Homeodomain‐interacting protein kinase 2 suppresses proliferation and aerobic glycolysis via ERK/cMyc axis in pancreatic cancer

To investigate the roles of the homeodomain‐interacting protein kinase (HIPK) family of proteins in pancreatic cancer prognosis and the possible molecular mechanism.

is not structurally related to the others outwith the catalytic domain. HIPK1-3 were first identified as Nkx1.2-interating proteins in yeast two-hybrid screening. Approximately 90% of the amino acid sequences that make up their kinase domains are conserved across HIPK1-3, and their noncatalytic region is also conserved. [3][4][5] HIPK family members are expressed in dynamic temporal and spatial patterns, suggesting their diverse and important roles during development. In disease models, HIPKs mediate key signalling pathways that regulating the response to various stress signals, including DNA damage, reactive oxygen species and hypoxia. 6 HIPK2 is the beststudied member of the HIPK family and is involved in the regulation of differentiation, proliferation and apoptosis of cells. [7][8][9] In cancer, HIPK2 is generally considered to be a potential tumour suppressor as it can promote apoptosis by phosphorylating tumour suppressor protein p53, leading to expression of proapoptotic genes. 10,11 HIPK2 also induces apoptosis by modulating molecules independently of p53, such as phosphorylation-dependent degradation of antiapoptotic transcriptional co-repressor CtBP. [12][13][14] Due to the structural similarities between HIPK1 and HIPK2, these two proteins share some redundant activities, and HIPK1 also regulates apoptosis via interacting with nuclear proteins. 15,16 HIPK3 has been reported to regulate the pathogenesis of type 2 diabetes. 17 To date, little is known about the roles of HIPK4 except for its capacity to phosphorylate p53 at Ser9. 18 These reports demonstrate that HIPK proteins might have important roles in cancer and are promising anticancer targets.
However, the expression status of HIPKs in pancreatic cancer and their indicative roles in prognosis have seldom been discussed. Thus, there is a need to uncover the role of HIPKs in pancreatic cancer and provide the possible molecular mechanism in the hope of discovering novel treatment targets for pancreatic cancer.
Ninety years ago, Otto Warburg published a body of work linking metabolism and cancer, demonstrating that tumour cells rapidly use glucose and convert the majority of it into lactate, which is known as the Warburg effect. Through aerobic glycolysis, tumour cells utilize glucose to meet the demand of uncontrolled proliferation. 19 Lactate produced by aerobic glycolysis creates an acidic environment that favours metastasis of cancer cells, as the extracellular matrix becomes destabilized under acidic conditions. 20 Activated oncogenes or loss of tumour suppressors could alter metabolism and induce aerobic glycolysis. In pancreatic cancer, mutation of KRAS oncogene induces metabolism reprogramming via induction of enhanced aerobic glycolysis. 21,22 Tumour suppressor gene TP53 could induce the expression of the TP53-induced glycolysis regulatory phosphatase (TIGAR), which is a regulator of aerobic glycolysis. 23 The cMyc oncogene is a master regulator that controls many aspects of cancer cell malignancy, and it is usually overexpressed in cancer cells that have been transformed into a more malignant phenotype. cMyc also induces enhanced aerobic glycolysis in cancer cells by induction of key glycolytic genes like glucose transporter 1 (GLUT1), hexokinase 2 (HK2) and lactate dehydrogenase A (LDHA). Thus, cMyc is considered to be a gatekeeper that balances uncontrolled proliferation and enhanced aerobic glycolysis. 24 In this regard, an understanding of cellular metabolic derangements in pancreatic cancer and uncovering the molecular mechanism underlying the derangement could lead to novel therapeutic approaches.
In the present study, we examined the indicative roles of HIPKs in pancreatic cancer prognosis and demonstrated that HIPK2 predicted better prognosis. In vitro studies demonstrated that overexpression of HIPK2 inhibited pancreatic cancer proliferation and aerobic glycolysis.
Mechanistic studies demonstrated that HIPK2 decreased cMyc protein stability via post-translational phosphorylation of cMyc via extracellular signal-regulated kinase (ERK). This work might provide novel predictive and treatment targets for pancreatic cancer.
The absorbance of each sample was measured at a wavelength of 450 nm using a microplate reader. Colony formation was performed as described previously. 25 Pancreatic cancer cells (n = 500) were seeded in six-well plates and cultivated for 10-14 days. The cells were fixed with 4% paraformaldehyde followed by staining with 1% crystal violet staining. The colonies were counted subsequently.

