Yes-associated protein 1 (YAP1) interacts with numerous transcription factors, including TEA-domain family proteins (TEAD) and p73. YAP1 is negatively regulated by the tumor suppressor Hippo pathway. In human cancers, the deregulation of the Hippo pathway and YAP1 gene amplification lead to the activation of YAP1, which induces epithelial–mesenchymal transition (EMT) and drug resistance. YAP1 inhibitors are expected to be useful in cancer therapy. On the other hand, in certain cancers, YAP1 upregulates p73-dependent gene transcription and behaves as a tumor suppressor. Moreover, as YAP1 regulates self-renewal and differentiation of tissue stem cells and plays an important role in tissue homeostasis, YAP1 activators may contribute to the regenerative medicine. With this in our mind, we screened for YAP1 activators by using human retinal pigment epithelial ARPE-19 cells expressing the TEAD-responsive fluorescence reporter under the coexpression of YAP1. From an extensive chemical compound library (n = 18,606) 47 candidate YAP1 activators were identified. These compounds were characterized to determine whether this assay provides bona fide YAP1 activators. Importantly, one YAP1 activator was effective against the human multiple myeloma IM-9 cells and chronic myeloid leukemia K562 cells.
Implications: YAP1 activation limits growth, induces apoptosis, and may be useful at suppressing hematological cancers. Mol Cancer Res; 16(2); 197–211. ©2017 AACR.
This article is featured in Highlights of This Issue, p. 185
Yes-associated protein 1 (YAP1) was identified as a protein interacting with yes tyrosine kinase (1). Thereafter, many YAP1-interacting proteins, such as ErbB4, p73, TEAD, SMAD, and Runx2, were reported (2–6). YAP1 does not directly bind to DNA but regulates gene transcription through the interaction with these molecules. YAP1 is negatively regulated by the tumor suppressor Hippo pathway (7, 8). YAP1 is phosphorylated by large tumor suppressor kinase (LATS) 1 and 2, the core kinases of the Hippo pathway. The phosphorylated YAP1 is recruited from the nucleus to the cytoplasm and undergoes degradation. In human cancers, the Hippo pathway is frequently deregulated and the YAP1 gene is amplified, so that YAP1 activity is enhanced (9). YAP1 upregulates cell cycle–promoting and antiapoptotic genes (10). Cancer cells with hyperactive YAP1 undergo epithelial–mesenchymal transition and acquire drug resistance (10). The activity of YAP1 correlates with short survival in cancer patients. Hence, YAP1 is regarded as a therapeutic target in cancer therapy (11). Among transcription factors interacting with YAP1, TEAD is the most important in the induction of epithelial–mesenchymal transition and drug resistance (12, 13). Accordingly, verteporfin and synthetic peptides that block the interaction between YAP1 and TEAD are shown to suppress cancer growth (14–16).
The role of YAP1 in cancer is twisting. YAP1 cooperates with p73, upregulates proapoptotic genes, and suppresses certain cancers (17). How YAP1 determines which fate as an oncoprotein or as a tumor suppressor to adopt is not fully understood, but the accumulating evidence suggests that Abelson murine leukemia viral oncogene homolog 1 (ABL1) is a key determinant (18). Upon DNA damage, ABL1 phosphorylates YAP1 at tyrosine 357 and promotes the interaction with p73. A recent study has revealed that multiple myeloma (MM) cells show DNA damage response but escape cell death, because YAP1 activity is low (19). When YAP1 is exogenously expressed or the Hippo pathway is suppressed to enhance YAP1 expression, MM cells die via the ABL1–YAP1–p73 axis. We can surmise that in cancers with the high expression of ABL1 and with the low expression of YAP1, YAP1 activators trigger apoptosis and are therapeutically effective.
YAP1 plays a crucial role in the regulation of tissue stem cells (20). YAP1 is essential for tissue repair and protection in intestine, liver, heart, and skin (21–26). YAP1 is also necessary to maintain neural stem cells in the brain (27). Based on these properties of YAP1, we infer that YAP1 activators are useful in the regenerative medicine.
We previously searched for YAP1 modulators by means of human osteosarcoma U2OS cells expressing green fluorescent protein (GFP)-tagged YAP1 (28). We used the subcellular localization of GFP-YAP1 as the readout and found that dobutamine inhibits YAP1 through β-adrenergic receptor. However, we have not yet found the compounds that activate YAP1. In this study, we established a new cell-based assay, in which the YAP1-dependent TEAD-responsive reporter activity is monitored. We performed a small chemical compound library screening with the use of this assay and obtained candidate YAP1 activators. We characterized these compounds to examine whether they indeed activate YAP1 and tested the idea that YAP1 activator can be used against MM and other blood cancer cells.
