The cholesterol metabolite 27 hydroxycholesterol facilitates breast cancer metastasis through its actions on immune cells

Abstract

Obesity and elevated circulating cholesterol are risk factors for breast cancer recurrence, while the use of statins, cholesterol biosynthesis inhibitors widely used for treating hypercholesterolemia, is associated with improved disease-free survival. Here, we show that cholesterol mediates the metastatic effects of a high-fat diet via its oxysterol metabolite, 27-hydroxycholesterol. Ablation or inhibition of CYP27A1, the enzyme responsible for the rate-limiting step in 27-hydroxycholesterol biosynthesis, significantly reduces metastasis in relevant animal models of cancer. The robust effects of 27-hydroxycholesterol on metastasis requires myeloid immune cell function, and it was found that this oxysterol increases the number of polymorphonuclear-neutrophils and γδ-T cells at distal metastatic sites. The pro-metastatic actions of 27-hydroxycholesterol requires both polymorphonuclear-neutrophils and γδ-T cells, and 27-hydroxycholesterol treatment results in a decreased number of cytotoxic CD8⁺T lymphocytes. Therefore, through its actions on γδ-T cells and polymorphonuclear-neutrophils, 27-hydroxycholesterol functions as a biochemical mediator of the metastatic effects of hypercholesterolemia.High cholesterol is a risk factor for breast cancer recurrence. Here the authors show that cholesterol promotes breast cancer metastasis via its metabolite 27-hydroxycholesterol (27HC) that acts on immune myeloid cells residing at the distal metastatic sites, thus promoting an immune suppressive environment.

Fig. 1

27HC is a biochemical mediator of the effects of obesity and elevated cholesterol on breast cancer metastasis. a A high-cholesterol diet (HCD) increases spontaneous metastasis of murine ER-negative Met1 cells (N = 10 each). b Mice fed a high-cholesterol diet (HCD) exhibit increased lung colonization of i.v. grafted Met1 mammary tumors (CD N = 7, HCD N = 8). c A high-fat diet (HFD) increases lung colonization of murine E0771 i.v. grafts in APOE3 mice (i.v. graft of cells). Pretreatment for 7 days prior to the graft and continued treatment throughout the study with daily placebo or atorvastatin (statin, 40 mg/kg) by oral gavage reduced metastatic colonization. Representative images of iRFP-expressing E0771 colonies (depicted as black) within the lungs are presented above quantified data (control diet (CD) placebo N = 11, statin N = 7, HFD placebo N = 8, statin N = 7, scale bar indicates 5 mm). d Chronic administration of exogenous 27-hydroxycholesterol (27HC) increases spontaneous metastasis to the lung as measured by relative fluorescence (inset) without impacting growth rate of the primary tumor (N = 3). e Metastatic effects of a HCD in MMTV-PyMT mice requires conversion to 27HC by CYP27A1 (wild-type CD N = 8, HCD N = 14, CYP27A1^−/− CD N = 3, HCD N = 9). Quantitative RT-PCR for PyMT transcript within the lung is presented in the graph and used to quantify metastasis since its expression should be restricted to mammary epithelial cells. f Chronic treatment with a small molecule inhibitor of CYP27A1, GW273297X (GW297X) decreases colonization of lung tissue from Met1 i.v. grafts (N = 4 each). g Lung colonization from E0771 i.v. grafts in APOE3 mice is increased by a HFD, but attenuated in mice chronically treated with GW297X (CD placebo N = 5, GW297X N = 3, HFD placebo N = 7, GW297X N = 6). Results are depicted as mean +/− SEM. Lines and asterisks denote statistical differences between groups (p < 0.05). [a, b, d, f: Unpaired two-tailed student’s t-test. c, e, g: One-way ANOVA followed by a Student Newman-Keuls multiple comparison test]

Fig. 2

Effects of 27HC are mediated by its cancer-cell extrinsic effects on the host. a Mice were pretreated with placebo or 27HC (20 mg/kg) daily for 5 days, at which point Met1 cells expressing iRFP were grafted (i.v.) and treatment ceased. At day 28, lungs were harvested, imaged and iRFP fluorescence quantified. Experimental timeline outlined above. Representative images of iRFP-expressing Met1 colonies (depicted as black) within the lungs are presented above quantified data (placebo N = 5, 27HC N = 4, scale bar indicates 5 mm). b Mice were grafted (i.v.) with Met1 cells expressing iRFP on day 0. On day 5, mice were treated daily with placebo or 27HC for either 5 days or for the remainder of the study (chronic 27HC). At day 28, lungs were harvested, imaged and iRFP fluorescence quantified. Experimental timeline outlined above quantified data (placebo N = 7, 5 days 27HC N = 7, chronic 27HC N = 6). c Pretreatment of naive mice (regime outlined in a) with the CYP27A1 inhibitor GW297X (100 mg/kg) reduces colonization of lung by E0771 cells, which is reversed by co-pretreatment with 27HC (N = 5/group). d Pretreatment of mice with GW297X (as in c) results in decreased Met1 colonies. For this experiment, mice were pretreated with placebo or GW297X (100 mg/kg) for 5 days at which point Met1 cells expressing iRFP were grafted (i.v.) and treatment ceased. Five days post-graft, lungs were harvested, imaged, and fluorescence quantified (placebo N = 9, GW297X N = 10). Results are depicted as mean +/− SEM. Lines and asterisks denote statistical differences between groups (p < 0.05). a, d: Unpaired two-tailed student’s t-test. b, c: One-way ANOVA followed by a Student Newman-Keuls multiple comparison test

