Hypomethylation‐mediated activation of cancer/testis antigen KK‐LC‐1 facilitates hepatocellular carcinoma progression through activating the Notch1/Hes1 signalling

Abstract Objectives Kita‐Kyushu lung cancer antigen‐1 (KK‐LC‐1) is a cancer/testis antigen reactivated in several human malignancies. So far, the major focus of studies on KK‐LC‐1 has been on its potential as diagnostic biomarker and immunotherapy target. However, its biological functions and molecular mechanisms in cancer progression remain unknown. Materials and Methods Expression of KK‐LC‐1 in HCC was analysed using RT‐qPCR, Western blot and immunohistochemistry. The roles of KK‐LC‐1 on HCC progression were examined by loss‐of‐function and gain‐of‐function approaches. Pathway inhibitor DAPT was employed to confirm the regulatory effect of KK‐LC‐1 on the downstream Notch signalling. The interaction of KK‐LC‐1 with presenilin‐1 was determined by co‐immunoprecipitation. The association of CpG island methylation status with KK‐LC‐1 reactivation was evaluated by methylation‐specific PCR, bisulphite sequencing PCR and 5‐Aza‐dC treatment. Results We identified that HCC tissues exhibited increased levels of KK‐LC‐1. High KK‐LC‐1 level independently predicted poor survival outcome. KK‐LC‐1 promoted cell growth, migration, invasion and epithelial‐mesenchymal transition in vitro and in vivo. KK‐LC‐1 modulated the Notch1/Hes1 pathway to exacerbate HCC progression through physically interacting with presenilin‐1. Upregulation of KK‐LC‐1 in HCC was attributed to hypomethylated CpG islands. Conclusions This study identified that hypomethylation‐induced KK‐LC‐1 overexpression played an important role in HCC progression and independently predicted poor survival. We defined the KK‐LC‐1/presenilin‐1/Notch1/Hes1 as a novel signalling pathway that was involved in the growth and metastasis of HCC.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is one of the most common and aggressive tumours worldwide. 1 Although successful partial hepatectomy and liver transplantation have markedly improved survival, patients with HCC typically have poor prognosis. Elucidation of the molecular mechanisms underlying HCC progression is critical for identifying novel therapeutic targets and improving the prognosis of HCC patients.
Cancer/testis (CT) antigens are a well-defined class of proteins that have been pursued as potential targets for anti-cancer therapy.
Cancer/testis antigens are typically restricted to gametes and trophoblasts in normal tissues but are aberrantly expressed in a range of human cancers. 2,3 Kita-Kyushu lung cancer antigen-1 (KK-LC-1), encoded by CT83 or Cxorf61 gene, has been characterized as a CT antigen. KK-LC-1 is expressed in 40% of non-small-cell lung cancer tissues, 4 81.6% of gastric cancer 5 and 75% of triple-negative breast cancer. 6 The expression rate of KK-LC-1 in human malignancies is markedly higher than the rates of other CT antigens. Correlated with Helicobacter pylori infection, 7 KK-LC-1 can be frequently detected even in early stage of gastric cancer 8 and at the pre-cancerous condition of the stomach. 9 KK-LC-1 is cancer cell selective and immunogenic, and expressed at high level and frequency in cancers, 6 making it an ideal cancer immunotherapy target. A recent study has implicated the utility of KK-LC-1 as a target for photodynamic therapy in malignancies. 10 However, the biological functions and underlying mechanisms of KK-LC-1 in human cancers remain obscure.
The Notch signalling is one of the well-conserved pathways that are involved in multicellular organism development and homeostasis. Activation of Notch signalling helps to modulate fundamental intracellular processes, including cell cycle regulation, cell differentiation and apoptosis. 11 Dysregulated Notch signalling could result in various diseases, such as leukaemia, Alagille syndrome, pulmonary arterial hypertension and ventricular septal defect. 12 Perturbations of Notch signalling as well as its regulatory molecules have been observed in various cancers. In breast cancer, JAG1 activates Notch signalling and promotes tumour metastasis. 13 Inhibition of Notch1/2 could suppress cell proliferation in renal cell carcinoma. 14 Notch1 overexpression leads to enhanced tumour growth in melanoma. 15 Notch signalling has been reported to be activated in HCC and promotes liver tumour formation. 16 Upon ligand binding, Notch undergoes a series of cleavage events to generate its activated form NICD.
γ-Secretase mediates the last step of the cleavage and fine-tunes the level of NICD in cytoplasm. 17 As a catalytic subunit of γ-secretase complex, presenilin-1 contributes to the intricate regulation of the Notch signalling. 18 A previous study mapped the landscape of CT genes in 19 cancer types 19 and suggested CT83, the gene that encodes KK-LC-1 protein, as a CT gene overexpressed in liver cancer. Here, we reported that upregulated KK-LC-1 played a role in HCC malignant phenotype in vitro and in vivo. With hypomethylated CpG islands leading to its overexpression in HCC, KK-LC-1 modulated the presenilin-1/ Notch1/Hes1 signalling and controlled cell proliferation, invasiveness and epithelial-mesenchymal transition (EMT). To our knowledge, we demonstrated for the first time that KK-LC-1 was involved in HCC progression and may serve as a prognostic predictor for HCC patients.

