Sphingosine-1-phosphate produced by sphingosine kinase 1 promotes breast cancer progression by stimulating angiogenesis and lymphangiogenesis

Abstract

Sphingosine-1-phosphate (S1P) is a pleiotropic bioactive lipid mediator that promotes breast cancer progression by diverse mechanisms that remain somewhat unclear. Here we report pharmacologic evidence of a critical role for sphingosine kinase 1 (SphK1) in producing S1P and mediating tumor-induced hemangiogenesis and lymphangiogenesis in a murine model of breast cancer metastasis. S1P levels increased both in the tumor and the circulation. In agreement, serum S1P levels were significantly elevated in stage IIIA human breast cancer patients, compared with age/ethnicity-matched healthy volunteers. However, treatment with the specific SphK1 inhibitor SK1-I suppressed S1P levels, reduced metastases to lymph nodes and lungs, and decreased overall tumor burden of our murine model. Both S1P and angiopoietin 2 (Ang2) stimulated hemangiogenesis and lymphangiogenesis in vitro, whereas SK1-I inhibited each process. We quantified both processes in vivo from the same specimen by combining directed in vivo angiogenesis assays with fluorescence-activated cell sorting, thereby confirming the results obtained in vitro. Notably, SK1-I decreased both processes not only at the primary tumor but also in lymph nodes, with peritumoral lymphatic vessel density reduced in SK1-I-treated animals. Taken together, our findings show that SphK1-produced S1P is a crucial mediator of breast cancer-induced hemangiogenesis and lymphangiogenesis. Our results implicate SphK1 along with S1P as therapeutic targets in breast cancer.

Figure 1. The SphK1/S1P axis in breast cancer progression

(A) 4T1 breast tumors were established by surgical implantation of 4T1-luc2 cells into the chest mammary fat pad under direct vision. The tumor burden was determined by bioluminescence technology. Representative IVIS images (left) and bioluminescent quantification (right) on the indicated days are shown. (B) Axillary lymph node metastases were measured by bioluminescence at the indicated times when breast tumor was removed for accurate measurement (left). Lung metastases were measured separately ex vivo (middle). The incidences of metastases to the axillary lymph nodes and the lung were determined on the indicated days (right). (C) mRNA was isolated from mammary fat pads of naïve BALB/c mice (gray filled bars), cultured 4T1-luc2 cells (black cross hatched bars), and primary tumors in the chest mammary fat pads formed by 4T1-luc2 cells (blue filled bars) 10 days after implantation. Expression of SphK1 and SphK2 was determined by QPCR and normalized to levels of GAPDH mRNA. Data are expressed as fold increases ± S.D. (D) S1P levels were determined by LC-ESI-MS/MS in the serum of sham surgery mice and in mice with 4T1-luc2 xenograft tumors 15 days after implantation. Data are expressed as fold increases ± SEM. (E) S1P was measured in serum from stage IIIA breast cancer patients and age/ethnicity-matched healthy volunteers (n = 5). *, P < 0.001; **, P < 0.05.

Figure 2. Inhibition of SphK1 decreases growth of 4T1-luc2 mammary tumors

(A) 4T1-luc2 cells were treated with the indicated concentration of SK1-I and total cell number was determined by measurement of luciferase activity. 4T1-luc2 cells were surgically implanted in mammary fat pads under direct vision. Tumor-bearing mice were randomized into two groups and injected i.p. with PBS or SK1-I (20 mg/kg) daily. Tumor volumes were measured on the indicated days (B) and tumor weights were determined after excision on day 18 (C) (n = 5 per group). (D) Tumor histology. Paraffin-embedded tumor sections were immunostained with Ki67 and counterstained with hematoxylin. Apoptotic cells were visualized by TUNEL staining. Scale bar, 50 µm. (E) S1P levels were determined by LC-ESI-MS/MS in mammary fat pads of naïve (open bar) and sham operated mice (black dotted bars), and in mammary tumors from mice that were treated with PBS (blue filled bars) or SK1-I (red hatched bars) on the indicated days. (F) SK1-I levels were determined on the indicated days by LC-ESI-MS/MS in mammary tumors from mice that were treated with SK1-I (red filled bars), and means ± SEM are shown. *, P < 0.01; **, P < 0.05.

