Transcription coactivator Yes-associated protein (YAP) plays an important role in the regulation of cell proliferation and apoptosis. Here, we identify a new role of YAP in the regulation of cellular senescence. We find that the expression levels of YAP proteins decrease following the replication-induced cellular senescence in IMR90 cells. Silencing of YAP inhibits cell proliferation and induces premature senescence. In additional experiments, we observe that cellular senescence induced by YAP deficiency is TEAD- and Rb/p16/p53–dependent. Furthermore, we show that Cdk6 is a direct downstream target gene of YAP in the regulation of cellular senescence, and the expression of Cdk6 is through the YAP–TEAD complex. Ectopic expression of Cdk6 rescued YAP knockdown-induced senescence. Finally, we find that downregulation of YAP in tumor cells increases senescence in response to chemotherapeutic agents, and YAP or Cdk6 expression rescues cellular senescence. Taken together, our findings define the critical role of YAP in the regulation of cellular senescence and provide a novel insight into a potential chemotherapeutic avenue for tumor suppression. Cancer Res; 73(12); 3615–24. ©2013 AACR.
The Hippo/MST signaling pathway has been well established as a tumor suppressor pathway and is involved in many diverse biologic processes including cell growth, cell death, and organ size control in organisms, ranging from Drosophila to mammals (1–5). As a major downstream target of the Hippo/MST1 pathway, transcription cofactor Yes-associated protein (YAP) has been reported to be phosphorylated and inhibited by Hippo pathway kinases (6–8). Recently, the role of YAP in oncogenesis has been delineated and includes the promotion of cell proliferation, antiapoptosis inducing epithelial–mesenchymal transition (EMT), and cancer development (9–12). Because YAP is unable to directly bind to DNA, the TEAD/TEF family of transcription factors is necessary for YAP's oncogenic activity through the formation of a stably transcriptionally active complex (11).
Cellular senescence is described as an antiproliferative program that leads to permanent growth arrest in the cell (13). Senescent cells undergo a stable growth arrest mediated by interplay between multiple pathways, most notably, the p16/Rb and p53/21 signal pathways (14–16). Currently, most of the cellular senescence regulation molecules identified have been shown to play roles in p16- or p53-dependent manners (17). Recently, it has been reported that the downregulation of p300/CBP–mediated chromatin-dependent alteration of S-phase progression defect induces cellular senescence independent of the activation of p53/p16 (18). Here, we identify a novel function of YAP in the regulation of cellular senescence. RNA interference–mediated YAP gene silencing led to a significant senescent phenotype in normal human fibroblast cells. We also show that YAP cooperates with TEAD transcription factors to control cyclin-dependent kinase 6 (CDK6) expression and expression of YAP or CDK6 inhibits cellular senescence. In addition, we observe that either p53 or p16 knockdown failed to, but the double knockdown of p53 and p16, rescue YAP deficiency–induced senescence. Knockdown of retinoblastoma (Rb) could significantly rescue YAP deficiency–induced senescence. Together, our results suggest that YAP deletion-induced cellular senescence is dependent on both p53 and p16 pathway. Consistently, we found that YAP2 knockdown reduced CDK6 expression and effectively increased anticancer drug-induced senescence in mesothelioma and liver cancer cells. Ectopic expression of YAP or CDK6 decreases drug-induced cancer cell senescence. Taken together, these findings show a new signaling pathway in the senescence process, which might be exploited to treat cancer by senescence induction through inhibiting YAP signaling.
Materials and Methods
All the short hairpin RNAs (shRNA) were cloned into plko.1-puro vectors. Targets sequences are as follows:
YAP rescue plasmid was generated by introducing 5 silent base pair mutations to the wild-type YAP cDNA using the Quick Change Site-Directed Mutagenesis Kit (Stratagene) with the following primers:
YAP rescue forward: 5′-GGAGATGGCAAAGACATCCTCGGGCCAAAGGTACTTCTTAAATC-3′
YAP rescue backward: 5′-GATTTAAGAAGTACCTTTGGCCCGAGGATGTCTTTGCCATCTCC-3′
IMR90, WI-38, and BJ cell lines (American Type Culture Collection) were maintained in Minimum Essential Medium supplemented with 10% FBS at 5% CO2 concentration.
Population doubling assays
Cells were passaged at 3-day intervals, plating 100,000 cells per 6-cm dish. At each passage, the cells were counted by a hemocytometer. Population doubling levels (PDL) were calculated using the equation PD = log (Nf/N0)/log 2, where Nf = number of final cells and N0 = number of initial cells. Cumulative PDLs were calculated by summing the PDLs from all passages. Data were expressed as cumulated PDL from 3 independent experiments.
