Tumor Suppressor HIPK2 Regulates Malignant Growth Via

Published OnlineFirst June 22, 2016; DOI: 10.1158/0008-5472.CAN-15-3310 Cancer Molecular and Cellular Pathobiology Research

Tumor Suppressor HIPK2 Regulates Malignant Growth via Phosphorylation of Notch1 Eun-Jung Ann1, Mi-Yeon Kim1,2, Ji-Hye Yoon1, Ji-Seon Ahn1, Eun-Hye Jo1, Hye-Jin Lee1, Hyun-Woo Lee3, Hyeok-Gu Kang3, Dong Wook Choi4, Kyung-Hee Chun3, Ji Shin Lee5, Cheol Yong Choi4, Adolfo A. Ferrando2,6,7, Keesook Lee1, and Hee-Sae Park1

Abstract

The receptor Notch1 plays an important role in malignant pro- maintained at a low level through proteasomal degradation. HIPK2 gression of many cancers, but its regulation is not fully under- phosphorylated the residue T2512 in Notch1-IC. Somatic muta- stood. In this study, we report that the kinase HIPK2 is responsible tions near this residue rendered Notch1-IC resistant to degradation, for facilitating the Fbw7-dependent proteasomal degradation of as induced either by HIPK2 overexpression or adriamycin treat- Notch1 by phosphorylating its intracellular domain (Notch1-IC) ment. In revealing an important mechanism of Notch1 stability, within the Cdc4 phosphodegron motif. Notch1-IC expression was the results of this study could offer a therapeutic strategy to higher in cancer cells than normal cells. Under genotoxic stress, block Notch1-dependent progression in many types of cancer. Notch1-IC was phosphorylated constitutively by HIPK2 and was Cancer Res; 76(16); 1–13. 2016 AACR.

Introduction tion through the proteasome (12–14). However, the kinase that phosphorylates the T2512 residue has not been conclusively Notch1 is a highly conserved transmembrane protein that plays identified. a crucial role in cancer development, cancer cell survival, prolif- In this study, we evaluated the crosstalk between HIPK2 and eration, and differentiation, and fate determination of cancer cells Notch1 signaling during tumorigenesis. We identified that che- (1). Notch1 signaling is aberrantly activated in breast cancer, and motherapeutic drug–induced HIPK2 phosphorylated the T2512 increased expression of Notch1 intracellular domain (Notch1-IC) residue in the Cdc4 phosphodegron (CPD) motif of Notch1-IC, is associated with poor survival in patients with various cancers, thus facilitating its degradation by Fbw7 ubiquitin ligase. We also including breast cancer (2–7). Moreover, proliferation of cells found that tissues of patients with breast cancer showed increased derived from these cancers can be suppressed by pharmacologic level of Notch1-IC and decreased levels of phosphorylated inhibition of Notch1. Therefore, preventing the generation of Notch1-IC T2512, HIPK2, and Fbw7. The chemotherapeutic Notch1-IC is a potential strategy for treating various cancers (8, 9). agent significantly decreased the growth, invasion, and tumori- Fbw7 binds to Notch1-IC via its WD40 domains and mediates its genic activity of Notch1-IC–expressing cells. However, this was ubiquitination and degradation by the proteasome system, which not observed in cells expressing Notch1-IC T2512A and the promotes proline, glutamic acid, serine, and threonine rich region somatic mutants Notch1-IC P2513L and Notch1-IC P2515 frame- (PEST) domain-dependent Notch1-IC degradation (10, 11). shift (P2515 fs) because of the decreased phosphorylation of Phosphorylation of the T2512 residue of Notch1-IC is important Notch1-IC. These data suggested that HIPK2-induced phosphor- for its recognition by Fbw7, which in turn facilitates its degrada- ylation of Notch1-IC at the T2512 residue plays an important role in cancer prevention and could be a potential biomarker for diagnosing breast cancer treatment. 1Hormone Research Center, School of Biological Sciences and Tech- nology, Chonnam National University, Gwangju, Republic of Korea. 2Institute for Cancer Genetics, Columbia University Medical Center, Materials and Methods New York, New York. 3Department of Biochemistry and Molecular Biology, Brain Korea 21 Project for Medical Sciences,Yonsei University Immunoblotting, qPCR, in vitro binding assay, immunofluo- 4 College of Medicine, Seoul, Korea. Department of Biological rescence, cell viability, and cell growth are described in Supple- Sciences, Sungkyunkwan University, Suwon, Republic of Korea. 5Department of Pathology, Chonnam National University Medical mentary Materials and Methods. School and Research Institute of Medical Sciences, Gwangju, Republic of Korea. 6Department of Pathology, Columbia University Medical Cell culture and transfection 7 Center, New York, New York. Department of Pediatrics, Columbia MDA-MB-231, MCF7 human breast cancer cells, and HEK293 University Medical Center, New York, New York. human embryonic kidney cells were purchased from ATCC in Note: Supplementary data for this article are available at Cancer Research 2008. Cells were cultured for less than 2 months before reinitiat- Online (http://cancerres.aacrjournals.org/). ing culture and were subjected to routine microscopic inspection þ þ Corresponding Author: Hee-Sae Park, Chonnam National University, 300, to check for stable phenotype. HIPK2 / and HIPK2 / mouse Youngbongdong, Bukku, Gwangju 500-757, Republic of Korea. Phone: 826- embryonic fibroblasts (MEF) were periodically authenticated by 2530-0021; Fax: 826-2530-2199; E-mail: [email protected] morphologic inspection and cytometric analysis. HEK293 and þ þ doi: 10.1158/0008-5472.CAN-15-3310 MCF7 cells and HIPK2 / and HIPK2 / MEFs were cultured in 2016 American Association for Cancer Research. DMEM supplemented with 10% FBS and 1% penicillin/

www.aacrjournals.org OF1

Ann et al.

streptomycin at 37 C and 5% CO2. MDA-MB-231 cells were vided by the Biobank of Chonnam National University Hwasun cultured in RPMI medium supplemented with 10% FBS and Hospital, a member of the Korea Biobank Network. The tissue 1% penicillin/streptomycin. For transfection, cells were cultured sections were incubated overnight with primary antibodies overnight until they reached 50% to 60% conﬂuence and were against Notch1-IC and HIPK2 at 4C. The sections were then subsequently transfected with appropriate expression vectors incubated with horseradish peroxidase–conjugated secondary containing speciﬁed combinations of DNA sequences using Lipo- anti-rabbit antibodies for 1 hour at 25C. The ABC Color Devel- fectamine 2000 reagent. opment Kit was procured from Invitrogen.

Luciferase reporter assay Statistical analysis The luciferase assay was conducted as described previously P-values were determined using two-tailed unpaired Student t (15). The cells were lysed using chemiluminescent lysis buffer test. P < 0.05 was considered statistically signiﬁcant. For all and were analyzed using a luminometer (Berthold). The activity experiments with error bars, SD was calculated to indicate the of the luciferase reporter protein in the transfected cells was variation within each experiment and data, and values represent normalized to the b-galactosidase activity in the same cells. mean SD.