Figure 3. SK1-I decreases tumor burden, lymph node and lung metastases, and circulating levels of S1P

4T1-luc2 cells were surgically implanted in mammary fat pads under direct vision. Tumor-bearing mice were randomized into two groups and injected i.p. daily with PBS or SK1-I (20 mg/kg) (n = 5). (A) Tumor burden was quantified by in vivo bioluminescence on the indicated days. Right panels show representative IVIS bioluminescent images. (B) S1P levels were determined by LC-ESI-MS/MS in serum of naïve (open bar) and sham operated mice (black dotted bars), mice bearing 4T1-luc2 tumors that were treated with PBS (blue filled bars) or SK1-I (red hatched bars) on the indicated days. (C) SK1-I levels were determined on the indicated days by LC-ESI-MS/MS in serum from the mice that were treated with SK1-I (red filled bars). Representative bioluminescent images (right) and quantification of regional lymph node (D) and lung (E) metastases were determined by bioluminescence 7 or 10 days after treatment with PBS or SK1-I, respectively. Data are expressed as mean ± SEM. *, P < 0.01; **, P < 0.05.

Figure 4. Inhibition of SphK1 suppresses in vitro hemangiogenesis and lymphangiogenesis

GFP expressing HUVECs and HLECs were cultured on reduced growth factor basement membrane matrix-coated 48 well plates and incubated for 6 hours with or without S1P (1 µM), Ang2 (500 ng/ml), or SK1-I (10 µM), as indicated. (A) Representative images. (B) Two random fields per well were photographed. Total tube length determined and means ± SD are shown. *, P < 0.01; **, P < 0.05.

Figure 5. Inhibition of SphK1 suppresses hemangiogenesis and lymphangiogenesis quantified by DIVAA/FACS

(A, B) FACS gating scheme to quantify blood endothelial cells (BEC) and lymphatic endothelial cells (LEC) from the same lymph node single cell suspension. (A) TER-119 and CD45 were used to gate out the RBCs and lymphocytes, respectively. CD31 and gp38 (podoplanin) expression allowed separation of LECs and BECs from other TER-119⁻CD45⁻ cells: fibroblastic reticular cells (FRCs) and double-negative stromal cells (Others). (B) LYVE-1 expression on the indicated populations or isotype control (IC) on the total population was quantified by FACS and mean values ± SEM are shown in the right panel. *, P < 0.001. (C-E) Quantification of BECs and LECs by DIVAA/FACS. (C) Angioreactors are closed at bottom end and filled with 20 µl of Matrigel. Addition of stimuli in the Matrigel enables directional migration of BECs and LECs into the angioreactors, which proliferate and form appropriate vessels (left). Hemangiogenesis was readily observable by the appearance of blood vessels in the angioreactors containing S1P (1 µM), or Ang2 (1 µg/ml) compared to PBS. (D) Cells were isolated from the angioreactors and BECs and LECs were analyzed by FACS with antibodies to TER-119, CD45, CD31 and gp38 as in (A). (E) Total number of BECs and LECs were calculated and means ± SEM are shown. *, P < 0.01; **, P < 0.05.

Figure 6. Inhibition of SphK1 suppresses hemangiogenesis and lymphangiogenesis in tumors and lymph nodes in vivo

4T1-luc2 cells implanted in chest mammary fat pads were allowed to form tumors for 2 days, which were quantified by IVIS. Mice with similar amounts of tumor were randomized and treated with either PBS or SK1-I (20mg/kg/day i.p.). Tumors (A, B) and regional axillary lymph nodes (C, D) were harvested after 4 days of treatment, minced, digested with collagenase, and BECs and LECs quantified by FACS. Total numbers of BECs or LECs in 4T1 tumors (B) and lymph nodes (D) are shown as means ± SEM. * P < 0.05. (E) Immunofluorescent staining for CD31 and LYVE-1. CD31⁺LYVE-1⁻ cells were considered as blood vessels, CD31⁺LYVE-1⁺ cells as lymphatic vessels (arrows). (F) Morphometric analysis of blood and lymphatic vessel densities were determined in the peritumoral regions after 7 days of treatment, and are shown as means± SEM. *, P < 0.05. Scale bar, 50 µm.