Plko shRNA plasmids or control vector were cotransfected with the lentiviral packaging plasmids VSVG and PCMV-dR8.12 into HEK293-T cells for virus production. Forty-eight hours after transfection, supernatant was filtered through a 0.45-μm filter, and used to infect cells. Seventy-two hours after infection, cells were selected with 1 μg/mL puromycinin culture medium.
Western blot analyses were conducted as described previously (19).
RNA isolation and quantitative real-time PCR assay
Total RNA was isolated from cells using TRIzol reagent (Invitrogen). The products from reverse transcription were preceded to real-time PCR (RT-PCR) with specific primers. Primer sequences used are available on request.
Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assays were conducted as described previously. The purified DNA was subjected to quantitative PCR (qPCR) with indicated ChIP primers as follows:
CDK6 ChIP-1 forward: TACTCTGGCGCTTTGTTGTG
CDK6 ChIP-1 backward: CGCTGTAGGTAGCAGAGGT
CDK6 ChIP-2 forward: ACTGGACCGGGCCTTTAG
CDK6 ChIP-2 backward: GAGAAGGTCTCTGTCCTC
CTGF ChIP forward: TCTGTGAGCTGGAGTGTG
CTGF ChIP backward: GCCAATGAGCTGAATGGAGT
Senescence-associated β-galactosidase (SA-β-gal) was detected using Cellular Senescence Assay Kit (Millpore; KAA002) following the manufacturer's protocol.
Cells were labeled with 10 mmol/L bromodeoxyuridine (BrdUrd; Sigma-Aldrich) in growth medium for 12 hours at 37°C. BrdUrd-labeled DNA was detected with mouse monoclonal anti-BrdUrd (GE Healthcare; RPN202) according to the manufacturer's protocol.
YAP expression levels correlate with senescence in IMR90 cells
It is well established that normal human diploid fibroblast strains senesce with multiple divisions (20). In our experiments, the human fibroblast cells IMR90 became completely senescent when they reached the PDL at 50 (Fig. 1A). At this stage, both the mRNA and protein levels of YAP were examined by quantitative real-time PCR (qRT-PCR) and Western blot analysis. Interestingly, the expression level of YAP was significantly downregulated in senescent IMR90 cells (PDL = 50) versus the young cells (PDL = 30). This finding suggests that the expression of this gene is associated with cellular senescence (Fig. 1B and C). Consistent with the results in IMR90 cells, we also found that YAP expression level decreased in aging cells in a similar human diploid fibroblast strain, WI-38 cells (Fig. 1D). Together, these results display that YAP expression is downregulated as cells become senescent.
IMR90 cells display senescence following YAP silencing
To further investigate the potential role of YAP in cellular senescence, we designed 2 lentivirus-based shRNAs against YAP and infected IMR90 cells for stable cell lines after selection. Both shRNAs knockdown YAP efficiently, particularly the sh-YAP #1 (Fig. 2A). Subsequent experiments were carried out using these stable cell lines. Compared with the control, silencing YAP significantly induced cellular senescence. SA-β-gal staining and BrdUrd incorporation were used as readouts for senescence. Knockdown of YAP increases the expression of SA-β-gal (Fig. 2B) and inhibits cell DNA replication (Fig. 2C). Furthermore, YAP downregulation also resulted in the inhibition of cell proliferation in a population doubling assay (PDL; Fig. 2D). To ensure our results held across multiple human fibroblast cell lines, we used BJ cells and also observed YAP knockdown results in a similarly senescent phenotype (Supplementary Fig. S1). To confirm that the effects were induced by YAP silencing, we constructed a rescue plasmid, which encoded a resistant form to sh-YAP#1 for higher efficiency. Not surprisingly, overexpression of YAP-Rescue was able to recover the senescent phenotype induced by YAP silencing (Fig. 2E).
YAP knockdown regulates senescence through TEAD and in Rb/p53/p16–dependent manner
As a transcriptional coactivator, YAP must bind transcription factors to induce gene expression, specifically, transcription factor TEAD family proteins interact with coactivator YAP and mediate YAP-induced gene expression (11, 21). The YAP–TEAD complex regulates a series of cell processes, including proliferation, apoptosis, EMT, and oncogenic transformation. We wanted to explore whether YAP regulation of senescence occurs in conjunction with transcription factor TEAD family proteins. In agreement with previous studies that TEAD functions as the main downstream target of YAP as to form a proproliferation signaling (11, 21), we found that knockdown of TEAD1/3/4 also induces an almost identical cellular senescent phenotype as YAP silencing (Fig. 3A–C). Further solidifying our hypothesis that YAP regulates cellular senescence in a TEAD-dependent manner, we found that the knockdown of both YAP and TEAD1/3/4 together did not lead to a further increased senescent phenotype, including SA-β-gal staining and BrdUrd incorporation as well as PDL (Fig. 3A–D).
