HINT2 downregulation promotes colorectal carcinoma migration and metastasis

Histidine triad nucleotide-binding 2 (HINT2), a member of the histidine triad proteins family, sensitizes cells to apoptosis in hepatocellular carcinoma. Here, we showed that HINT2 expression is lower in primary colorectal cancer (CRC) and metastasis tissues than in normal colorectal tissues, and that HINT2 abundance is inversely correlated with CRC tumor stage. Treating CRC cells with 5-aza-2′-deoxycytidine, a demethylating agent, upregulated HINT2, suggesting HINT2 downregulation is caused by methylation of the gene promoter. HINT2 downregulation increased tumor migration and invasion in vitro, promoted CRC cell metastasis in vivo, and increased expression of epithelial-to-mesenchymal transition (EMT) markers. Furthermore, HINT2 downregulation depended on hypoxia inducible factor (HIF)-2α-mediated transcriptional activation of zinc finger E-box-binding homeobox 1 (ZEB1). These results suggest that HINT2 downregulation promotes HIF-2α expression, which induces EMT and enhances CRC cell migration and invasion. HINT2 may thus a useful clinical indicator of CRC progression and metastasis risk.


INTRODUCTION
Colorectal cancer (CRC) is one of the most common malignant cancers, and approximately 50-60% of patients present with metastases at initial diagnosis [1][2][3][4]. Because advanced metastatic CRC remains largely incurable, there is an urgent need to elucidate the molecular mechanism underlying this aggressive cancer.
Histidine triad nucleotide-binding 2 (HINT2) is a member of the histidine triad superfamily that performs a range of biological functions. HINT2 is localized exclusively in the mitochondrial matrix and is a tumor suppressor in human hepatocellular carcinoma [5]. HINT2 also regulates lipid metabolism, glucose homeostasis, and mitochondrial respiration [6]. Mitochondria not only play a role in co-opting the regulatory pathways that inactivate anoikis, but also regulate metastatic cell survival [7][8][9]. Furthermore, quantitative real-time polymerase chain reaction (qRT-PCR) results revealed HINT2 downregulation in colon carcinoma as compared to normal tissue [10]. Thus, HINT2 appears to be associated with CRC progression.
The present study analyzed the association of HINT2 with CRC metastasis. We found that downregulating HINT2 induces EMT in CRC cells and promotes CRC migration and metastasis in vitro and in vivo, respectively. HINT2 inhibits CRC cell invasion and migration by inducing EMT through ZEB1-mediated CDH1 inhibition via hypoxia inducible factor (HIF)-2α. ZEB1 knockdown via siRNA rescued the effects of HINT2 downregulation in CRC cells.

HINT2 expression in CRC tissue and cell lines
We selected 46 human CRC cases with paracancer normal colorectal tissues. qRT-PCR and immunohistochemistry (IHC) analyses revealed that HINT2 was downregulated in CRC as compared to normal colorectal tissue ( Figure 1A-1H), consistent with a previous study [10]. Multivariate analysis showed that HINT2 mRNA levels was downregulated in CRCs without lymph node metastasis than that in lymph node positive ones(P < 0.01; Figure 1I). IHC results in CRC liver and lymph node metastasis samples showed that HINT2 expression was low or absent in metastases ( Figure 1C and 1F-1G). A semi-quantitative analysis revealed that HINT2 expression is progressively lost with increasing tumor stage ( Figure 1M).
To determine whether DNA methylation was involved in transcriptional regulation of HINT2, LoVo and SW620 cells were treated with 5-aza-2′-deoxycytidine, a DNA demethylating agent. Treatment increased HINT2 levels in Lovo (3.4-fold) and SW620 (2.7-fold) cells ( Figure 1L). These results suggested that DNA methylation might be involved in HINT2 downregulation.

