Fusobacterium nucleatum upregulates the immune inhibitory receptor PD-L1 in colorectal cancer cells via the activation of ALPK1

Abstract

Fusobacterium nucleatum is a Gram-negative oncobacterium that is associated with colorectal cancer. The molecular mechanisms utilized by F. nucleatum to promote colorectal tumor development have largely focused on adhesin-mediated binding to the tumor tissue and on the pro-inflammatory capacity of F. nucleatum. However, the exact manner in which F. nucleatum promotes inflammation in the tumor microenvironment and subsequent tumor promotion remains underexplored. Here, we show that both living F. nucleatum and sterile F. nucleatum-conditioned medium promote CXCL8 release from the intestinal adenocarcinoma HT-29 cell line. We determined that the observed pro-inflammatory effect was ALPK1-dependent in both HEK293 and HT-29 cells and that the released F. nucleatum molecule had characteristics that match those of the pro-inflammatory ALPK1 ligand ADP-heptose or related heptose phosphates. In addition, we determined that not only F. nucleatum promoted an ALPK1-dependent pro-inflammatory environment but also other Fusobacterium species such as F. varium, F. necrophorum and F. gonidiaformans generated similar effects, indicating that ADP-heptose or related heptose phosphate secretion is a conserved feature of the Fusobacterium genus. By performing transcriptional analysis of ADP-heptose stimulated HT-29 cells, we found several inflammatory and cancer-related pathways to be differentially regulated, including DNA mismatch repair genes and the immune inhibitory receptor PD-L1. Finally, we show that stimulation of HT-29 cells with F. nucleatum resulted in an ALPK1-dependent upregulation of PD-L1. These results aid in our understanding of the mechanisms by which F. nucleatum can affect tumor development and therapy and pave the way for future therapeutic approaches.

Keywords: ADP-heptose; ALPK1; DNA damage response; Fusobacterium; Fusobacterium nucleatum; PD-L1; colorectal cancer; immune checkpoint inhibitor therapy.

Figure 1.

Fusobacterium species induce CXCL8 secretion from intestinal epithelial cells. HT-29 intestinal epithelial cells were stimulated with (a) living Fusobacterium species or (b) sterile conditioned culture medium from Fusobacterium strains for 24 h, after which the CXCL8 secretion was measured. As controls, bacterial medium, cell culture medium (DMEM) or purified flagellin was used. Values represent the mean ± SEM of three independent experiments performed in duplicate. Ratio paired t tests comparing stimulations to DMEM or GMM were used for statistical analysis. *p < 0.05, **p < 0.01.

Figure 2.

The Fusobacterium-released pro-inflammatory factor is a small carbohydrate. Cell-free bacterial conditioned culture medium from F. nucleatum 23,726 was subjected to heat inactivation, sodium periodate treatment and size-based fractionation. (a) HT-29 cells were stimulated with the F. nucleatum 23,726 culture medium or the various treated culture media and CXCL8 release after 24 h of stimulation was measured (b) HEK293T cells were treated similarly as in a). Values represent mean ± SEM of three independent experiments performed in duplicate. Ratio paired t test comparing stimulations to conditioned medium were used for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3.

Fusobacterium species induce NF-κB activation in an ALPK1-dependent manner in HEK293 cells. a) the genetic mutation in the ALPK1 gene in CRISPR-Cas9 edited HEK293 ALPK1^−/− cells was validated using sanger sequencing. b) HEK293 wildtype cells, HEK293 ALPK1^−/− cells supplemented with an empty vector control or ALPK1^−/− cells transfected with an ALPK1 complementation plasmid were stimulated with TNF, adp-heptose, bacterial culture medium or bacterial conditioned culture medium of various Fusobacterium species. NF-κB activation was measured after 5.5 h stimulation and calculated as fold increase over the bacterial medium control. Values represent mean ± SEM of three independent experiments performed in triplicate. Unpaired t tests were used for statistical analysis of HEK293 cells compared to HEK293 ALPK1^−/− cells. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4.

Fusobacterium-conditioned culture medium and live co-culture induce limited CXCL8 in HT-29 ALPK1^−/− cells. (a) the genetic mutation in the ALPK1 gene in CRISPR-Cas9 edited HT-29 ALPK1^−/− cells was validated using sanger sequencing. (b) CXCL8 concentrations of HT-29 ALPK1^−/− cells stimulated with bacterial culture medium or indicated controls for 24 h. c) CXCL8 concentrations of HT-29 ALPK1^−/− cells stimulated with living bacterial cultures at an MOI of 100 or indicated controls for 24 h. Values represent mean ± SEM of two or three independent experiments performed in triplicate. Ratio paired t-test comparing stimulations to GMM or DMEM were used for statistical analysis.

