Tumour suppressive long non‐coding RNA AFDN‐DT inhibits gastric cancer invasion via transcriptional regulation

Abstract Emerging evidence has revealed that dysregulation of lncRNA is associated with the initiation and progression of cancer. However, the function of these lncRNAs in cancer remains largely unexplored. Here, we reported that AFDN‐DT, an lncRNA that is repressed in gastric cancers (GC), functions as a tumour suppressor by inhibiting cell growth and metastasis through transcriptional repression of genes involved in metastasis. Using in vitro and in vivo models, we demonstrated that overexpression of AFDN‐DT inhibited the proliferation and metastasis of GC. We found that AFDN‐DT was located at the nucleus and interacted with the chromatin in gastric cells. Further, ChIRP‐seq experiments and RNA‐seq analysis revealed that AFDN‐DT directly bound to the promoter regions and regulated the expression of genes essential for malignant transformation. Moreover, we demonstrated that DNA hypermethylation could repress AFDN‐DT expression and treatment with DNA methylation inhibitors restored its expression. Collectively, the results of our study demonstrated the tumour suppressive role of AFDN‐DT in GC and elucidated the transcription regulatory role of tumour suppressive lncRNAs, which can serve as potential prognostic markers for GC.

prognostic values in several forms of cancer. 6,13 Latest evidence on tumorigenesis has revealed that lncRNAs play a vital role in the tumorigenesis of many forms of cancers, 6,13 and several oncogenic lncRNAs, such as NEAT1, 7,14 MALAT1 15,16 and PVT1, 17 are well established. In addition, certain other lncRNAs, such as GAS5 18 and ZFAS1, 19 have been shown to act as tumour suppressors during carcinogenesis.
AFDN divergent transcript (AFDN-DT), also known as MLLT4 antisense RNA 1 (MLLT4-AS1), is an lncRNA located upstream of the Afadin (AFDN) gene. 20 AFDN, otherwise known as MLLT4 and AF6, has been identified as a fusion partner in Lysine Methyltransferase 2A gene (KMT2A). 21 In our previous study, we showed that high levels of AFDN-DT, but not AFDN, correlated with a favourable outcome in patients with GC 20 and suggested that AFDN-DT might be tumour suppressor, even though the mechanism of AFDN-DT in GC remains completely unknown.
In the present study, we revealed that AFDN-DT inhibits cell growth and invasion of GC cells. Moreover, RNA pull-down experiments and RNA-seq analysis revealed that AFDN-DT directly interacts with transcription factors and histones, and it is involved in transcriptional regulation. Moreover, we demonstrated that DNA hypermethylation is mainly responsible for the repressed expression of AFDN-DT, and treatment with DNA methylation inhibitors restored the expression of AFDN-DT.

| Cell line and lentiviral transfection
HGC27 cell line was purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences and cultured in RPMI 1640 (Hyclone) supplemented with 10% foetal bovine serum (FBS) (Biological Industries). The full-length mRNA of AFDN-DT was cloned into the Lenti-CMV vector, and the lentivirus was produced as previously described. 22 The lentivirus exhibiting AFDN-DT overexpression was transfected into HGC27 cells, and the positively transduced cells were selected using puromycin.

| Quantitative reverse transcription PCR (RT-qPCR)
Total RNA was extracted using TRIzol reagent (Life Technologies)

| Animal models
Twenty male BALB/c-nu/nu mice (Shanghai SLAC Laboratory Animal Co., Ltd.) aged 6 weeks, were divided into two groups and were subcutaneously injected with 1 × 10 7 AFDN-DT or vector-transfected HGC27 cells. The tumour volume was assessed 18 days after injection. All the mice were euthanized on day 34, and the tumour weight was evaluated.

| Wound healing assay
AFDN-DT or vector transduced HGC27 cells (5 × 10 5 cells/well) were seeded into 6-well plates (Corning). Wounds were made by scratching the adherent cells on the plate with a sterile 200 μl pipette tip (12 hours after seeding), replaced with fresh culture medium and then cultured for 24 hours. The migration ability was evaluated by analysing the migration of the cells into the wounded area.