Materials and Methods
DNA constructions and virus productions
pCIneoFLAG-His6 (pCIneoFH), pClneoFLAG-His6-FLAG (pClneoFHF), pCIneoMyc, pCIneoLuc, pCIneomCherry, and pQCXIP-EF were described previously (28–32). NheI/EcoRI fragment was isolated from pEGFP-C2 (Clontech Laboratories) and ligated into NheI/EcoRI sites of pCIneo (Promega) to generate pCIneoEGFPC2. pIRES2-EGFP (Clontech Laboratories) was cut with NotI, filled in, and digested with NheI. pLL3.7 vector was cut with EcoRI, filled in, and digested with NheI. The fragment from pIRES-EGFP was ligated into pLL3.7 to replace GFP with IRES-GFP and the resulting vector was named pLL3.7 K122. PCR was performed with the primers (H3244, 5′-caattggcagaaatcggtactggctttccatc-3′ and H3245, 5′-acgcgtgaattccgcgttatcgctctgaaagta-3′) on pHTN-Halo-tag vector (Promega) to amplify Halo-tag. The PCR product was digested with MfeI/MluI and ligated into EcoRI/MluI of pCIneomCherry to generate pCIneomCherry-Halo. pCMV SPORT human YAP (MHS1010-7508628) was purchased from Open Biosystems. The coding region was amplified by PCR with the primers (H2130, 5′-acgcgtcccgggcagcagccgccgcctcaa-3′ and H2352, 5′-gatatcataaccatgtaagaaagctt-3′), digested by MluI/EcoRV and ligated into MluI/Smal sites of pCIneoFH to generate pCIneoFH-YAP1. MluI/NotI fragment from pCIneoFH-YAP1 was ligated into MluI/NotI sites of pCIneoLuc and pCIneomCherry-Halo to generate pCIneoLuc-YAP1 and pCIneomCherry-Halo-YAP1. NheI/NotI fragment from pCIneomCherry-Halo-YAP1 was ligated into XbaI/NotI of pQCXIP-EF to generate pQCXIP-mCherry-Halo-YAP1. pCIneoFH-YAP1 was digested with NotI, filled in, and cut with NheI. The isolated fragment was ligated into NheI/SmaI sites of pLL3.7 K122 to generate pLL3.7 K122 FH-YAP1. NcoI/BglII from pEYFP-C1 (Clontech Laboratories) was ligated into NcoI/BamHI of pGL3 to replace luciferase with YFP and to generate YFP reporter. BamHI/Sall and Sall/HindIII fragments from 8xGT-IIC-δ51LucII luciferase reporter (a gift from Hiroshi Sasaki; RIKEN RDB08067; ref. 33) were ligated into BglII/HindIII sites of the YFP reporter to generate pTEAD-responsive-promoter YFP, which was digested with NotI and partially digested with HindIII. The resulting fragment and HindIII/XbaI fragment from H2B-mCherry (a gift from Robert Benezra; Addgene plasmid, 20972; ref. 34) were ligated into NotI/XbaI sites of pLL3.7 K122 FH-YAP1 and pLL3.7 K122 to obtain pLL3.7 K122 FH-YAP1-ires-GFP-TEAD-responsive-promoter-H2B-mCherry reporter (Fig. 1A) and pLL3.7 K122 control-TEAD-responsive-promoter-H2B-mCherry, respectively. pCMV FLAG-human YAP1 was described previously (35). pCIneoMyc- and pCIneoFLAG-LATS1 were generated from pCGN HA-Warts (a gift from Hiroyuki Saya) using PCR with the primers (H1921, 5′-acgcgtatgaggcctaagacctttcc-3′ and H1922, 5′-gtcgactaaacatatactagatcgcga-3′), while pCIneoMyc-LATS2 was generated from pcDNA LATS2-FLAG (a gift from Tadashi Yamamoto) with the primers (H1923, 5′-acgcgtatgaggccaaagacttttcc-3′ and H1924, 5′-gtcgactacacgtacacaggctggc-3′). pCIneoLuc-PP1A and pCIneoLuc-PP2A were described previously (30). pCMV SPORT human TEAD4 (MHS1010-58163) was purchased from Open Biosystems. PCR was performed with the primers (H2710, 5′-gaattcgagggcacggccggcaccat-3′ and H2710, 5′-gtcgactcattctttcaccagcctg-3′). The PCR product was ligated into EcoRI/Sall sites of pCIneoLuc to generate pCIneoLuc-TEAD4. pCMV alkaline phosphatase is a gift from Sumiko Watanabe. PCR was performed with the primers (H3159, 5′-gcggccgcttaagtgaacaactagtgcca-3′ and H3160, 5′-caattgagatctttcacaaattttgtaatc-3′) on piLenti-siRNA-GFP (Applied Biological Materials Inc.) to amplify GFP-2A-puromycin. The product was digested with SpeI/MfeI and ligated into NheI/EcoRI of pLL3.7 to generate pLL3.7-GFP-2A-puro. PCR was performed with the primers (H3161, 5′-gcggccgcccccttcaccgagggcctattt-3′ and H3162, agatctagactattctttcccctgcactgt-3′) on pLKO1-shYAP2 (a gift from Kunliang Guan; Addgene plasmid, 27369; ref. 6). The product was digested with NotI/XbaI and ligated into the same sites of pLL3.7-GFP-2A-shYAP2. Human p73 cDNA was obtained from Open Biosystem (40125802). The coding region was amplified by PCR with the primers (H2074, 5′-acgcgtatggcccagtccaccgccac-3′ and H2075, 5′-gtcgactcagtggatctcggcctcc-3′), digested by MluI/Sall, and ligated into pCIneoFHF to generate pCIneoFHF-p73. pGL2 MDM2 reporter was a gift of Hitoshi Okazawa (Tokyo Medical and Dental University; ref. 36). PCR was performed with the primers (H2781, 5′-gacgtcgacgtgcggtctctctctgtt-3′ and H2782, 5′-gcggccgctattctagaaattcagggccgggattctc-3′) on piPSC-Nanog (SBI System Biosciences) to amplify 2A self-cleaving peptide and with the primers (H2862, 5′-gctagccccaacatgcctgaaccctctaagtct-3′ and H2863, 5′-gtcgaccttgtacagctcgtccatgccgcc-3) on H2B-mCherry. The former product was digested with SalI/NotI and ligated into the same sites of pCIneoFH to generate pCIneoFH-2A, and the later product was digested with NheI/Sall and ligated into the same sites of pCIneoFH-2A to generate pCIneo-H2B-mCherry-2A. SmaI/NotI fragment from pDsRed-N1 (Clontech Laboratories) was ligated into the same sites of pIRES (Clontech Laboratories) to generate pIRES-DsRedN1. NheI/PmlI fragment from pLL3.7 K122 and PmlI/MfeI fragment from pIRES-DsRedN1 were ligated into NheI/EcoRI sites of pLL3.7 to generate pLL3.7-ires-DsRedN1. NheI/NotI fragment from pCIneo-H2B-mCherry-2A was ligated into the same sites of pLL3.7-ires-DsRedN1 to generate pLL3.7-H2B-mCherry-2A. PCR was performed with the primers (H3112, 5′-tctagactagttggtaaagccaccatggaa-3′ and H3113, 5′-ctagtcgacggcgatctttccgcccttctt-3′) on pGL3 (Promega) to amplify luciferase. The product was digested with XbaI/NotI and ligated into the same sites of pLL3.7-H2B-mCherry-2A to generate pLL3.7-H2B-mCherry-2A-luciferase.
