Epigenetic Silencing of Tumor Suppressor lncRNA NKILA: Implication on NF-κB Signaling in Non-Hodgkin’s Lymphoma

The long non-coding RNA (lncRNA) NKILA, localized to 20q13.31, is a negative regulator of NF-κB signaling. As a CpG island is embedded in the promoter region of NKILA, NKILA is hypothesized as a tumor suppressor lncRNA reversibly silenced by promoter DNA methylation in non-Hodgkin’s lymphoma (NHL). Methylation-specic PCR (MSP) and quantitative bisulte pyrosequencing were performed to detect the methylation of NKILA in normal peripheral blood buffy coats, normal tonsils tissue, NHL cell lines and NHL primary samples. SU-DHL-6 cells were treated with 5-Aza-2'-deoxycytidine for reversal of methylation-associated NKILA silencing. Tumor suppressor properties and biological function of NKILA were demonstrated by knockdown of NKILA in SU-DHL-1 cells.


Background
Non-Hodgkin's lymphoma (NHL) encompasses a heterogeneous group of diseases, including B-cell, T-cell, and natural killer (NK)-cell lymphoma based on the origin and lineage of tumor cells [1]. B-cell NHL comprises more than 70% of all lymphomas while T-cell NHL accounts for 10-15% of all lymphomas [2]. Diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma are the most two common subtypes of NHL [3] whereas NK-cell lymphoma is an aggressive subtype rare in Western countries [4]. The incidence of NHL was 6.7/100,000 among males and 4.7/100,000 among females worldwide, which ranked the 8th and 10th of all types of cancers respectively [5]. The clinical presentations of NHL patients include painless lymphadenopathy, organomegaly and presence or absence of B symptoms (night sweats, weight loss > 10% and fever with temperature > 38°C) [6].
Long non-coding RNA (lncRNA) is broadly de ned as a class of non-coding RNA measuring > 200 nucleotides [14,15]. LncRNAs play essential roles in regulating multiple biological processes, including cellular metabolism, organogenesis, and carcinogenesis [16][17][18]. LncRNA, NKILA (NF-KappaB Interacting LncRNA), localized to 20q13.31, was rstly found to be downregulated in breast cancer. Overexpression of NKILA resulted in inhibition of metastasis and increase of apoptosis by repression of NF-κB signaling activity in breast cancer cells, indicating the tumor suppressor property of NKILA [19]. Moreover, NF-κB signaling pathway is constitutively activated and hence implicated in the pathogenesis of NHL. However, the function of NKILA in lymphoma remains unknown.
As a CpG island is present at the promoter region of NKILA, we postulated that NKILA is a tumor suppressor lncRNA reversibly silenced by promoter DNA methylation in NHL. Herein, the methylation of NKILA will be studied in NHL, and the role of NKILA in lymphomagenesis will also be investigated.

Patient samples
One hundred and two formalin xed, para n-embedded (FFPE) or fresh frozen diagnostic lymph node biopsy tissues, including 56 DLBCL, and 26 MCL and 20 PTCL cases, were acquired from ve hospitals in Hong Kong (Queen Mary Hospital, Kwong Wah Hospital, Princess Margaret Hospital, United Christian Hospital and Pamela Youde Nethersole Eastern Hospital). The diagnosis of lymphoma was based on the WHO (World Health Organization) classi cation [20]. Eleven FFPE tonsil tissues were also obtained from healthy individuals undergoing tonsillectomy. Our study was approved by the Institutional Review Board of Queen Mary Hospital.
Afterwards, the cells were harvested for DNA and RNA extraction.
DNA and RNA extraction DNA from NHL cell lines and healthy peripheral blood was extracted with DNA Blood Mini kit (Qiagen). DNA extraction from frozen patient biopsies was conducted with automated DNA extraction system (DNA Tissue Kit from Qiagen). DNA extraction from FFPE tissues performed by using QIAamp DNA FFPE Tissue Kit (Qiagen). Total RNA was extracted with Direct-zol™ RNA MiniPrep kit (Zymo Research).
Methylation-speci c polymerase chain reaction (MSP) Sodium bisul te conversion was conducted with EpiTect Bisul te Kit (Qiagen). Afterwards, MSP was performed in bisul te-treated DNA with two sets of primers, which were speci c to unmethylated (U-MSP) or methylated (M-MSP) DNA sequence. MSP primers were designed at the CpG island upstream to NKILA gene by online tool Methprimer (http://www.urogene.org/methprimer/). Details of primer sequence and PCR condition for MSP were listed in Table 1  Quantitative reverse transcription polymerase chain reaction (qRT-PCR) Total RNA was isolated with the mirVanaTM miRNA Isolation Kit (Invitrogen), followed by reverse transcription with SuperScript® III (Invitrogen). qRT-PCR was performed with SYBR® Select Master Mix (ABI), and the human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the endogenous control. The relative quantity of NKILA expression was calculated by the method of 2-ΔΔCt and normalized against the endogenous control. The primer sequences for NKILA and GAPDH were listed in Table 1 [19,21].

