Dietary vitamin B3 supplementation induces the antitumor immunity against liver cancer via biased GPR109A signaling in myeloid cell

Abstract

The impact of dietary nutrients on tumor immunity remains an area of ongoing investigation, particularly regarding the specific role of vitamins and their mechanism. Here, we demonstrate that vitamin B3 (VB3) induces antitumor immunity against liver cancer through biased GPR109A axis in myeloid cell. Nutritional epidemiology studies suggest that higher VB3 intake reduces liver cancer risk. VB3 supplementation demonstrates antitumor efficacy in multiple mouse models through alleviating the immunosuppressive tumor microenvironment (TME) mediated by tumor-infiltrating myeloid cell, thereby augmenting effectiveness of immunotherapy or targeted therapy in a CD8⁺ T cell-dependent manner. Mechanically, the TME induces aberrant GPR109A/nuclear factor κB (NF-κB) activation in myeloid cell to shape the immunosuppressive TME. In contrast, VB3 activates β-Arrestin-mediated GPR109A degradation and NF-κB inhibition to suppress the immunosuppressive polarization of myeloid cell, thereby activating the cytotoxic function of CD8⁺ T cell. Overall, these results expand the understanding of how vitamins regulate the TME, suggesting that dietary VB3 supplementation is an adjunctive treatment for liver cancer.

Graphical abstract

Figure 1

Dietary VB3 supplementation reduces the risk of human liver cancer (A) Multivariable logistic regression models to analyze the relationship between dietary B group vitamins intake and digestive system cancers in the NHANES database; error bars represent 95% confidence intervals. (B) Multivariable Cox regression models to analyze the relationship between dietary B group vitamins intake and liver cancer in the UK Biobank database; error bars represent 95% confidence intervals. (C) Kaplan-Meier survival curves for VB3 supplementation with overall cancer-related mortality outcomes in the NHANES database.

Figure 2

Dietary VB3 supplementation inhibits liver cancer growth (A–C) The schedule of establishing subcutaneous liver cancer models and VB3 treatment (A). Tumor volumes (B) and tumor weights (C) of Hepa1-6 (n = 8), LPC-H12 (n = 7), and H22 (n = 6) were measured. (D–F) The schematic of the experimental design and timeline for DEN+CCl4-indued primary liver cancer model and VB3 treatment (D). Representative images of liver were shown (E). Total tumor number and liver weight were quantified (F, n = 12). (G) Representative histopathology images of the liver. Scar bar, 210 μm. (H) Representative images of AFP and Ki67 staining. Scar bar, 100 μm. (I) Tumor volumes and tumor weights were observed in the prevention (P) and treatment (T) model of VB3 intervention. Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by two-way ANOVA (B and I), one-way ANOVA (C and I), or Student’s t test (F). ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Figure 3

VB3 remodels immunosuppressive TME to activate adaptive immune response (A–D) The frequency of infiltrated immune cells in the TME was analyzed from Hepa1-6 (A, n = 8), LPC-H12 (B, n = 7), H22 (C, n = 6), and DEN+CCl4 tumor model (D, n = 12) after VB3 treatment. (E) Representative images of CD8, INOS, Arg-1, Ly6G, cleaved caspase-3, and TUNEL staining form Hepa1-6 tumors. Scar bar, 100 μm for IHC and 60 μm for IF. (F) VB3-induced changes in tumor-infiltrated CD8⁺ T cells were determined (n = 6), including activation (naive-like CD8⁺ T, CD8⁺CD44⁻CD62L⁺; effector CD8⁺ T, CD8⁺CD44⁺CD62L⁻), cytotoxic function (Perforin, GZMB, IFN-γ, and TNF-α), proliferation (Ki67), and exhaustion (PD-1, TIM-3, LAG-3, and TIGIT). (G) VB3-induced CD4⁺ T component changes in the TME were determined (n = 6). CD4⁺ T_con, CD4⁺FOXP3⁻, CD4 T_reg, CD4⁺FOXP3⁺. (H and I) Tumor volumes (H) and tumor weights (I) were observed in VB3-treated mice after CD8⁺ T cells depletion (n = 6). (J) The proportion of infiltrated immune cells in tumors was determined (n = 6). Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by one-way ANOVA (A, B, C, I, and J), Student’s t test (D, F, and G) or two-way ANOVA (H). ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Figure 4

VB3 acts on myeloid cells to boost antitumor effect of CD8⁺ T cells (A and B) Tumor volumes (A) and tumor weights (B) were observed in VB3-treated mice after myeloid cells depletion (n = 6). (C and D) The proportion (C) and cytotoxic function (D) of CD8⁺ T cells were explored (n = 6). (E–G) The myeloid cells were sorted by flow cytometry in tumor tissues. The mRNA expression of Arg-1, CD163, IL-10, VEGF, CCL2, CSF-1, CXCL1, CXCL2, and CXCL5 was detected (n = 6). (H) Primary BMDM, G-MDSC, or RAW264.7 were treated with LPC-H12 CM and VB3 for 12 h. The mRNA expression of Arg-1, CD163, IL-10, VEGF, CCL2, CSF-1, CXCL1, CXCL2, and CXCL5 was detected (n = 3). (I and J) Primary BMDM and G-MDSC were treated with LPC-H12 CM and VB3 for 24 h. The polarization ratio of G-MDSCs or M1-like TAMs and M2-like TAMs were explored by flow cytometry. Representative gating images were shown (I). The polarization ratio was quantified (J, n = 3). (K and L) Carboxyfluorescein succinimidyl ester (CFSE)-labeled CD8⁺ T cells were co-incubated with VB3-primed myeloid cells for 48 h. The proliferation rates of CD8⁺ T cells were detected by flow cytometry. Representative gating images were shown (K). The proliferation ratio was quantified (L, n = 4–5). Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by two-way ANOVA (A), one-way ANOVA (B, C, D, H, J, and L), or Student’s t test (E–G). ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Figure 5

