12-O-Tetradecanoylphorbol-13-acetate (TPA) is anti-tumorigenic in liver cancer cells via inhibiting YAP through AMOT

TPA stimulates carcinogenesis in various types of cancers. However, we found that TPA inhibits transformative phenotypes in liver cancer cells via the translocation of YAP from the nucleus, where it functions as a transcriptional co-factor, to the cytoplasm. Such effects led to a separation of YAP from its dependent transcription factors. The inhibitory effects of TPA on YAP were AMOT dependent. Without AMOT, TPA was unable to alter YAP activity. Importantly, the depletion of YAP and AMOT blocked the TPA-reduced transformative phenotypes. In sum, TPA has been established as an anti-tumorigenic drug in liver cancer cells via YAP and AMOT.

All IF and WB were performed by conventional methods, and the protocols are available elsewhere.
Co-immunoprecipitation (co-IP). Co-IP was performed as described previously 30  Statistical analysis. Tests to examine the differences between groups included Student's t test and one-way ANOVA; p < 0.05 was regarded as statistically significant.

TPA inhibits transformative phenotypes and YAP activity in liver cancer cells.
We found that the cell-proliferation and colony-formation capacities of Bel-7402 and Bel-7404 cells could be dose-dependently inhibited by increasing concentrations of TPA (Fig. 1A,B). By contrast, Caspase 3/7 activities could be dose-dependently induced (Fig. 1C). These results suggested that TPA inhibits the transformative phenotypes of liver cancer cells. Because YAP is a transcription co-factor 17 and its activity relies on its dependent transcription factors (including TEAD and CREB) 31,32 , we tested whether TPA treatment influences the transcriptional activities of TEAD and CREB. We found that YAP-dependent TEAD and CREB activities could be dose-dependently reduced by TPA, as measured using a TEAD-Gal4/pUAS-LUC and an HULC-promoter luciferase reporter, which contains Scientific RepoRts | 7:44940 | DOI: 10.1038/srep44940 a CREB-responses element, before and after treating cells with increasing concentrations of TPA (Fig. 1D,E). TEAD target genes, CTGF and ANKRD1 33,34 , and CREB target genes, MCAM and HULC 35,36 , were also found to be downregulated by TPA treatment (Fig. 1F). Moreover, the inhibitory efficacies of TPA on the mRNA levels of these genes were abolished when YAP was depleted (Fig. 1F), further demonstrating that TPA can inhibit the activities of YAP-dependent transcription factors.
Then, we investigated whether TPA directly affected the phosphorylation of YAP (p-YAP), a hallmark of YAP inactivation 37 . We found that p-YAP was dose-dependently elevated, while total YAP levels were unaffected by increasing concentrations of TPA (Fig. 1G). Further, we found that TPA was able to shuttle YAP from the nucleus, where YAP exerts its major function on tumorigenesis 38 , to the cytoplasm (Fig. 1H); these effects were also dose-dependent, suggesting YAP can be directly inhibited by TPA. The bar graphs are shown as the percent of cells in each of three categories (nuclear > cytoplasm, nuclear = cytoplasm, and nuclear < cytoplasm) from 100 randomly counted cells. The data are shown as the mean ± SD from three independent experiments. The data from cells treated with DMSO infected with or without GFP-sh are arbitrarily set to 100% (except Fig. 1H). *p < 0.05 and **p < 0.01, as analyzed using one-way ANOVA. Specifically in Fig 1H, comparisons of the percent of the cells categorized into the "nuclear > cytoplasm" group among different treatments, as indicated, are also statistically analyzed.
