcr) of apc, analagous to truncating mutations typically found in human colorectal cancers (Haramis et al., 2006; Hurlstone et al., 2003). We compared mTORC1 activity in apc^mcr/mcr zebrafish to that in wild-type and heterozygous zebrafish. Homozygous Apc mutation activates mTORC1, as assessed by western blotting of whole embryo lysates for phosphorylation of ribosomal protein S6, a well-established readout of mTORC1 activity (Laplante and Sabatini, 2012). Phosphorylated S6 is increased in homozygous mutants at 3 and 4 days post fertilization (dpf) compared with pooled wild-type and heterozygous fish, without a change in total S6 (Fig. 1A). To assess tissue-specific changes in mTORC1 activity, we also performed immunohistochemistry on larvae fixed at 3 dpf. Immunostaining for phosphorylated S6 reveals broadly increased mTORC1 activity in multiple mesodermal and endodermal derivatives (Fig. 1B). These data indicate that an oncogenic Apc mutation activates mTORC1.

To test whether aberrantly active mTORC1 contributes to apc^mcr/mcr phenotypes, we asked whether mTORC1 inhibition rescues phenotypes in these fish. Torin1, a direct inhibitor of mTOR kinase activity (Thoreen et al., 2009), was added to developing zebrafish embryos at 24 hours post fertilization (hpf) and survival was assessed daily. Vehicle-treated apc^mcr/mcr zebrafish larvae died 3-4 days post fertilization (dpf) whereas treatment with Torin1 extends their survival by up to 24 hours, with ~40% of Torin1-treated larvae surviving at 4 dpf compared with ~10% of vehicle-treated controls (P<0.0001; Fig. 2A). We also assessed the effect of adding Torin1 at different developmental stages. Adding Torin1 at 0–6 hpf or 24 hpf had similar effects on survival: Approximately 3.5 times as many Torin1-treated apc^mcr/mcr zebrafish were alive at 4 dpf compared with vehicle-treated apc^mcr/mcr zebrafish when treatment was started at either 0–6 hpf or 24 hpf (Fig. 2B). Treatment at 36 hpf or later was less effective (Fig. 2B). Importantly, Torin1 treatment reduced mTORC1 activity in homozygous apc^mcr/mcr zebrafish to levels similar to those in vehicle-treated wild-type/heterozygous fish (Fig. 2C), highlighting the role for aberrant mTORC1 activation resulting from oncogenic Apc mutation. To address whether the effects of Torin1 are specifically mediated by mTORC1 inhibition, we also treated apc^mcr/mcr zebrafish with rapamycin, an allosteric inhibitor of mTORC1 that acts through a distinct mechanism and through a separate binding site in the mTORC1 complex (Sarbassov et al., 2006). Rapamycin similarly extends survival of apc^mcr/mcr zebrafish (Fig. 2D). These data suggest that aberrant mTORC1 activation resulting from oncogenic Apc mutation contributes to early lethality in apc^mcr/mcr zebrafish.

The apc^mcr/mcr zebrafish have dramatically reduced circulation compared with wild-type fish at 2–3 dpf (supplementary material Movies 1–4) (Hurlstone et al., 2003). We used high speed videomicroscopy to measure blood flow through the dorsal aorta; approximately half of vehicle-treated fish had no circulation through the dorsal aorta and ~75% had a flow rate of less than 0.2 mm/second (Fig. 3A). Torin1 treatment improves circulation, resulting in a flow rate greater than 0.2 mm/second in ~50% of mutants, and reducing the percentage of mutants with no circulation to ~25% (Fig. 3A). Rapamycin similarly doubled the percentage of mutants with flow greater than 0.2 mm/second and halved the percentage of mutants with no circulation (Fig. 3B). Representative movies showing blood flow through the dorsal aorta in wild-type fish and each of the three categories used for scoring apc^mcr/mcr zebrafish can be found in supplementary material Movies 1–4. Torin1 and rapamycin did not increase the heart rate of the apc^mcr/mcr zebrafish (Fig. 3C,D). These data indicate that mTORC1 activation in apc^mcr/mcr zebrafish contributes to the reduced circulation phenotype.