Quantitative bisul te pyrosequencing
The promoter region of NKILA overlapped with the amplicon of MSP was ampli ed in the bisul te converted DNA with methylation-unbiased primers. Primer sequences and condition for PCR were listed as follows: (1:1000, Cell signaling) at 4°C overnight with gently rotation. The membrane was then washed and incubated with HRP-linked secondary antibodies, anti-mouse (1:3000, Cell signaling) and anti-rabbit (1:3000, Cell signaling) for one hour, followed by TBST washing. The ECL HARP substrate was used on the membrane before developed in X-ray lms.
Nuclear and cytosol fractionations of SU-DHL-1 cells were prepared by the Nuclear/Cytosol Fractionation Kit (BioVision), according to the manufacturers' instructions. The samples were then performed Western blot.

Statistical Analysis
The difference of cell proliferation and cell death between SU-DHL-1 cells transfected with NKILA siRNA and Silencer Negative Control were compared by Student's t-test. The difference of NKILA methylation frequency in different subtypes of NHL primary samples was analyzed by χ 2 test. All P-values were 2sided. P < 0.05 was considered as signi cant difference.

Results
Methylation of NKILA in normal healthy controls and NHL cell lines NKILA was reported to repress NF-κB signaling pathway [16-18] which is constitutively activated and implicated in lymphomagenesis. There is a CpG island at the promoter region of NKILA. Hence, the methylation status of NKILA was investigated by MSP in the bisul te-converted DNA of normal healthy controls, including 10 peripheral blood buffy coats and 11 normal tonsil tissues, in addition to 10 NHL cell lines. Direct sequencing of M-MSP products from methylated positive control DNA demonstrated that all unmethylated cytosines were converted into thymidines after PCR, whereas all methylated cytosines remained unchanged, indicating the complete bisul te conversion and speci city of MSP (Fig. 1a). By MSP, methylation of NKILA was absent in all of normal peripheral blood buffy coats and normal tonsil tissues (Fig. 1b). Amongst NHL cell lines, NKILA was completely methylated (MM) in SU-DHL-6 and completely unmethylated (UU) in GRANTA-519, JEKO-1, MINO, REC-1, SP-53, KARPAS-299 and SU-DHL-1. However, neither U-MSP nor M-MSP signals were observed in SUP-T1 cells (Fig. 1c). Furthermore, the methylation status of NKILA in NHL cell lines detected by MSP was veri ed by quantitative bisul te pyrosequencing. SU-DHL-6 cells with complete methylation of NKILA had a mean methylation percentage of 69.4%. In contrast, NHL cell lines with completely unmethylation of NKILA had a mean methylation percentage ranging from 5.0-6.5%, which con rmed the methylation status detected by MSP (Fig. 1d). These results indicated that NKILA was methylated in a tumor-speci c manner in NHL cells.