Scientific RepoRts | 7:44940 | DOI: 10.1038/srep44940 TPA separates YAP from its dependent transcription factors. Because YAP activity relies on its transcription factors 16 , we performed IF and found that TPA treatments led to YAP translocation from the nucleus to the cytoplasm ( Fig. 2A). However, the nuclear localization of YAP-dependent transcription factors, including TEAD, CREB, p73 and Runx2, was not altered in Bel-7402 cells ( Fig. 2A). Similarly, data from fractionation experiments indicated that TPA treatments led to gradually increasing cytoplasmic accumulation of YAP, whereas decreased nuclear expression of YAP was caused by increasing concentrations of TPA in both Bel-7402 and Bel-7404 cells (Fig. 2B). Further, co-IP experiments demonstrated that TPA treatments dissociated endogenous YAP from endogenous CREB, Runx2, TEAD and p73 in Bel-7402 and Bel-7404 cells (Fig. 2C). In Bel-7402 cells co-transfected with exogenous YAP-FLAG and TEAD4-Myc, CREB-HA, or Runx2-HA, we also found that TPA treatments inhibited the interactions between YAP-FLAG and its dependent exogenous transcription factors (Fig. 2D,F). Taken together, TPA separates YAP from its dependent transcription factors. Runx2-HA (F). Exogenous YAP-FLAG was immuno-precipitated by anti-FLAG antibodies, and co-immunoprecipitations of TEAD4-Myc, CREB-HA and Runx2-HA were measured by WB using the indicated antibodies. The representative WB images from three independent experiments are shown in the upper panel. The relative ratios are shown in the lower panel. The data from the "DMSO" group are arbitrarily set to 100%. The data are shown as the mean ± SD from three independent experiments. *p < 0.05 and **p < 0.01, as analyzed using Student's t test ( Fig. 2A and C) and one-way ANOVA ( Fig. 2B and D-F). Specifically in Fig. 2A, comparisons of the percent of the cells categorized into the "nuclear > cytoplasm" group between different treatments, as indicated, are also statistically analyzed. AMOT overexpression has similar effects as TPA. We have previously reported that AMOT overexpression causes the inhibition of YAP 28 . Here, cytoplasmic accumulation of YAP could be found in Bel-7402 and Bel-7404 cells successfully transfected with exogenous AMOT-HA, whereas nuclear accumulation of YAP was observed in the cells without successful transfection (Fig. 3A), demonstrating that AMOT overexpression can drive YAP from the nucleus to the cytoplasm. We also found that AMOT overexpression dose-dependently suppressed the interactions between exogenous YAP-FLAG and TEAD4-Myc, CREB-HA or Runx2-HA ( Fig. 3B-D). Further, the amounts of CREB, p73, Runx2 and TEAD in the immunoprecipitates that were pulled down by anti-YAP antibodies were greatly reduced when AMOT was overexpressed (Fig. 3E). Moreover, AMOT concentrations of AMOT-HA by a TEAD-Gal4/pUAS-LUC and HULC-promoter reporter luciferase system, respectively. All the data from cells transfected with Empty vectors are arbitrarily set to 100%. The data are shown as the mean ± SD from three independent experiments. *p < 0.05 and **p < 0.01. The data in Fig. 3A and E were analyzed using Student's t-test and the data in Fig. 3B-D and F-H was analyzed using one-way ANOVA. Specifically in Fig. 3A, comparisons of the percent of the cells categorized into the "nuclear > cytoplasm" group between indicated treatments are also statistically analyzed.
Scientific RepoRts | 7:44940 | DOI: 10.1038/srep44940 overexpression reduced the mRNA levels of CTGF, ANKRD1, MCAM and HULC; however, the depletion of YAP blocked such effects (Fig. 3F), suggesting that AMOT controls TEAD and CREB target gene transcription, possibly via YAP. Furthermore, the transcriptional activities of TEAD and CREB measured by luciferase-based experiments were dose-dependently reduced in Bel-7402 and Bel-7404 cells transfected with increasing concentrations of exogenous AMOT-HA compared to those from the control cells (Fig. 3G,H). These results suggested that, similar to the effects of TPA, AMOT overexpression is capable of inhibiting YAP-dependent transcriptional activity.