Heterozygous apc^mcr/+ zebrafish had enlarged livers by 3 dpf due to an increase in the number of hepatocytes (Fig. 4) (Goessling et al., 2008). To investigate whether this phenotype can be rescued by mTORC1 inhibition, we crossed heterozygous apc^mcr/+ zebrafish to lfabp:RFP fish, which express red fluorescent protein under control of the liver-specific liver fatty acid binding protein (lfabp) promoter (Her et al., 2003). About 50% of resulting progeny are expected to be heterozygous apc^mcr/+ and 50% are expected to be wild type at the Apc locus (apc^+/+). We scored liver size in the resulting RFP+ progeny and found that ~50% had normal-size livers and 50% had enlarged livers, as previously reported (Goessling et al., 2008). Treatment with Torin1 or rapamycin greatly reduced the percentage of fish with enlarged livers and increased the percentage with normal livers (Fig. 4A,B). Genotyping confirmed a strong correlation between heterozygous Apc mutation and liver enlargement in vehicle-treated fish (Fig. 4C,D). However, when Torin1-treated fish with normal-size livers were genotyped, approximately half were found to be apc^mcr/+, confirming that Torin1 restores normal liver size in heterozygous Apc mutant zebrafish (Fig. 4C,D). To quantify the effect of Torin1 treatment on liver hyperplasia, the number of RFP-positive cells per liver was measured by flow cytometry. The apc^mcr/+ fish had ~75% more cells per liver and this was restored to wild-type levels by Torin1 (Fig. 4E). Torin1 had no effect on the number of hepatocytes in wild-type zebrafish. These data demonstrate a crucial role for mTORC1 in liver hyperplasia resulting from oncogenic Apc mutation.

We have found that aberrant mTORC1 activation contributes to early lethality, reduced circulation and liver hyperplasia resulting from oncogenic Apc mutation. Apc loss of function also activates Wnt/β-catenin signaling, which plays a major role in Apc mutant phenotypes. Thus we asked whether combined inhibition of mTORC1 and Wnt/β-catenin signaling could rescue additional phenotypes. Defective morphogenesis of the anterior-posterior axis resulting in aberrant body curvature is one of the most prominent phenotypes in apc^mcr/mcr zebrafish (Fig. 5) (Hurlstone et al., 2003). Treatment with either Torin1 or the Wnt inhibitor XAV939 (Huang et al., 2009b) alone did not affect body curvature, but combining Torin1 and XAV939 resulted in apc^mcr/mcr zebrafish with straight bodies similar to wild type (Fig. 5A). Approximately 95% of vehicle-, Torin1- or XAV939-treated apc^mcr/mcr zebrafish had curved body axes, whereas ~40% of apc^mcr/mcr zebrafish treated with both Torin1 and XAV939 were straight (Fig. 5B). Similarly, rapamycin improved body curvature in combination with XAV939 but had no effect on its own (Fig. 5C). Representative images of scoring categories used in Fig. 5B and 5C are shown in 5D. These data suggest that Apc mutation activates mTORC1 and Wnt/β-catenin signaling independently and highlight the importance of their combined activation. Furthermore, these findings demonstrate that simultaneous inhibition of both pathways can rescue Apc mutant phenotypes that are not rescued by targeting either pathway alone.

To confirm our findings in zebrafish using an established mammalian model of polyposis, we examined GSK-3 and mTORC1 activity in Apc^min mice, which develop intestinal adenomas through loss of heterozygosity (LOH) for Apc. Phospho-S6 staining was strongly induced in polyps compared with adjacent normal tissue (Fig. 6A), similar to previous reports (Fujishita et al., 2008; Metcalfe et al., 2010). Furthermore, phosphorylation of glycogen synthase, an endogenous GSK-3 substrate, was markedly reduced in polyps, but not normal epithelium, from Apc^min mice (Fig. 6B). These data are consistent with a model in which loss of APC reduces GSK-3 activity, which in turn activates both mTORC1 and Wnt/β-catenin signaling.