HINT2 downregulation promotes CRC cell migration and invasion
HINT2 was knocked down in SW480 cells via shRNA. HINT2 expression changes were confirmed by western blotting (Figure 2A). To investigate whether EMT induction by HINT2 contributed to CRC cell migration and invasion, we first performed transwell migration assays in control and SW480 stable HINT2 knockdown cells. Cell migration was increased in knockdown cells (P < 0.01) ( Figure 2D-2E). A wound-healing assay confirmed that HINT2 knockdown increased cell motility (P < 0.01) ( Figure 2B-2C). A quantitative in vitro Matrigel invasion assay showed that HINT2 knockdown increased SW480 cell invasion (P < 0.01; Figure 2F-2G). We also transfected HINT2 into SW620 cells and confirmed its overexpression ( Figure 3A). primary cancer (B and E), and metastasis tissues (C: lymph node MT, F: liver MT), and quantification (G) of the above. Arrows indicate liver metastasis and normal liver tissue, original magnification, ×200. (H) RT-qPCR analysis of HINT2 expression; data represent mean ± SD, n = 46 per group, *P < 0.05 H. RT-qPCR analysis of HINT2 expression; data represent mean ± SD, group 1 represents tissue without metastasis (n = 20), and group 2 represents tissue with metastasis (n = 26), *P < 0.05 (I). HINT2was detected in CRC cell lines via realtime PCR (J) and western blotting (K) Bar graph shows SW620 cell-normalized HINT2 expression in CRC cell lines. Hint2 was detected in Lovo and SW620 cells via real-time PCR after treatment with control (DMSO) or 5-aza-dC (L). Relative HINT2 protein intensity in cancer tissues relative to tumor stage (M) *P < 0.05, **P < 0.001 compared to SW620 cells. All data are representative of at least three independent experiments.
Oncotarget 13523 www.impactjournals.com/oncotarget Using the same migration and invasion assays described above, we found that HINT2 overexpression decreased SW620 cell migration and invasion ( Figure 3B-3F). Overall, these results support the hypothesis that HINT2 downregulation promotes CRC cell migration and invasion.

HINT2 induces EMT
HINT2-downregulated SW480 cells changed from tightly packed to disseminated and diffusely organized ( Figure 4A). To determine whether EMT molecular alterations occurred in these cells, expression of the epithelial marker E-cadherin (CDH1), and the mesenchymal markers, vimentin and N-cadherin (CDH2), were measured via western blotting. CDH1 was decreased while CDH2 and vimentin were increased in HINT2 knockdown SW480 and HT-29 cells as compared to controls ( Figure 4D-4E).
To investigate the role of HINT2 in regulating CDH1 transcription, a luciferase reporter assay was performed. HT-29 and SW480 cells were transfected with pGL3/CDH1 along with a Renilla reporter plasmid. CDH1 luciferase activity was decreased (P < 0.01) in HINT2 knockdown cells compared with controls ( Figure 4B). RT-PCR confirmed these results ( Figure 4C). Taken together, these data imply that HINT2 induces an anti-EMT gene expression profile in CRC cell lines.
ZEB1 promotes migration and invasion by inducing EMT in colon cancer [9,10]. To determine whether the negative correlation between HINT2 and Results were plotted as the average number of migrated cells from six random microscope fields. Transwell assay showing SW480 cell invasion (F) and qualification (G) after 48-h incubation (×200 magnification). Results were plotted as the average number of invading cells from six random microscope fields. All data are representative of at least three independent experiments. **P < 0.01, ***P < 0.001 compared to controls. www.impactjournals.com/oncotarget CDH1 is mediated by ZEB1, we performed a luciferase reporter assay using human pGL3/CDH1. Stable HINT2 knockdown or control SW480 and HT-29 cells were transfected with control or ZEB1 siRNA, then transfected with the CDH1 promoter vector pGL-CDH1, and relative luciferase activity was detected. HINT2 downregulation repressed CDH1 transcription. When both ZEB1 and HINT2 were downregulated, there was no change in relative luciferase activity ( Figure 5C).
A previous study showed that HINT2 knockout mice had increased levels of HIF-2α and reactive oxygen species [20]. We performed a qRT-PCR assay to measure HIF-2α expression in HT-29 and SW480 cells. We found that the HIF-2α expression increased in HT-29 and SW480 cells after down-regulation of Hint2, which is consistence with previously report ( Figure 5E). HIF-1α and HIF-2α share potential binding sequences in the ZEB1 promoter. HIF-1α reportedly binds exclusively to hypoxia response element (HRE) 3 (634-629nt) [21]. We mutated HRE1 (521-516 nt) and HRE2 (529-524 nt) and found no changes in ZEB1 relative luciferase activity upon changes in HINT2 expression ( Figure 5F).

