CD36-mediated activation of endothelial cell apoptosis by an N-terminal recombinant fragment of thrombospondin-2 inhibits breast cancer growth and metastasis in vivo

Abstract

Thus far the clinical benefits seen in breast cancer patients treated with drugs targeting the vascular endothelial growth factor (VEGF) pathway are only modest. Consequently, additional antiangiogenic approaches for treatment of breast cancer need to be investigated. Thrombospondin-2 (TSP-2) has been shown to inhibit tumor growth and angiogenesis with a greater potency than the related molecule TSP-1. The systemic effects of TSP-2 on tumor metastasis and the underlying molecular mechanisms of the antiangiogenic activity of TSP-2 have remained poorly understood. We generated a recombinant fusion protein consisting of the N-terminal region of TSP-2 and the IgG-Fc1 fragment (N-TSP2-Fc) and could demonstrate that the antiangiogenic activity of N-TSP2-Fc is dependent on the CD36 receptor. We found that N-TSP2-Fc inhibited VEGF-induced tube formation of human dermal microvascular endothelial cells (HDMEC) on matrigel in vitro and that concurrent incubation of anti-CD36 antibody with N-TSP2-Fc resulted in tube formation that was comparable to untreated control. N-TSP2-Fc potently induced apoptosis of HDMEC in vitro in a CD36-dependent manner. Moreover, we could demonstrate a CD36 receptor-mediated loss of mitochondrial membrane potential and activation of caspase-3 in HDMEC in vitro. Daily intraperitoneal injections of N-TSP2-Fc resulted in a significant inhibition of the growth of human MDA-MB-435 and MDA-MB-231 tumor cells grown in the mammary gland of immunodeficient nude mice and in reduced tumor vascularization. Finally, increased serum concentrations of N-TSP2-Fc significantly inhibited regional metastasis to lymph nodes and distant metastasis to lung as shown by quantitative real-time alu PCR. These results identify N-TSP2-Fc as a potent systemic inhibitor of tumor metastasis and provide strong evidence for an important role of the CD36 receptor in mediating the antiangiogenic activity of TSP-2.

Fig. 1

N-TSP2-Fc inhibits HDMEC tube formation on Matrigel in a CD36-dependent manner. HDMEC were seeded on Matrigel in the presence of VEGF. Cultures were treated with either PBS (a), or IgG-Fc (b), or N-TSP2-Fc (c). In d HDMEC were incubated with both N-TSP2-Fc and blocking CD36 antibody (CD36 ab). (Bar = 200 μm). Quantitative image analysis of tube length (e) revealed that treatment with N-TSP2-Fc significantly reduced the average tube length as compared to control-treated HDMEC. After preincubation of HDMEC with a blocking CD36 antibody (CD36 ab) the average tube length was comparable in N-TSP2-Fc- and control-treated HDMEC. Treatment with IgG-Fc did not have any significant effects on the average tube length. Data are expressed as mean ± SEM (n = 3). *** P < 0.001

Fig. 2

N-TSP2-Fc induces apoptosis of HDMEC by CD36 receptor-mediated loss of mitochondrial membrane potential and activation of caspase-3. Percentage of apoptotic HDMEC (a) was measured by flow cytometry. Treatment with N-TSP2-Fc resulted in a significant increase of HDMEC apoptosis (a). Concurrent addition of a blocking CD36 antibody (CD36 ab) with N-TSP2-Fc inhibited this effect (a). The percentage of mitochondrial membrane potential (b) was decreased after treatment of HDMEC with N-TSP2-Fc. Simultaneous incubation of N-TSP2-Fc with anti-CD36 antibody (CD36 ab) did not induce loss of mitochondrial membrane potential when compared to control HDMEC (b). Active caspase-3 was analyzed in HDMEC by western blot analysis (c). Image analysis (d) of the bands revealed that caspase-3 activity was increased by incubation of HDMEC with N-TSP2-Fc. Caspase-3 activity of control-treated HDMEC or HDMEC incubated with N-TSP2-Fc and concurrent addition of anti-CD36 antibody (CD36 ab) was comparable (d). In all experiments a–d, HDMEC were incubated with PBS (control), or with 40 μg/ml of N-TSP2-Fc, or with 40 μg/ml of N-TSP2-Fc plus 10 μg/ml of blocking CD36 antibody. All experiments were performed in the presence of VEGF (20 ng/ml). Results are expressed as mean + SEM (n = 3). ** P < 0.01, *** P < 0.001

Fig. 3

Systemic treatment with N-TSP2-Fc inhibits in vivo growth and angiogenesis of MDA-MB-435 tumors grown in the mammary gland. PBS (control) or 5 mg per kg body weight of N-TSP2-Fc were daily injected intraperitoneally, beginning at day 6 after tumor cell implantation. MDA-MB-435 tumor growth was inhibited by treatment with N-TSP2-Fc as compared with PBS control treatment (a). Computer-assisted image analysis of CD31-stained cryostat tumor sections revealed a significant reduction of the number of blood vessels per mm² tumor area (b), of the average vessel size (c), and of the relative tumor area covered by blood vessels (d) in mice treated with N-TSP2-Fc. CD31-stained blood vessels were evaluated in three different ×10 fields in sections obtained from five different tumors for each treatment group. Data are expressed as mean + SEM. *** P < 0.001

Fig. 4

N-TSP2-Fc inhibits lymph node and lung metastasis. Quantitative real-time PCR was used to analyze human alu repeats from MDA-MB-435 tumor cells present in genomic murine DNA that was extracted from lymph node (a) and lung tissue (b). Lymph node (a) and lung (b) metastasis were analyzed at two and 5 weeks after tumor cell implantation in mice treated either with PBS (control; n = 10) or N-TSP2-Fc (NTF; n = 10). To measure the background signal in murine lymph node and lung tissue, quantitative real-time alu PCR was performed in non-tumor-bearing mice (mice w/o tumor). Data revealed an increase of the percentage of human DNA in murine DNA obtained from lymph node and lung tissue of tumor-bearing mice at two and 5 weeks after tumor cell transplantation when compared to mice without tumors indicating the metastatic spread of MDA-MB-435 tumor cells to the lymph nodes and lung. Treatment with N-TSP2-Fc resulted in a significant reduction of the percentage of human DNA in murine DNA obtained from lymph node and lung tissue at two and 5 weeks after transplantation of MDA-MB-435 tumor cells in the mammary gland of nude mice. Values represent mean + SEM. * P < 0.05, ** P < 0.01, *** P < 0.001