Protein stability assay Results Cells were incubated overnight until they reached 50% to 60% Adriamycin regulates Notch1 and HIPK2 signaling in breast fl con uence. Next, 0.1 mmol/L cycloheximide was added to the cancer cells medium to inhibit the cells from synthesizing new proteins. Adriamycin and N-[(3,5-difluorophenyl)acetyl]-L-alanyl-2- The cells were treated with cycloheximide at different time points phenyl]glycine-1,1-dimethylethyl ester (DAPT) combination (0, 0.5, 1, 2, 4, and 6 hours) and were lysed in RIPA buffer. therapy remarkably reduced the viability of MDA-MB-231, MCF7, Endogenous Notch1-IC levels were determined by immunoblot- and HEK293 cells (Supplementary Fig. S1A–S1C). ting with anti-Notch1-IC antibody (1:3,000 dilution). Western blot analysis showed that Notch1-IC protein level in adriamycin-induced HIPK2 was dramatically decreased in a dose- Immunocomplex kinase assay and time-dependent manner in MDA-MB-231 cells (Fig. 1A and The kinase assay was conducted as described previously (16). B). In addition, other genotoxic agents, cisplatin, camptothecin, Cultured cells were harvested and lysed with buffer A. Immuno- and etoposide, also decreased Notch1-IC levels in MDA-MB-231 complex kinase assays were conducted by incubating the immu- cells (Supplementary Fig. S1D). We detected high levels of nopellets with 0.5 mg of substrate proteins in 20 mL of reaction Notch1-IC, Hes1, and Hes5 and a low level of HIPK2 in breast buffer for 30 minutes at 30 C. The phosphorylated substrates cancer cell lines. Interestingly, adriamycin treatment significantly were separated by SDS-PAGE and were quantified using the Fuji decreased the Notch1-IC level and the expression of its target BAS-2500 Phosphorimager. genes. Furthermore, adriamycin treatment increased the expression of HIPK2 (Fig. 1C), suggesting that adriamycin-induced Tumor xenografts upregulation of HIPK2 was inversely correlated to Notch1 All studies involving nude mice were approved by the Animal signaling. Care and Use Committee of Yonsei University Medical School To determine the inverse correlation between Notch1-IC and (approval number 2014-0026), and were performed in specific HIPK2 in breast cancer cells, we performed immunohistochem- pathogen-free facilities in accordance with the Guidelines for the ical analyses of these two proteins in 61 breast cancer samples Care and Use of Laboratory Animals of YUMS. The mice received using tissue microarrays (Fig. 1D). A highly significant negative subcutaneous injection of cells (density, 1 106 cells/mL) in each correlation between Notch1-IC and HIPK2 was observed in these flank under 150 mL saline:Zoletil:Rompun (7:1:1) anesthesia. The breast cancer tissues, in which 38 samples (62%) had low levels mice were randomly divided into groups (n ¼ 5 per group) that of HIPK2 and 43 samples (70%) had high levels of Notch1-IC. were untreated, or treated with intraperitoneal injections every 3 Expression of Notch1-IC was significantly higher than that of days (seven cycles) with 1.5 mg/kg adriamycin (Santa Cruz HIPK2 in breast cancer tissues, highlighting an evident inverse Biotechnology). The body weights were measured at 1-week correlation between HIPK2 and Notch1-IC expressions (Fig. 1D intervals and tumor sizes (from palpable tumor formation until and Supplementary Fig. S1E; Supplementary Table S1). Transcrip- termination) were measured every 2 to 3 days using calipers. tional activity of Notch1-IC in HEK293 cells was inhibited by Tumor volume was calculated using the following formula: length adriamycin in a dose-dependent manner (Supplementary Fig. 2 – width 0.5236. The mice were sacrificed in a 7.5% CO2 S1F S1H). Furthermore, we observed that adriamycin inhibited chamber, and their tumors were harvested for immunoblotting Notch1-IC–induced expression of luciferase reporter gene in and other analyses. MDA-MB-231 and MCF7 cells (Supplementary Fig. S1I and S1J). These results implied that adriamycin inhibited Notch1 Patient study and IHC signaling in breast cancer cells. All patient tumor samples were obtained under Institutional Review Board (IRB)–approved protocols (Chonnam National Adriamycin-induced HIPK2 suppresses Notch1 signaling University Hwasun Hospital IRB #2013-092). Written informed We determined whether adriamycin modulated Notch1 sig- consent was obtained from all patients. All samples were deiden- naling by inducing HIPK2. Knockdown of HIPK2 with shRNA tified, and protected health information was reviewed according constructs (shHIPK2#1 and shHIPK2#2) reduced HIPK2 levels in to the Health Insurance Portability and Accountability Act guide- a dose-dependent manner (Supplementary Fig. S2A and S2B). lines. The biospecimens and data used for this study were pro- HIPK2 knockdown prevented adriamycin-induced suppression of

OF2 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

Notch1-IC transcriptional activity in HEK293 and MDA-MB-231 cating that HIPK2 suppressed Notch1-mediated gene expression cells (Supplementary Fig. S2C and Fig. 2A). As expected, expres- (Supplementary Fig. S2L). sion of Hes1 decreased after adriamycin treatment (Supplemen- tary Fig. S2D). Adriamycin-induced suppression of Hes1 was HIPK2 regulates proteasomal degradation of Notch1-IC restored by HIPK2 knockdown (Fig. 2B). Furthermore, adriamy- Endogenous Notch1-IC was more effectively bound to RBP-Jk þ þ þ cin decreased the levels of endogenous Hes1 and Hes5 in HIPK2 / in HIPK2 / MEFs than in HIPK2 / MEFs (Fig. 3A). Further- þ MEFs but not in HIPK2 / MEFs (Fig. 2C; lanes 2 and 4). more, we observed that adriamycin-mediated accumulation of Collectively, these results suggested that HIPK2 induction was HIPK2 suppressed the endogenous interaction between Notch1- important in suppressing Notch1 signaling. IC and RBP-Jk compared with that in HIPK2 knockdown cells Next, we determined whether HIPK2 regulated the transcription- (Fig. 3B). Coimmunoprecipitation assay confirmed that HIPK2 alactivity of Notch1-IC. Ectopic expression of HIPK2 attenuated the interrupted the physical association between Notch1-IC and RBP- Notch1-IC transcriptional activity in a dose-dependent manner Jk (Supplementary Fig. S3A and S3B). Collectively, these (Supplementary Fig. S2E–S2G). HIPK2 suppressed the transcrip- results established that HIPK2 inhibited the Notch1-IC–RBP-Jk tional activity of Notch1-IC in MDA-MB-231 and MCF7 cells interaction. (Supplementary Fig. S2H and S2I). Notch1-induced expression of Surprisingly, HIPK2 significantly downregulated the levels of luciferase reporter gene was higher in HIPK2 knockdown cells than endogenous (Fig. 3A and B; lanes 1 and 2) and overexpressed in control cells (Fig. 2D). Transcriptional activity of Notch1-IC was Notch1-IC (Supplementary Fig. S3A and S3B; lane 6). Next, we þ þ higher in HIPK2 / MEFs than in HIPK2 / MEFs (Fig. 2E). determined whether HIPK2 regulated the level of Notch1-IC. WSB1, an E3 ubiquitin ligase, is involved in the degradation of Ectopic expression of HIPK2 reduced the level of Notch1-IC in HIPK2 (17). HIPK2 suppressed the transcriptional activity of a dose-dependent manner (Supplementary Fig. S3C); however, Notch1-IC; however, WSB1 expression blocked HIPK2-induced this was not observed with the ectopic expression of HIPK2– suppression of Notch1-IC transcriptional activity (Supplementary K221R mutant (Supplementary Fig. S3D). The steady-state level Fig. S2J). Notch1-IC transcriptional activity was inhibited by and half-life of endogenous Notch1-IC were higher in cells treated HIPK2, but not by kinase-dead HIPK2-K221R (Supplementary with shHIPK2 than in cells treated with shCon (Fig. 3C). In Fig. S2K). Furthermore, HIPK2 significantly reduced the expres- addition, the steady-state level and half-life of endogenous þ þ sion of the Notch1 target genes Hes1, Hes5, p27, and Myc, indi- Notch1-IC were higher in HIPK2 / MEFs than in HIPK2 /

ADR : 0 0.5 421 (mmol/L) A B ADR : 0 3 6 12 24 (h)