Methylation and expression of NKILA in NHL cell lines
To explore the relationship between promoter DNA methylation and the expression of NKILA, semiquantitative RT-PCR of NKILA was performed in NHL cell lines. As demonstrated by the DNA gel, no expression of NKILA was detected in SU-DHL-6 cells that was completed methylated for NKILA. Conversely, expression of NKILA was observed in other cell lines completely unmethylated for NKILA (Fig. 2a).
Furthermore, to study whether promoter DNA methylation was associated with reversible silencing of NKILA, SU-DHL-6 cells, which were completely methylated for NKILA, were treated with a demethylating agent, 5-AzadC for 7 days. Upon treatment with 5-AzadC, the promoter of NKILA was demethylated as illustrated by the emergence of U-MSP signal (Fig. 2b), with re-expression of NKILA (Fig. 2c). Hence, these data suggested that reversible silencing of NKILA was mediated by promoter DNA methylation in NHL cells.

Methylation of NKILA in primary NHL samples
To investigate the methylation of NKILA in NHL primary samples, MSP was performed with bisul teconverted DNA in primary samples, including 26 mantle cell lymphoma (MCL), 56 DLBCL and 20 peripheral T-cell lymphoma (PTCL). MSP results showed that no methylation of NKILA was detected in primary MCL samples. However, NKILA was found to be methylated in 29 (51.79%) DLBCL and 4 (20%) PTCL cases (Fig. 3a-c), hence preferentially methylated in DLBCL than MCL (P < 0.0001) and PTCL (P = 0.007).
NKILA inhibiting IkBα phosphorylation and NF-κB activation NKILA has been reported to suppress NF-κB signaling pathway by blocking IkBα phosphorylation in breast cancer, non-small cell lung cancer and nasopharyngeal carcinoma [19,22,23]. To elucidate NKILA function in lymphoma, NKILA-targeted siRNA was used to knock down NKILA and non-targeting siRNA was used as a control in SU-DHL-1. The qRT-PCR result con rmed NKILA was knocked down at 24-hours and 48-hours post-transfection (Fig. 4a). To further evaluate NKILA function in NF-κB signaling pathway, we examined IkBα phosphorylation and p65 nucleus translocation after NKILA knock-down, with augmentation of the NF-κB signaling by TNFα [19]. The result showed that NKILA knock-down led to increase of IkBα phosphorylation in SU-DHL-1 at both 24-hours and 48-hours post-transfection (Fig. 4b), which was associated with enhanced p65 translocated into nucleus compared to the control. Moreover, enhanced nuclear translocation of phosphorylated p65 ser536 was observed after NKILA knock-down (Fig. 4c). These results collectively indicated NKILA negatively regulated NF-κB signaling pathway by inhibiting IkBα phosphorylation, and reduced nuclear translocation of total and phosphorylated p65.
Effect of knock-down of NKILA on SU-DHL-1 cells As NKILA plays as a negative regulator in NF-κB signaling pathway, its tumor suppressor function in lymphoma was further explored by examining cell proliferation and cell death in SU-DHL-1 cells that were completely unmethylated for NKILA. Knock-down of NKILA led to a signi cantly increased cell proliferation rate compared to the control (Fig. 5a). Furthermore, knock-down of NKILA resulted in reduced cell death in SU-DHL-1 (Fig. 5b). These results supportively indicated NKILA acted as a tumor suppressor lncRNA in SU-DHL-1.