| Patient samples
We obtained paired HCC and adjacent non-tumorous liver tissues from 60 patients who underwent surgical resection in the First

| RNA extraction and quantitative real-time polymerase chain reaction (RT-qPCR)
TRIzol reagent (Invitrogen, Carlsbad, CA, USA; Thermo Fisher Scientific) was used to extract total RNA from tissues and cells. cDNA synthesis was performed with PrimeScript RT Master Mix (TaKaRa, Dalian, China), and RT-qPCR was performed using TB Green Premix Ex Taq (TaKaRa) on the Applied Biosystems 7900HT Fast Real-Time

Conclusions:
This study identified that hypomethylation-induced KK-LC-1 overexpression played an important role in HCC progression and independently predicted poor survival. We defined the KK-LC-1/presenilin-1/Notch1/Hes1 as a novel signalling pathway that was involved in the growth and metastasis of HCC.
PCR System (Applied Biosystems, Foster City, CA, USA; Thermo Fisher Scientific). GAPDH was used as the internal control. The 2 -ΔΔCt method was employed to analyse the relative expression. The PCR primers used are presented in Table S1.

| Western blot analysis and coimmunoprecipitation
Western blot was performed using an SDS-PAGE Electrophoresis System as described previously 20

| Immunohistochemistry analysis and immunofluorescence
Expression of KK-LC-1, PCNA, E-cadherin, vimentin and NICD1 was evaluated using an immunohistochemical method described previously. 20 The intensity of immunohistochemical staining was set as 0 for negative, 1 for weak, 2 for medium and 3 for strong. The percentage of stained area was scored as 0 for 0%-10%, 1 for 11%-25%, 2 for 26%-50%, 3 for 51%-75% and 4 for 76%-100%. The overall staining score was calculated by multiplying the intensity score and staining area score as previously described. 20 Immunofluorescence was carried out as previously reported. 20

| Construction of stable cell lines
To observe the knockdown effects of KK-LC-1, HCCLM3 and  Table S1.

| Cell migration and invasion assays
Cell migration and invasion were evaluated with transwell inserts coated without Matrigel (for the migration assay) or with Matrigel (for the invasion assay) as reported previously. 21

| Animal studies
All animal studies were approved by the Institutional Animal Care and

| Bisulphite sequencing PCR (BSP)
Genomic DNA was treated with bisulphite using the Methylation-Gold Kit (Zymo Research Corporation) according to the manufacturer's instructions. We predicted CpG island of the CT83 gene locus (between −2000 and +400 bp from transcription start site, NC_000023.11: c116465033-116462634) by MethPrimer 22 (http:// www.urogene.org/methprimer/) and designed primers for quantitative methylation analysis. Sequences of primers are listed in Table   S1. The coverage of product of the primer was between −90 and +250 bp (NC_000023.11: c116463123-116462783) from transcription start site of the CT83 gene and contained 17 CpG sequences.
The modified DNA was amplified using the specific sequencing PCR primers. The 341-bp PCR products were cloned into pMD.19-T Vector (TaKaRa). At least ten independent colonies of each sample were sequenced by GeneCreate (Wuhan, China).