| Transwell migration assay
Cell migration was determined using the transwell membranes (Corning). Briefly, the upper transwell chamber was coated with 60 μL of Matrigel (BD) for 2 hours prior to the invasion assays.
AFDN-DT or vector transduced HGC27 cells (2 × 10 5 cells/well) in serum-free RPMI 1640 were seeded into the upper transwell chamber, and 600 μL of RPMI 1640 medium supplemented with 10% FBS was added to the lower chamber. After incubation for 24 hours, the cells that adhered to the upper surface of the membrane were removed. Meanwhile, the invaded or migrated cells, which adhered to the lower surface, were stained with 0.1% crystal violet and measured by optical microscopy.

| RNA sequencing
RNA was extracted using TRIzol reagent (Life Technologies) according to the manufacturer's instructions. The RNA-seq library was constructed using the VAHTS mRNA-seq V2 Library Prep Kit for Illumina® (NR601, Vazyme) according to the manufacturer's instructions. The sequencing was performed using Illumina's Novaseq platform according to manufacturer's instruction. The raw RNA-seq data have been deposited at the GEO database (GSE139326).

| Chromatin isolation by RNA purification (ChIRP)
ChIRP was performed as previously described. 23 The AFDN-DT transduced HGC27 cells were fixed with 1% formaldehyde for

| Bioinformatics analysis
All the sequenced reads were mapped to the hg38 genome using HISAT2, 24 and the sequence alignment mapping (SAM) files were sorted using samtools. 25 The expression of the genes was quantified using the htseq-count, 26 and the differentially expressed genes were identified using the DEseq2. 27 Gene ontology analysis was performed using the DAVID Functional Annotation Bioinformatics Microarray Analysis database. Motif enrichment analysis was performed using the MEME suite. 28

| DNA methylation analysis
Information on DNA methylation and mRNA expression in the gastric tumour samples was obtained from TCGA cohort. 29

| Treatment with DNA methylation inhibitor
HGC27 cells were treated with or without the DNA methylation inhibitor, decitabine, at a final concentration of 5 μmol/L for 24 hours.

| Statistical analysis
All experiments were independently performed in triplicate. GraphPad Prism 7.0a was used for the statistical analyses. Data are presented as the mean ± SD. Statistical significance was determined using Student's t test. A P-value < .05 indicated statistical significance.

| Inhibition of cell growth and invasion of GC cells by AFDN-DT in vitro
To determine the function of AFDN-DT in GC, we firstly over-

| Interaction of AFDN-DT with histone proteins in GC cells
Next, to determine the mechanism underlying AFDN-DT-mediated growth and metastasis inhibition, we first determined the distribution of AFDN-DT lncRNA in the cytoplasm and nucleus. We found that AFDN-DT was mainly expressed in the nucleus, suggesting its potential function in transcriptional regulation ( Figure 3A). Next, we performed an RNA pull-down assay coupled with LC-MS/MS to determine the interactome of AFDN-DT in the HGC27 cells. As shown in Figure 3B, silver-staining results of the proteins pulled down by AFDN-DT suggested that AFDN-DT can specifically bind to certain proteins. Moreover, the LC-MS/MS results revealed 14 proteins that could directly interact with AFDN-DT with high confidence ( Figure 3C).
We found that these target proteins were mainly involved in nucleosome organizations, such as HIST1H4A, HIST1H1C, HIST1H2BK and H3F3C. Notably, the identified targets were core histone proteins, suggesting the possible transcriptional regulatory function of AFDN-DT in GC, and this needs to be explored in future studies.

