A novel isoform of Homeodomain-interacting protein kinase-2 promotes YAP/TEAD transcriptional activity in NSCLC cells

Homeodomain-interacting protein kinase-2 (HIPK2) can either promote or inhibit transcription depending on cellular context. In this study, we show that a new HIPK2 isoform increases TEAD reporter activity in NSCLC cells. We detected HIPK2 copy number gain in 5/6 (83.3%) NSCLC cell lines. In NSCLC patients with high HIPK2 mRNA expression in the Human Protein Atlas, the five-year survival rate is significantly lower than in patients with low expression (38% vs 47%; p = 0.047). We also found that 70/78 (89.7%) of NSCLC tissues have moderate to strong expression of the N-terminal HIPK2 protein. We detected and cloned a novel HIPK2 isoform 3 and found that its forced overexpression promotes TEAD reporter activity in NSCLC cells. Expressing HIPK2 isoform 3_K228A kinase-dead plasmid failed to increase TEAD reporter activity in NSCLC cells. Next, we showed that two siRNAs targeting HIPK2 decreased HIPK2 isoform 3 and YAP protein levels in NSCLC cells. Degradation of the YAP protein was accelerated after HIPK2 knockdown in NSCLC cells. Inhibition of HIPK2 isoform 3 decreased the mRNA expression of YAP downstream gene CTGF. The specific HIPK2 kinase inhibitor TBID decreased TEAD reporter activity, reduced cancer side populations, and inhibited tumorsphere formation of NSCLC cells. In summary, this study indicates that HIPK2 isoform 3, the main HIPK2 isoform expressed in NSCLC, promotes YAP/TEAD transcriptional activity in NSCLC cells. Our results suggest that HIPK2 isoform 3 may be a potential therapeutic target for NSCLC.


INTRODUCTION
Homeodomain-interacting protein kinase-2 (HIPK2) can either promote or inhibit transcription depending on cellular context [1]. HIPK2 is canonically considered as a tumor suppressor because it phosphorylates p53 at Ser46 in response to ultraviolet radiation, which promotes apoptosis [2][3][4]. Hipk2 was shown to be a haploinsufficient tumor suppressor gene in a radiation-induced mouse lymphoma model [5]. HIPK2 may also promote DNA damage repair pathways and protect cells against genome instability induced by radiation [6,7]. Accumulating evidence suggests that HIPK2 may also play an oncogenic role. The sole Drosophila member of the HIPK family, Hipk, promotes Notch signaling transduction in Drosophila eye development in a kinase-dependent manner [8]. Human HIPK2, mapped to chromosome 7q32-q34 [9], is frequently amplified and overexpressed in pilocytic astrocytoma [10].

Research Paper
Oncotarget 174 www.oncotarget.com HIPK2 expression is significantly higher in cervical cancer than in healthy tissue [11]. HIPK2 is also significantly higher in aggressive meningiomas than in benign meningiomas [12]. HIPK2 is positively associated with cell growth in androgen-receptor-positive prostate cancer cells [13]. In addition, HIPK2 is transcriptionally regulated by nuclear factor erythroid 2 and HIPK2 knockdown increased the sensitivity to cisplatin in non-small cell lung cancer (NSCLC) cells [14].
Hipk promotes the transcriptional activity of Yorkie (Yki), the ortholog of Yes-associated protein (YAP), to induce target gene expression and tissue growth in Drosophila [15][16][17]. Hipk regulates Yki in a kinasedependent manner [15,16]. These studies suggest that the closest human homolog of Hipk, HIPK2, may have the same role in mammalian cells. For instance, HIPK2 promotes abundance and activity of YAP in a kinase-dependent fashion in 293T cells [15]. Further investigation, particularly in the context of human cancer cells, is required to define the mechanism of HIPK2-YAP regulation.
HIPK2 was found to exist as several different splice variants in mammals [18]. Because the function of these splice variants may be quite different, the cellular context of these isoforms may play different roles in cancer. In this study, we focused on analysis of the main HIPK2 isoform expressed in non-small-cell lung cancer (NSCLC), which consists of adenocarcinoma and squamous cell carcinoma.