IB: Notch1-IC Notch1-IC IB: Notch1-IC Notch1-IC

IB: HIPK2 HIPK2 IB: HIPK2 HIPK2

IB: β-Actin β-Actin IB: β-Actin β-Actin

MDA-MB-231 MDA-MB-231

C MDA-MB-231 MCF7 HEK293 D High High High Low ADR : - + - + - +

IB: Notch1-IC Notch1-IC Notch1-IC WT

IB: HIPK2 HIPK2

Low Low Low High IB: Hes1 Hes1

IB: Hes5 Hes5 HIPK2

IB: β-Actin β-Actin

Case 1 Case 2 Case 3 Case 4

Human breast cancer tissue

Figure 1. Adriamycin (ADR) affects Notch1 signaling and cell viability of breast cancer cells. A, Western blotting of Notch1-IC and HIPK2 in MDA-MB-231 cells treated with the speciﬁed dose of adriamycin for 12 hours. B, Western blotting of Notch1-IC and HIPK2 in MDA-MB-231 cells treated with 2 mmol/L adriamycin for the indicated time. C, Western blotting of Notch1-IC, HIPK2, Hes1, and Hes5 in HEK293 cells treated with 2 mmol/L adriamycin for 12 hours. D, immunohistochemical staining of Notch1-IC and HIPK2 in serial tissue arrays of 61 breast cancer samples. Notch1-IC and HIPK2 levels in each sample were semiquantiﬁed as high or low in a double-blind manner, according to the standards presented.

www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF3

Ann et al.

A B Myc-Notch1-IC : - + + + MDA-MB-231 ADR : - - + + shHIPK2 : - - - + 70 4xCSL-Luc

60 IB: Hes1 Hes1 P = <0.001 50 2.3 (P = 0.001) P 3.4 ( = <0.001) IB: Myc Myc-Notch1-IC 40 P = 0.001 2.5 (P = 0.005) 1.7 (P = 0.032) 30 RLU (fold) IB: HIPK2 HIPK2 20 β-Actin 10 IB: β-Actin

0 Notch1-IC : - + + + + - - ADR : - - + + - + - shHIPK2 : - - - + + - +

C D 4xCSL-Luc E 4xCSL-Luc HIPK2+/+ HIPK2–/– 16 16 P = 0.038 P 14 = 0.044 ADR : - + - + 14 12 12 IB: Hes1 Hes1 10 10 8 8

IB: Hes5 Hes5 6 6 RLU (fold) RLU (fold) 4 4 IB: HIPK2 HIPK2 2 2 0 0 Notch1-IC : - + - + - - IB: β-Actin β-Actin Notch1-IC : + +

shCon shHIPK2 HIPK2+/+ HIPK2–/–

Figure 2. Adriamycin (ADR) suppresses the transcriptional activity of Notch1 by inducing HIPK2. A, luciferase reporter analysis of the transcriptional activity of Notch1-IC in the presence or absence of HIPK2 in MDA-MB-231 cells at 12 hours after adriamycin treatment. Error bars correspond to pooled data from three independent experiments. B, Western blotting of Hes1 in HEK293 cells transiently transfected with the indicated plasmids and treated with 2 mmol/L adriamycin for 12 hours. C, Western blotting of Hes1, Hes5, and HIPK2 in HIPK2þ/þ and HIPK2/ MEFs treated with 2 mmol/L adriamycin for 12 hours. D, luciferase reporter analysis of the transcriptional activity of Notch1-IC in HEK293 cells transfected with Notch1-IC, shCon, and shHIPK2. Error bars correspond to pooled data from three independent experiments. E, luciferase reporter analysis of the transcriptional activity of Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs transfected with Notch1-IC. Error bars correspond to pooled data from three independent experiments. RLU, relative luciferase units.

MEFs treated with cycloheximide with or without adriamycin restored HIPK2-induced decrease in the transcriptional activity (Supplementary Fig. S3E and Fig. 3D). The half-life of ectopically of Notch1-IC (Supplementary Fig. S4A). In addition, Western expressed Notch1-IC decreased in the presence of HIPK2 (Sup- blotting showed that Fbw7DF restored HIPK2-induced decrease in plementary Fig. S3F). These results indicated that Notch1-IC was Notch1-IC level (Supplementary Fig. S4B). Endogenous interac- rapidly degraded in the presence of HIPK2. tion between Notch1-IC and Fbw7 was signiﬁcantly higher in þ þ Proteasome inhibitor MG132 enhanced the transcriptional HIPK2 / MEFs than in HIPK2 / MEFs (Fig. 3G). Physical þ þ activity of endogenous Notch1-IC in HIPK2 / MEFs (Fig. 3E). association between Fbw7 and Notch1-IC was observed in cells Furthermore, the level of endogenous Notch1-IC was higher in expressing a constitutive active mutant of HIPK2 (HIPK2KD), but HIPK2 / MEFs (Fig. 3F; lanes 1 and 3). Moreover, treatment with not in a dominant-negative mutant of HIPK2 (HIPK2KDKR; MG132 increased the level of endogenous Notch1-IC by inhibit- Supplementary Fig. S4C). Furthermore, levels of ubiquitinated ing its proteasomal degradation (Fig. 3F; lane 2). Treatment with Notch1-IC increased in cells expressing wild-type (WT) HIPK2 but MG132 restored HIPK2-induced inhibition of Notch1-IC tran- not in cells expressing kinase-dead HIPK2-K221R mutant (Sup- scriptional activity (Supplementary Fig. S3G). Furthermore, treat- plementary Fig. S4D). These results indicated that the kinase ment with the proteasome inhibitors, MG132, ALLN, or lacta- activity of HIPK2 was crucial for the Fbw7 ubiquitin ligase- cystin, restored the adriamycin-mediated decrease in the level of mediated degradation of Notch1-IC. endogenous Notch1-IC (Supplementary Fig. S3H). These results indicated that HIPK2 decreased the stability of endogenous HIPK2-induced phosphorylation of Notch1-IC at the T2512 Notch1-IC through a proteasome-dependent pathway. residue promotes its degradation We determined the involvement of Fbw7 by using a dominant- Reciprocal GST pull-down assay showed that HIPK2 negative Fbw7 mutant lacking the F-box (Fbw7DF). Fbw7DF was physically associated with Notch1-IC (Supplementary

OF4 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

HIPK2+/+ HIPK2–/– ADR + ADR + A B shCon shHIPK2

IP: IP:

IB: RBP-Jk RBP-Jk IB: RBP-Jk RBP-Jk

HIPK2 HIPK2

Notch1-IC Cell Cell Notch1-IC lysates lysates RBP-Jk RBP-Jk

β-Actin β-Actin

CHX :0 0.51 2 4 6 (h) C D CHX :0 0.51 2 4 6 (h)

ADR + shCon Notch1-IC HIPK2+/+ + ADR Notch1-IC

ADR + shHIPK2 Notch1-IC HIPK2–/– + ADR Notch1-IC

IB : Notch1-IC IB: Notch1-IC

β β IB: -Actin -Actin IB: β-Actin β-Actin

β IB: β-Actin β-Actin IB: β-Actin -Actin

ADR + shCon HIPK2+/+ + ADR 1.2 ADR + shHIPK2 1.2 HIPK2–/–+ ADR

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4 t 1/2 = 1 hr t = 1.5 hr 0.2 1/2 0.2

t levels Relative Notch1-IC 1/2 = 4.1 hr t

Relative Notch1-IC levels levels Relative Notch1-IC = 3.2 hr 0.0 1/2 0.0 –0.2 –0.2 0 0.5 6421 64210.50 Hours after CHX treatment Hours after CHX treatment

E F G HIPK2+/+ HIPK2–/– 4xCSL-Luc P < 0.001 3.5 IP: +/+ –/– 3.0 HIPK2 HIPK2 2.5 MG132 : - + - + IB: Fbw7 Fbw7 2.0 1.5 IB: Notch1-IC Notch1-IC Notch1-IC RLU (fold) 1.0