Figure 5.

Adp-heptose promotes a largely pro-inflammatory and pro-carcinogenic signature in HT-29 cells. a) Heatmap of the significantly up- and downregulated genes in HT-29 cells stimulated with adp-heptose compared to non-stimulated (no stim) cells. b) volcano plot of the up- and downregulated genes resulting from adp-heptose stimulation, a number of highly upregulated genes is indicated. c) MSigDB Hallmark over-representation gene set analysis of top five up- and downregulated sets of genes. d) gene set enrichment analysis (GSEA) plots using MSigDB gene sets. The top half shows the running enrichment score for the top five most differentially regulated gene sets (indicated by different colors), colored bars indicate the position of each of the members of the gene sets along the ranked gene list (rank in ordered dataset). Bottom half shows the ranked list metric as bar plot (log2 fold change), which measures the degree of correlation of genes with the gene sets (left for positive and right for negative correlation).

Figure 6.

The impact of adp-heptose stimulation on DNA mismatch repair gene expression. a) gene expression in fragments per kilobase of transcript per million mapped reads (FPKM) of MSH2, MSH6, MLH1 and PMS2 from RNA-sequencing results of non-stimulated, flagellin-stimulated or adp-heptose-stimulated HT-29 cells from three independent experiments. Statistical testing under these three conditions was performed using the empirical Bayesian hierarchical modeling approach encoded in R package EBSeq. ** adjusted p value < 0.01, *** adjusted p value < 0.001. b) HT-29 cells were stimulated with adp-heptose, flagellin or non-stimulated (NS) for 2.5 h and expression of indicated genes was assessed by rt-qPCR. Values are indicated as log2 fold change over the non-stimulated control. Statistical testing was performed by comparing adp-heptose or flagellin to control stimulation. c) HT-29 cells were stimulated with medium control or bacterial conditioned culture medium of indicated strains for 2.5 h and expression of indicated genes was assessed by rt-qPCR. Values are indicated as log2 fold change over the non-stimulated control. Statistical testing was performed by comparing bacterial strains to medium control stimulation. For b and c, values represent mean ± SEM of three independent experiments performed in triplicate and an paired t test was used for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 7.

(Continued).

Figure 7.

Adp-heptose and F. nucleatum upregulate PD-L1 in HT-29 cells in an ALPK1-dependent manner. (a) gene expression in fragments per kilobase of transcript per million mapped reads (FPKM) of PD-L1 from RNA-sequencing results of non-stimulated, flagellin-stimulated or adp-heptose-stimulated HT-29 cells from three independent experiments. Statistical testing under these three conditions was performed using the empirical Bayesian hierarchical modeling approach encoded in R package EBSeq. *** adjusted p value < 0.001. (b) HT-29 cells or HT-29 ALPK1^−/− cells were stimulated with adp-heptose, flagellin or non-stimulated (NS) for 2.5 h and expression of PD-L1 was assessed by rt-qPCR. Values are indicated as log2 fold change over the non-stimulated control. Statistical testing was performed by comparing adp-heptose or flagellin to control stimulation. (c) HT-29 cells or HT-29 ALPK1^−/− cells were stimulated with bacterial conditioned culture medium from indicated strains or medium control. Values are indicated as log2 fold change over the non-stimulated control, statistical analysis was performed by comparing Fusobacterium strains to the medium control. (d) PD-L1 expression in HT-29 cells or HT-29 ALPK1^−/− cells stimulated with ADP-L-heptose (500 ng/ml) for the indicated amount of time. Values are indicated as log2 fold change over the non-stimulated control, was performed by comparing adp-heptose conditions to control stimulation. E) PD-L1 expression in HT-29 cells or HT-29 ALPK1^−/− cells stimulated for 6 h with indicated amounts of ADP-L-heptose. Values are indicated as log2 fold change over the non-stimulated control, statistical analysis was performed by comparing adp-heptose conditions to control stimulation. For B-E, values represent mean ± SEM of three independent experiments performed in triplicate. A paired t test was used for statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 8.

Adp-heptose-mediated upregulation of PD-L1 is dependent on NF-κB. a) adp-heptose stimulation results in the phosphorylation of IκBα in HT-29 as determined by western blot analysis. b) inhibition of NF-κB by BAY 11–7082 abrogates adp-heptose-mediated upregulation of PD-L1 expression. Values are indicated as fold change over the non-stimulated control, statistical analysis was performed by comparing adp-heptose conditions to control stimulation. An unpaired t test was used for statistical analysis. ***p < 0.001. c) schematic overview of the effects of fusobacterium-derived adp-heptose in colorectal cancer cells.