Antibodies and reagents
Rat monoclonal anti-YAP1 was described previously (28). Other antibodies and reagents were obtained from commercial sources: mouse anti-PARP (51-6639GR) and mouse anti-BAX (610982) (BD Pharmingen); mouse anti-α-tubulin (T9026), mouse anti-β-actin (A1978), Hoechst 33342, propidium iodide, trypan blue solution, thiazolyl blue tetrazolium bromide, and SIGMAFAST 3,3′-diaminbenzidine tablets (D4293) (Sigma-Aldrich); anti-DYKDDDDK-tag antibody (014-22383), anti-DYKDDDDK-tag beads (016-22784), Phos-tag acrylamide and imatinib (Wako Pure Chemical Industries); mouse anti-phospho-Histone H2A.X (Ser139) (05-636) (Merck Millipore); rabbit monoclonal anti-YAP1 (14074), rabbit monoclonal anti-YAP1/TAZ (D24E4; 8418), rabbit polyclonal anti-phospho-S127 YAP1 (4911), rabbit polyclonal anti-phospho-S909/S871 LATS1/2 (9157), and rabbit monoclonal anti-phospho-T1079/T1041 LATS1/2 (D57D3) (8654) (Cell Signaling Technology); and d-luciferin (Summit Pharmaceuticals International).
Cell cultures and transfection
HEK293FT and ARPE-19 cells were cultured in Dulbecco's Modified Eagle Medium containing 10% fetal bovine serum (FBS) and 10 mmol/L Hepes-NaOH pH7.4 under 5% CO2 at 37°C. Human MM IM-9 and human chronic myeloid leukemia K562 cells were cultured in RPMI1640 medium supplemented with 10% FBS, and 10 mmol/L Hepes-NaOH pH7.4. HEK293FT cells were purchased from Invitrogen. ARPE-19 and IM-9 cells were obtained from Department of Ophthalmology and Department of Hematology, Tokyo Medical and Dental University, respectively. Cell authentication was performed in 2016 using the short tandem repeat analysis by JCRB Cell Bank (National Institutes of Biomedical Innovation, Health and Nutrition). Mycoplasma contamination was tested by use of e-Myco Mycoplasma PCR Detection Kit (iNtRON Biotechnology). DNA transfection was performed using Lipofectamine 2000 (Invitrogen) and polyethylenemine “Max” (Polysciences). ARPE-19 reporter cells and control reporter cells were prepared using pLL3.7 K122 FH-YAP1-TEAD-responsive-promoter-H2B-mCherry and pLL3.7 K122 control-TEAD-responsive-promoter-H2B-mCherry. HEK293FT cells expressing mCherry-Halo-YAP1 was generated by use of the retrovirus vector, pQCXIP-mCherry-Halo-YAP1, and puromycin. YAP1-depleted and luciferase-expressing IM-9 cells were prepared by use of pLL3.7-GFP2A-shYAP2 and pLL3.7-H2B-mCherry-2A-luciferase, respectively.
Chemical library screening
For the first screening, we used 18,606 chemical compounds owned by the Tokyo Medical and Dental University Chemical Biology Screening Center. ARPE-19 reporter cells were plated at 3.5 × 103 cells/well in a 96-well plate. Each plate contained 90 μL medium. 24 hours later, each compound was applied to the cells to a final concentration of 10 μmol/L. The original stock, whose concentration was 10 mmol/L in DMSO, was diluted at 1:100 by the medium and 10 μL of the diluted sample was added to each plate. As a negative control, DMSO was similarly diluted and added to the cells. Seventy-two hours later, the cells were washed with PBS and fixed with 4% formaldehyde/PBS for 15 min. Then, the formaldehyde/PBS was changed to PBS containing 1 g/L Hoechst 33342 to visualize the nuclei. Green and red fluorescence intensities of 200 cells in each well were measured by use of ArrayScan VTI (Thermo Fisher Scientific). The ratio of mCherry intensity over GFP intensity was calculated by using BioApplication Compartment Analysis V4 (Thermo Fisher Scientific). The formula for Z-score calculations is as follows: Z-score = (the result obtained for each cell − the median result obtained for each assay plate) / normalized interquartile range (NIQR, namely 0.7413 × (the result corresponding to the first 25% when ranked in order—the result corresponding to the first 75% when ranked in order) of each assay plate. Each assay plate contained 80 samples. In the first screening, we selected 124 compounds that gave Z-scores larger than 3.5 as positive samples. We subsequently applied 124 compounds to ARPE-19 cells and reanalyzed them. In the second screening, we used the same experimental method as in the first screening but compared mCherry/GFP ratio of each sample with that of the DMSO-treated control sample to select 47 compounds that augmented mCherry/GFP ratio by more than 1.5-fold as positive samples. In the third screening, we evaluated the effects of 47 compounds on a YAP1-depedent TEAD reporter assay in HEK293FT cells and YAP1 phosphorylation in ARPE-19 reporter cells. The overview of the screenings is shown in Supplementary Fig. S1.