| Statistical analysis
Statistical analysis was performed using the SPSS 21.0 (IBM Corporation, Armonk, NY, USA) and GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA). Wilcoxon signed-rank test was used to compare the expression level of KK-LC-1 between HCC samples and adjacent non-tumorous liver specimens. Other quantitative variables were assessed by Student's t test. Correlations between KK-LC-1 and clinicopathologic characteristics of HCC patients were analysed using Fisher's exact test. Spearman's rank correlation coefficient was used to demonstrate the relationship between CT83 DNA methylation level and expression level of CT83. Patient survival was compared with the Kaplan-Meier method, and the significance was determined by the log-rank test. Cox proportional hazards regression model was used for multivariate survival analysis. Values are expressed as mean ± standard error of the mean. Statistical significance was concluded at *P < 0.05, **P < 0.01, ***P < 0.001.

| KK-LC-1 is frequently overexpressed in HCC and serves as a prognostic indicator
To understand the role of KK-LC-1 in liver cancer, we first determined the expression levels of KK-LC-1 in 60 pairs of HCC samples and adjacent non-tumorous liver specimens by RT-qPCR. We observed that KK-LC-1 was frequently elevated in HCC tissues compared with that in the matched non-tumorous liver samples ( Figure 1A Collectively, these data indicated that KK-LC-1 was frequently overexpressed in HCC and associated with HCC clinicopathological characteristics. KK-LC-1 may be used as a potential prognostic indicator for HCC.

| KK-LC-1 is required for the maintenance of malignant phenotypes of HCC cells
To understand the functional role of KK-LC-1 in HCC cells, we first performed knockdown of KK-LC-1 via shRNA in SMMC7721 and HCCLM3 with high levels of KK-LC-1. Three shRNAs (sh1-KK-LC-1, sh2-KK-LC-1 and sh3-KK-LC-1) were designed to knock down KK-

LC-1 expression. Expression level of KK-LC-1 was determined by
RT-qPCR and Western blot; sh1-KK-LC-1 was the most effective one F I G U R E 1 KK-LC-1 is frequently overexpressed in HCC and serves as a prognostic indicator. A, The expression level of KK-LC-1 mRNA was compared between 60 paired HCC (T) and adjacent non-tumorous (N) tissues using RT-qPCR. GAPDH was used for normalization. B, KK-LC-1 mRNA expression level was elevated in 60 paired HCC tissues compared to that in adjacent non-tumorous samples. Wilcoxon signed-rank test was adopted to analyse the difference between the groups. The data are shown as the mean ± SE. ***P < 0.001. ( Figure 2A,B). Meanwhile, cells with stable lentiviral KK-LC-1 overexpression in Focus and Huh7 were also constructed and named as KK-LC-1-OE cells subsequently (Figure 3A,B). We then performed a variety of in vitro assays to evaluate the effects of KK-LC-1 knockdown or overexpression on cell functions, including proliferation, migration and invasion. Using CCK-8, EdU and colony formation assays, we observed a dramatic reduction in the proliferative capacity of sh-KK-LC-1 SMMC7721 and HCCLM3 cells ( Figure 2C-E). On the contrary, overexpression of KK-LC-1 in Focus and Huh7 cells exhibited a significant increase in proliferative capacity ( Figure 3C-E).
Additionally, transwell assays were performed, and we found that KK-LC-1 promoted HCC cell migration and invasion ( Figure 2F and Figure 3F).