0.5 β β IB: -Actin -Actin Fbw7 0.0 Cell - - MG132 : + + lysates HIPK2 HIPK2+/+ HIPK2–/–

β-Actin

Figure. 3. HIPK2 decreases the stability of Notch1-IC through proteasome-dependent degradation. A, endogenous coimmunoprecipitation of RBP-Jk with Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs. B, endogenous coimmunoprecipitation of RBP-Jk with Notch1-IC in HEK293 cells expressing shCon and shHIPK2, and treated with 2 mmol/L adriamycin for 12 hours. C, levels of Notch1-IC in HEK293 cells transfected with shCon and shHIPK2 and treated with 2 mmol/L adriamycin for 12 hours and 100 mmol/L cycloheximide (CHX) for the indicated time were determined by Western blotting. D, levels of Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs treated with 2 mmol/L adriamycin for 12 hours and 100 mmol/L cycloheximide for the indicated time were determined by Western blotting. C and D, we quantiﬁed the intensity of each band using a densitometer and plotted relative intensities. Data are expressed as means SD from three independent experiments. E, luciferase reporter analysis of the transcriptional activity of Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs treated with 5 mmol/L MG132 for 6 hours. Error bars correspond to pooled data from three independent experiments. F, lysates of HIPK2þ/þ and HIPK2/ MEFs treated with MG132 for 6 hours were analyzed by Western blotting with antibodies against Notch1-IC and b-actin. G, endogenous coimmunoprecipitation of Fbw7 with Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs. www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF5

Ann et al.

HIPK2+/+ HIPK2–/– ADR + A Mock ADR shHIPK2 B

IP: IP:

IB: Notch1-IC Notch1-IC IB: HIPK2 HIPK2

HIPK2 HIPK2

Notch1-IC Cell Notch1-IC Cell lysates lysates RBP-Jk β-Actin

β-Actin

CPD Consensus motif Consensus Sequence : S/T-x-x-x-[E,S,T] C Myc-HIPK2 : - WT K221R D

Notch1 T2512 )/77363((63 Kinase assay GST-Notch1-IC-P32 Presenilin1 T116 ,<773)7(('7 T62 Cyclin E ,3773'.(('' T380 //7733466*. 32 Kinase assay GST-Fbw7-P Myc T58 /37733/6636 Jun T239 *(7733/663, SREBP1 T456 7/7733366'$ IB: Myc Myc-HIPK2 SV40 large T antigen T701 3377333((3(

E F G MG132 HIPK2 –/– HIPK2 Binding motif Consensus Sequence [S/T]-P or P-[S/T] Notch1-IC Notch1-IC WT T2512A GFP-HIPK2 : - WT K221R GFP-HIPK2 : - + - + hNotch1(T2512) /493(+3)/7363(6'42519 mNotch1(T2487) /493(+3)/7363(6'42494 IB: pT2512 pT2512 IB: pT2512 pT2512 xNotch1(T2480) /493(+3)/7363(6'42485 IP: Notch1-IC gNotch1(T2546) /493'+3)/7363(6'42553 IP : Myc dNotch (T2657) /947/'6<37363(63*2664 IB: HIPK2 HIPK2 IB: Myc Myc-Notch1-IC CPD Consensus motif

IB: GFP IB: β-Actin β-Actin GFP-HIPK2

H I DMSO MG132 7 WT 4xCSL-Luc Myc-Notch1-IC : + + W W M M T2512A 6 HA-ub : - + + + + + P = 0.005 Flag-Fbw7 : - - + + + + 5 GFP-HIPK2 : - + - + - + 4

3 RLU (fold) 2 IB: HA Notch1-IC (Ub)s 1

Notch1-IC : – + + – HIPK2 : – – + +

HIPK2 –/– IP: Myc

Figure 4. HIPK2 regulates Notch1-IC by phosphorylating its T2512 residue. A, endogenous coimmunoprecipitation of Notch1-IC with HIPK2 in HEK293 cells transiently transfected with the indicated plasmids and then treated with 2 mmol/L adriamycin (ADR) for 12 hours. B, endogenous coimmunoprecipitation of HIPK2 with Notch1-IC in HIPK2þ/þ and HIPK2/ MEFs. C, immunocomplex kinase analysis to determine HIPK2-induced phosphorylation of recombinant GST-Notch1-IC or GST-Fbw7 in HEK293 cells transfected with the indicated plasmids. Cell lysates were treated with anti-Myc antibody. Results of 32P-autoradiography after SDS-PAGE and corresponding protein loading are shown at the bottom. D, CLUSTALW alignment of mammalian CPDs. E, CLUSTALW alignment of the CPD consensus motif and HIPK2- binding motif of Notch1 sequences from various species; human Notch1 CPD and HIPK2-binding site (T2512) are shown in red. F, endogenous Western blotting of phosphorylated Notch1-IC T2512 in HIPK2/ MEFs transiently transfected with the indicated plasmids. G, Western blotting of phosphorylated Notch1-IC T2512 in HEK293 cells transiently transfected with the indicated plasmids and then treated with 5 mmol/L MG132 for 6 hours. H, luciferase reporter analysis of the transcriptional activity of Notch1-IC in HIPK2/ MEFs transfected with Notch1-IC, Notch1-IC T2512A, and HIPK2. Error bars correspond to pooled data from three independent experiments. I, in vivo Notch1-IC ubiquitination assay in HEK293 cells transiently transfected with the indicated plasmids and then treated with DMSO or MG132 for 6 hours.

OF6 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

Cell motility assay

A B 0 hr 12 hr 24 hr P < 0.001 35 Mock P 120 30 < 0.001 Mock Notch1-IC shHIPK2 100 Notch1-IC 25 +shHIPK2

80 20 Notch1-IC n.s. 15 60

10 40

Apoptotic cells (%) shHIPK2 5 20 Relative scratch area (%) 0 0 - - - 01224 ADR : + + + Notch1-IC + shHIPK2 Time (h) MOCK Notch1-IC Notch1-IC T2512A

shNotch1

C Transwell cell migration assay D Matrigel invasion assay Mock Mock ADR Mock ADR ADR n.s. 180 Mock n.s. 140 160 ADR P < 0.001 120 shCon shCon 140 P < 0.001 100 120 80 100 60 80 40 60

shHIPK2 Cell number (%) 40 Cell number (%) 20 shHIPK2 20 0 shCon shHIPK2 0 shCon shHIPK2

E F MDA-MB-231 Notch1-IC MCF7 Notch1-IC b-Actin P n.s. b-Actin 250 200 < 0.001 P < 0.001 180 n.s. 200 160 140 P < 0.001 150 120 P < 0.001 100 100 80 60 50 40 Colony formation (%) formation (%) Colony Colony formation (%) 20 0 0

: RDA : --++- + ADR : --++- + MOCK Notch1-IC Notch1-IC MOCK Notch1-IC Notch1-IC T2512A T2512A

shNotch1 shNotch1

G # 1 # 7 # 9 # 10 # 16 # 23 # 33 # 35 # 37

TN N T TN TN TN TN TN TN TN

IB: Notch1-IC Notch1-IC

IB: Notch1-ICpT2512 Notch1-ICpT2512

IB: HIPK2 HIPK2

IB: Fbw7 Fbw7

IB: β-Actin β-Actin

T: Tumor Tissue N: Nomal Tissue

H n = 54, P < 0.001 n = 54, P = 0.001 n = 54, P < 0.001 n = 54, P < 0.001

3.5 3.5 3.5 3.5

3.0 3.0 3.0 3.0

2.5 2.5 2.5 2.5

2.0 2.0 2.0 2.0

1.5 1.5 1.5 1.5

1.0 1.0 1.0 1.0 Notch1-ICpT2512 0.5 0.5 0.5 0.5 Relative expression level expression Relative

0.0 0.0 Fbw7 Relative expression level 0.0 0.0 HIPK2 Relative expression level Notch1-IC Relative Notch1-IC Relative expression level Normal Tumor Normal Tumor Normal Tumor Normal Tumor

www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF7

Ann et al.

