Although the association of thrombosis and occult cancer has been recognized for >40 years, the role of thrombin in enhancing the malignant phenotype has only been recently appreciated. To better understand the mechanisms involved we performed an Affymetrix gene display on 2 tumor cell lines, B16F10 and UMCL to look for genes which could be activated by thrombin and contribute to the promalignant phenotype. Cathepsin D (CD) was upregulated 2–4 fold, confirmed by mRNA and protein analysis. We focused on CD because its upregulation has been associated with human malignancy and poor prognosis, particularly with breast CA as well ovarian, prostate, bladder and melanoma. Thrombin (0.5 u/ml for 24 hrs) upregulated mRNA 1.5 – 2.3 fold and protein 1.5 – 4.2 fold in 8 different cell lines: human breast MCF7 and MDAMB-231, cervical HeLa, prostate PC3, primary HUVEC and brain microvascular endothelial cells as well as murine melanoma B16F10 and mammary 4T1. Thrombin also increased secretion of CD 3.1 – 8.2 fold in MCF7, HeLa and PC3 cell lines. We next examined the effect of CD (shRNA) knock down (KD) and thrombin in 4T1 cells. Thrombin (0.5 u/ml, 24 hrs) enhanced chemotaxis ∼2 fold in 4T1 cells. This was completely inhibited when CD was knocked down with siRNA but not with empty vector (EV) or scrambled vector (SV). We next tested the effect of secreted tumor cell CD (4T1 conditioned media) on angiogenesis (neoangiogenesis branching) following addition to a chick chorioallantoic membrane (CAM). The angiogenesis index from 24 hr conditioned media of wild type (WT) and KD cells was 46 ± 4 and 19 ± 3 respectively, n=4, p=0.03. A 2.1 fold inhibition was also obtained with anti-CD Ab vs irrelevant IgG, n=3, p=0.001. This prompted testing for the effect of pure enzymatically-active CD (2 ng) on the CAM. This resulted in a 2.1 fold increase in angiogenesis which was completely inhibited by the proteolytic inhibitor, Pepstatin A (Pep A), n=10, p<0.005. Pep A alone had no effect. To further define the mechanism of enhanced angiogenesis, we measured the effect of CD on HUVEC chemotaxis in the presence and absence of Pep A. CD enhanced chemotaxis of HUVEC 2.2 fold, which was also completely inhibited by Pep A, n=4, p<0.01. CD KD in a HUVEC-matrigel tube assay revealed similar results, n=3. Since proteolysis was required for angiogenesis we hypothesized that CD could be activating metalloproteinases (MMP) required for endothelial tube formation. We therefore performed zymography. This revealed an ∼2 fold upregulation of MMP-2 and MMP-9 compared to PBS which was completely inhibited by Pep A as well as a generic MMP inhibitor, GM6001, n=3. CD activation of angiogenesis and 4T1 tumor cell chemotaxis suggested that in vivo growth and metastasis could be affected by CD. Accordingly, 1 × 105 4T1 cells were injected into the flank of syngeneic mice. CD KD revealed a 3–4 fold inhibition of tumor growth on days 19 and 25 respectively, n=5. Hirudin, a specific thrombin inhibitor significantly inhibited WT and EV cells 2–3 fold on days 19 and 25 respectively (n=5), which is compatible with thrombin upregulation of CD. Both CD and thrombin also contributed to spontaneous metastasis with CD KD vs EV vs (EV + hirudin) of 6.4 ± 0.9, 1.6 ± 0.5 and 1.4 ± 0.4 nodules respectively on day 27, n=5, p<0.02. We conclude that thrombin-induced CD contributes to the malignant tumor phenotype by inducing tumor cell migration, growth, metastasis and angiogenesis. CD-induced angiogenesis requires its proteolytic activation of MMP-2 and MMP-9.

Author notes

Disclosure: No relevant conflicts of interest to declare.

Sign in via your Institution