Quantitative reverse transcription-PCR (qRT-PCR)
qRT-PCR analysis was performed using SYBR Green (Roche) and ABI7500 Real-Time PCR system (Applied Biosystems). The used primers are as follows: 5′-cagcacagcaaattctccaa-3′ and 5′-tggattttgagtcccaccat-3′ for YAP1; 5′-ggcaggggagagtgatacaga-3′ and 5′-gaagccaattctcacgaaggg-3′ for MDM2; 5′-gggcatctggaccctcctac-3′ and 5′-tcctttcacctggaggacag-3′ for FAS; 5′-atgttttctgacggcaacttc-3′ and 5′-atcagttccggcaccttg-3′ for BAX; 5′-gtccttcgtgtgggctacat-3′ and 5′-cgaggatcttcggttgacat-3′ for LATS1; 5′-ttcatccaccgagacatcaa-3′ and 5′-ctccatgctgtcctgtctga-3′ for LATS2; 5′-ccaatgacaacgcctcctg-3′ and 5′-tggtgcagccagaaagctc-3′ for CTGF; 5′-agcctcgcatcctatacaacc-3′ and 5′-ttctttcacaaggcggcactc-3′ for CYR61; and 5′-ccactcctccacctttgac-3′ and 5′-accctgttgctgtagcca-3′ for GAPDH.
RNA interferences were performed by use of Lipofectamine RNAiMAX (Life Technologies). The double-strand RNAs (dsRNAs) used are as follows: human YAP1#1, s20366, human YAP1#2, s20367, human LATS1, s17392, human LATS2, s25505, and Silencer Select Negative Control No. 2 siRNA (Ambion).
A total of 4 × 106 HEK293FT cells and HEK293FT cells expressing mCherry-Halo-YAP1 were plated on the collagen-coated 6-cm dish and transfected 0.6 μg 8xGT-IIC-δ51LucII luciferase reporter and 0.6 μg pCMV alkaline phosphatase. Twenty-four hours later, the cells from one dish were replated into 24 wells in 24-well plate, treated with 10 μmol/L each compound for 18 hours, and harvested. The cells were lysed in 200 μL of the lysis buffer (25 mmol/L Tris-HCl pH7.4, 2 mmol/L DTT, 2 mmol/L EDTA, 4 mmol/L EGTA, 4 mmol/L MgCl2, 0.1%(w/v) Triton X-100, and 10 mg/L (Amidinophenyl)methanesulfonyl fluoride (APMSF). The luciferase activity and the alkaline phosphatase activity were assayed with PicaGene (Toyo Ink) and CDP-Star (Roche) and measured by ARVO MX (PerkinElmer). The alkaline phosphatase assay was performed in 90 mmol/L glycine buffer pH10.5 with 0.1 mmol/L ZnCl2.
ARPE-19 reporter cells were plated at 2 × 105 cells/6-cm dish. Forty-eight hours later, the cells were harvested, washed with ice-cold PBS twice, suspended in 200 μL of the extraction buffer (10 mmol/L Hepes-NaOH pH8.0, 1.5 mmol/L MgCl2, 10 mmol/L KCl, 0.34 mol/L sucrose, 10% (v/v) glycerol, 0.05% (v/v) NP40, and 10 mg/L APMSF), kept on ice for 5 minutes, and suspended up-down with 200 μL yellow tip five times. Suspension (66.6 μL) was saved as the total lysate. The remaining sample was centrifuged at 800 × g for 5 minutes at 4°C. The supernatant was further centrifuged at 20,000 × g for 10 minutes and the final supernatant was used as the cytosol fraction. The pellet was resuspended in 133.3 μL of the extraction buffer and centrifuged again at 800 × g for 4 minutes. The pellet was resuspended in 133.3 μL of the extraction buffer and used as the nuclear fraction.
ARPE-19 reporter cells were plated at 1.5 × 105 cells/well in a 6-well plate and treated with 10 μmol/L each compound. Twenty-four hours later, the cells were harvested, and the cell lysates were analyzed by use of Phos-tag acrylamide and PolyVinylidene DiFluoride membranes. The signals were measured with ImageJ.
LATS kinase assay
A total of 2 × 106 HEK293FT cells were plated in a 6-cm dish. Twenty-four hours later the cells were transfected with either pCIneoMyc-LATS1 or pCIneoMyc-LATS2. Twenty-four hours after the transfection, the cells were replated at 4 × 105 cells/well in a 12-well plate and treated with 10 μmol/L IBS003031. Twenty-four hours later, the cells were harvested and the cell lysates were immunoblotted with the indicated antibodies. The sequences around Ser909 of LATS1 and Ser871 of LATS2 and around Thr1079 of LATS1 and Thr1041 of LATS2 are identical, and the phosphorylated LATS1 and LATS2 are detected with the same antibodies.
