Podoplanin promotes tumor growth, platelet aggregation, and venous thrombosis in murine models of ovarian cancer

Abstract

Background: Podoplanin (PDPN) is a sialylated membrane glycoprotein that binds to C-type lectin-like receptor 2 on platelets resulting in platelet activation. PDPN is expressed on lymphatic endothelial cells, perivascular fibroblasts/pericytes, cancer cells, cancer-associated fibroblasts, and tumor stromal cells. PDPN's expression on malignant epithelial cells plays a role in metastasis. Furthermore, the expression of PDPN in brain tumors (high-grade gliomas) was found to correlate with an increased risk of venous thrombosis.

Objective: We examined the expression of PDPN and its role in tumor progression and venous thrombosis in ovarian cancer.

Methods: We used mouse models of ovarian cancer and venous thrombosis.

Results: Ovarian cancer cells express PDPN and release PDPN-rich extracellular vesicles (EVs), and cisplatin and topotecan (chemotherapies commonly used in ovarian cancer) increase the expression of podoplanin in cancer cells. The expression of PDPN in ovarian cancer cells promotes tumor growth in a murine model of ovarian cancer and that knockdown of PDPN gene expression results in smaller primary tumors. Both PDPN-expressing ovarian cancer cells and their EVs cause platelet aggregation. In a mouse model of venous thrombosis, PDPN-expressing EVs released from HeyA8 ovarian cancer cells produce more frequent thrombosis than PDPN-negative EVs derived from PDPN-knockdown HeyA8 cells. Blood clots induced by PDPN-positive EVs contain more platelets than those in blood clots induced by PDPN-negative EVs.

Conclusions: In summary, our findings demonstrate that the expression of PDPN by ovarian cancer cells promotes tumor growth and venous thrombosis in mice.

Keywords: murine models of cancer; ovarian cancer; platelet; podoplanin; venous thrombosis.

Figure 1.. PDPN expression in ovarian cancer cells and tumors.

(A) Human [fallopian tube epithelium (FTE), endothelial cell line G1S1, fibrosarcoma cell line HT1080, ovarian cancer cell lines A2780, CAOV3, HeyA8, IGROV1, OVCAR3, OVCAR4, OVCAR5, OVCAR8, OVCAR432, and SKOV3] and (B) murine (pericyte-like cell line10T1/2, ovarian cancer cell lines ID8, and IG10) whole-cell lysates were Western blotted for PDPN. PDPN protein band is ~37 kDa, while GAPDH (~35 kDa) and β actin (~42 kDa) were used as loading controls. (C) PDPN expression on HeyA8 after incubation for various time intervals with cisplatin (6μM) and topotecan (5nM). Vinculin (~120 kDa) was used as a loading control. Representative immunohistochemistry (IHC) for PDPN in (D) human ovarian cancer tumor specimen and (E) a section of syngeneic tumor nodules induced by ID8 murine ovarian cancer cells in mice (Scale bar, 50 μm).

Figure 2.. Effect of PDPN-positive human ovarian cancer cells and their EVs on platelet aggregation.

Representative platelet aggregation tracing induced by incubating (A) ovarian cancer cells (5 × 10⁶ cells) or (C) their EVs (100 ng) in platelet-rich plasma (5 × 10⁸ platelets/mL). The mean maximum platelet aggregation was compiled from several aggregation studies with (B) ovarian cancer cells and (D) their EVs (n ≥ 7 and * = p ≤ 0.05).

Figure 3.. Effect of PDPN knockdown on platelet aggregation induced by cells and EVs.

Expression of PDPN on HeyA8, HeyA8-shControl, and HeyA8-shPDPN as detected by Western blotting of (A) the whole cell–lysate or (E) lysate from their EVs. PDPN protein band is ~37 kDa, while Vinculin (~120 kDa) and Alix (~96 kDa) were used as loading controls. PDPN expression on (B) HeyA8-shControl and HeyA8-shPDPN and (F) their EVs were detected by flow cytometry. Aggregation of human platelets induced by HeyA8-shPDPN and HeyA8-shControl cells (5 × 10⁶ cells) or their EVs (100 ng). (C,G) Representative optical aggregation tracings and (D,H) mean maximum platelet aggregation (n ≥ 6 and * = p ≤ 0.05).

Figure 4.. Role of PDPN on tumor growth in a murine model of ovarian cancer.

(A) Human and murine PDPN mRNA was quantified relative to 18s rRNA by RT-qPCR of RNA samples extracted from tumor nodules induced by human HeyA8 cells in Nu/Nu mice (n = 4 tumor nodules/group). (B) Representative IHC for PDPN in tumor nodules resected from tumor-bearing mice. Scale bar = 50 μm. (C) Average tumor nodules’ total weight and (D) ascites volume was compared between mice injected with HeyA8-shControl and HeyA8-shPDPN ovarian cancer cells. n = 9–10 mice/ group. (E) Kaplan Meier plot comparing the overall survival in ovarian cancer patients with low (n=469) or high (n=186) PDPN expression. HR=1.49 (1.2–1.84), p = 0.00029. * = p ≤ 0.05.

Figure 5.. PDPN expression on EVs promotes venous thrombosis.

(A) Western blot analysis of PDPN expression in EVs collected from media of transduced HeyA8 cells. Cell media was centrifuged at different speeds to isolate small EVs and medium + large EVs. Alix (~96 kDa) was used as a loading control for small EVs. (B) Representative angiographic images of IVC stenosis and thrombosis. (Ba) Following IVC stenosis surgery, severe stenosis of the infrarenal IVC is notable on angiography (red asterisk). (Bb) Thrombus formation and (Bc) extension in this model recapitulates the clinical imaging findings of a filling defect within the IVC (green lines) and formation of collateral veins (Bb, Bc). Red dot: IVC stenosis; green line: thrombus length. (C) Frequency of IVC thrombosis induced by injection of EVs from HeyA8-shControl (13 out of 13) and HeyA8-shPDPN (7 out of 15). (D) Representative clot structure and composition were detected by immunofluorescence microscopy after staining thrombi for platelets (red), and RBCs (green autofluorescence). (E) Quantification of thrombus area covered by platelets in blood clots induced by injection of EVs from HeyA8-shControl and HeyA8-shPDPN.