| KK-LC-1 is associated with EMT process in HCC
Previous reports have indicated that a number of CT antigens induce EMT in human malignancies. 23 Therefore, we speculated that KK-LC-1 might be involved in the EMT process. sh-KK-LC-1 HCCLM3 cells exhibited cobblestone-like epithelial morphology, whereas the F I G U R E 2 KK-LC-1 knockdown suppresses HCC cell proliferation, migration and invasion in vitro. A, KK-LC-1 expression in SMMC7721 and HCCLM3 was stably knocked down using shRNA. RT-qPCR was performed to assess the knockdown efficiency. B, Western blot was conducted to evaluate the effects of sh-KK-LC-1 on SMMC7721 and HCCLM3 cells. C, CCK-8 assays were used to analyse cell proliferation in HCC cells with KK-LC-1 knockdown. D, EdU assays were performed to examine the effects of KK-LC-1 knockdown on HCC cell proliferation ability. Proportion of EdU-positive cells was determined and statistically analysed. E, Colony formation assays were conducted in sh-KK-LC-1 HCC cells and control cells. Number of colonies was counted and shown as histogram. F, Transwell assays were used to analyse HCC cell motility. Migrated and invaded cells were quantified and statistically analysed. The data are shown as the mean ± SE. *P < 0.05, **P < 0.01, ***P < 0.001.

| KK-LC-1 promotes HCC growth and metastasis in vivo
To explore the effects of KK-LC-1 on HCC in vivo growth, we established a xenograft mice model to test HCC cell proliferation. HCC cells with KK-LC-1 knockdown or overexpression were inoculated subcutaneously into nude mice. Our results showed that the growth of tumours from sh-KK-LC-1 cells was significantly inhibited compared with that of the control xenografts. In contrast, the growth of tumours from KK-LC-1-OE cells was markedly accelerated compared with that of the controls (Figure 5A-C). Western blot and immunohistochemical analysis of the xenografts showed that KK-LC-1 knockdown or overexpression was maintained in the xenografts ( Figure 5D,E). PCNA staining of the xenografts indicated lower proliferation rate in mice inoculated with sh-KK-LC-1 HCCLM3 cells and higher proliferation rate in mice injected with KK-LC-1-OE Huh7 cells ( Figure 5E). Wondering whether KK-LC-1 was associated with EMT markers in HCC in vivo, we performed E-cadherin and vimentin immunostaining analysis on tumours derived from xenograft mice models. As indicated in Figure 5E Figure 5F). Given that EMT is an early phenotype contributing to tumour metastasis, we next constructed tail vein injection mice model to evaluate the effect of KK-LC-1 on HCC metastasis. In this model system, the incidence of both intrahepatic and pulmonary metastases in the sh-KK-LC-1 group was significantly decreased compared with the control group, whereas the KK-LC-1-OE group exhibited higher intrahepatic and pulmonary metastatic capacity compared with the control group ( Figure 6A-C).