HIPK2 gene gain and amplification in NSCLC cell lines
We analyzed HIPK2 copy number across the database of the Cancer Cell Line Encyclopedia [19] in cBioPortal [20] and found HIPK2 gain and amplification in human cancer cell lines (Supplementary Table 1). According to the database, 82 (49.4%) of 166 NSCLC cell lines have gain or amplification, 61 (36.7%) have two copies and 23 (13.9%) have loss of HIPK2 copy number. To verify HIPK2 amplification, we analyzed 6 lung cancer cell lines (A549, H460, H522, H1299, H2170, SKLU1) and normal LP9 cells using fluorescent in situ hybridization (FISH). We used green fluorescein-labeled HIPK2 FISH probe to determine the copy number of HIPK2 and chromosome enumeration probe 7 (CEP7) to determine the copy number of chromosome 7, where the HIPK2 gene is located ( Figure 1A). We detected gain of HIPK2 copy number ( Figure 1B) in lung cancer cell lines compared to the normal LP9 cells.

High HIPK2 mRNA expression correlates with low survival rate in NSCLC patients
The survival curve of HIPK2 mRNA expression for patients with NSCLC is shown in the database of the Human Protein Atlas (https://www.proteinatlas.org/ ENSG00000064393-HIPK2/pathology/lung+cancer). The five-year survival rates are 47% for patients with low HIPK2 mRNA expression and 38% for those with high expression (p = 0.047) ( Figure 1C).

HIPK2 isoform 3 inhibition decreased YAP protein level in NSCLC cells
In human NSCLC H2030 cells, the siRNAs targeting exons 7 and 2 of HIPK2 reduced the protein level of HIPK2 isoform 3 (110 kDa) and decreased YAP protein and its downstream CTGF protein levels. The siRNA that targets exon 15 of HIPK2 did not reduce the protein level of HIPK2 isoform 3 and increased the YAP protein level ( Figure 3A). Accordingly, HIPK2 isoform 3 may have a role in controlling the Hippo-YAP pathway.

Forced-expression of HIPK2 isoform 3 increased TEAD reporter activity in NSCLC cells
In H2030 cells, forced-expression of HIPK2 isoform 3 significantly increased Hippo reporter activity (p < 0.001), whereas forced-expression of HIPK2 isoform 1 significantly reduced TEAD reporter activity (p < 0.001) ( Figure 3B). To understand the mechanism of HIPK2 isoform 3 in promoting TEAD reporter activity, we Oncotarget 176 www.oncotarget.com performed mutagenesis of HIPK2 isoform 3 at Lys228 (K228) to generate mutated HIPK2 isoform 3 (K228A). We transfected lung cancer H2030 and H1975 cells with the plasmid construct and measured TEAD reporter activity in the cells. Forced HIPK2 isoform 3 expression showed increased TEAD reporter activity in these cells compared to control cells (p < 0.001) ( Figure 3C). Expressing the mutation construct K228A decreased TEAD reporter activity compared to wild-type cells (p < 0.001) ( Figure 3C). Our results indicate that Lys228 (K228) is the kinase active site of HIPK2 isoform 3 in regulating TEAD reporter activity.

TBID decreased cancer stem cells and TEAD reporter activity in NSCLC cells
To test if TBID can reduce the population of cancer stem cells, we performed side population and 3D tumorsphere analysis in NSCLC cells. In our previous studies, NSCLC cell lines A549 and H460 showed side population characteristics [24,25]. After treatment with 10 µM TBID (based on the IC 50 (6.7 µM, 72 hrs) in H460 cells and (10.8 µM, 72 hrs) in A549 cells), the side population decreased from 5.5% to 1.7% in A549 cells, and from 5.7% to 0.95% in H460 cells ( Figure 5A). TBID reduced tumorsphere size and number in A549 and H460 cells ( Figure 5B). It also significantly decreased TEAD reporter activity in a dose-dependent manner in H460 cell lines (p < 0.001) ( Figure 5C).