Fig. 3

Metastatic effects of 27HC require polymorphonuclear neutrophils. a Myeloid-derived immune cells are required for the colonizing effects of 27HC, as clodronate-liposome depletion of phagocytic cells ablates its effects. Mice were pretreated prior to cellular graft (i.v.) with placebo or 27HC and/or liposomes containing either PBS or clodronate as outlined in Fig. 2a. Representative images of iRFP-expressing Met1 colonies (depicted as black) within the lungs are presented above quantified data for both Met1 and E0771 mammary cancer models (Met1 in order from left to right N = 5, 3, 3, 4, E0771 in order from left to right N = 3, 3, 4, 4, scale bars indicate 5 mm). b 27HC treatment of metastasis bearing mice results in the recruitment of inflammatory monocytes (IMs) and polymorphonuclear neutrophils (PMNs) to the lung, as assessed by FACS analysis (in order from left to right N = 3, 3, 6). For comparison, one group of mice received only an i.v. PBS injection (no cells). c 27HC increases in vitro migration of bone marrow-derived neutrophils in a CXCR2-dependent manner, as the small-molecule inhibitor of CXCR2, SB225002 attenuates its effects (N = 3/group). d Colonizing effect of 27HC requires activation of CXCR2. Mice were pretreated with indicated ligands for 5 days prior to engraftment with Met1 cells expressing iRFP (as in Fig. 2a). Representative images of iRFP-expressing Met1 colonies (depicted as black) within the lungs are presented above quantified data (N = 6/group, scale bar indicates 5 mm). e PMNs are required for the colonizing effects of 27HC, as immune-depletion of PMNs with an antibody against Ly6G (α-Ly6G) ablates its effects. Mice were pretreated with indicated ligands for 5 days prior to engraftment with Met1 cells expressing iRFP (as in Fig. 2a). Representative images of iRFP-expressing Met1 colonies (depicted as black) within the lungs are presented above quantified data (placebo groups N = 4, isotype control 27HC group N = 6, α-Ly6G 27HC group N = 7, scale bar indicates 5 mm). Results are depicted as mean +/− SEM. Lines and asterisks denote statistical differences between groups (p < 0.05). [One-way ANOVA followed by a Student Newman-Keuls multiple comparison test]

Fig. 4

27HC increases presence of γδ-T cells and PMNs, and decreases presence of cytotoxic CD8+ T cells within tumors and metastatic lungs. a Bone marrow-derived neutrophils were co-cultured with splenocytes in the presence or absence of 27HC for 4 days. At this time, cells were analyzed by flow cytometry for the presence of γδ-T cells (N = 5/group). b The 27HC stimulated increase in γδ-T cells is not apparent when only splenocytes were cultured. c–e Early-stage E0771 tumors from mice treated with placebo or 27HC were analyzed by flow cytometry for γδ-T cells, PMNs and cytotoxic T cells (placebo N = 10, 27HC N = 9). f–h Early-stage metastatic lungs (Met1) from mice treated with placebo or 27HC were analyzed by flow cytometry (N = 8/group). i γδ-T cells are required for the colonizing effects of 27HC, as immune-depletion of γδ-T cells with an antibody against TCRγδ (α-TCRγ/δ) ablates its effects on Met1 colonization of the lungs (N = 8 except for isotype control 27HC group where N = 10). Mice were pretreated with indicated ligands for 5 days prior to engraftment with Met1 cells expressing iRFP (as in Fig. 2a). j γδ-T cells are required for the colonizing effects of 27HC, as 27HC fails to increase colonization in Tcrd^−/− mice (from left to right N = 13, 12, 12, 13). k Immune depletion of CD8⁺ cells (α-CD8) increases colonization, and exogenous 27HC treatment does not further increase colonization of Met1 cells to the lungs (from left to right: N = 9, 7, 8, 8). l Scatterplot analysis of TCGA data of 421 invasive breast cancer tumors for mRNA expression of CD66b, IL-17, CD3e, and PRF1 (perforin), parsed by median expression of CYP7B1. Results are depicted as mean +/− SEM, with the exception of a, b, and l where the horizontal line depicts the mean. Asterisks denote statistical differences between groups (p < 0.05). [a–h: Unpaired two-tailed student’s t-test. i–k: One-way ANOVA followed by a Student Newman-Keuls multiple comparison test. l: Mann–Whitney test]