The importance of enhanced Wnt/β-catenin activity as a mediator of Apc loss-of-function phenotypes is firmly established from studies in model organisms and in colorectal carcinomas in humans. However, reports that early neoplastic lesions can appear before nuclear localization of β-catenin is detectable have raised the possibility that APC regulates additional effectors that contribute to Apc mutant phenotypes in parallel with, but independent of, nuclear β-catenin. We show that truncating mutations in Apc, which are homologous to those found in human colorectal cancers and FAP, cause marked activation of mTORC1 and that mTORC1 activation is important for the resulting phenotypes, including lethality in the larva stage, impaired circulation, defects in morphogenesis and liver hyperplasia. These findings support a crucial role for mTORC1 downstream of oncogenic Apc mutations.

These findings are consistent with a signaling pathway in which APC directly enhances GSK-3 activity (Valvezan et al., 2012). We previously showed that APC directly facilitates GSK-3 activity in vitro and that knockdown of APC reduces GSK-3 activity toward endogenous substrates, including glycogen synthase, in cultured cells. Multiple functional parallels between APC and GSK-3 support this model. Thus, APC and GSK-3 play similar roles in bone morphogenic protein (BMP) signaling (Fuentealba et al., 2007; Fukuda et al., 2010; Miclea et al., 2011), ERK signaling (Park et al., 2006; Wang et al., 2006; Zhai et al., 2007), mitosis (Dikovskaya et al., 2012; Hanson and Miller, 2005; Happel et al., 2009; Wakefield et al., 2003), stem cell homeostasis (Huang et al., 2009a; Qian et al., 2008) and embryonic development in C. elegans (Rocheleau et al., 1997; Schlesinger et al., 1999), Drosophila (McCartney et al., 2001; Yu et al., 1999) and zebrafish (Hurlstone et al., 2003; Lee et al., 2007). Inoki et al. showed that GSK-3 suppresses mTORC1 activity (Inoki et al., 2006), and this was confirmed in vivo in mouse bone marrow (Huang et al., 2012; Huang et al., 2009a); thus our model predicts that inhibition of GSK-3 caused by Apc loss of function should activate mTORC1 in vivo. We observed that glycogen synthase phosphorylation is substantially reduced in intestinal polyps from Apc^min mice (Fig. 6B), which supports our model that Apc mutation reduces GSK-3 activity. We also confirm that mTORC1 is markedly activated in Apc mutant polyps (Fig. 6A), consistent with previous reports (Fujishita et al., 2008; Metcalfe et al., 2010). Furthermore, mTORC1 inhibitors reduce adenoma size and number, as well as mortality, in Apc mutant mice, demonstrating the importance of mTORC1 activation for tumorigenesis resulting from Apc mutation (Fujishita et al., 2008; Koehl et al., 2010; Metcalfe et al., 2010).

In contrast to adult Apc^min/+ mice [as well as adult apc^mcr/+ zebrafish (Haramis et al., 2006; Hurlstone et al., 2003)], the intestines in apc^mcr/mcr homozygous mutant larvae develop with reduced cell number, disorganization of the epithelium and narrowing of the intestinal lumen (Faro et al., 2009; Phelps et al., 2009b). Neither Torin1 nor rapamycin treatment rescued this phenotype (data not shown), perhaps because the defect is not associated with increased proliferation and/or because other factors independent of mTORC1 activation play a key role in this mutant phenotype.