DISCUSSION
In this study, we detected increased HIPK2 DNA copy number in five of the six NSCLC cell lines analyzed. The N-terminal HIPK2 was expressed in 89% of NSCLC tissues and the C-terminal HIPK2 was only expressed in 11% of NSCLC tissues. The novel HIPK2 isoform (110 kDa) was detected as the dominant HIPK2 isoform expressed in NSCLC cell lines (6/6) and tissues (13/24), and the HIPK2 isoform 1 (150 kDa) was only detected at lower levels in one of the six NSCLC cell line (H1975) and not detected in NSCLC tissues (0/24) (Supplementary  Figures 2, 3). These results suggest the novel HIPK2 isoform may play an oncogenic role in NSCLC. Thus, we cloned this novel HIPK2 isoform and investigated its relationship with YAP.
Since there were currently two validated isoforms in the NCBI, we named this new HIPK2 isoform as HIPK2 isoform 3. It contains 13 exons: 1-12, 13a and 13b. Compared to HIPK2 isoform 1, HIPK2 isoform 3 has an extra exon 13b and lacks exons 14 and 15. The protein product of the cloned HIPK2 isoform 3 was detected at 110 kDa by western blotting. Furthermore, the siHIPK2s (exon 2, 7) decreased the110 kDa bands in H2030 cell  0.001, t-test). TEAD reporter activity was significantly higher after forcedexpression of HIPK2 isoform 3 (Iso3) (p < 0.001, t-test). (C) TEAD reporter activity in H2030 and H1975 cell lines was significantly lower after forced-expression of HIPK2 isoform 3 K228A mutation than after forced-expression of HIPK2 isoform 3 (p < 0.001, t-test). Error bars indicate standard deviations; *** p ≤ .001. www.oncotarget.com lines. We found that forced overexpression of the HIPK2 Isoform 3 promotes TEAD reporter activity in H1975 and H2030 cell lines. Two siRNAs targeting HIPK2 decreased HIPK2 Isoform 3 and YAP protein levels in NSCLC cells. In previous work, we used a YAP protein stability assay with extracellular signal-regulated kinase (ERK) inhibition in the squamous lung cancer cell line H2170 [26]. Degradation of YAP protein was accelerated after HIPK2 knockdown in H2170 cells. The protein level of YAP/TEAD downstream gene CTGF was decreased by the two siHIPK2s. These results suggest that HIPK2 isoform 3 promotes YAP/TEAD transcriptional activity and may play an oncogenic role in NSCLC.
HIPK2 isoform 1 has an auto-inhibitory domain (AID) and a YH domain [27,28]. HIPK2 isoform 1 kinase activity may be inhibited by the AID [27,29,30] and/or YH domain [31]. Deletions of the AID and YH domains abolished the ability of HIPK2 isoform 1 to suppress the Wnt/ß-catenin pathway, suggesting that the integrity of the C-terminal region of HIPK2 isoform 1 is important for the recruitment of transcriptional corepressors [31]. These findings are similar to our findings that HIPK2 isoform 3 (without AID and YH domains) increases TEAD reporter activity and siHIPK2 specific to isoform 1 increases YAP protein, whereas overexpression of HIPK2 isoform 1 suppresses TEAD reporter activity in H2030 cells. Our data suggest that the AID and YH domains of HIPK2 isoform 1 negatively regulate YAP/TEAD activity ( Figure 5D).
Our findings using HIPK2 isoform 1 plasmid are the opposite of those of Poon et al. in the 293T cell line [16]. Several explanations for this difference are possible. One is that context-dependent function of HIPK2 isoform 1 in Wnt signaling has been reported [1,32,33], and a positive or negative role of HIPK2 on Wnt signaling depends on the availability of particular T-cell factor (TCF) proteins in the cell. HIPK2 isoform 1 would inhibit the pathway upon phosphorylation of an activator type TCF, but would activate it upon phosphorylation of a repressor type TCF. A similar context-dependent function of HIPK2 isoform 1 may exist in the Hippo signaling pathway. In addition, Oncotarget 180 www.oncotarget.com the AID of HIPK2 isoform 1 inhibit potential YAP phosphorylation by KD of HIPK2 and affect its stability depending on the relative cellular level of HIPK2 isoform 3 [15,17] (Figure 5D). For instance, forced expression of HIPK2 isoform 1 may induce a certain level of TEAD reporter activity in the absence of HIPK2 isoform 3. Conversely, forced expression of HIPK2 isoform 1 may reduce the relative high level of TEAD reporter activity in the presence of high level HIPK2 isoform 3.
Recently, we discovered the first crystal structure of HIPK2 and found that its catalytic lysine (K228) is critical for binding HIPK2 to a small molecule HIPK2 inhibitor [34]. TBID, the first and selective HIPK2 inhibitor, is also predicted to interact with HIPK2 at K228 [23]. In this study, HIPK2 isoform 3 site-mutagenesis K228A failed to increase TEAD reporter activity, suggesting that the positive YAP regulation of isoform 3 is dependent on its kinase activity. TBID treatment reduced TEAD reporter activity of NSCLC cells in a dose-dependent manner. In addition, TBID reduced cancer stem cell populations and inhibited tumorsphere formation of NSCLC cells. Thus, these data suggest that K228 in the kinase domain is a critical site for HIPK2 oncogenic function and future development of HIPK2-targeting therapy such as small molecules.
In summary, this study suggests that HIPK2 isoform 3 promotes YAP/TEAD transcriptional activity and it may play an oncogenic role in NSCLC. Our results also suggest that HIPK2 isoform 3 may be a potential therapeutic target for NSCLC.

