Fig 2: METTL3 stabilize DEK mRNA via m6A modification in GC. (A,B) RIP-qPCR assays were used to detect the enrichment of METTL3 binding DEK in MGC80-3 and AGS cells (*, vs. IgG, p < 0.05; #, vs. anti-METTL3 in NC group, p < 0.05). (C,D) MeRIP-qPCR assay assessing m6A enrichment of DEK mRNA when METTL3 was overexpressed in MGC80-3 and AGS cells (n = 3, **: p < 0.01). (E) Luciferase reporters of wild type and mutation DEK were constructed. (F,G) Mutation of the m6A recognition sequence or knockdown of METTL3 increased luciferase activity in MGC80-3 and AGS cells (n = 3, *: p < 0.05). (H,I) METTL3 was overexpressed in MGC80-3 and AGS cells, and the half-life of DEK mRNA was detected by qPCR (n = 3, **: p < 0.01).

Fig 3: METTL3 regulated DEK expression in GC. (A) DEK m6A modification site prediction pattern map predicted by m6A2Target and SRAMP database (Both pre-mRNA genomic sequences (with introns) and mature mRNA sequences (containing 3- and 5-terminal non-coding sequences) of DEK were predicted at the SRAMP site. Site 1 was not predicted from the pre-mRNA genome sequence). (B) The GEPIA database predicts the correlation of METTL3 with DEK expression. (C) The ENCORI website predicts METTL3 in association with DEK expression. (D) Representative images of METTL3 and DEK immunohistochemical staining in tumor sections from the same patient. (E) The graph shows the correlation of METTL3 with DEK IHC scores and gives the Pearson correlation coefficient (R), p-value, as well as simple numbers (n) in the upper corner. (F,G) METTL3 was overexpressed or knocked down in MGC80-3 and AGS cells, DEK mRNA levels were detected by qPCR, and protein levels were detected by Western blot (n = 3, **: p < 0.01).

Fig 4: Screening of immune-related histone chaperones and clustering of 369 TCGA-HCC samples based on HSPA8 and DEK. (A) Correlation analysis between 36 known histone chaperones and the four ESTIMATE indices, including stromal score, immune score, ESTIMATE score and tumor purity. (B,C) Comparison of the mRNA expression of HSPA8 or DEK between tumor and normal tissues from the TCGA database. (D,E) Comparison of the protein level of HSPA8 or DEK between tumor and normal tissues from the HPA database. Scale bar, 100 µm. (F,G) Correlation analysis between HSPA8 or DEK and the abundance of tumor-infiltrating lymphocytes by TISIDB. (H) TCGA-HCC samples are divided into two clusters according to HSPA8 and DEK via consensus clustering. (I) Spearman rank correlation analysis between the expression of HSPA8 and DEK. *** p < 0.001.

Fig 5: DEK is highly expressed in GC tissues and cells. (A) TCGA database analysis of DEK levels in GC tissues and normal tissues. (B) DEK expression levels in STAD are predicted at the UALCAN. (C) Representative images of DEK immunohistochemistry (IHC) in normal tissue of GC (left) and GC tissue (right). (D) Western Blot assay detected DEK protein in normal tissues and GC tissues in 6 pairs of samples. (E) The Western Blot signal for each patient was normalized with GAPDH and presented as a histogram. (F) qPCR experiments to analyze DEK mRNA levels in GES1, HGC27, MGC80-3, AGS, and SGC7901 cell lines and normalized to GAPDH. (G,H) Western blot assay to detect DEK protein expression in GES1, HGC27, MGC80-3, AGS, and SGC7901 cell lines. Western Blot signals were normalized with Actin. Among them, the differentiation degree of HGC27, MGC8-3, AGS and SGC7901 cells increased in turn.