Previous studies have suggested that mTOR expression is increased by Wnt/β-catenin signaling (Fujishita et al., 2008; Metcalfe et al., 2010). However, our data suggest that APC also regulates mTORC1 by enhancing GSK-3 activity independently of β-catenin (Figs 5, 6) (Valvezan et al., 2012); although these two mechanisms are distinct, they are not exclusive of each other. Nevertheless, our observations further suggest that experimental manipulations of upstream Wnt pathway components do not necessarily give information on β-catenin function because Wnts, APC, Axin and GSK-3 also regulate mTORC1 (Fig. 1) (Inoki et al., 2006; Valvezan et al., 2012). Thus, although results from previous studies examining the effects of Dkk-1 on zebrafish cardiac development (Hurlstone et al., 2003) and Dvl2 in mouse intestine (Metcalfe et al., 2010) support a role for canonical Wnt signaling, they do not distinguish between effects mediated by β-catenin and mTORC1.

The importance of combined Wnt/β-catenin and mTORC1 signaling is also supported by our finding that aberrant body curvature in apc^mcr/mcr zebrafish is rescued by combining mTORC1 inhibition with a Wnt pathway inhibitor, as neither inhibitor alone achieves rescue. We find that hyperplasia of the liver in heterozygous apc^mcr zebrafish is partially rescued by mTORC1 inhibition, and previous work has demonstrated that this phenotype is also partially rescued by β-catenin knockdown (Goessling et al., 2008), again indicating a role for both β-catenin and mTORC1 in this hyperplastic phenotype. Taken together, these data suggest that activation of both mTORC1 and Wnt/β-catenin signaling contributes to multiple phenotypes resulting from oncogenic Apc mutation.

In further support of a role for mTORC1 downstream of APC, apc^mcr zebrafish are partially phenocopied by loss of function in TSC components: thus knockdown of Tsc1 activates mTORC1 and causes similar defects in body curvature and genetic inactivation of Tsc2 similarly activates mTORC1, results in early lethality and causes liver hyperplasia similar to apc^mcr heterozygotes (DiBella et al., 2009; Goessling et al., 2008; Hurlstone et al., 2003; Kim et al., 2011). Importantly, liver hyperplasia in Tsc2 mutants is rescued by rapamycin, suggesting that the hyperplasia is caused by mTORC1 activation. The Tsc2 phenotype is not identical to apc^mcr, however. This could be due to persistence of maternal Tsc2 gene expression relative to Apc as well as the important role of β-catenin and possibly other effectors downstream of APC.

Extracolonic manifestations in FAP include polyps of the gastric fundus and small bowel, osteomas, lipomas, desmoid tumors, adrenal cortical adenomas, congenital hypertrophy of the retinal pigment epithelium (CHRPE) and hepatoblastomas (Rustgi, 2007), suggesting parallels with other hereditary disorders that cause neoplasias in multiple organs, including tuberous sclerosis, Peutz-Jeghers syndrome and Cowden’s syndrome (Crino et al., 2006; Inoki et al., 2005; Rustgi, 2007). Peutz-Jeghers and Cowden’s syndromes are hamartomatous polyposis syndromes that involve multiple organ systems and are mediated in part through activation of mTORC1. Peutz-Jeghers syndrome is caused by mutations in LKB, an upstream regulator of AMPK and TSC2, and Cowden’s syndrome is caused by mutations in PTEN that also activate mTORC1 (Inoki et al., 2005). Although each of these syndromes is clinically and pathologically distinct, it is interesting to note that they all share mTORC1 activation, which might be an early molecular step in tumor formation in each case. This parallel also suggests that mTORC1 inhibitors might be effective in controlling some of the extracolonic manifestations of FAP. In contrast, the differences in these syndromes could be due to the fact that APC, LKB and PTEN each regulate multiple other effectors.

In conclusion, APC negatively regulates mTORC1 and oncogenic Apc mutations disrupt this function, resulting in constitutive mTORC1 activation in both zebrafish and mammalian models. Several resulting phenotypes are partially rescued by mTORC1 inhibition, demonstrating a widespread role for mTORC1 activation downstream of APC mutation. Combined inhibition of mTORC1 and Wnt signaling rescues additional phenotypes, highlighting the importance of the unique combination of active mTORC1 and Wnt that results from oncogenic APC mutation. Combined inhibition of Wnt and mTORC1 signaling could be an effective strategy for treatment of colorectal cancers resulting from Apc mutation.