Tissue samples and immunocytochemistry
Tissue microarray sections contained fresh lung tumor and adjacent normal lung tissues from patients with NSCLC who were undergoing surgical resection of the primary tumor. Primary human NSCLC samples from 115 patients were fixed in formalin and embedded in paraffin in 4-μm tissue microarray sections. In 10 of these patients, a small amount of normal lung tissue had been obtained simultaneously to serve as controls. All human tissue samples were obtained and analyzed in accordance with procedures approved by the institutional review board of the University of California, San Francisco (IRB H8714-22942-01). Tissue microarrays were stained with hematoxylin and eosin for general morphology analysis. For IHC analysis, endogenous peroxidase was quenched for 15 min. at room temperature with 3% H2O2 in methanol in each lung section. Sections were blocked with 4% normal goat serum in PBS with 0.2% Triton for 2 hrs at room temperature before incubation overnight at 4°C with the properly diluted antibodies. The N-terminal specific anti-HIPK2 antibody (ab28507; abcam) and the C-terminal specific anti-HIPK2 antibody (#5091S; Cell Signaling, Beverly, MA, USA) were used. Three independent researchers blindly scored positivity, and the data represent the samples that were scored positive by all three individuals. The following scoring system was used: −, no stain; +, weak staining (≤ 10% stained cellularity considered as positive); ++, moderate staining (> 10% but ≤ 30% stained cellularity considered as positive); +++, strong staining (> 30% but ≤ 50% stained cellularity considered as positive). All scoring was done under an objective lens (×20) with a Zeiss Axioscop 2 microscope (Carl Zeiss, Jena, Germany) and photomicrographs were obtained with a Carl Zeiss AxioCam MrC5 camera under a 20× or 40× objective lens.

RNA interference
Cells were seeded in a 6-well plate with fresh media without antibiotics 24 hrs before transfection, with a target of 30-50% confluency at the time of transfection. HIPK2 Exon2 siRNA (ID: s57556 catalog No: 4392420), HIPK2 Exon7 siRNA (ID: hss120797 catalog No: 1299001), HIPK2 Exon15 siRNA (ID: 109493 catalog No: AM16708) and control siRNA were purchased from Thermo Scientific (Waltham, MA, USA). Cells were transfected with 100 nmol/l of siRNA using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. After siRNA transfection, the plates were incubated for 48 hrs at 37°C before further analysis.

Semi-quantitative RT-PCR
Total RNA from the various cell lines was isolated using the RNeasy extraction method (Qiagen, Valencia, Oncotarget 181 www.oncotarget.com CA, USA). First-strand cDNA was synthesized from total RNA by iScript cDNA synthesis (Bio-Rad, Hercules, CA, USA) according to the manufacturer's instructions. Taqman RT-PCR analysis was performed on cDNA in a 384-well plate, using a Prism 7900HT Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Primers and Taqman probes for human HIPK2 and human GUSB were purchased from Applied Biosystems. The expression of target gene in each sample was assayed in triplicate and normalized to human GUSB for mRNA expression analysis. Decreased transcriptional levels of HIPK2 were calculated by dividing the transcriptional levels measured in HIPK2 siRNA samples from those in the control samples.

Western blotting
NSCLC cell lines were seeded in a 6-well plate as 500,000 cells/well and cultured at 37°C supplied with 5% CO2 without antibiotics for 24 hrs. Cells were then transfected with HIPK2 siRNA (siHIPK2) or control siRNA (siCtrl). Cell lysates were then immunoblotted using the N-terminal specific anti-HIPK2 antibody (ab28507; abcam; Immunogen: a peptide corresponding to residues 37-86 of human HIPK2) and the C-terminal specific anti-HIPK2 antibody (#5091S; Cell Signaling, Beverly, MA; Immunogen: a peptide corresponding to residues surrounding Gln1045 of human HIPK2)., anti-YAP (#4912S; Cell signaling), anti-GAPDH (#2188S; Cell signaling) or anti-β-actin (#4970; Cell Signaling) antibodies. NIH ImageJ was used to quantify the intensity of western blot bands, and the relative protein expression levels were calculated by normalizing with GAPDH protein levels.