| KK-LC-1 activates the Notch1 signalling in HCC
Identification of the downstream signalling mediators of KK-LC-1 is crucial for understanding the mechanisms underlying KK-LC-1 regulation of HCC cells. Previous reports have indicated the significant role of Notch1 signalling in hepatic oncogenesis. 24,25 Recent work has also implicated the activation of Notch1 signalling in CT antigen-mediated cancer progression. 28 As shown in Figure 7A Immunohistochemical analysis of the xenografts indicated the same correlation between KK-LC-1 and NICD1 in vivo ( Figure 7B). To confirm the activation of Notch1 signalling by KK-LC-1 in liver cancer, we treated Huh7 cells with 5 μmol/L of N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butylester (DAPT; Selleck, Houston, TX, USA) for three days to suppress the release of NICD1. Western blot analysis showed that DAPT treatment inhibited KK-LC-1-induced NICD1 release and Hes1 expression, and reversed the altered levels F I G U R E 8 KK-LC-1 physically interacts with presenilin-1 to regulate the Notch1 signalling in HCC. A, presenilin-1 was detected in cell lysates immunoprecipitated with an anti-KK-LC-1 antibody. B, KK-LC-1 was detected in cell lysates immunoprecipitated with an anti-presenilin-1 antibody. C, Double immunofluorescence staining of KK-LC-1 and presenilin-1 was conducted to show that KK-LC-1 and presenilin-1 colocalized with each other. Cell nuclei were stained with DAPI. D, presenilin-1 was knocked down using shRNA in Huh7 cells. E, presenilin-1 was knocked down using shRNA in Huh7 KK-LC-1-OE cells. F, Western blot was performed to evaluate the expressions of NICD1, Hes1, E-cadherin and vimentin in Huh7 KK-LC-1-OE cells with presenilin-1 knockdown. G, CCK-8 assays for Huh7 KK-LC-1-OE cells with presenilin-1 knockdown. H, EdU assays were conducted to assess the proliferation ability of Huh7 KK-LC-1-OE cells with presenilin-1 knockdown. Proportion of EdU-positive cells was determined and statistically analysed. I, Transwell assays were used to analyse HCC cell motility in Huh7 KK-LC-1-OE cells with presenilin-1 knockdown. Migrated and invaded cells were quantified and statistically analysed. The data are shown as the mean ± SE. *P < 0.05, **P < 0.01, ***P < 0.001 of EMT markers in KK-LC-1-OE Huh7 cells ( Figure 7C). Rescue experiments also showed that DAPT treatment reduced the elevated proliferative, migrative and invasive ability of KK-LC-1-OE Huh7 cells ( Figure 7D-F). These results concluded that KK-LC-1 facilitated HCC progression through activating Notch1 signalling.

| KK-LC-1 physically interacts with presenilin-1 to regulate the Notch1 signalling in HCC
We investigated the mechanism by which KK-LC-1 regulated the Notch1 signalling. Presenilin-1 is a catalytic subunit of the γ-secretase complex, which is an endoprotease complex that catalyses the S3 cleavage of Notch1. As the core component of γ-secretase, presenilin-1 functions as the upstream modulator of Notch1 signalling in various physiological and pathophysiological conditions. Increased expression of presenilin-1 has recently been correlated to HCC cell proliferation. 29 Therefore, we hypothesized that presenilin-1 may be involved in KK-LC-1-mediated liver cancer progression.
We first investigated whether KK-LC-1 could interact with presenilin-1 in HCCLM3 cells. As shown in Figure 8A anti-presenilin-1 antibody ( Figure 8B). Double immunofluorescence staining further suggested that both proteins were colocalized in the cytoplasm ( Figure 8C). To further address the role of presenilin-1 in KK-LC-1-mediated HCC progression, we knocked down the expression of presenilin-1 using shRNA in Huh7 cells ( Figure 8D) and Huh7 KK-LC-OE cells ( Figure 8E). We also showed that the KK-LC-1-induced Notch1 activity and EMT were restored after presenilin-1 knockdown ( Figure 8F). Rescue experiments showed that the malignant phenotypes in KK-LC-1-OE Huh7 cells were ameliorated by presenilin-1 silencing, as observed using the CCK-8, EdU and transwell assays ( Figure 8G-I). Taken together, these results demonstrated that KK-LC-1 promoted Notch1-mediated HCC progression through physically interacting with presenilin-1.

| CpG hypomethylation induces expression of KK-LC-1 in HCC
Previous works suggested that CT83, the gene that encodes KK-LC-1 protein, is highly hypomethylated in the basal form of breast cancer. 6,30 Therefore, we wondered whether this epigenetic regulatory mechanism existed in HCC. We analysed two gene methylation microarrays (GSE37988 and GSE57956) and two gene expression microarrays (GSE14520 and GSE25097) from GEO database to explore the relationship between CT83 DNA methylation levels and its ex- Importantly, KK-LC-1 protein was also induced in these two cell lines treated with 5-Aza-dC ( Figure 9C). To confirm the results in tissue samples, we performed MSP analysis on three paired HCC specimens ( Figure 9D). The methylation rate of CT83 in tumour tissues was markedly lower than that in the corresponding non-tumorous tissues. We subsequently applied BSP analysis to examine the methylation status of the region upstream of CT83 in one matched HCC tissue and non-tumorous specimen ( Figure 9E,F). According to the BSP results, we verified lower CpG methylation level in HCC tissue with higher expression level of CT83 as compared to adjacent nontumorous liver tissue. Collectively, these data affirmed that the activation of KK-LC-1 in HCC is attributed to the CpG hypomethylation.