| Involvement of AFDN-DT in transcriptional regulation through directly binding to the regulatory regions
To investigate the direct targets of transcription regulation by AFDN-DT, we performed the ChIRP using biotinylated antisense DNA probe sets ( Figure 4A). We identified 9162 AFDN-DT targets that directly bind to regions on the chromatin (Table S1)

| Involvement of AFDN-DT regulated genes in transcriptional regulation
To further understand the effects of AFDN-DT on transcriptional regulation, we performed RNA-seq experiments in HGC27 cells with or without AFDN-DT overexpression ( Figure 5A and Table S2).
Following this experiment, we overlapped the genes that were differentially expressed in cells with AFDN-DT overexpression, and the directly bound genes were identified using ChIRP-seq. We identified 415 genes that were directly regulated by AFDN-DT (Figure 5B-C   and Table S3). EGR1, FOS and JUN and their corresponding proteins were all down-regulated by overexpression of AFDN-DT ( Figure 5C).
This observation was further validated by RT-qPCR using HGC27 cells with or without AFDN-DT overexpression ( Figure 5D). Furthermore, the gene ontology analysis revealed that the genes directly regulated by AFDN-DT were mainly involved in transcriptional regulation (Table S4).

| Repression of AFDN-DT by DNA hypermethylation in GC
To determine the mechanism underlying the deregulation of AFDN-DT in GC, We analysed DNA methylation, which is an inheritable epigenetic modification, of the AFDN-DT gene in patients with GC from TCGA cohorts. 29 We found that the DNA methylation of AFDN-DT negatively correlated with AFDN-DT expression in the patients ( Figure 6A). To confirm the direct link between AFDN-DT DNA methylation and AFDN-DT expression, we treated HGC27 cell lines with decitabine, a DNA methylation inhibitor, for 24 hours.
Subsequently, we found that its expression was significantly restored upon inhibition of DNA methylation and the growth of HGC27 cells was significantly repressed (Figure 6B-D). Taken together, these observations showed the mechanism underlying the down-regulation of AFDN-DT in GC.  Many studies on genes coding for tumour suppressive proteins as well as microRNAs have established that DNA hypermethylation contributes to tumour suppression, which eventually leads to carcinogenesis in gastric tissues. 35,36 Here, we reported that DNA hypermethylation of the AFDN-DT promoter down-regulated the expression of AFDN-DT in the tumour samples. Furthermore, we found that demethylation by decitabine increased the expression of AFDN-DT. However, it is also possible that decitabine-induced inhibition of cell growth might be due to the restored expression of other tumour suppressor genes. Therefore, the role of the AFDN-DT-mediated regulation network and its association F I G U R E 6 Down-regulation of AFDN-DT is contributed by DNA hypermethylation in gastric cancer. A, The expression of AFDN-DT was negatively correlated with the DNA methylation of AFDN-DT. The expression of AFDN-DT and the normalized beta value in gastric cancer patient in TCGA cohorts were plotted. B, Decitabine decreased the DNA methylation on the promoter region of AFDN-DT. The DNA methylation was determined by meDIP-qPCR using antibodies against 5-mC. Primers against the CpG island located at the promoter region of AFDN-DT were used. C, DNA demethylation restored the expression of AFDN-DT in HGC27 cells. Relative expression of AFDN-DT was determined with or without Decitabine treatment. D, DNA demethylation induces HGC27 cell growth inhibition. Decitabine was used to inhibit the DNA methylation. Cell growth after 24 h treatment of Decitabine was determined by CCK8 assay. Data represent the mean of five replicates ± SD, **P < .001, ***P < .0001

| D ISCUSS I ON
with other tumour suppressive factors should be explored in the future.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N
YXL designed the study, performed experiments and wrote the manuscripts; JW, LW and CYM performed experiments; KX performed the bioinformatic analysis; PX designed the study, interpreted the results and wrote the manuscript. All authors read and approved the final manuscript.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
The study was approved by the institutional research ethics committee of Songjiang Hospital Affiliated to Shanghai Jiaotong University School of Medicine (Preparatory Stage).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets used and/or analysed during the current study are available from the corresponding author on reasonable request.