| cMyc protein stability assessment
Pancreatic cancer cells were treated with cycloheximide (CHX; 100 μg/mL) and harvested for the indicated times, followed by Western blotting and quantification of the blot bands.

| Glycolysis measurement
Glucose utilization and glycolysis were measured using glucose uptake assay, lactate production assay and Seahorse extracellular flux analyzer.

| Immunohistochemistry
Expression of HIPK2 in pancreatic cancer patients' samples was assessed by immunohistochemistry. Paraffin sections were incubated for 1 hour at 70°C, deparaffinized in xylene and rehydrated in graded ethanol. The slides were neutralized with 3% H 2 O 2 for 30 minutes.
The antigen retrieval was processed with citrate buffer (pH 6.0) in an incubator at 95°C. After antigen retrieval, the slides were incubated with primary and secondary antibodies. HIPK2 antibody (ab28507; Abcam) was used at a dilution of 1:50. The sections were stained with 3,3-diaminobenzidine, terminated in PBS and counterstained with haematoxylin.

| Tissue specimens
The clinical tissue samples used in this study were histopathologically and clinically diagnosed at Fudan University Shanghai Cancer Center (FUSCC). Prior patient consent and approval from the Institutional Research Ethics Committee were obtained.

| Statistical analyses
Statistical analyses were performed using SPSS version 17.0 (IBM Corp., Armonk, NY, USA) using independent Student's t test (twotailed) or one-way analysis of variance. Logistic regression was used to determine the correlation between HIPK2, GLUT1, HK2 and LDHA expression level and clinicopathological characteristics in the TCGA cohorts. Statistical significance was based on two-sided P < 0.05.

| Decreased HIPK2 expression indicates worse prognosis of pancreatic cancer
To  Table S1. We validated the roles of HIPK2 in pancreatic cancer OS in a FUSCC cohort using immunohistochemical staining. The scoring of HIPK2 staining was indicated as Low HIPK2 and High HIPK2 ( Figure S2).
Consistent with the TCGA cohort, patients with lower HIPK2 expression displayed shorter OS ( Figure 1C). Subsequently, we assessed expression of HIPK2 in pancreatic cancer and adjacent para-tumour tissues. Quantitative PCR demonstrated that expression of HIPK2 was higher in para-tumour samples than in tumour samples ( Figure 1D). Immunohistochemical staining demonstrated that HIPK2 expression was significantly lower in tumour than para-tumour samples ( Figure 1E,F). Collectively, these results demonstrated that HIPK2 might play negative roles in pancreatic cancer prognosis prediction and function as a tumour suppressor.

| HIPK2 negatively regulates proliferation of pancreatic cancer cells
To validate the roles of HIPK2 in vitro, we measured expression of HIPK2 in pancreatic cancer cell lines. HIPK2 expression was lower in PANC-1 and SW1990 cells, but it was higher in normal epithelial cell line HPDE (Figure 2A). Thus, we overexpressed FLAG-tagged

| HIPK2 functions as a negative regulator of aerobic glycolysis in pancreatic cancer
It is well known that enhanced aerobic glycolysis promotes uncontrolled proliferation of cancer cells; thus, we investigated whether HIPK2 negatively regulated aerobic glycolysis. In the process of aerobic glycolysis, cancer cells utilize glucose to produce lactate, instead of the mitochondrial respiration pathway. Glycolysis and mitochondrial respiration can be measured by Seahorse extracellular flux analyzer, with ECAR to reflect aerobic glycolysis and OCR to indicate mitochondrial respiration. HIPK2 overexpression decreased glycolysis and glycolytic capacity, which was reflected by ECAR measurement ( Figure 3A,B). Mitochondrial respiration was impaired in the process of aerobic glycolysis, resulting in a reduction of oxygen consumption that was reflected by OCR measurement. We observed that HIPK2 overexpression enhanced OCR, suggesting its positive roles in mitochondrial respiration ( Figure 3C,D). Collectively, these results demonstrate that HIPK2 is a negative regulator of aerobic glycolysis.