Fig. S5A and S5B). Coimmunoprecipitation assay further con- HIPK2 can strongly regulate apoptosis by p53-dependent and firmed the interaction between Notch1-IC and HIPK2 (Supple- -independent pathways (20). We found that adriamycin treatment þ þ mentary Fig. S5C). Adriamycin treatment increased the interac- increases the phosphorylation of Notch1-IC T2512 in both p53 / tion between HIPK2 and Notch1-IC, whereas HIPK2 knockdown and p53 / HTC116 and HIPK2–Notch1 pathway is regulated in inhibited this interaction (Fig. 4A). We also observed that independent of p53 (Supplementary Fig. S6G). We found that Notch1-IC coimmunoprecipitated with endogenous HIPK2 in reintroduction of HIPK2 WT into HIPK2 / MEFs led to increased þ þ HIPK2 / MEFs but not in HIPK2 / MEFs (Fig. 4B). Immuno- phosphorylation of Notch1-IC T2512 residues (Fig. 4F). blotting demonstrated that HIPK2 accumulation was inversely Next, we examined the effects of various kinases such as HIPK, correlated with the level of Notch1-IC (Fig. 4A and B). Notch1- CaMK, AGC, and MAP kinase families on the phosphorylation of IC strongly interacted with the kinase domain of HIPK2 but Notch1-IC T2512. Phosphorylation of Notch1-IC T2512 not with the C-terminal of HIPK2 or with kinase-dead HIPK2– increased with the overexpression of HIPK2 and MEKK1 but not K221R mutant (Supplementary Fig. S5D). Similarly, Fbw7 with the overexpression of other kinases (Supplementary Fig. S6H interacted with the kinase domain of HIPK2 but not with the and S6I). We performed a luciferase assay with Notch1-IC, C-terminal of HIPK2 or with kinase-dead HIPK2-K221R mutant MEKK1, and 4XCSL-Luc in HIPK2 / MEFs. We found that (Supplementary Fig. S5E). HIPK2 could bind to Notch1-IC MEKK1 inhibits the transcriptional activity of Notch1-IC in mutants containing the homopolymer repeat of glutamine HIPK2-independent manner (Supplementary Fig. S6J). HIPK2 (OPA) or PEST domain but not to a Notch1-IC mutant contain- phosphorylated the T2512 residue in WT Notch1-IC but not in the ing the RAM-ANK domains (Supplementary Fig. S5F). We found Notch1-IC T2512A mutant (Fig. 4G). We also observed that that Notch1-IC was predominantly located in the nucleus and HIPK2 overexpression suppressed the transcriptional activity of that HIPK2 was predominantly located in the nuclear compart- WT Notch1-IC but did not suppress the transcriptional activity of ments as dots (Supplementary Fig. S5G). Furthermore, we the Notch1-IC T2512A mutant in HIPK2 / MEFs (Fig. 4H). We observed that Fbw7 formed a trimeric complex with Notch1- further observed that the Notch1-IC T2512A mutant was resistant IC and HIPK2 (Supplementary Fig. S5H). to HIPK2-mediated degradation, indicating that HIPK2-induced The in vitro kinase assay showed that GST-Notch1-IC, but not phosphorylation of Notch1-IC was crucial for its degradation GST-Fbw7, was phosphorylated by HIPK2 (Fig. 4C). Immuno- (Supplementary Fig. S6K). HIPK2 and Fbw7 facilitated the ubi- blotting with antibodies against phosphorylated serine/threonine quitination of WT Notch1-IC but not of the Notch1-IC T2512A confirmed the phosphorylation of Notch1-IC by WT HIPK2 mutant (Fig. 4I). These results indicated that the HIPK2-mediated but not by kinase-dead HIPK2–K221R mutant (Supplementary phosphorylation of Notch1-IC was crucial for its degradation. Fig. S6A). Notch1-IC contains the CPD motif in the PEST domain (Fig. 4D; ref. 18). HIPK2 preferentially phosphorylates serine/ HIPK2 represses tumorigenesis by downregulating Notch1-IC threonine residues in [S/T]-P or P-[S/T] motif (19). In silico studies We observed that cells expressing the Notch1-IC T2512A have shown that, in vertebrates, the C-terminus of Notch1-IC mutant but not WT Notch1-IC were resistant to adriamycin- contains a possible conserved threonine residue. This motif is induced apoptosis (Fig. 5A). Notch1-IC expression or HIPK2 present at T2512 of human Notch1-IC (Fig. 4E). Site-directed knockdown moderately stimulated the migration of MDA-MB- mutagenesis showed that replacement of threonine at position 231 cells. Cells coexpressing Notch1-IC and shHIPK2 showed 2512 of Notch1-IC by alanine decreased its phosphorylation by significant migration compared with control cells (Fig. 5B). HIPK2 (Supplementary Fig. S6B). Next, we generated the anti- HIPK2 knockdown reversed adriamycin-induced inhibition of body specific for phospho-T2512 and confirmed its sensitivity MDA-MB-231 cell migration (Fig. 5C). Adriamycin-induced inhi- using dot blot assay (Supplementary Fig. S6C). Furthermore, we bition of MDA-MB-231 cell migration was also significantly examined adriamycin-mediated phosphorylation of Notch1-IC reversed by the expression of Notch1-IC T2512A mutant (Sup- T2512 by HIPK2 by using an antibody against phosphory- plementary Fig. S7A). Adriamycin significantly reduced cell inva- lated Notch1-IC T2512 (Supplementary Fig. S6D and S6E). We sion. However, HIPK2 knockdown prevented adriamycin- speculated whether adriamycin-mediated phosphorylation of induced suppression of cell invasion (Fig. 5D). Adriamycin- þ þ Notch1-IC T2512 is similarly regulated in HIPK2 / MEFs and induced inhibition of MDA-MB-231 cell invasion was significant- HIPK2 / MEFs. We found that adriamycin induced the phos- ly restored by the expression of Notch1-IC T2512A mutant þ þ phorylation of Notch1-IC T2512 in HIPK2 / MEFs but not in (Supplementary Fig. S7B). Adriamycin treatment dramatically HIPK2 / MEFs (Supplementary Fig. S6F). decreased the colony-forming activity of cells expressing the

Figure 5. HIPK2-mediated downregulation of Notch1-IC suppresses the tumorigenesis of breast cancer. A, HEK293 cells transfected with shNotch1, GFP, Myc-Notch1-IC, and Myc-Notch1-IC T2512A, and treated with 2 mmol/L adriamycin for 12 hours were fixed and stained with DAPI. Apoptotic nuclei of GFP-positive cells were quantified using a fluorescence microscope. Data are expressed as mean SD of values from three independent experiments. B, photomicrographs of wound healing assay of MDA-MB-231 cells expressing either Notch1-IC or shHIPK2. C, photomicrographs of transwell motility assay of MDA-MB-231 cells expressing shCon and shHIPK2, and treated with adriamycin for 12 hours. Bar graph shows the mean number of cells per filter. Data are expressed as mean SD of values from three independent experiments. Representative colonies are shown. D, photomicrographs from Matrigel invasion assay of MDA-MB-231 cells expressing shCon and shHIPK2, and treated with adriamycin for 12 hours. Bar graph shows the mean number of cells per filter. Data are expressed as mean SD of values from three independent experiments. Representative colonies are shown. E and F, colony-forming assay of Notch1-silenced MCF7 (E) and MDA-MB-231 (F) cells expressing Myc-Notch1-IC WT and Myc-Notch1-IC T2512A and treated with adriamycin for 12 hours. Cells were further incubated for 14 days, fixed with 4% paraformaldehyde, and stained with 0.5% crystal violet. Data are expressed as mean SD of values from three independent experiments. Representative colonies are shown. G, Western blotting of Notch1-IC, phosphorylated Notch1-IC T2512, HIPK2, and Fbw7 in breast cancer tissues (T) and adjacent normal tissues (N). b-Actin was used as the loading control. H, comparison of the relative levels of Notch1-IC, phosphorylated Notch1-IC T2512, HIPK2, and Fbw7 in breast cancer tissues and normal tissues by using paired t test (data are expressed as mean SD; n ¼ 54, , P < 0.001).