Coimmunoprecipitation and LUMIER assay
HEK293FT cells were plated at 8 × 105 cells/well in a 6-well plate and were transfected with various combinations of expression vectors. After overnight culture, IBS003031 was added to a final concentration of 10 μmol/L. Twenty-four hours later, the cells were harvested and lysed in the lysis buffer (25 mmol/L Tris-HCl pH7.4, 150 mmol/L NaCl, 10 mmol/L MgCl2, 2 mmol/L EDTA, 10% (v/v) glycerol, 1% (v/v) Triton X-100, 10 mg/L APMSF, 10 mg/L leupeptin, and 10 mg/L pepstatin A). After centrifugation at 20,000 × g for 10 minutes, the supernatant was collected and the immunoprecipitation was performed with anti-DYKDDDDK-tag beads. For LUMIER assay, HEK293FT cells were plated at 8 × 105 cells/well in a 6-well plate and were transfected with various combinations of the expression vectors. Twenty-four hours later, the cells from each well were replated into 6 wells of a 24-well plate and treated with DMSO or 10 μmol/L IBS003031 for 24 hours. The immunoprecipitation was performed with anti-DYKDDDDK tag beads. The immunoprecipitates were washed vigorously and the luciferase activity was measured as described for the reporter assay.
Cell proliferation assay
The colorimetric MTT assay was performed with thiazolyl blue tetrazolium bromide. The insoluble formazan was dissolved with 50 μL DMSO, and the absorbance at 570 nm was measured with SmartSpec 3000 (Bio-Rad).
Lactate dehydrogenase (LDH) assay
K562 cells were plated at 3 × 106 cells/dish in a 6-cm dish and transfected with control siRNA or YAP1#1 siRNA. Forty-eight hours later, the cells were replated at 3 × 104 cells/well in a 96-well plate. The cells were treated with either control DMSO or 10 μmol/L IBS003031 for 48 hours. As K562 cells, although hematopoietic cells, attached to the plate, 50 μL cell-free medium was collected and LDH activity was assayed using 2-p-iodophenyl-3-p-nitrophenyl tetrazolium chloride, N-methylphenazonium methyl sulfate, nicotinamide adenine nucleotide, and lactic acid (37). All these reagents were obtained from Tokyo Chemical Industry Co., Ltd.
Detection of sub-G1 population
IM-9 cells were harvested and fixed in ice-cold 70% (v/v) ethanol, washed with PBS, and resuspended in PBS containing 10 mg/L propidium iodide and 1 g/L RNaseA. The sub-G1 population was evaluated with FACS Calibur (BD Biosciences). The data were analyzed with BD CellQuest Pro Software.
All animal experiments were approved by the Tokyo Medical and Dental University Animal Care and Use Committee. A total of 2 × 106 IM-9 cells carrying luciferase were injected intraperitoneally into female C.B17/Icr-scidJcl mice (5 weeks old) (Clea Japan. Inc.). Tumor progression was monitored by bioluminescence imaging on IVIS Lumina system (PerkinElmer). Control DMSO or IBS003031 (5 mg/kg body weight) was intraperitoneally daily. To evaluate the effect of IBS003031 on YAP1 in vivo, 3.5 × 106 IM-9 cells were subcutaneously transplanted with Matrigel (BD Biosciences). Twenty-eight days later, after tumor formation, DMSO or IBS003031 (5 mg/ kg body weight) was intraperitoneally injected. One day later, mice were sacrificed and tumors were excised. Tumors were fixed with 4% (w/v) formaldehyde in PBS for overnight and dehydrated in 70% (v/v) ethanol for another overnight. After embedded in paraffin blocks and sliced into 4-μm slices, the samples were deparaffined and heated in citrate buffer to activate antigens. After endogenous peroxidase was inactivated, the samples were blocked with 3% (w/v) BSA in PBS and immunostained with rabbit monoclonal anti-YAP1 antibody. Signals were labeled by using R.T.U. VECTASTAIN kit (Vector Laboratories Inc.) and SIGMAFAST 3,3′-diaminobenzidine tablets.
Statistical analyses were performed with the Student t test for the comparison between two samples and analysis of variance with the Dunnett test for the multiple comparison using the GraphPad Prism 5.0 (GraphPad Software).
The generation of ARPE-19 cells stably expressing the TEAD-responsive reporter
YAP1 interacts with various transcription factors including TEAD, SMAD, Runx2, and p73. In the in vitro reporter assays, YAP1 enhances the reporter activities that respond to these transcription factors. However, the recent study using ChIP-seq demonstrates that TEAD is associated with 75% of YAP1 peaks and that TEAD is the most prominent partner of YAP1 (38). Moreover, the TEAD reporter assay exhibits the most remarkable activity and is the easiest to monitor. We expressed mCherry-fused histone2B (H2B-mCherry) under the TEAD-responsive promoter and FLAG-tagged YAP1 linked to the internal ribosome entry site (IRES) and GFP under the cytomegalovirus promoter (CMV) in immortalized human retinal pigment epithelial ARPE-19 cells (Fig. 1A). We evaluated mCherry signals driven by the TEAD-responsive promoter. YAP1 knockdown suppressed mCherry signals, supporting that mCherry signals reflect YAP1 activity and that the ARPE-19 reporter cells can be used for the screening of YAP1 activators (Fig. 1B).
The identification of the compounds that enhance the H2B-mCherry expression
We applied 18,606 chemical compounds to ARPE-19 reporter cells and cultured them for 72 hours to select the compounds that changed the expression of H2B-mCherry (Supplementary Fig. S1). We obtained 124 putative enhancers in the initial screening, which gave Z-scores higher than 3.5 (Fig. 1C). For the second selection, we directly compared these compounds with the control and selected 47 compounds that augmented the signals by more than 1.5-fold. To confirm that the compounds enhanced the signal via YAP1, we applied the compounds to ARPE-19 control reporter cells expressing only the TEAD-responsive reporter. The effects of the compounds were remarkably attenuated, supporting that the exogenously expressed YAP1 is required for the compounds to enhance H2B-mCherry signal (Supplementary Fig. S2A). Considering the risk that the concentration employed might be too high, we applied representative enhancers at the lower dose. IBS003031 upregulated H2B-mCherry signals at 1 μmol/L and 3 μmol/L, whereas IBS012851 was active at 3 μmol/L (Supplementary Fig. S2B). Moreover, when YAP1 was knocked down in the reporter cells, the effects of the compounds were abolished (Fig. 1D). These findings corroborate that the compounds enhance H2B-mCherry signal via YAP1.