| HIPK2 downregulates cMyc and cMyctargeted glycolytic genes
In HIPK2-overexpressing PANC-1 and SW1990 cells, we measured the changes in cMyc levels. Quantitative PCR and Western blotting showed a significant reduction in cMyc protein levels, while the mRNA levels decreased only slightly ( Figure 4A,B).

| HIPK2 regulates aerobic glycolysis via cMyc
HIPK2 induces a decrease in cMyc protein levels, suggesting that cMyc mediates HIPK2-induced aerobic glycolytic regulation. We rescued cMyc expression in HIPK2-overexpressing cells ( Figure 5A). By assessing glucose uptake and lactate generation, we demonstrated that cMyc WT attenuated the decrease in glucose uptake caused by HIPK2 overexpression in PANC-1 and SW1990 cells ( Figure 5B). Cancer cells utilize glucose to produce lactate via glycolysis, which can be measured by lactate F I G U R E 1 HIPK2 indicates better prognosis and is downregulated in pancreatic cancer. Kaplan-Meier analysis using TCGA pancreatic adenocarcinoma cohort demonstrated that patients with lower HIPK2 expression had better OS than those with lower HIPK2 expression (A). DFS analysis with TCGA pancreatic adenocarcinoma cohort demonstrated that HIPK2 levels could serve as a predictor of tumour recurrence, and patients with higher HIPK2 expression had better DFS (B). By using FUSCC pancreatic cancer cohort and immunohistochemical staining, we demonstrated that higher HIPK2 expression indicated better prognosis (C). Quantitative PCR analysis demonstrated that HIPK2 expression was higher in adjacent tumour samples than the tumour samples (D). Immunohistochemical staining confirmed that HIPK2 expression was higher in adjacent normal tumour samples (E and F)  Figure 5D).

| HIPK2 regulates cMyc stability via suppressing ERK activation
As observed above, we demonstrated that HIPK2 decreased cMyc protein levels more than mRNA levels, suggesting that HIPK2 regulates cMyc post-translationally. By assessing the half-life of cMyc, F I G U R E 3 HIPK2 functions as a negative regulator of aerobic glycolysis in pancreatic cancer. Diagram of ECAR results obtained by Seahorse extracellular flux analyzer to determine the impact of HIPK2 on aerobic glycolysis (A). Glycolysis was quantified by measuring ECAR in the presence or absence of glucose injection. Glycolytic capacity was calculated as the difference between ECAR following the injection of oligomycin and the basal ECAR reading. Further analysis of ECAR demonstrated that HIPK2 introduction into PANC-1 and SW1990 cells decreased glycolysis and glycolytic capacity of the cells (B). Representation of OCR measurement with Seahorse analyzer to confirm the role of HIPK2 in mitochondrial respiration (C). Oligomycin, a complex V inhibitor, is injected into the reaction system to derive ATP-linked respiration. FCCP, a protonophore, is added to collapse the inner membrane gradient, allowing the electron transport chain to function at its maximal rate, and the maximal respiration capacity is derived by subtracting non-mitochondrial respiration from the FCCP rate. OCR analysis results demonstrated that HIPK2 overexpression in PANC-1 and SW1990 cells increased ATP production and maximal respiration, suggesting its positive roles in mitochondrial respiration (D) F I G U R E 4 HIPK2 downregulated cMyc and cMyc-targeted glycolytic genes. HIPK2 overexpression decreased cMyc protein levels in PANC-1 and SW1990 cells (A). HIPK2 had a small impact on cMyc mRNA levels (B). Real-time PCR demonstrated that HIPK2 decreased cMyc-targeted glycolytic genes, including GLUT1, HK2 and LDHA (C). Introduction of HIPK2 into PANC-1 and SW1990 cells decreased GLUT1, HK2 and LDHA protein levels (D). In TCGA cohort of pancreatic cancer patients, we observed a negative and significant correlation between HIPK2 with GLUT1, HK2 and LDHA, respectively (E-G) we demonstrated that HIPK2 overexpression decreased the protein stability of cMyc in PANC-1 and SW1990 cells ( Figure 6A,B). We measured changes in ERK1/2 activation in HIPK2-overexpressing cells, and introduction of HIPK2 attenuated activation of ERK1/2 ( Figure 6C). ERK phosphorylated Ser62 at cMyc and stabilized it; thus, we measured changes in Ser62-phosphorylated cMyc. In HIPK2 overexpressing cells, by immunoprecipitating cMyc wand subsequent Western blot analysis with pS62-cMyc antibody, we observed a decrease in phosphor-S62 cMyc, suggesting HIPK2 could regulated cMyc stability via ERK-mediated phosphorylation ( Figure 6D). To confirm this, we overexpressed ERK2 in HIPK2-overexpressing cells and observed an increase in cMyc protein levels, but the dominant negative form of ERK2 K52R could not rescue the decrease in cMyc caused by HIPK2 ( Figure 6E).