HINT2 downregulation promotes CRC metastasis in vivo
Control and HINT2 knockdown SW480 cells were injected into mice. Liver metastases were observed in five mice from the experimental group while no visible metastases formed in any control mice ( Figure 6A-6B). Histologically, metastatic nests were found in all mice injected with HINT2 knockdown SW480 cells, and in three mice from the control group. Metastatic foci in the knockdown group were larger than those in the control group as measured by HE staining ( Figure 4C). IHC Cells were photographed under a phase contrast microscope. Transwell assay showing SW620 cell migration (D and E) (×200 magnification) after 48-h incubation. Results were plotted as the average number of migrated cells from six random microscope fields. Transwell assay showing SW620 cell invasion (F) and qualification (G) (×200 magnification) after 48-h incubation. Results were plotted as the average number of invading cells from six random microscope fields. All data are representative of at least three independent experiments. **P < 0.01, ***P < 0.001 compared to controls.
Oncotarget 13525 www.impactjournals.com/oncotarget results showed that HINT2 was absent in the metastatic nests and tiny foci of the knockdown group, while HINT2 staining was strong in the control group ( Figure 4D). Together, these results indicate that ZEB1 mediates HINT2-dependent EMT regulation through HIF-2α.

DISCUSSION
Metastasis is the most significant contributor to cancer death, and the EMT is a fundamental event in this process [22][23][24]. Silencing EMT-related genes is an attractive therapeutic approach for improving treatment outcomes [25][26][27]. The present study found that HINT2 downregulation promotes CRC cell migration and invasion by enhancing the ZEB1-mediated EMT response through HIF-2α.
HINT2 expression is generally downregulated in human hepatocellular carcinoma cells and HINT2 appears to sensitize these cells to apoptosis [10]. We hypothesized that HINT2 downregulation also promotes CRC progression. We confirmed that HINT2 expression was lower in CRC tissues than in normal colon mucosa, especially during metastasis, and that HINT2 abundance was inversely correlated with CRC tumor stage. Two cell lines derived from metastases exhibited lower HINT2 levels, suggesting that HINT2 loss promotes this process. We observed that HINT2 knockdown induced cell infiltration through transwell membranes, while HINT2 overexpression inhibited this process. In mice injected with CRC cells, HINT2 downregulation induced the formation of more and larger colonies in the liver, indicating that reduced HINT2 promoted CRC metastasis by promoting cell migration and invasion.
The EMT is important for tumor cells to acquire migratory and invasive properties [28,29], and was recently implicated in several types of invasive human malignancies [30][31][32][33], including CRC [34]. Hence, EMT and its markers, including CDH1, are useful diagnostic biomarkers or therapeutic targets in CRC. In the present study, HINT2 downregulation enhanced CDH2 and vimentin expression and suppressed CDH1 in CRC cells. The outcome of a luciferase reporter assay using the Oncotarget 13526 www.impactjournals.com/oncotarget CDH1 promoter supports transcriptional regulation as the major mechanism underlying HINT2-dependent EMT marker expression changes. Therefore, our work provides evidence that HINT2 downregulation may enhance CRC cell migration and invasion by inducing EMT.
Given the pivotal role of ZEB1 in downregulating CDH1 and inducing the loss of cell polarity, ZEB1 drives the EMT and cancer progression [40,41]. We assessed the importance of the interaction between HINT2 and ZEB1 in regulating EMT in CRC cells. We knocked down ZEB1 in stable HINT2-silenced HT-29 and SW480 cells and then examined cell morphology and EMT marker expression. The effects of HINT2 on the EMT response were rescued by ZEB1 knockdown. The luciferase assay further demonstrated that HINT2 downregulation induced CDH1 transcription in a ZEB1-dependent manner in CRC cells.
HIF-2α is an important angiogenic factor with prognostic value, and high HIF-2α expression in CRC promotes tumor progression [42]. We confirmed the results of a previous study showing that HINT2 knockout mice had increased HIF-2α levels. To evaluate whether ZEB1 mediated HINT2-dependent regulation of the EMT program through HIF-2α, the promoter sequences of ZEB1 were analyzed using bioinformatics methods. As transcription factors, HIF-1α and HIF-2α regulate downstream genes by binding HRE sites in their promoter regions. We found three potential HRE sites in the proximal promoter (approximately 1000 nt upstream; the start codon ATG defined as 0) of the ZEB1 gene: HRE-1 at and CDH1 transcriptional inhibition (C) caused by HINT2 downregulation. HINT2 knockdown or control SW480 and HT-29 cells were transfected si-ZEB1 or si-Con (control), then transfected with the CDH1 promoter vector, pGL-CDH1, and relative luciferase activation was detected. Data from control cells were set at 100%. Western blot analysis of HINT2, ZEB1, and EMT markers in SW480 control, HINT2-knockdown, and HINT2/ZEB1 double knockdown cells (D) HINT2 and ZEB1 double knockdown nearly rescued the effect of HINT2 silencing alone on EMT. Real-time PCR analysis of HIF-2α in HT-29 and SW480 control and HINT2-knockdown cells (E) HINT2 knockdown enhanced HIF-2α expression. Activation of plu567 or pluc567-mut in SW480 cells with or without HINT2 knockdown (n = 3 replicate experiments) (F) Transwell assay showing SW480 cell migration (G) and invasion (H) after 48-h incubation (×200 magnification). Results were plotted as the average number of migrated cells from six random microscope fields. *P < 0.05, ***P < 0.001 compared to controls.
Oncotarget 13527 www.impactjournals.com/oncotarget 517-521 nt; HRE-2 at 525-529 nt; and HRE-3 at 630-634 nt. HIF-1α directly regulated ZEB1 expression through HRE-3. Mutation of both HRE1 and HRE2 revealed no changes in ZEB1 relative luciferase activity after altering HINT2 expression, and demonstrated that HINT2 downregulation leads direct HIF-2α activation of ZEB1. This is the first evidence that HINT2 downregulation promotes CRC metastasis through HIF-2α-mediated regulation of ZEB1 expression. HINT2 might be a useful clinical biomarker, indicating CRC progression and metastasis risk. supplemented with 10% FBS. All cells were purchased from the Type Culture Collection cell bank (Chinese Academy of Sciences, Beijing, P.R. China) and maintained in a 5% CO 2 humidified atmosphere at 37°C.