Luciferase reporter assays
The TEAD binding site was originally reported as the GTIIC site (ACATTCCA) in Hela cell [35]. The 8xGTIIC-luciferase was a gift from Stefano Piccolo (Addgene plasmid # 34615; http://n2t.net/addgene:346 15;RRID:Addgene_34615). The 8×GTIIC plasmid was transfected in NSCLC cell lines. HIPK2 knockdown assays require co-transfection of HIPK2 siRNA or control siRNA with 8xGTIIC plasmid into the cells in a 24-well plate. The Renilla luciferase pRL-TK plasmid (Promega, Madison, WI, USA), whose expression is driven by the housekeeping thymidine kinase gene promoter, was co-transfected to normalize for transfection efficiency. All transfection experiments were performed using Lipofectamine2000 (Invitrogen) in accordance with the manufacturer's instructions. After 48 hrs cells were lysed and luciferase assays were performed following the manufacturer's instructions. Results are expressed as fold induction, which is the ratio of luciferase activity induced in 8xGTIIC-transfected cells relative to basal luciferase activity in control transfected cells. All experiments were performed in triplicate. Means and standard deviations were calculated.

Protein degradation assay
Human NSCLC H2170 cells with HIPK2 knockdown by siRNA were treated with 100 μg/ml cycloheximide, the inhibitor of protein synthesis, and harvested at the time points of 0, 1, 2, 4, 6 and 8 hrs. Total proteins were extracted and expression of YAP was analyzed by western blotting.

Tumorsphere cultures
Tumorspheres were cultured in Stem Cell media (Gibco, Grand Island, NY, USA) following the manufacturer's instructions. When spheres reached 150-200 lM, they were broken up by trituration with a 26-gauge needle 10 times. After seven days, the secondary sphere formation was assayed with (TBID 10 µM) or without drug inhibitor present. The spheres with diameter larger than 60 µm in 9 fields were counted.

Side population assay
The compound efflux ability of cancer stem cells comes from the increased expression of ATP-binding cassette (ABC) transporters within the cell membrane and serves as the basis of an important flow-cytometrybased cell-sorting assay called the side population (SP) assay [36,37]. Side population staining was performed as described [38]. Briefly, 1 × 10 6 /ml cells were resuspended in RPMI 1640 containing 2% fetal bovine serum and labeled with H33342 (Sigma-Aldrich GmbH, Steinheim, Germany) at a concentration of 5.0 μg/ml for 60 min in a 37°C water bath, either alone or with 100 μM verapamil hydrochloride (Sigma-Aldrich GmbH) and 10 µM TBID (Pharmazeutische Chemie, Germany) for 72 hours. Tubes were gently inverted every 20 min and then centrifuged at 400 × g for 5 min at 4°C. The pellets were resuspended in cold PBS containing 2 µg/ ml propidium iodide. The resuspended cells were passed through a 40-μm mesh filter and maintained at 4°C in the dark until flow cytometry analysis was done on a BD FACS Aria II. Hoechst dye was excited at 355 nm (32), and fluorescence was measured at two wavelengths using a 450/50-nm (blue) band-pass filter and a 670/30-nm (33) long-pass edge filter. Cells were displayed on dot-plots gated on live cells, PI negative, and viewed in a Hoechst Blue versus Hoechst Red dot-plot to visualize the side population. As a positive control, ABC transporter inhibitor verapamil, at the bottom-left corner in the flow cytometry plots, disappeared or faded out. The position where the tail disappeared was used as a control to gate the area of side population cells. www.oncotarget.com

Statistical analysis
The data shown represent mean ± SD. The chisquare independence test was used to compare IHC results between the staining intensity of HIPK2 and YAP in the same NSCLC tumors or cell lines. Student's t-test was used to compare gene expression results and luciferase reporter activities between experimental and control groups. p values < 0.05 were considered significant.

Author contributions
YD conceived, designed, and performed the experiments, analyzed the data and wrote the paper. HK analyzed the data, performed the experiments, wrote the paper and revised the manuscript. YLY performed the experiments, analyzed the data and wrote the paper. YW, SL, GC, YW, ZX analyzed the data. GC and YW performed the experiments. ZX designed and performed the experiments. UK and DMJ analyzed the data and revised the manuscript. LY conceived the experiments, analyzed the data and revised the manuscript.