GPR109A is required for the antitumor effect of VB3 (A) The Kaplan-Meier survival analysis for liver cancer patients based on GPR109A mRNA expression. (B) The GPR109A expression on non-immune cells (CD45⁻), other immune cells (CD45⁺CD11b⁻), and myeloid cells (CD45⁺CD11b⁺) was detected by flow cytometry in tumor tissues (n = 6). (C) The proportion of GPR109A⁺ myeloid cells in TME was explored at day 5, 10, and 15 after tumor implantation (n = 6). (D) Representative images of CD11b and GPR109A staining in DEN+CCl4 model. Scar bar, 60μm. (E) The Kaplan-Meier survival analysis for GPR109A expression and MDSCs or macrophages abundance of liver cancer patients. (F) The GPR109A expression on myeloid cells were determined in DEN+CCl4-induced tumor tissues after VB3 treatment (n = 6). (G and H) The mRNA expression (G, n = 6) and protein expression (H) of GPR109A in sorted tumor-associated myeloid cells were explored. (I–K) Tumor volumes (I) and tumor weights (J) in WT and GPR109A−/− mice were observed after VB3 treatment (n = 6). The frequency of infiltrated immune cells in TME was analyzed (K). (L–N) Tumor volumes (L) and tumor weights (M) in WT mice were observed with VB3 and MPN treatment (n = 6). The frequency of different infiltrated immune cells in TME was analyzed (N). (O) Representative gating images of GPR109A expression on CD45⁻ cells, CD45⁺CD11b⁻ cells, and CD45⁺CD11b⁺ cells. Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by one-way ANOVA (B, C, J, K, M, and N), Student’s t test (F and G), or two-way ANOVA (I and L). ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Figure 6

Biased β-Arrestin/GPR109A axis activation in myeloid cells is responsible for antitumor effect of VB3 (A) Myeloid cells from WT mice or RAW264.7 were treated with/without LPC-H12 CM and VB3 for 12 h. The GPR109A mRNA expression in cells was detected (n = 3). (B) Myeloid cells from WT mice or RAW264.7 were treated with LPC-H12 CM and 50 μM VB3 for 0, 3, 6, 12, and 24 h. The protein expression of GPR109A, P-P65, and P65 was detected. (C and E) Myeloid cells from WT or GPR109A−/− mice were treated with/without Hepa1-6 CM and VB3 (50 μM) for 12 h. The mRNA expression of GPR109A, Arg-1, IL-10, VEGF, CCL2, CSF-1, CXCL1, CXCL2, and CXCL5 was explored (C, n = 3). The protein expression of GPR109A, P-P65, and P65 was detected (E). (D and F) Myeloid cells from WT mice were treated with MPN (50 μM) previously for 4 h and treated with Hepa1-6 CM and VB3 (50 μM) for another 12 h. The mRNA expression of GPR109A, Arg-1, IL-10, VEGF, CCL2, CSF-1, CXCL1, CXCL2, and CXCL5 was detected (D, n = 3). The protein expression of GPR109A, P-P65, and P65 was evaluated (F). (G–I) Scheme of myeloid cells treatment and coincubation. Myeloid cells from WT mice were previously stimulated with CM for 12 h before treatment with VB3 (50 μM), ACTD (50 μM), and CHX (50 μM) for another 12 h (G). The mRNA expression (H, n = 3) and protein expression (I) of GPR109A were detected. (J) Myeloid cells from WT mice were treated with LPC-H12 CM and VB3 for 12 h. The binding between GPR109A and β-Arrestin1or β-Arrestin2 was determined. (K and L) β-Arrestin1/2-knockdown RAW264.7 cells were treated with LPC-H12 CM and VB3 (50 μM) for 12 h. The mRNA expression of GPR109A, Arg-1, IL-10, VEGF, CCL2, CSF-1, CXCL1, CXCL2, and CXCL5 was detected (K). The protein expression of GPR109A, P-P65, and P65 was explored (L). Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by one-way ANOVA. ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.

Figure 7

VB3 augments the immune/targeted therapy against liver cancer (A–C) Tumor volumes (A) and tumor weights (B) were observed (n = 6) with/without VB3 and anti-PD-L1 antibody treatment. The frequency of infiltrated immune cells in TME was analyzed (C, n = 6). (D–F) Tumor volumes (D) and tumor weights (E) were measured (n = 8–9) with/without VB3 and lenvatinib treatment. The frequency of different infiltrated immune cells in the TME was explored (F, n = 7–8). Data are presented as means ± SEM. Experiments were conducted independently in triplicates or more. Statistical analysis was performed by two-way ANOVA (A and D) or one-way ANOVA (B, C, E, and F). ns, p > 0.05; ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001.