Human colorectal cancer samples
Human CRC specimens (including one with liver metastatic tissue) were obtained from 46 CRC patients treated in the Department of Oncological Surgery, Fujian Provincial Hospital, Fuzhou, China. All patients had consented to tissue collection for research at the time of surgery. All specimens were diagnosed as CRC by the Department of Pathology, Fujian Provincial Hospital. No patient had received chemotherapy, immunotherapy or radiotherapy prior to specimen collection. Tissue samples were divided into two portions: the first portion was frozen at −80°C for mRNA isolation and the second was used for pathological examination. This research was approved by the Ethics Committee of Fujian Medical University.

Immunohistochemistry (IHC)
Paraffin-embedded tissue sections were dewaxed and heated in a pressure cooker. Endogenous peroxidase activity was blocked using H 2 O 2 . Sections were incubated with normal serum buffer for 30 min and then with primary antibodies (diluted 1:100) for 1 h at room temperature. Sections were then incubated with horseradish peroxidase-conjugated secondary antibodies before a DAB kit (BosterBio, Wuhan) was used to develop the color reaction. Sections were counter-stained with hematoxylin and eosin (HE), and dehydrated using xylene and alcohol before mounting. Images were acquired with a Leica2000 fluorescence microscope. IHC staining intensity was measured using Image-Pro Plus software (Media Cybernetics, Rockville, MD, USA) in three randomly selected high-power (×400) fields per section.