The effect on the TEAD reporter activity
We next confirmed the effects of these candidate compounds on the conventional TEAD-responsive reporter activity in HEK293FT cells. We prepared HEK293FT cells stably expressing YAP1 and after the confirmation that YAP1 overexpression did not show any significant effect on the cell viability (data not shown), performed the reporter assay. 6 compounds (IBS003031, IBS012851, CBI001316, IBS002470, IBS007607, and IBS014765) enhanced the reporter activity more than 1.75-fold (Fig. 2A, black columns; Fig. 2B, black dots). We designated these compounds as TEAD reporter enhancers (TRE). 6 compounds (IBS011345, IBS006077, IBS011342, IBS006808, IBS004557, and IBS001781) rather suppressed the reporter activity (Fig. 2A, white columns; Fig. 2B, white dots). We speculate that the discrepancy between the ARPE-19-based assay and the conventional reporter assay might be caused by the differences of cells and the duration of the treatment of the compounds.
The effects on the phosphorylation of YAP1
YAP1 is negatively regulated by the canonical Hippo pathway via the phosphorylation (8). To directly address whether and how compounds affect the phosphorylation of YAP1, we treated APRE-19 reporter cells with the compounds and ran the cell lysates on the Phos-tag SDS-PAGE. The lowest bands were detected with anti-FLAG antibody but not with anti-phospho-S127 YAP1 antibody (Supplementary Fig. S3A). The upper three bands were detected with both the antibodies. Moreover, LATS1/2 silencing increased the lowest band (Supplementary Fig. S3A, an arrowhead). These findings support that the lowest band corresponds to the unphosphorylated YAP1. We quantified signals by use of ImageJ and calculated the relative amount of the unphosphorylated YAP1. Among 6 TRE compounds (Fig. 3A and B, arrows), IBS014765 did not increase the unphosphorylated YAP1 (Fig. 3A and B, stars). Other 5 TRE compounds did increase unphosphorylated YAP1 (Fig. 3A and B). IBS003031 and IBS012851 remarkably decreased the phosphorylated YAP1 signal (Fig. 3A, white arrowheads). These compounds were speculated to activate YAP1 via the canonical Hippo pathway, which increases phosphorylated YAP1 and decreases unphosphorylated YAP1. Interestingly, the structures of IBS003031 and IBS012851 are similar (Fig. 3C). We have recently reported that ethacridine, which was originally identified as a TAZ activator, reduces phosphorylated YAP1 and increases unphosphorylated YAP1 (30). Ethacridine also has an acridine backbone. We speculate that these compounds activate YAP1 through the common target molecule. After we confirmed that IBS003031 enhanced the expression of YAP1 target genes, CTGF and CYR61, in U2OS cells, we decided to focus on IBS003031 (Supplementary Fig. S4).
The effect of IBS003031 on LATS1/2 activities
We postulated that IBS003031 might activate YAP1 through the inhibition of the phosphorylation by LATS kinases or the facilitation of the dephosphorylation. We first tested whether IBS003031 could inhibit LATS1/2 activities. We expressed Myc-LATS1 and Myc-LATS2 in HEK293FT cells and treated the cells with IBS003031. The phosphorylation at S909/S871 and T1079/T1041 of LATS1/2 is essential for the activities of LATS1/2 (39). We immunoblotted the cells lysates with the antibodies against phosphorylated-S909/S871 and -T1079/T1041 of LATS1/2. IBS003031 increased the expression of Myc-LATS1 and augmented the phosphorylation at S909 and T1079 (Fig. 4A, left). IBS003031 slightly increased the expression of Myc-LATS2 and enhanced the phosphorylation at S871 and T1041 (Fig. 4A, right). We next tested whether IBS003031 blocked the binding of YAP1 to LATS1/2. We expressed luciferase-fused YAP1 with FLAG-LATS1 and LATS2-FLAG, conducted the immunoprecipitation with anti-DYKDDDDK-tag antibody beads, and measured the coimmunoprecipitated luciferase activity (Fig. 4B). Consistent with the result in Fig. 3A, IBS003031 enhanced the expression of FLAG-LATS1 and LATS2-FLAG, and consequently increased the amount of FLAG-LATS1 and LATS2-FLAG in the immunoprecipitates (Fig. 4B, the bottom immunoblottings). Luc-YAP1 that was coimmunoprecipitated with FLAG-LATS1 and LATS2-FLAG was also higher under the treatment with IBS003031. Based on these findings, we conclude that it is unlikely that IBS003031 decreases YAP1 phosphorylation through inhibiting the activation of LATS kinases or blocking the interaction of YAP1 and LATS kinases.
The effect of IBS003031 on the dephosphorylation of YAP1
We next hypothesized that IBS003031 might promote the dephosphorylation of YAP1. We treated ARPE-19 reporter cells expressing FLAG-YAP1 with IBS003031 and analyzed the cell lysates on the Phos-tag SDS-PAGE (Fig. 5A). Under the treatment of IBS003031, phosphorylated YAP1 decreased and concomitantly unphosphorylated YAP1 increased in a time-dependent manner. We also performed the subcellular fractionation. IBS003031 increased unphosphorylated YAP1 in both the cytoplasm and the nucleus, supporting that YAP1 is dephosphorylated in the cytoplasm and then transported into the nucleus (Fig. 5B).
