The anti-cancer effect of an ω-3 diet is dependent on bone marrow cells with functional GPR120
The anti-cancer effect of an ω-3 diet was previously found to be dependent on functional GPR120 in the host (6), but the host mechanism is yet to be defined. We hypothesized that ω-3 FAs in fish oil act through GPR120 functional immune cells from the bone marrow that migrate to the tumor microenvironment. To test this hypothesis we performed a bone marrow transplant experiment with FVB GPR120 WT or GPR120 KO mice and tail vein-injection of green fluorescent protein labeled bone marrow from GPR120 WT or GPR120 KO mice. After 3 weeks we implanted MycCap cells subcutaneously. Once the MycCap tumors reached 30–50 mm3, the mice were fed an isocaloric ω-3 or ω-6 diet (Fig. 1A). When irradiated GPR120 WT mice were injected with GPR120 WT bone marrow, the ω-3 diet compared with the ω-6 diet significantly inhibited tumor growth and final tumor volume (514.8 ± 88.6 mm3 in ω-6 diet versus 168.4 ± 26.5 mm3 in ω-3 diet, p = 0.0007) and decreased final tumor weight (0.77 ± 0.13 g in ω-6 diet versus 0.31 ± 0.05 g in ω-3 diet, p = 0.002) (Fig. 1B and F). However, when GPR120 KO bone marrow cells were injected into irradiated GPR120 WT mice, feeding the ω-3 diet compared to the ω-6 diet did not reduce tumor growth and final tumor volume (337.2 ± 80.9 mm3 in ω-6 diet versus 328.5 ± 74.8 mm3 in ω-3 diet) and final tumor weight (0.73 ± 0.14 g in ω-6 diet versus 0.63 ± 0.12 g in ω-3 diet) (Fig. 1C and F). Previously, an ω-3 diet was found not to inhibit tumor growth in MycCaP tumors grown in global GPR120 KO mice (Liang P, 2019). However, in the present experiment, after injecting GPR120 WT bone marrow into irradiated GRP120 KO mice, the ω-3 compared to the ω-6 diet inhibited allograft tumor growth and tumor volume (829.3 ± 190.6 mm3 in ω-6 diet versus 291.3 ± 94.1 mm3 in ω-3 diet, p = 0.035) and tumor weight (1.70 ± 0.40 g in ω-6 diet versus 0.61 ± 0.21 g in ω-3 diet, p = 0.04), (Fig. 1D and F). When GPR120 KO bone marrow cells were injected into GPR120 KO recipient mice, feeding the ω-3 diet compared to the ω-6 diet did not reduce tumor growth and final tumor volume (220.6 ± 66.7 mm3 in ω-6 diet versus 259.3 ± 79.7 mm3 in ω-3 diet) and tumor weight (0.43 ± 0.15 g in ω-6 diet versus 0.48 ± 0.12 g in ω-3 diet) (Fig. 1E and F). Throughout the experiments there was no difference in mouse body weight or food intake between the groups (Supplementary Fig. 1).
An ω-3 diet reduced the number of GPR120 wild-type bone marrow-derived M2 macrophages in MycCaP allografts
Previously we reported that feeding an ω-3 diet decreased M2-like macrophage tumor infiltration 6. For the bone marrow transplantation experiments described above, we sought to determine the effect of the ω-3 vs ω-6 diet on bone marrow-derived immune cells infiltrating the MycCap allograft tumors, and the dependence on GPR120. By flow cytometry, 70–80% of the CD45 + cells in the MycCaP tumors were green fluorescent protein positive (GPF+) cells and thus derived from the infused bone marrow (data not shown). Flow cytometry was performed on the tumor GFP + CD45 + cells to determine the populations of bone marrow-derived immune cells in the tumor tissue. In the recipient GPR120 WT mice that received GPR120 WT bone marrow, the number of F4/80 + CD206 + M2 polarized macrophages was significantly decreased from 55.8 ± 2.3% (ω-6 diet) to 45.2 ± 2.1% (ω-3 diet) (p = 0.0036) (Fig. 2A). However, there was no significant difference between the ω-3 and ω-6 groups in the number of F4/80 + CD206 + M2 polarized macrophages in the tumors of the recipient GPR120 WT mice that received GPR120 KO bone marrow (Fig. 2B). In the GPR120 KO mice that received GPR120 WT bone marrow, the number of F4/80 + CD206 + M2 polarized macrophages was significantly decreased from 52.2 ± 1.1% (ω-6 diet) to 43.2 ± 3.3% (ω-3 diet) (p = 0.035) (Fig. 2C). However, there was no significant difference between the ω-3 and ω-6 groups in the number of F4/80 + CD206 + M2 polarized macrophages in the tumors in the GPR120 KO mice that received GPR120 KO bone marrow (Fig. 2D). There was no significant difference between the ω-3 or ω-6 diet groups in the other immune cells infiltrating the allografts such as F4/80 + CD11b + total macrophages, F4/80 + CD68 + M1 macrophages (Fig. 3A-D), CD11b + Gr1 + MDSCs, F4/80-CD11b + Gr1 + neutrophils, CD8+ or CD4 + T cells and B220 + B cells (Supplementary Fig. 2). We isolated CD206+ (M2) macrophages from the MycCaP tumors using magnetic beads to study the effect of the ω-3 diet on gene expression of M2-like macrophage markers, Gene expression of CD206, Arg1, MMP-9, CCL2, IL-10 and TNF-α were significantly decreased in isolated CD206 + cells from allograft tissue in the ω-3 group compared to the ω-6 group from recipient GPR120 WT or KO mice that received GPR120 WT bone marrow (Fig. 3A,C), but not from recipient GPR120 WT or KO mice that received GPR120 KO bone marrow (Fig. 3B,D).