TPA simultaneously induces the translocation of both YAP and AMOT from the nucleus to the cytoplasm. Next, we tested whether the interaction between AMOT and YAP can be affected by TPA. We found that TPA dose-dependently enhanced the AMOT-YAP interaction in both Bel-7402 and Bel-7404 cells (Fig. 4A). IF experiments revealed the gradual and simultaneous translocation of both YAP and AMOT from the nucleus to the cytoplasm (Fig. 4B). By cytosolic and nuclear fractionation experiments, the gradual and simultaneous translocation of YAP and AMOT from nucleus to the cytoplasm in both Bel-7402 and Bel-7404 cells was also observed (Fig. 4C). These results suggested that the changes in the subcellular localization of YAP and AMOT AMOT is essential for the TPA-induced inhibition of YAP. Next, we sought to determine whether AMOT is essential for the TPA-induced inhibition of YAP. We found that YAP translocation from the nucleus to the cytoplasm was blocked when AMOT was knocked down by two independent shRNAs against AMOT compared to the control in both Bel-7402 and Bel-7404 cells (Fig. 5A).
Further, cytosolic and nuclear fractionation experiments demonstrated that the knockdown of AMOT could only increase the nuclear expression of YAP while decreasing the cytoplasmic expression of YAP, but AMOT knockdown blocked the TPA-induced translocation of YAP from the nucleus to the cytoplasm in both Bel-7402 and Bel-7404 cells (Fig. 5B). Moreover, the TPA-induced reduction of TEAD and CREB transcriptional activities was abolished when AMOT was knocked down (Fig. 5C). Further, the TPA-reduced downregulation of the mRNA levels of CTGF, ANKRD1, MCAM and HULC in the control cells could not be observed in Bel-7402 and . The data are shown as the mean ± SD from three independent experiments. The data from cells infected with GFP-sh and treated with DMSO are arbitrarily set to 100% (except Fig. 5A). **p < 0.01. The data from Fig. 5A and B-D were analyzed using Student's t-test and one-way ANOVA, respectively. Specifically in Fig 5A, comparisons of the percent of the cells categorized into the "nuclear > cytoplasm" group between different treatments, as indicated, are also statistically analyzed.

TPA-reduced transformative phenotypes rely on YAP and AMOT.
Here, we sought to determine whether the TPA-reduced transformative phenotypes in liver cancer cells depend on YAP and AMOT. We found that TPA-reduced cell proliferation and colony formation capacities could be blocked by the knockdown of YAP and AMOT, respectively (Fig. 6A,B). Furthermore, TPA-induced Caspase 3/7 activities could also be abolished by the knockdown of YAP and AMOT, respectively (Fig. 6C). However, TPA still had inhibitory effects on the transformative phenotypes of liver cancer cells, even when the WW domain containing transcription regulator 1 (TAZ), a homolog of YAP, was depleted ( Supplementary Fig. S1A-D), thereby excluding the possibility that the effects generated by TPA occur via a TAZ-dependent mechanism.
In vivo xenograft experiments also demonstrated that the TPA efficacy with respect to tumor growth inhibition was much reduced when YAP or AMOT was knocked down (Fig. 6D). These results demonstrate that TPA inhibits transformative phenotypes in liver cancer cells, possibly via AMOT and YAP.

Discussion
Previous studies have demonstrated that TPA enhances cell migration through activating protein kinase C (PKC) in several types of tumor cells 4,6,7 . As a PKC activator, TPA also inhibits the apoptosis induced by Fas/FasL in human promyelocytic leukemia cells 9 . Although TPA has been reported to be pro-tumorigenic 39 , its exact effects remain controversial. A study from Gong et al. 40 has suggested that TPA exerts opposing roles on cell proliferation, possibly via regulating the Hippo/YAP pathway in a cell type-dependent manner. This may be because TPA can activate different PKC isoforms in different types of cells. For example, TPA affects the nPKC isoform to activate LATS, a natural inhibitor of YAP, in Swiss3T3, MEF and A549 cells, whereas TPA affects the cPKC isoform to  Fig. 6D). The data from cells "infected with GFP-sh and treated with DMSO" are arbitrarily set to 100%. **p < 0.01. The data from Fig. 6A-C were analyzed using Student's t-test, and the data from Fig. 6D were analyzed using one-way ANOVA.