Embryos were raised at 28.5°C in standard E3 medium (Westerfield, 1993). Pronase (30 μg/ml; Roche #10165921001) was added at 24 hpf to digest the chorion and washed out at 48 hpf. The apc^mcr zebrafish and primers used for genotyping were described previously (Hurlstone et al., 2003). Lfabp:rfp zebrafish were described previously (Her et al., 2003). Rapamycin (LC Laboratories #R-5000), Torin1 (Tocris Bioscience #4247) and/or XAV939 (Sigma #X3004) were added directly into the fish medium starting at 24 hpf unless otherwise indicated. Medium was changed and fresh treatments were added every 24 hours. Videomicroscopy and quantification of blood flow rate was performed in unanesthetized zebrafish larvae at 60 hpf as previously described (Hoage et al., 2012). Flow cytometry was performed as previously described (Goessling et al., 2008) except that embryos were dissociated in 0.25% trypsin (Westerfield, 1993). For immunofluorescence, 3-dpf zebrafish embryos were fixed in 4% paraformaldehyde for 2 hours at room temperature, washed in PBST, dehydrated in an ascending ethanol series and submitted to the Perelman School of Medicine Cancer Histology Core at the University of Pennsylvania for paraffin embedding and sectioning. Immunofluorescence staining for S6 and phospho-S6 was performed as in Apc^min mouse intestinal sections (see below). Zebrafish husbandry and egg procurement were carried out in accordance with the guidelines of the University of Pennsylvania Institutional Animal Care and Use Committee.

Zebrafish embryos were lysed on ice in 5 μl/embryo of buffer containing 1% NP-40, 20 mM Tris pH 8.0, 50 mM NaCl, 2.5 mM EDTA, 1 mM DTT, protease inhibitor cocktail (Sigma #P8340) diluted 1:100, phosphatase inhibitor cocktails #2 (Sigma #P5726) and #3 (Sigma #P0044) also diluted 1:100 each. An equal volume of standard 2× Laemmli Sample Buffer was added and samples were heated at 95°C for 5 minutes and then centrifuged at 14,000 rpm for 5 minutes at 4°C. Supernatants were collected for SDS-PAGE and western blot analysis using the following antibodies purchased from Cell Signaling Technology: S6 ribosomal protein (#2317) and phospho-S6 ribosomal protein Ser235/236 (#4858).

Apc^min mice were purchased from The Jackson Laboratory (stock #002020) and have been described previously (Moser et al., 1993). For immunofluorescence, intestines were fixed in 10% phosphate-buffered formalin overnight, washed in PBST, dehydrated in an ascending ethanol series and submitted to the Perelman School of Medicine Molecular Pathology and Imaging Core at the University of Pennsylvania for paraffin embedding and sectioning. Paraffin-embedded sections were washed in xylene and rehydrated in a descending ethanol series followed by boiling in 10 mM sodium citrate pH 6.0 and blocking in Starting Block buffer (Thermo Scientific #37539). Antibodies to S6, phospho-S6 Ser235/236 or phospho-glycogen synthase Ser641 (Cell Signaling Technology #3891) were added overnight at 4°C followed by incubation with Cy2- or Cy3-conjugated secondary antibodies.

We thank Anil Rustgi, John Lynch, Ben Stanger, Jonathan Epstein, Shannon Fisher and all members of the Klein laboratory for helpful suggestions. We also thank Roshan Jain, Santanu Banerjee, Eric De Groh and Michael Granato for help with videomicroscopy; Elizabeth Hexner for help with statistical analysis; Tianying Jiang for help with zebrafish sectioning and staining; and Manimegalai Muthumani for help with zebrafish husbandry and mating.