An ω-3 diet decreased expression of cholesterol transporter genes in GPR120 wild-type M2 macrophages isolated from MycCaP allografts
Based on recent studies reporting that cholesterol efflux from TAMs leads to cancer progression, we sought to determine if the ω-3 diet inhibits cholesterol efflux from tumor infiltrating M2-like macrophages 27,28,31. We found that the ω-3 diet as compared to the ω-6 diet significantly reduced gene expression of cholesterol transporters Abca1, Abca6 and Abcg1 in M2 (CD206+) macrophages isolated from MycCaP allograft tumors in irradiated GPR120 WT and KO mice given GPR120 WT bone marrow (Fig. 3E,G). However, there was no significant effect on cholesterol transporters in M2 (CD206+) macrophages isolated from MycCaP tumors in GPR120 WT or KO mice given GPR120 KO bone marrow (Fig. 3F,H).
Docosahexaenoic acid (DHA) decreased MycCaP Conditioned Media-Induced Cholesterol Efflux from M2 Macrophages
Recent studies in ovarian and lung cancer reported that cancer cells are able to induce transition of TAMs to the M-2 phenotype by altering macrophage lipid metabolism 27,28. In the study by Goossens P. et al, co-culture of the ovarian cancer cell line ID-8 with bone marrow derived macrophages (BMDMs) increased cholesterol efflux in lipid rafts of BMDMs and increased IL-4-mediated BMDM reprogramming 27. To examine the role of ω-3 fatty acids affecting prostate cancer-induced alterations in macrophage cholesterol metabolism we utilized an in-vitro model in which Raw 264.7 macrophages were cultured with MycCaP cell conditioned media (CM). MycCaP CM significantly upregulated gene expression of cholesterol transporters Abca1, Abca6 and Abcg1 and cholesterol synthesis-related genes Acat1 and Acat2 and reduced membrane cholesterol levels in Raw 264.7 macrophages, and these effects (excluding Acat2 expression) were inhibited by the ω-3 fatty acid DHA (Fig. 4A-D). MycCaP CM also induced Raw 264.7 macrophages to have characteristics of M2 macrophages (increased expression of Arg 1, CD206, and MMP9) (Fig. 4E). DHA inhibited expression of M2 markers (CD206, Arg 1, MMP9, IL10 and CCL2) in Raw 264.7 macrophages exposed to MycCaP conditioned media and IL-4 (Fig. 4F).
To determine if the inhibitory effects of DHA on cholesterol efflux from M2-like macrophages is mediated through GPR120, we utilized an ex-vivo assay using BMDMs from GPR120 WT and KO mice. MycCap CM increased expression of the cholesterol transporters Abca1 and Abca6 and lowered membrane cholesterol levels in GPR120 WT and KO BMDMs (Supplementary Fig. 3A,B,E). Addition of DHA inhibited these effects in the GPR120 WT BMDMs but had no significant effect in the GPR120KO BMDMs (Supplementary Fig. 3C,D,F). MycCap CM increased IL-4 induced Arg 1 and CD206 expression in GPR120 WT and KO BMDMs and this effect was inhibited by DHA in the GPR120 WT BMDMs but not the GPR120 KO BMDMs (Supplementary Fig. 3G-L).
Inhibition of prostate cancer development in transgenic Hi-Myc mice by an ω-3 diet is dependent on functional host GPR120
To examine the role of GPR120 on the cancer preventive effects of an ω-3 diet we applied the dietary intervention to the Hi-Myc mouse model. Hi-Myc mice develop murine prostatic intraepithelial neoplasia (PIN) as early as 2 to 4 weeks, and transition to invasive adenocarcinoma between 6 and 9 months 21. Hi-Myc GPR120 WT mice and Hi-Myc GPR120 KO mice were fed isocaloric (30% Kcal fat) ω-3 or ω-6 diets, sacrificed at 6-months, and the prostate lobes examined by a pathologist blinded to group assignment. In the Hi-myc GPR120 WT mice, the ω-3 group had a significant 22.6% lower progression to prostate cancer in the lateral prostate lobes at 6-months compared to the ω-6 group (Fig. 5B,C), (77.4%, n = 53 vs. 54.8%, n = 63, p = 0.009). However, there was no significant difference in the progression to prostate cancer in the lateral lobes at 6-months in the ω-3 vs ω-6 diet in the hi-myc GPR120 KO mice (Fig. 5D,E), (78%, n = 41 vs 78.4%, n = 37, p = 0.972). For the above mouse experiments there was no significant difference between the ω-3 and ω-6 groups in the progression to prostate cancer at 6-months in the ventral, anterior, and dorsal prostate lobes (Supplementary Table 3). Throughout the feeding experiments, there was no significant difference in caloric intake or mouse weights between the groups (Supplementary Fig. 4). Examples of H&E immunostains of benign, PIN, and adenocarcinoma are shown in Fig. 5F, G, and H.