The effect of IBS003031 on the interaction of YAP1 with protein phosphatase (PP) 1A and PP2A
PP1A and PP2A are involved in the dephosphorylation of YAP1 (40, 41). We examined whether IBS003031 could strengthen the interaction of YAP1 with PP1A and PP2A. We expressed luciferase-fused PP1A and PP2A with FLAG-YAP1 in HEK293FT cells and measured the luciferase activity coimmunoprecipitated with FLAG-YAP1. IBS003031 slightly enhanced the interaction of YAP1 with PP1A but the difference was not significant, while it significantly increased the interaction between YAP1 and PP2A (Fig. 5C). We further confirmed that IBS003031 enhanced the interaction between YAP1 and PP2A in U2OS cells (Supplementary Fig. S5). In conclusion, IBS003031 is supposed not to suppress the phosphorylation of YAP1 by LATS kinases but is likely to promote the dephosphorylation of YAP1.
The effect of IBS003031 on MM cells in vitro
YAP1 induces apoptosis in MM cells (19). In normal cells, upon DNA damage, ABL1 is released from 14-3-3 and enters the nucleus (42). The nuclear ABL1 phosphorylates YAP1 at tyrosine, enhances the association of YAP1 with p73, and upregulates p73-dependent proapoptotic gene transcription (18). In MM cells, due to the low expression of YAP1, the ABL1–YAP1–p73 axis is compromised (19). Certain MM cells express phosphorylated H2A.X termed as γ-H2A.X, which is a hallmark of DNA damage response, but escape apoptosis because of the low expression of YAP1 (43). However, when YAP1 is exogenously expressed or YAP1 is activated by the suppression of the Hippo pathway, MM cells begin to die. We presumed that YAP1 activators would trigger apoptosis in MM cells. In this study, we used IM-9 cells and treated them with various doses of IBS003031 and performed the MTT assay. IBS003031 decreased the viability in the dose-dependent manner (Fig. 6A, left). IM-9 cells expressed γ-H2A.X with no treatment, which means that IM-9 cells show DNA damage response at the basal condition (Fig. 6A, 0 μmol/L in the immunoblot for γ-H2A.X). IBS003031 enhanced Bax expression in a dose-dependent manner (Fig. 6A, the immunoblot for Bax). The treatment with 6 and 9 μmol/L IBS003031 apparently decreased γ-H2A.X expression, but as α-tubulin expression was also reduced, we speculated that the decrease of γ-H2A.X expression might be due to the decrease of the total cell number. We also confirmed in FACS analysis that IBS003031 increased sub-G1 population in a dose-dependent manner (Fig. 6B, top). YAP1 depletion reduced the IBS003031-induced sub-G1 population (Fig. 6B, bottom). Trypan blue exclusion assay further supported that IBS003031 caused cell death more strongly in parent IM-9 cells than in YAP1-depleted IM-9 cells (Fig. 6C, left). In the immunoblotting, IBS003031-induced Bax expression was attenuated by YAP1 depletion (Fig. 6C, right). Imatinib, ABL1 inhibitor, reduced the effect of IBS003031 in the MTT assay (Fig. 6D). All these findings support that IBS003031 induces apoptosis in IM-9 cells through ABL1 and YAP1. We also observed that IBS003031 suppressed cell viability and induced cell death in human myeloid leukemia K562 cells depending on YAP1 (Supplementary Fig. S6).
The effect of IBS003031 on p73-dependent transcription
IBS003031 increased the amount of MDM2, FAS, and BAX in the mRNA level, which are p73 target genes (Fig. 7A). Accordingly, in HEK293FT cells, IBS003031 stimulated MDM2-promoter luciferase reporter activity and YAP1 knockdown antagonized the effect of IBS003031 (Fig. 7B). Moreover, IB003031 enhanced the coimmunoprecipitation of FLAG-p73 and GFP-YAP1 from HEK293FT cells (Fig. 7C).
The effect of IBS003031 on MM cells in vivo
We transplanted luciferase-expressing IM-9 cells into the immunocompromised mice. Twenty days after transplantation, control DMSO or IB003031 (a final concentration of 5 mg/kg body weight) was injected intraperitoneally daily for 1 week (Fig. 8A). Tumor growth was evaluated in the bioluminescence imaging (Fig. 8B). The trend of the suppression of the growth by IBS003031 was observed (Fig. 8C). In the parallel experiment, we subcutaneously transplanted IM-9 cells and after tumor was formed, DMSO or IBS003031 was injected intraperitoneally (Supplementary Fig. S7A). One day later, tumor was isolated and endogenous YAP1 was immunostained. IBS003031 indeed increased nuclear YAP1 (Supplementary Fig. S7B).
In this study, we expressed YAP1 and TEAD-responsive fluorescent protein reporter in human retinal pigment ARPE-19 cells and used them to detect small chemical compounds that enhance the reporter activity (Fig. 1). ARPE-19 cells are immortalized cells, but show contact inhibition, suggesting that the Hippo pathway is intact. In the initial screening by using 18,606 compounds, we tested 80 compounds as one set in a 96-well plate and calculated Z-score in each set. The first screening gave 124 compounds with Z-score higher than 3.5 (Fig. 2). In the second screening, we directly compared them with the control and further selected 47 compounds that enhanced the reporter by more than 1.5. These compounds did not enhance the fluorescence reporter in ARPE-19 cells expressing only TEAD-responsive reporter, supporting that the compounds upregulated the reporter via exogenously expressed YAP1. In the conventional reporter assay using HEK293FT cells and TEAD-responsive luciferase reporter, 6 compounds attenuated the reporter activity (Fig. 2). ARPE-19 reporter cells were exposed to the compounds for 72 h, whereas HEK293FT cells were treated with the compounds for 18 h. The reason why 6 compounds show the unexpected effect may be the difference of the experimental condition.