| D ISCUSS I ON
Even though significant progress has been made in the diagnosis, surgery and chemotherapy of pancreatic cancer, there has been no satisfactory improvement in 5-year survival of the disease.
Therefore, discovery of novel treatment targets and uncovering the underlying molecular mechanism have become increasingly urgent.
In the present study, we indicated that HIPK2 serves a marker for predicting better prognosis of pancreatic cancer. Mechanistic studies have suggested that HIPK2 negatively regulates proliferation and aerobic glycolysis in pancreatic cancer via the ERK/cMyc axis.
The best-characterized tumour-suppressive role of HIPK2 is its modification of p53 post-translational phosphorylation and involvement in p53-induced apoptosis. Due to its impact on p53-mediated apoptosis regulation, HIPK2 is considered to be a tumour suppressor. 26 In pancreatic cancer, genetic mutation of TP53 is a high frequency mutation, but as a negative regulator of p53, the roles of HIPK2 have seldom been reported. 27 In our present study, we reported HIPK2 to be a tumour suppressor, which is consistent with reports in other types of cancer. One reason that accounts for poor prognosis of pancreatic cancer is its intrinsic resistance to chemotherapy and radiotherapy, which could induce the apoptotic process. 28 53,54 In this study, we demonstrated that HIPK2 regulated cMyc stability via post-translational modifications including phosphorylation. As HIPK2 is also a serine/threonine kinase, and cMyc protein stability can be regulated by serine and threonine phosphorylation, there is the possibility that HIPK2 interacts with and phosphorylates cMyc, which causes cMyc destabilization. 55 The possibility that HIPK2 interacts with cMyc is that they share the same interacting partners such as mitogen-activated protein kinase kinase kinase 1 (MEKK1) and FBW7(F-Box And WD Repeat Domain In conclusion, we demonstrated that decreased HIPK2 expression is an unfavourable marker for pancreatic cancer OS. Subsequent F I G U R E 7 Representation of the working model. Decreased HIPK2 in pancreatic cancer cells promoted activation of ERK signalling, leading to increased phosphorylation of cMyc at Ser62 and the resultant cMyc stability, which ultimately increased glycolysis and cell proliferation in vitro studies demonstrated that HIPK2 regulated proliferation and aerobic glycolysis in pancreatic cancer cells. In-depth mechanistic studies suggested that HIPK2 attenuated activation of ERK and resulted in destabilization of cMyc ( Figure 7). Collectively, the present study uncovered novel predictive and treatment targets of pancreatic cancer.

ACK N OWLED G EM ENTS
This research was supported by the National Science Foundation

CO N FLI C T O F I NTE R E S T
All authors state that they have no conflicts of interest.