It has been shown that transcription factors p53 and cell-cycle regulator p16 are key regulators of cellular senescence and that most other regulators of senescence have been shown to depend on the p16 or p53 pathways (14, 17). Therefore, to further elucidate the senescence pathway, our next question dealt with whether YAP-regulated senescence was indeed controlled by p16/p53. In our SA-β-gal staining experiments, silencing either p16 or p53 did not rescue the senescent phenotype caused by YAP knockdown (Fig. 4A), which indicated YAP deficiency–induced senescence is not dependent on p53 or p16. We also found that the individual knockdown of p53 or p16 could not rescue the YAP knockdown-induced senescent phenotype by conducting population doubling assay experiments (Fig. 4B). However, double knockdown of p53 and p16 partially rescued YAP knockdown-induced senescence (Fig. 4C), suggesting that YAP antagonized senescence through both the p53 and p16 pathways.
CDK6 plays a role in cellular senescence as a downstream target of YAP–TEAD complex
Most genes—including p53/p21 and p16/Rb—regulate senescence through control of the cell cycle (22–24). As expected, YAP knockdown significantly reduced the S-phase and increased the G1-phase in the IMR90 cells (Supplementary Fig. S2A). We next examined the molecular mechanism of YAP–TEAD–mediated cell-cycle control. From our microarray data (Supplementary Table S1), we identified CDK6, which has been previously characterized for its important role in regulating cell-cycle progression as a potential target (25–27). In our experiments, YAP or TEAD knockdown significantly inhibited CDK6 mRNA level and protein expression in IMR90 cells (Fig. 5A and B). Because other cell-cycle regulation genes have been previously identified as senescence regulators, we also examined the expression of these regulation genes and did not find any significant changes between control and YAP knockdown cells (Supplementary Fig. S2B). This result was in accordance with our microarray data. Our next step involved determining whether CDK6 is the direct downstream target gene of the YAP–TEAD complex. We characterized 2 conserved TEAD-binding sites (GGAATG) in the promoter region of human CDK6 gene (Fig. 5C). Further ChIP experiments confirmed the binding of YAP to these 2 sites, whereas we used the already identified YAP–TEAD target CTGF as a positive control (Fig. 5D). Furthermore, overexpression of CDK6 can partially rescue the increased senescence induced by YAP knockdown (Fig. 5E). We also found that YAP knockdown decreases the phosphorylation level of Rb in IMR90 cells (Fig. 5B) and knockdown of Rb reduce YAP deficiency–induced IMR90 cell senescence (Fig. 5F). Together, these results suggest YAP controls cellular senescence through transcriptionally regulating the expression of CDK6.
YAP plays a critical role in tumor cell senescence
Recently, many groups have shown the antitumor role of senescence in cancer treatment (28–31). Therefore, our next area of interest was examining the role of YAP in regulating senescence in cancer cells. It has been reported that the tumor suppressor gene, LATS2, is a homozygous deletion in a mesothelioma cell line NCI-H2052 and results in the constitutive activation of YAP (32). Consistently, we observed the much lower pS127-YAP in NCI-H2052 cells (Fig. 6A) and, compared with the LATS2 wild-type mesothelioma cell line NCI-H2452, NCI-H2052 was more resistant to pemetrexed (a putative drug for the mesothelioma treatment)–induced senescence (Fig. 6B). Similarly, we observed hydrogen peroxide could induce pS127-YAP phosphorylation in IMR90 cells, indicating MST/LATS kinases might be involved in the oxidation-induced cellular senescence (Supplementary Fig. S3). We next found that reduced expression of YAP in NCI-H2052 cells directly induces cellular senescence and increases sensitivity to drug-induced senescence (Fig. 6C). As expected, we found that CDK6 is decreased (Fig. 6D), and proliferation is also reduced (Fig. 6E) in the YAP knockdown NCI-H2052 cells, indicating that CDK6 may play an important role in YAP-controlled cancer cell senescence. YAP is overexpressed in many liver cancer cells and has been closely linked to hepatocellular carcinoma development. Therefore, we next wanted to characterize the role of YAP following drug-induced senescence in liver cancer. We used 2 liver cancer cell lines—HepG2 and Bel7402—in which YAP is highly expressed (33). Knockdown of YAP significantly increased the number of senescent cells in these 2 cell lines upon pemetrexed treatment (Fig. 6F and G). Consistently, expression of YAP or Cdk6 significantly inhibits senescence in pemetrexed-treated cancer cells (Supplementary Fig. S4A and S4B). In sum, these results show that YAP also plays a critical role in tumor cell senescence and inhibition of YAP could be a new therapeutic avenue for liver cancers.