The phosphorylation by LATS kinases is best characterized and the most prominent regulatory mechanism of YAP1. According to the canonical Hippo pathway, YAP1 activators are predicted to decrease YAP1 phosphorylation and increase the amount of unphosphorylated YAP1. Among 6 compounds that remarkably enhanced the reporter activity in the conventional reporter assay, IBS003031 and IBS012851, faithfully to the canonical Hippo pathway, increased the amount of unphosphorylated YAP1 and decreased that of phosphorylated YAP1 (Fig. 4). Intriguingly, the structures of both compounds and ethacridine that we previously reported as a YAP1/TAZ activator are similar. We speculate that these compounds target the same or the related molecule.
In this study, we wanted to conclude whether bona fide YAP1 activators can be obtained through the screening with ARPE-19 reporter cells. To facilitate the analysis, we focused on IBS003031 that remarkably increased the amount of unphosphorylated YAP1 and decreased that of phosphorylated YAP1. IBS003031 enhanced the expression of LATS kinases. IBS003031 did not inhibit the activities of LATS kinases and did not block the interaction between YAP1 and LATS kinases (Fig. 3). We could conclude that the IBS003031-induced upregulation of unphosphorylated YAP1 is not caused by the inhibition of the phosphorylation. On the other hand, IBS003031 enhanced the interaction of YAP1 with protein phosphatases and induced the dephosphorylation of YAP1 (Fig. 4 and Supplementary Fig. S5). The subcellular fractionation experiment revealed that unphosphorylated YAP1 increased not only in the nucleus but also in the cytoplasm under the treatment with IBS003031. This finding suggests that the import of unphosphorylated YAP1 into the nucleus is limited and that a slight increase in the nuclear unphosphorylated YAP1 may be sufficient to upregulate gene transcription. The precise mechanism how IBS003031 enhances the interaction between YAP1 and protein phosphatases is not clear. Apoptosis-stimulating of p53 protein 2 functions as a scaffold for YAP1 and PP1A to promote the interaction of both proteins (40). α-Catenin inhibits the interaction between YAP1 and PP2A (41). Proteomic studies have revealed that the protein network underlying the Hippo pathway includes several interactions with serine and threonine protein phosphatases (44). IBS003031 may work through these interactions.
YAP1 is primarily thought to be an oncoprotein (9). The expression of nuclear YAP1 correlates with poor prognosis in cancer patients. Numerous reports demonstrate that YAP1 activation leads to oncogenesis. Meanwhile, YAP1 is implicated in p73-dependent transcription of proapoptotic genes (17). YAP1 depletion suppresses cell death and promotes tumor formation in certain cancers. ABL1 is important for YAP1 to adopt the fate as a tumor suppressor (18). DNA damage induces the nuclear localization of ABL1, which phosphorylates YAP1. Tyrosine-phosphorylated YAP1 interacts with p73 and upregulates p73-dependent transcription. A recent study has revealed the tumor suppressive aspect of YAP1 in MM cells (19). MM cells, in which YAP1 expression is low, do not undergo apoptosis in response to DNA damage, but begin to die when YAP1 is exogenously expressed or upregulated by the suppression of mammalian Ste20-like kinase 1. Based on this report, we speculated that YAP1 activator could induce cell death in MM cells. As expected, IBS003031 triggers cell death in human MM IM-9 cells (Fig. 5). YAP1 depletion and ABL1 inhibitor blocked this effect. IBS003031 strengthened the interaction between YAP1 and p73 (Fig. 6). All these findings indicate that IBS003031 induces apoptosis in IM-9 cells via the ABL1–YAP1–p73 axis. We further confirmed that IBS003031 suppressed the growth of IM-9 cells in vivo and increased the nuclear YAP1 (Fig. 7 and Supplementary Fig. S7). The importance of the ABL1–YAP1–p73 axis as a barrier against hematological malignancies was discussed in the previous paper (19). We found that IBS003031 induced cell death in human myeloid K562 cells via YAP1. Therefore, YAP1 activation may be useful to control not only MM cells but also other blood cancer cells (Supplementary Fig. S6). The study including animal experiments will be awaiting.
In conclusion, the compound screening by use of ARPE-19 reporter cells provided us with a novel YAP1 activator. We have also demonstrated the possibility that YAP1 activation can be a choice in the treatment of MM cells.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: J. Maruyama, X. Jiang, M. Ishigami-Yuasa, Y. Hata
Development of methodology: J. Maruyama, K. Inami, F. Michishita, K. Nakagawa
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Maruyama, K. Inami, F. Michishita, H. Iwasa, K. Yamamoto, Y. Hata
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Maruyama, K. Inami, F. Michishita, H. Kagechika
Writing, review, and/or revision of the manuscript: J. Maruyama, X. Jiang, H. Iwasa, M. Ishigami-Yuasa, H. Kagechika, Y. Hata
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Nakagawa, H. Kagechika, N. Miyamura, J. Hirayama, H. Nishina
This work was supported by research grants from the Japan Society for the Promotion of Science (JSPS; 26460359 to H. Iwasa and 26293061 to Y. Hata) and the Mitsubishi Foundation (26138).
The authors acknowledge Hideyuki Saya, Tadashi Yamamoto, Sumiko Watanabe, and Hiroshi Sasaki for materials.
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.