suppress LATS in HEK293A cells. Thereby, one drug might cause opposite effects via the same signaling pathway. However, which PKC isoform mediates TPA-induced simultaneous translocation of AMOT and YAP from the nucleus to cytoplasm in liver cancer cells remains unknown and must be investigated in future studies. Unlike the pro-tumorigenic roles of TPA in other types of tumor cells, TPA might act as an anti-tumorigenic agent in the human liver cancer cell line HepG2 8 . Consistent with this possibility, we report here that TPA is able to suppress transformative phenotypes in two human liver cancer cell lines, Bel-7402 and Bel-7404, further suggesting that TPA might be used as an anti-tumor drug in the treatment of liver cancer.
YAP translocation from the nucleus to the cytoplasm might dissolve the YAP-TEAD complex 41 , which promotes cell proliferation and maintains survival programs by inducing the expression of target genes, such as CyclinD1, CyclinE and CTGF 42,43 . We found that TPA treatment leads to the separation of YAP from a series of its dependent transcription factors, including TEAD, CREB, Runx2 and p73, whose nuclear localization is not affected by TPA. Notably, these transcription factors play critical roles in the promotion or suppression of liver tumorigenesis 20,29 . Because YAP functions as a co-transcription factor, the loss of YAP in the nucleus decreases the transcriptional activities and subsequent target gene expression of its dependent transcription factors. Although YAP-dependent transcription factors are either pro-tumorigenic (for example, TEAD and CREB) or anti-tumorigenic (for example, Runx2 and p73), we believe the major roles of YAP in liver cancer cells are pro-tumorigenic because YAP is highly up-regulated in liver cancer tissues compared to the corresponding adjacent normal liver 19 ; YAP knockout restricts liver development 25 ; and YAP depletion leads to impaired transformative phenotypes in liver cancer cells 28 . These results might explain why TPA is anti-tumorigenic in liver cancer cells.
AMOT has been identified as a potential component of Hippo signaling, and numerous studies have reported that AMOT inhibits YAP activity 24,44 . The interaction between YAP and AMOT has also been well established 45 , and the PPxY motifs of AMOT are essential for interaction with the WW domains within the YAP protein 46 . Our findings have demonstrated that without AMOT, TPA is unable to alter the subcellular localization of YAP or the activities of YAP-dependent transcription factors. Therefore, AMOT is critical for maintaining the efficacy of TPA in treating YAP-dependent liver cancer.
In the present and our previous studies 47 , we found that endogenous AMOT is in the nucleus, whereas exogenous AMOT-HA was observed to be excluded from the nucleus, as most other reports 28,45 have described. Protein synthesis occurs in the cytoplasm and especially on the rough endoplasmic reticulum (ER) 48 . Consequently, exogenous AMOT-HA protein should be abundantly translated on the ER, thus leading to AMOT-HA accumulation in the cytoplasm. It also takes time to transfer AMOT-HA, via an unknown mechanism, to its correct subcellular localization, possibly in the nucleus. We hypothesize that concurrently, the cytoplasmic accumulation of AMOT-HA induces YAP translocation from the nucleus to the cytoplasm, ultimately resulting in its inhibitory effect on YAP. This might explain why AMOT-HA exists mainly in the cytoplasm and might explain the observation that AMOT and YAP appear to bind each other better when they co-localize in the cytoplasm than in the nucleus. The relevant mechanism should be investigated further in the future. Moreover, we have further tested the specificity of the anti-AMOT antibody used in the present study. We found the nuclear staining went away when knocked down AMOT with shRNA, as measured by IF experiment (Fig. S2A). The knockdown efficiencies of AMOT by shRNAs targeting AMOT were also analyzed by WB experiments (Fig. S2B). These results indicated that the nuclear signals that recognized by this anti-AMOT antibody represent genuine endogenous AMOT in the nucleus of Bel-7402 and Bel-7404 cells.
In conclusion, our findings demonstrate that TPA is anti-tumorigenic in liver cancer cells via an AMOT and YAP-dependent mechanism (Fig. 6E). Simultaneously increasing AMOT function might enhance the efficacy of TPA in treating liver cancer.