Our results suggest that the transcriptional cofactor YAP plays an important role in the maintenance of cell proliferation and resistance to cellular senescence. Interestingly, we also found that YAP expression is closely correlated with senescence. This raises the question of how to control YAP expression during cell aging. It has been shown that YAP level can be mediated by ubiquitination and miRNAs such as miR-375 (12, 34). In our study, however, we also detected the downregulation of YAP mRNA in senescent cells, which suggests the potential role of transcription factors or epigenetics modification enzymes in YAP transcription initiation regulation. In another line of our experiments, we found that ectopic expression of YAP in IMR90 cells (Supplementary Fig. S5) and tumor cells (Supplementary Fig. S4) could inhibit senescence, further confirming the important role of YAP in the regulation of cellular senescence. However, YAP expression fails to rescue the senescent phenotype in the old passaged cells (PDL = 47; Supplementary Fig. S5), which indicates that YAP plays a role of the antisenescence in the relatively senescent cells.
A recent study showed that YAP downregulation increased senescence in a colorectal carcinoma HCT116 cell line in a p53- and p21-dependent manner (35). In our study, we found that the YAP–TEAD signaling pathway that regulates senescence to be partially dependent of the key regulators p16 and p53. Most importantly, we found CDK6 as a novel downstream target of the YAP–TEAD transcriptional active complex. In the rescue experiments, ectopic expression of CDK6 partially recovered the senescent phenotype. Taken together, our data confirm that CDK6 is required for YAP-controlled senescence regulation and suggests other molecules may be involved in this process. At the same time, we observed that knockdown of Rb significantly reduced YAP deficiency–induced senescence in IMR90 cells (Fig. 5F) and YAP deficiency caused a reduced Rb phosphorylation in both IMR90 cells (Fig. 5B) and tumor cells (Supplementary Fig. S6A), indicating that YAP-CDK6 signaling regulates senescence is Rb-dependent. Consistently, we also found that YAP knockdown or pemetrexed treatment reduced phosphorylation of Rb protein in liver cancer cells (Supplementary Fig. S7A and S7B), which further support our conclusion that YAP-CDK6 signaling occurs in the regulation of senescence.
The accumulated reactive oxygen species (ROS) and DNA damage are usually coupled with cellular senescence (36, 37). However, in our experiments, we did not observe the changes of DNA response or ROS in YAP knockdown-induced senescent cells (Supplementary Fig. S7A–S7C). The telomere serves a critical function in the cell and previous research has shown telomere shortening is common in senescent cells (28, 38). In our experiments, however, no telomere shortening was observed in the YAP knockdown senescent IMR90 cells (unpublished data), an interesting and potentially worthwhile avenue for further exploration.
In sum, we characterized YAP as a new negative regulator of cellular senescence through transcriptionally regulating the expression of CDK6 gene (Fig. 7). Furthermore, we found that YAP plays a prooncogenic role in cancer through its antisenescence function, which may provide a novel insight into cancer chemotherapeutic treatment.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: Q. Xie, J. Chen, S. Meng, Z. Yuan
Development of methodology: Q. Xie, J. Chen
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Chen, H. Feng, S. Peng, J. Li
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Q. Xie, J. Chen, U. Adams, Z. Yuan
Writing, review, and/or revision of the manuscript: Q. Xie, U. Adams, Z. Yuan
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Chen, U. Adams, Y. Bai, J. Huang
Study supervision: Z. Yuan
Other: L. Huang
This work was supported by the Ministry of Science and Technology of China (973-2012CB910701 and 973-2009CB918704 to Z. Yuan), and the National Science Foundation of China (81125010 and 81030025) to Z. Yuan.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The authors thank members of the Yuan laboratory for critical reading of the article and helpful discussions.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received October 4, 2012.
- Revision received February 26, 2013.
- Accepted March 15, 2013.
- ©2013 American Association for Cancer Research.