OF8 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

empty vector and Notch1-IC compared with that of cells expres- expression of Notch1-IC P2513L and Notch1-IC P2515fs sing the Notch1-IC T2512A mutant in endogenous Notch1- mutants significantly restored adriamycin-mediated inhibition silenced MDA-MB-231 and MCF cells (Fig. 5E and F). of MDA-MB-231 cell migration (Supplementary Fig. S9B). We collected 54 samples of breast cancer tissue along with the Expression of Notch1-IC P2513L and Notch1-IC P2515fs adjacent normal breast tissue from patients with breast cancer and mutants significantly restored adriamycin-induced inhibition performed immunoblotting to compare the levels of Notch1-IC, of MDA-MB-231 cell invasion (Supplementary Fig. S9C). Adria- phosphorylated Notch1-IC T2512, HIPK2, and Fbw7 between mycin inhibited the colony-forming activity of cells expressing these tissues, using b-actin as a loading control. Levels of Notch1- WT Notch1-IC but not in cells expressing Notch1-IC P2513L IC were higher in 40 breast cancer tissue samples than in corre- and Notch1-IC P2515fs mutants (Supplementary Fig. S9D sponding normal breast tissue samples (Fig. 5G and H). Of the 40 and S9E). pairs (both breast tumor and normal breast tissues), 27 normal Mice injected with cells expressing WT Notch1-IC, Notch1-IC breast tissue samples showed relatively higher levels of phosphor- P2513L, and Notch1-IC P2515fs mutants showed significant ylated Notch1-IC T2512, HIPK2, and Fbw7 (Fig. 5G and H; tumor growth compared with mice injected with control cells. Supplementary Table S2). Thus, analyses of breast tissue samples Adriamycin dramatically reduced tumor growth in mice showed that levels of Notch1-IC were elevated in tumor tissues injected with cells expressing the empty vector and WT and that Notch1-IC levels were elevated in tumor tissues and the Notch1-IC. However, no effect was observed on tumor growth levels of phosphorylated Notch1-IC T2512, HIPK2, and Fbw7 in mice injected with cells expressing Notch1-IC P2513L and were negatively correlated with those of Notch1-IC. Notch1-IC P2515fs mutants (Fig. 7A–C). Adriamycin treatment significantly increased the levels of phosphorylated Notch1-IC T2512 and HIPK2 and decreased the levels of Natural mutations in the CPD motif of Notch1-IC make it Notch1-IC in xenograft tumor derived from Notch1-IC–expres- resistant to HIPK2-induced phosphorylation sing cells but not in cells expressing Notch1-IC P2513L and Information present in a publicly available cancer genome Notch1-IC P2515fs mutants (Fig. 7D). These findings suggested database (COSMIC) indicates that 229 of the 304 somatic muta- that Notch1-IC induced tumorigenesis both in vitro and in vivo tions in the Notch1 PEST domain are concentrated within the and that adriamycin-mediated phosphorylation of Notch1-IC CPD motif in patients with hematopoietic and lymphoid malig- T2512 by HIPK2 could effectively prevent the tumorigenesis of nancies (21). We selected two major natural mutations P2513L breast cancer (Fig. 7E and Supplementary Fig. S10). and P2515fs from the somatic mutations present in the C- terminal CPD region of Notch1-IC (Fig. 6A and B). We confirmed that both Notch1-IC P2513L and Notch1-IC P2515fs could bind Discussion to RBP-Jk and increase the transcriptional activity of Notch1-IC Recently, several studies have identified the stability and activ- target genes, Hes1 and Hes5 (Supplementary Fig. S8A–S8C). ity of Notch1-IC in various cancers (5, 22, 23). The Fbw7-medi- Overexpression of HIPK2 suppressed the transcriptional activity ated degradation of Notch requires phosphorylation of multiple of WT Notch1-IC but did not suppress the transcriptional activity residues. The phosphorylation event at T2512 residue is required of Notch1-IC P2513L and Notch1-IC P2515fs mutants (Supple- for polyubiquitination and proteasome-dependent degradation mentary Fig. S8D). We observed that WT Notch1-IC produced a of the protein by Fbw7 (12–14). In this study, we showed that strong phosphorylation signal; however, no signal was detected chemotherapeutic agent–induced HIPK2 directly phosphorylates for Notch1-IC P2513L and Notch1-IC P2515fs mutants (Sup- Notch1-IC at the T2512 residue, stimulating its proteasome- plementary Fig. S8E). Overexpression of HIPK2 increased the dependent degradation by Fbw7 ubiquitin ligase. phosphorylation of Notch1-IC T2512 but did not increase the HIPK2 is a key regulator of cancer cell death in response to phosphorylation of Notch1-IC P2513L and Notch1-IC P2515fs genotoxic stress and prevents cancer progression (17, 24–27). To mutants in HEK293 cells (Fig. 6C). our knowledge, this is the first study to show an association Coimmunoprecipitation assay showed that HIPK2 increased between Notch1-IC and HIPK2 in clinical breast cancer samples. the physical association between Notch1-IC and Fbw7 but did not We observed that the levels of phosphorylated Notch1-IC T2512, increase the physical association between Notch1-IC P2513L and HIPK2, and Fbw7 were significantly higher in normal breast tissue Notch1-IC P2515fs mutants and Fbw7 (Fig. 6D). Moreover, than in the tumor tissue. In addition, expression of Notch1-IC was Notch1-IC P2513L and Notch1-IC P2515fs mutants were resis- inversely correlated with the levels of phosphorylated Notch1-IC tant to HIPK2-mediated degradation (Fig. 6E). HIPK2 and Fbw7 T2512, HIPK2, and Fbw7. facilitated the ubiquitination of Notch1-IC but not of Notch1-IC Human T-ALL cell lines and tissue samples from primary P2513L and Notch1-IC P2515fs mutants (Fig. 6F). These results tumors in patients harbor activating mutations in the extracellular indicated that Notch1-IC P2513L and Notch1-IC P2515fs heterodimerization domain and/or the C-terminal PEST domain mutants were resistant to HIPK2-induced phosphorylation, estab- of Notch1 (28–30). We selected two natural mutants Notch1-IC lishing that HIPK2-induced phosphorylation of Notch1-IC was P2513L and Notch1-IC P2515fs with somatic mutations in the crucial for its degradation. C-terminal CPD region of Notch1-IC. We found that Notch1-IC P2513L and Notch1-IC P2515fs mutants were resistant to Biological aspects of somatic mutations in the CPD motif of HIPK2-induced phosphorylation. Furthermore, after adriamycin Notch1-IC treatment, HIPK2 significantly decreased cancer growth and inva- We observed that cells expressing Notch1-IC P2513L and sion by increasing the apoptosis of cancer cells expressing Notch1- Notch1-IC P2515fs mutants but not WT Notch1-IC were IC; however, this was not observed for cancer cells expressing resistant to adriamycin-induced apoptosis (Supplementary Fig. Notch1-IC P2513L and Notch1-IC P2515fs mutants. Thus, the S9A). We performed transwell migration assays and found that results of this study provide novel insights into the functional

www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF9

Ann et al.