Luciferase reporter assay
To examine transcriptional regulation of CDH1 promoters, the CDH1 promoter sequence was amplified from normal human genomic DNA and subcloned into a pGL3-basic luciferase reporter vector (Promega) to generate the CDH1 promoter/firefly luciferase reporter construct, pGL3-CDH1.
48 h after transfection, luciferase activities were measured using a Dual-Luciferase Reporter Assay System (Promega), and results were normalized to Renilla luciferase activity. Firefly luciferase activity in cells transfected with sh-HINT2 was represented as the percentage of activity relative to that of cells transfected with control shRNA. All experiments were repeated at least in triplicate.

Quantitative real-time PCR (qRT-PCR)
Total RNA was extracted from cells using TRizol (Invitrogen) according to the manufacturer's instructions. cDNA was synthesized using a PrimeScript ® RT reagent kit (Takara Biotechnology). mRNA levels were detected in triplicate with SYBR ® Premix Ex Taq (Takara). Transcript levels are expressed relative to levels observed in experimental or control cells and tissues, which was set to 1.0. Relative gene expression was calculated using the comparative threshold value (ΔΔCt) method. Primers

Western blotting
Total proteins were extracted from cells or tissue. Equal amounts of protein were resolved using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12.5%), transferred onto a nitrocellulose membrane, and then blocked with 5% non-fat dry milk in phosphate-buffered saline (PBS)-Tween 20 containing 1% (wt/vol) bovine serum albumin (BSA). After blocking, membranes were incubated with polyclonal antibodies against HINT2, CDH1, CDH2, Vimentin and GAPDH, followed by immunoglobulin G (IgG) secondary antibody at 4°C overnight. Immunoreactive proteins were then detected using an enhanced chemiluminescence detection system (Amersham, GE Healthcare Bio-Sciences, Pittsburgh, PA, USA). Equal protein loading was confirmed using GAPDH. The experiment was performed three times in triplicate.

Transwell migration and invasion assays
Cells (5 × 10 5 /mL, 200 μL/well) in serum-free medium were loaded into the upper chambers of Transwell inserts (BD Biosciences, San Jose, CA, USA). The lower chambers were filled with medium containing serum. After 48-h incubation, cells on the upper surfaces of the inserts were removed, and cells on the lower surface of the membrane were fixed with 100% methanol and stained with 0.1% crystal violet. To minimize bias, cells from at least six randomly selected fields were counted and averaged under ×200 magnification, and images were captured using an inverted light microscope. Each study was conducted in triplicate in three independent experiments. For the invasion assay, 5 × 10 6 /mL (200 μL/well) cells were loaded into the top of a 24-well Matrigel-coated insert (8-μm pores; BD Biosciences) according to the manufacturer's protocol. After 48 h, non-invading cells were removed using cotton swabs. Cells that had migrated to the bottom of the membrane were fixed with 100% methanol and stained with 0.1% crystal violet. Invading cells were counted in the same manner as in migration assays.

Wound-healing assay
Cells were seeded in 6-well plates and incubated until 90% confluent. The wound was generated by scratching the monolayer surface with a 200-μL pipette tip, and cells were then washed three times with PBS to remove cell debris. Images were obtained by microscopy at 0 and 24 h, and distances covered by migrated cells were quantified.

Animal models
A liver metastasis assay was used to determine the effects of HINT2 on metastasis. Our study was performed in accordance with institutional guidelines, and was approved by the Animal Experimentation Committee of the Central Institute for Experimental Animals, Fujian Medical University. We bred female BALB/c NOD mice to 4-5 weeks of age. Colon cancer cells were harvested with 0.25% trypsin-EDTA solution, washed, and suspended in serum-free medium at 2 × 10 5 cells/mL. Experimental liver metastases were generated via intrasplenic injection of 1 × 10 5 cancer cells (50 μL of cell suspension) and splenectomy. Mice were sacrificed six weeks later and liver metastases enumerated immediately without fixation. Numbers of metastatic liver clones in each liver, and small foci were compared in sh-HINT2 group and sh-control group.

Statistical analysis
Each study was repeated at least three times. Data are presented as the mean ± SD. All statistical analyses were conducted with SPSS software version 13.0 (SPSS Inc., Chicago, IL, USA). Between-group and amonggroup comparisons were conducted using Student's t-tests and analysis of variance (ANOVA), respectively. P < 0.05 was considered statistically significant.