| D ISCUSS I ON
Cancer/testis antigens have emerged as a class of tumour-associated antigens with promising clinical implications as biomarkers and immunotherapeutic targets for malignancies. 31,32 Many CT antigens have been revealed to be involved in liver cancer progression. 33,34 KK-LC-1, a CT antigen with remarkably higher expression rate than that of other CT antigens, possesses a great advantage in cancer targeted therapies. 10 However, functional and mechanistic investigation of KK-LC-1 in human malignancies has never been reported before. In this study, we demonstrated for the first time that elevated KK-LC-1 played an important role in HCC progression.
Presenilin was first identified as the site of missense mutations leading to early-onset familial Alzheimer's disease. 36,37 The association of presenilin-1 with human malignancies has also been implicated. 38,39 A recent study has proposed the role of presenilin-1 in facilitating HCC progression. As reported, presenilin-1 level is elevated in HCC, accompanied by increased γ-secretase activity.
Presenilin-1 promotes cell proliferation in HCC. 29 Presenilin-1-dependent γ-secretase is a crucial proteinase for Notch1 activation. 40 γ-secretase is a complex comprising Presenilin, anterior pharynxdefective 1, nicastrin and Presenilin enhancer-2. Presenilin is recognized as the catalytic subunit and is essential for the activity of γ-secretase complex. Presenilin, together with other components, forms the functional enzyme to catalyse the processing of Notch. 41 In the present study, we found that KK-LC-1 physically bound to presenilin-1 and affected the downstream Notch signalling.
The Notch pathway is crucial for tissue and organ development and has been shown to participate in HCC progression. 42,43 In mammals, four types of Notch receptors (Notch1-4) have been described. Among these receptors, Notch1 is dominantly expressed in hepatocytes. 44 High Notch1 expression is associated with large tumour size, microvascular invasion and HCC metastasis. 45 Notch1 overexpression predicts an unfavourable survival outcome for liver cancer patients. 46 Our results showed that manipulation of KK-LC-1 was correlated with altered Notch activity. Inhibition of the Notch signalling by DAPT decreased the enhanced malignant phenotypes induced by KK-LC-1 overexpression, indicating that KK-LC-1 augmented HCC growth and metastasis through the Notch1 signalling.
Hypomethylated CpG islands have been reported to be associated with the activation of CT genes in several types of cancers.
CT45, a 6-member family of X-linked CT genes, is upregulated in epithelial ovarian cancer. Elevated level of CT45 expression is demonstrated to be induced by promoter hypomethylation. 47 In lung adenocarcinoma, PIWIL1 has been identified as an extremely highly expressed CT gene. Promoter DNA hypomethylation of PIWIL1 causes its overexpression. 48 In the present study, we found that hypomethylated CpG islands caused the upregulation of CT83, which is the gene that encodes KK-LC-1. We detected a lower rate of CT83 methylation in HCC samples than that in the paired non-tumorous tissues. As confirmed by BSP results, decreased CpG methylation level was associated with the reactivation of CT83 in HCC. 5-Aza-dC treatment in HCC cells with low CT83 level led to elevated CT83 expression.
In conclusion, we demonstrated that hypomethylation-induced KK-LC-1 facilitated HCC progression. KK-LC-1 was suggested as a prognostic indicator for HCC. Based on the mechanistic experiments, we found that KK-LC-1 physically interacted with presenilin-1 to promote the Notch1/Hes1 pathway in HCC, contributing to a better understanding of the intricate regulatory network underlying HCC pathogenesis. Targeting the KK-LC-1/presenilin-1/Notch1/ Hes1 pathway could be a candidate for molecularly targeted therapy in HCC.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interests.