ANGPTL4-Mediated Promotion of Glycolysis Facilitates the Colonization of Fusobacterium nucleatum in Colorectal Cancer

Figures

• 

  Figure 1.. Enrichment of F. nucleatum in colorectal cancer cells is related to enhanced glycolysis. A, Abundance of Fusobacterium genus (g_Fusobacterium) calculated with RNA-seq data was derived from TCGA-COAD cohort. The median and sample size are shown below each dataset. NT, tumor-adjacent normal tissue; PT, primary tumor. B, Representative images and quantification of F. nucleatum (FN) stained with FISH in coculture with DLD1 or NCM460. Human cell nuclei were stained with DAPI. n = 10 samples. C, Quantification of F. nucleatum stained with FISH in coculture with different colorectal cancer cell lines or NCM460. n = 8 samples. D, Relative ECAR of different cell lines cultured with or without F. nucleatum. n = 8 samples. E and F, Quantification of stained F. nucleatum biomass in coculture with DLD1 (E) or SW480 (F), with or without treatment of 2DG (10 mmol/L). n = 8 samples. Representative images of stained F. nucleatum and DLD1 cocultures are shown in E. G, Planktonic growth of F. nucleatum measured by OD600nm under the treatment of vehicle or 2DG. n = 8 samples. H, Xenograft tumor volume under different conditions. Mice were treated with F. nucleatum and with PBS vehicle (n = 7) or 2DG…

• 

  Figure 2.. F. nucleatum promotes glycolysis via inducing ANGPTL4 expression in colorectal cancer cells. A, GSEA of RNA-seq data compared the transcriptome of DLD1 cultured with or without F. nucleatum (FN). ES, enrichment score; NES, net enrichment score. n = 3 samples. B, Selection of the DEGs between FN (−) and FN (+) groups that were also related to the enrichment of gene sets of Hallmark_Hypoxia and/or Hallmark_Glycolysis. Numbers of overlapped genes are indicated. FC, fold change; q, adjusted P value. C, Heatmap for DEGs selected in B. n = 3 samples. D, qPCR analysis of the selected DEGs in DLD1 cultured with or without F. nucleatum. n = 3 samples. E, Spearman correlation between the abundance of Fusobacterium genus (g_Fusobacterium) and expression level of ANGPTL4 in primary tumor tissues (n = 469) of TCGA-COAD patients. Data were acquired from TCGA database and Poore and colleague's study (21). The orange line represents the best-fit linear regression, and the dotted line represents the 95% confidence intervals. F, Overall survival plots of patients with high (n = 68) and low (n = 68) ANGPTL4 expression levels (quartile group cutoff) analyzed in the GEPIA database. Log-rank test. HR,…

• 

  Figure 3.. The colonization of F. nucleatum in colorectal cancer cells depends on ANGPTL4. A–C, Representative images and quantification of F. nucleatum (FN) stained with FISH in coculture with DLD1. The impact of ANGPTL4 knockdown (A), ANGPTL4 overexpression (B), or treatment with 2DG (10 mmol/L)/rhANGPTL4 (5 μg/mL; C) on F. nucleatum colonization in cocultures was evaluated. n = 8 samples. D, qPCR measurement of ANGPTL4 mRNA level in DLD1 cultured with or without F. nucleatum, treated with or without 2DG (10 mmol/L; n = 6 samples). E, ANGPTL4 expression level in xenograft tumors under different conditions. Mice were treated with AAV-delivered shRNA for nontarget control (shCtrl) or ANGPTL4 knockdown (shANGPTL4), and with or without F. nucleatum. n = 8 for shCtrl and shCtrl + FN group; n = 9 for shANGPTL4 group; and n = 10 for shANGPTL4 + FN group. F, Xenograft tumor volume under different conditions. G, Representative images of tumors under different conditions. H, Relative F. nucleatum abundance in xenograft tumors quantified with qPCR. The data are presented as mean ± SEM. Each circle represents an individual sample. Samples were collected from three independent experiments in A–D and…

• 

  Figure 4.. ANGPTL4 facilitates F. nucleatum colonization by enhancing glucose uptake. A, mRNA levels of genes encoding glucose transporter (GLUT) 1–14 revealed by the RNA-seq data (n = 3 samples). FN, F. nucleatum. B, Western blot analysis of ANGPTL4 and GLUT1 in DLD1 cultured with or without F. nucleatum. The bands of ANGPTL4 in the top panel were observed at 35 and 50 kDa according to the manufacturer's instruction. C, Spearman correlation between expression levels of ANGPTL4 and GLUT1 in the primary tumor tissues (n = 469). Data were derived from TCGA-COAD cohort. The orange line represents the best-fit linear regression, and the dotted line represents the 95% confidence intervals. D, Glucose uptake of DLD1 cultured with or without F. nucleatum determined by 2-NBDG fluorescent signals. n = 12 samples. E–G, The impact of ANGPTL4 overexpression on GLUT1 mRNA level (E; n = 6 samples), protein level (F), and glucose uptake (G; n = 12 samples) of DLD1. H–J, The impact of ANGPTL4 knockdown on GLUT1 mRNA level (H; n = 6 samples), protein level (I), and glucose uptake (J; n = 12 samples) of DLD1. K, Representative images and quantification of FISH-stained F. nucleatum in coculture with DLD1,…