A B P2515 Upper Nonsense N=218 aerodigestive Breast Central Missense

tract The Others 1% nervous Q2504 3% 2% system Coding silent Skin 2% 2% Frameshift

Phosphodegron P2513 Lung P2506

5% E2507 S2514 P2518 H2508 Q2520 Q2502 T2512 L2511 E2516 F2510 P2509

HQLQVPEHPFLpTPSPESPDQW Notch1-IC Haematopoietic and lymphoid tissue 85%

NH2 RAM ANK TAD PEST Domains

PEST (N = 304) Unique mutated notch1 in tissue

MG132 C MG132 D Myc-Notch1-IC : - WT WT M1 M1 M2 M2 Notch1-IC Notch1-IC Notch1-IC Flag-Fbw7 : - + + + + + + WT P2513L P2515*fs GFP-HIPK2 : - - + - + - +

GFP-HIPK2 : - + - + - + IB: Flag Flag-Fbw7

IB: Notch1-ICpT2512 Notch1-ICpT2512 IP: Myc

IP: Myc IB: Myc Myc-Notch1-IC

IB: Myc Myc-Notch1-IC IB: Flag Flag-Fbw7

IB: GFP GFP-HIPK2 IB: GFP GFP-HIPK2

M1: Myc-Notch1-IC P2513L M2: Myc-Notch1-IC P2515*fs

E F DMSO MG132 Myc-Notch1-IC : + + + + Myc-Notch1-IC : + + W W M1 M1 M2 M2 GFP-HIPK2 : - + ++ +++ HA-ub : - + + + + + + + Flag-Fbw7 : - - + + + + + + IB: Myc Myc-Nothc1-IC GFP-HIPK2 : - + - + - + - +

IB: Myc Myc-Nothc1-IC P2513L

Notch1-IC (Ub)s IB: Myc Myc-Notch1-IC P2515*fs IB: HA

IB: GFP GFP-HIPK2

IP: Myc

Figure 6. Somatic mutations in the CPD motif of Notch1-IC make it resistant to HIPK2. A, pie chart representing the distribution of mutated Notch1 in various tissues (data from COSMIC). B, mutations in the CPD motif of Notch1 in leukemia cells. C, Western blotting of phosphorylated Notch1-IC T2512 in HEK293 cells transiently transfected with the indicated plasmids and then treated with 5 mmol/L MG132 for 6 hours. D, coimmunoprecipitation of anti-Flag with anti-Myc in HEK293 cells transiently transfected with the indicated plasmids and then treated with 5 mmol/L MG132 for 6 hours. E, Western blotting of Myc-Notch1-IC, Myc-Notch1-IC P2513L, Myc-Notch1-IC P2515fs, and GFP-HIPK2 in HEK293 cells transiently transfected with the individually indicated plasmids. F, in vivo Notch1-IC ubiquitination assay in HEK293 cells transiently transfected with the indicated plasmids and then treated with DMSO or MG132 for 6 hours.

effects of somatic mutations on tumorigenesis and responses to ubiquitination and subsequent rapid degradation (10, 11, 31, anticancer drugs. 32). However, the kinase that phosphorylates at the T2512 residue Several research groups have shown that Fbw7 binds to phos- in Notch1 phosphodegron site has not been deﬁnitively identi- phorylated Notch-IC via its WD40 domains, and mediates its ﬁed. Moreover, phosphorylation of T2512 in Notch1-IC reduces

OF10 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

B 1,800 1,600 Mock ) 3 1,400 Mock + ADR A Notch1-IC 1,200 Notch1-IC + ADR Mock Notch1-IC Notch1-IC Notch1-IC Notch1-IC P2513L Notch1-IC P2513L + ADR P2513L P2515*fs 1,000 Notch1-IC P2515*fs 800 Notch1-IC P2515*fs + ADR ADR : - + - + - + - + 600 * 400 * * : P < 0.001 200 Tumor volume (mm 0

9630 181512 21 (day) Time (days)

2.5 C P < 0.001 n.s. n.s. D Mock Notch1-IC Notch1-IC Notch1-IC P2513L P2515*fs

2.0 ADR : - + - + - + - +

IB: Myc Myc-Notch1-IC 1.5 P < 0.001 IB: Notch1-ICpT2512 Notch1-ICpT2512 1.0

IB: HIPK2 HIPK2

Tumor weight (g) 0.5 IB: Fbw7 Fbw7

0.0 IB: β-Actin β-Actin

Genotoxic stress (UV, IR, E chemotherapeucs)

Cytoplasm

HIPK2

Notch1-IC

HIPK2

Notch1-IC

Notch1-IC Hes1,5 SMRT off Hes1,5 on P SKIP T2512 MAML RBP-Jk p21 Notch1-IC RBP-Jk p21 c-Myc c-Myc CyclinD1,3 CyclinD1,3 p27 p27 etc. etc.

Tumorigenesis Tumorigenesis

Nucleus 26S Proteasome

Figure 7. Physiologic effects of somatic mutations in the CPD motif of Notch1-IC. A, xenograft growth of MDA-MB-231-SQ cells expressing the indicated constructs and treated with adriamycin. Tumors excised from nude mice, and statistical analysis of mean tumor volume and tumor weight. Tumor volumes (mm3) were measured on indicated days (n ¼ 5 per group). Error bars, mean SD; , P < 0.001. Inhibition of tumor growth by adriamycin was measured by calculating tumor volume (B) and tumor weight (C). Representative images of the dissected tumors are shown in A. D, cell lysates of random tumors selected from eight groups were analyzed by immunoblotting with antibodies against Myc, phosphorylated Notch1-IC T2512, HIPK2, Fbw7, and b-actin. E, HIPK2 suppresses Notch1 signaling by phosphorylating its T2512 residue, which in turn initiates the proteasome-mediated degradation of Notch1-IC by Fbw7 in the nucleus. Notch1-IC forms a trimeric complex with Fbw7 and HIPK2. HIPK2 enhances the degradation of Notch1-IC via the Fbw7-dependent proteasome pathway. www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF11

Ann et al.

migration, tumor growth, and metastasis. Further studies are Writing, review, and/or revision of the manuscript: E.-J. Ann, J.-S. Ahn, required to determine that identification of other kinases that E.-H. Jo, H.-J. Lee, H.-S. Park phosphorylate and promote proteasomal degradation of Notch1- Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.-J. Ann, J.-H. Yoon, J.-S. Ahn, E.-H. Jo, IC T2512 in various cancer cells. D. W. Choi, C.Y. Choi, H.-S. Park In conclusion, we showed that HIPK2 phosphorylated the Study supervision: J.-H. Yoon, C.Y. Choi T2512 residue of Notch1-IC and facilitated its proteasomal deg- Other (provided reagents and supervised Dr Miyeon Kim's research related to radation through Fbw7. Given the frequent inactivation of HIPK2 this project): A.A. Ferrando in breast cancer, these findings provide a better understanding of the roles of HIPK2 and Notch1 in cancer development. Further- Acknowledgments more, these findings could provide essential insights into the We thank R. Kopan (Washington University Medical School) for the Notch regulation of Notch1-IC by HIPK2, and thus help in identifying constructs and B.E. Clurman (Fred Hutchinson Cancer Research Center) for the useful diagnostic and therapeutic targets for treating breast cancer. Notch and Fbw7 constructs.

Disclosure of Potential Conﬂicts of Interest Grant Support No potential conﬂicts of interest were disclosed. This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by Authors' Contributions the Ministry of Science, ICT and Future Planning (2014R1A4A1003642 Conception and design: E.-J. Ann, K. Lee, H.-S. Park to H.-S. Park). Development of methodology: E.-J. Ann, H.-S. Park The costs of publication of this article were defrayed in part by the Acquisition of data (provided animals, acquired and managed patients, payment of page charges. This article must therefore be hereby marked provided facilities, etc.): J.-H. Yoon, J.-S. Ahn, H.-J. Lee, H.-W. Lee, H.-G. Kang, advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate K.-H. Chun, J.S. Lee, H.-S. Park this fact. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.-J. Ann, M.-Y. Kim, J.-H. Yoon, E.-H. Jo, K.-H. Chun, Received December 8, 2015; revised June 8, 2016; accepted June 13, 2016; H.-S. Park published OnlineFirst June 22, 2016.