Discussion
Several observations were made in this study. Firstly, we showed that NKILA was methylated in NHL cell lines and NHL primary samples but unmethylated in normal controls, hence methylated in tumor-speci c manner in NHL. This was consistent with tumor-speci c pattern of methylation of tumor suppressor protein-coding genes, such as p16 and p15 [24], and non-coding tumor suppressor miRNAs, such as miR-342-3p [25] and miR-1250-5p [26] in NHL. However, this contrasted with the tissue-but not tumor-speci c pattern of methylation, such as miR-373 [27] and miR-127 [28], were shown to be methylated in both normal counterparts and tumor cells, hence likely unimportant in carcinogenesis.
Secondly, in primary NHL samples, NKILA was frequently methylated in DLBCL but not MCL or PTCL samples. Indeed, NHL is highly heterogenous with different genetic and epigenetic features [31]. For instance, microarray-based DNA methylation study in 367 hematological neoplasms demonstrated that promoter DNA hypermethylation was more frequent in precursor B and T lymphoid neoplasias and mature B-cell lymphomas of germinal center origin (such as DLBCL, FL and Burkitt's lymphoma) than mature T-cell lymphomas such as PTCL [32]. Therefore, differential methylation of NKILA might be accounted by the difference in cell of origin and their inherent pathogenetic mechanisms.
Thirdly, to our knowledge, this is the rst study that reported promoter DNA methylation mediated the reversible silencing of NKILA in NHL, which is evidenced by an inverse correlation between the NKILA methylation and its expression in NHL cell lines, and re-expression of NKILA upon demethylation treatment in cell line with complete methylation of NKILA. Apart from promoter DNA methylation, NKILA have been reported to be regulated by other mechanisms. For instance, Huang et al [29] showed that, in cytotoxic T lymphocytes, NKILA transcription induced by antigen stimulation was mediated by increase of the acetylation of histones (H4ac, H3K27ac, and H3K9ac) at the promoter region of NKILA, suggesting the expression of NKILA regulated by histone modi cation. In addition, in breast cancer cells, some oncogenic miRNAs, such as miR-103 or miR-107, could directly target and downregulate the expression of NKILA, which was con rmed by luciferase reporter assay [19].
Fourthly, NKILA was rst shown to suppress the NF-κB signaling in breast cancer. It could directly bind to NF-κB/IκB complex and mask the phosphorylation site of IκB, thereby suppressing IKK-induced IκB phosphorylation and hence NF-κB activity [19]. As constitutive NF-κB activation is the hallmark of many lymphoid malignancies, including Hodgkin lymphoma, DLBCL and multiple myeloma [30] and hence implicated in lymphomagenesis, it is essential to comprehend the precise function of NKILA and its interaction with NF-κB in lymphoma. To the best of our knowledge, this is the rst study showing the role of NKILA in lymphoma. Previous studies reported that phosphorylation p65 at ser536 led to enhanced transcriptional activity of NF-κB [31][32][33]. We observed that knockdown of NKILA led to upregulation of NF-κB signaling pathway by promoting IκBα phosphorylation, and consequent nucleus translocation of total p65 and phosphorylated p65 in lymphoma cells. Furthermore, as a transcription factor, NF-κB regulates expression of multiple downstream effectors, enhancing cell proliferation, inhibiting cell apoptosis, or promoting cell migration and invasion [34]. Herein, in NHL cells, down-regulation of NKILA resulted in increase of cellular proliferation and decrease of cell death, consistent with the tumor suppressor role of NKILA in multiple tumors, including melanoma, lung cancer, rectal cancer, laryngeal cancer, and breast cancer [19,22,[35][36][37][38]. In addition, NKILA has also been shown to inhibit tumor invasion and migration in epithelial cancers, such as breast cancer, hepatocellular carcinoma, and tongue squamous cell carcinoma [19,38,39]. Collectively, these results suggested that the NKILA could inhibit cellular proliferation and induce cell death by suppressing the NF-κB signaling pathway in NHL cells.

Conclusion
Taken together, epigenetic silencing of lncRNA NKILA was mediated by promoter DNA methylation in a tumor-speci c manner in NHL. Frequent hypermethylation of NKILA was preferentially detected in DLBCL. NKILA exerted its tumor suppressive property by inhibition of cellular proliferation and increase of cell death in NHL cells in association with suppression of NF-κB signaling pathway.

Declarations
The study has been approved by the Institutional Review Board of Queen Mary Hospital (UW 05-269 T/932) and written informed consent has been obtained from patients for the participation of this study.

Consent for publication
All the co-authors consent to publish the work in Cell Communication & Signaling.
Availability of data and materials