• 

  Figure 5.. F. nucleatum colonization enhances the binding of HIF1α with ANGPTL4 and the level of acetylation of H3K27ac. A, ChIP-qPCR analysis of the level of H3K27ac and HIF1α binding within the promoter region of ANGPTL4, in DLD1 cultured with or without F. nucleatum (FN). Nonspecific DNA immunoprecipitation in isotype IgG control was not detectable (qPCR CT value >45) in all samples. Top, peak regions of H2K27ac and HIF1α binding within ANGPTL4 locus determined with ChIP-seq datasets in Cistrome Data Browser (n = 3 samples). TSS, transcription start site. B–G, Overall H3K27ac level determined by Western blot in different cell lines with indicated treatment. Gentamicin (300 μg/mL) and metronidazole (200 μg/mL) were added into the cocultures at the same time of F. nucleatum inoculation when antibiotics treatment was indicated in B and C. F. nucleatum or P. gingivalis (PG) was inoculated at an MOI of 5:1 in B–E and at an MOI of 5:1, 10:1, or 20:1 in F and G as indicated. H, Representative images of FISH-stained F. nucleatum (purple) and IHC-stained H3K27ac (brown) in paraffin-embedded colon adenocarcinoma tissues (n = 27). Human cell nuclei were stained with DAPI (gray) in FISH images and…

• 

  Figure 6.. F. nucleatum–induced ANGPTL4 expression is dependent on H3K27ac. A–D, Overall H3K27ac level (A and C) and ANGPTL4 mRNA (B and D) quantification in DLD1 (A and B) or SW480 (C and D) cultured with or without F. nucleatum (FN), treated with DMSO vehicle or A-485 (1 μmol/L; n = 6 samples in B and n = 3 samples in D). E, ChIP-qPCR analysis of the level of HIF1α binding within the promoter region of ANGPTL4 in DLD1 cocultured with F. nucleatum and treated with DMSO vehicle or A-485 (1 μmol/L). Nonspecific DNA immunoprecipitation in isotype IgG control was not detectable (qPCR CT value >45) in all samples (n = 3 samples). F, Representative images and quantification of FISH-stained F. nucleatum in coculture with DLD1, under the treatment of DMSO vehicle, A-485 (1 μmol/L), or A-485 (1 μmol/L) + rhANGPTL4 (5 μg/mL). Human cell nuclei were stained with DAPI (n = 8 samples). G and H, Western blot analysis of ANGPTL4 and GLUT1 in DLD1 (G) or SW480 (H) cultured with or without F. nucleatum, treated with DMSO vehicle or A-485 (1 μmol/L). The bands of ANGPTL4 at the top were observed at 35 and 50 kDa according to the manufacturer's instruction. I, The effect of A-485 (1 μmol/L) treatment on…

• 

  Figure 7.. HDAC activity is suppressed in colorectal cancer cells cocultured with F. nucleatum. A,In situ HDAC activity measured in DLD1 or NCM460 cultured with or without F. nucleatum (FN); 1 μmol/L TSA was used as a positive control for HDAC activity inhibition. deAc-FdL, deacetylated FLUOR DE LYS substrate (n = 4 samples). B–D, Nuclear protein was extracted from DLD1 or NCM460 cultured with or without F. nucleatum or P. gingivalis (PG). HDAC activity (B; n = 4 samples) or H3K27ac level (C and D) was quantified, respectively. E, mRNA levels of genes encoding HDAC1–11 and SIRT1–7 revealed by the RNA-seq data (n = 3 samples). F, Spearman correlation between the abundance of Fusobacterium genus (g_Fusobacterium) and the expression level of HDAC1–3 in the primary tumor tissues from TCGA-COAD patients (n = 469). The orange line represents the best-fit linear regression, and the dotted line represents the 95% confidence intervals. G and H, HDAC1–3 mRNA (G; n = 6 samples) and protein (H) levels quantified in DLD1 cultured with or without F. nucleatum. I, The deacetylation activity of recombinant human HDAC1 or HDAC2 (2 μg per assay well) treated with indicated conditioned medium (CM); 1 μmol/L…