References 1. Artavanis-Tsakonas S, Rand MD, Lake RJ. Notch signaling: cell fate 15. Ann EJ, Kim HY, Choi YH, Kim MY, Mo JS, Jung J, et al. Inhibition of Notch1 control and signal integration in development. Science 1999;284: signaling by Runx2 during osteoblast differentiation. J Bone Miner Res 770–6. 2011;26:317–30. 2. Pece S, Serresi M, Santolini E, Capra M, Hulleman E, Galimberti V, et al. 16. Mo JS, Ann EJ, Yoon JH, Jung J, Choi YH, Kim HY, et al. Serum- and Loss of negative regulation by Numb over Notch is relevant to human glucocorticoid-inducible kinase 1 (SGK1) controls Notch1 signaling by breast carcinogenesis. J Cell Biol 2004;167:215–21. downregulation of protein stability through Fbw7 ubiquitin ligase. J Cell 3. Reedijk M, Odorcic S, Chang L, Zhang H, Miller N, McCready DR, et al. Sci 2011;124:100–12. High-level coexpression of JAG1 and NOTCH1 is observed in human 17. Choi DW, Seo YM, Kim EA, Sung KS, Ahn JW, Park SJ, et al. Ubiquitination breast cancer and is associated with poor overall survival. Cancer Res and degradation of homeodomain-interacting protein kinase 2 by WD40 2005;65:8530–7. repeat/SOCS box protein WSB-1. J Biol Chem 2008;283:4682–9. 4. Stylianou S, Clarke RB, Brennan K. Aberrant activation of notch signaling in 18. Welcker M, Clurman BE. FBW7 ubiquitin ligase: a tumour suppressor at the human breast cancer. Cancer Res 2006;66:1517–25. crossroads of cell division, growth and differentiation. Nat Rev Cancer 5. Rizzo P, Miao H, D'Souza G, Osipo C, Song LL, Yun J, et al. Cross-talk 2008;8:83–93. between notch and the estrogen receptor in breast cancer suggests novel 19. Yamada D, Perez-Torrado R, Filion G, Caly M, Jammart B, Devignot V, therapeutic approaches. Cancer Res 2008;68:5226–35. et al. The human protein kinase HIPK2 phosphorylates and down- 6. Dievart A, Beaulieu N, Jolicoeur P. Involvement of Notch1 in regulates the methyl-binding transcription factor ZBTB4. Oncogene the development of mouse mammary tumors. Oncogene 1999;18: 2009;28:2535–44. 5973–81. 20. Wei G, Ku S, Ma GK, Saito S, Tang AA, Zhang J, et al. HIPK2 represses beta- 7. Koch U, Radtke F. Notch and cancer: a double-edged sword. Cell Mol Life catenin-mediated transcription, epidermal stem cell expansion, and skin Sci 2007;64:2746–62. tumorigenesis. Proc Natl Acad Sci U S A 2007;104:13040–5. 8. Al-Hussaini H, Subramanyam D, Reedijk M, Sridhar SS. Notch signaling 21.ForbesSA,BindalN,BamfordS,ColeC,KokCY,BeareD,etal. pathway as a therapeutic target in breast cancer. Mol Cancer Ther COSMIC: mining complete cancer genomes in the catalogue of somat- 2011;10:9–15. ic mutations in cancer. Nucleic Acids Res 2011;39(Database issue): 9. Garber K. Notch emerges as new cancer drug target. J Natl Cancer Inst D945–50. 2007;99:1284–5. 22. Real PJ, Ferrando AA. NOTCH inhibition and glucocorticoid therapy in 10. Hubbard EJ, Wu G, Kitajewski J, Greenwald I. sel-10, a negative regulator of T-cell acute lymphoblastic leukemia. Leukemia 2009;23:1374–7. lin-12 activity in Caenorhabditis elegans, encodes a member of the CDC4 23. Wang Z, Banerjee S, Li Y, Rahman KM, Zhang Y, Sarkar FH. Down- family of proteins. Genes Dev 1997;11:3182–93. regulation of notch-1 inhibits invasion by inactivation of nuclear factor- 11. Gupta-Rossi N, Le Bail O, Gonen H, Brou C, Logeat F, Six E, et al. Functional kappaB, vascular endothelial growth factor, and matrix metalloproteinase- interaction between SEL-10, an F-box protein, and the nuclear form of 9 in pancreatic cancer cells. Cancer Res 2006;66:2778–84. activated Notch1 receptor. J Biol Chem 2001;276:34371–8. 24. Rinaldo C, Prodosmo A, Siepi F, Soddu S. HIPK2: a multitalented partner 12. O'Neil J, Grim J, Strack P, Rao S, Tibbitts D, Winter C, et al. FBW7 mutations for transcription factors in DNA damage response and development. in leukemic cells mediate NOTCH pathway activation and resistance to Biochem Cell Biol 2007;85:411–8. gamma-secretase inhibitors. J Exp Med 2007;204:1813–24. 25. D'Orazi G, Cecchinelli B, Bruno T, Manni I, Higashimoto Y, Saito S, 13. Thompson BJ, Buonamici S, Sulis ML, Palomero T, Vilimas T, Basso G, et al. et al. Homeodomain-interacting protein kinase-2 phosphorylates The SCFFBW7 ubiquitin ligase complex as a tumor suppressor in T cell p53 at Ser 46 and mediates apoptosis. Nat Cell Biol 2002;4:11–9. leukemia. J Exp Med 2007;204:1825–35. 26. Hofmann TG, Moller A, Sirma H, Zentgraf H, Taya Y, Droge W, et al. 14. Demarest RM, Ratti F, Capobianco AJ. It's T-ALL about Notch. Oncogene Regulation of p53 activity by its interaction with homeodomain-interact- 2008;27:5082–91. ing protein kinase-2. Nat Cell Biol 2002;4:1–10.

OF12 Cancer Res; 76(16) August 15, 2016 Cancer Research

HIPK2 Negatively Regulates Notch1-IC

27. Dauth I, Kruger J, Hofmann TG. Homeodomain-interacting protein kinase 30. Li N, Fassl A, Chick J, Inuzuka H, Li X, Mansour MR, et al. Cyclin C is a 2 is the ionizing radiation-activated p53 serine 46 kinase and is regulated haploinsufﬁcient tumour suppressor. Nat Cell Biol 2014;16:1080–91. by ATM. Cancer Res 2007;67:2274–9. 31. Oberg C, Li J, Pauley A, Wolf E, Gurney M, Lendahl U. The Notch 28. Weng AP, Ferrando AA, Lee W, Morris JPt, Silverman LB, Sanchez-Irizarry C, intracellular domain is ubiquitinated and negatively regulated by the et al. Activating mutations of NOTCH1 in human T cell acute lympho- mammalian Sel-10 homolog. J Biol Chem 2001;276:35847–53. blastic leukemia. Science 2004;306:269–71. 32. Wu G, Lyapina S, Das I, Li J, Gurney M, Pauley A, et al. SEL-10 is an inhibitor 29. Roy M, Pear WS, Aster JC. The multifaceted role of Notch in cancer. Curr of notch signaling that targets notch for ubiquitin-mediated protein Opin Genet Dev 2007;17:52–9. degradation. Mol Cell Biol 2001;21:7403–15.

www.aacrjournals.org Cancer Res; 76(16) August 15, 2016 OF13

Tumor Suppressor HIPK2 Regulates Malignant Growth via Phosphorylation of Notch1

Eun-Jung Ann, Mi-Yeon Kim, Ji-Hye Yoon, et al.

Cancer Res Published OnlineFirst June 22, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-3310

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/06/22/0008-5472.CAN-15-3310.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2016/07/28/0008-5472.CAN-15-3310. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.