Podoplanin expression in primary brain tumors induces platelet aggregation and increases risk of venous thromboembolism

Abstract

Venous thromboembolism (VTE) is common in patients with brain tumors, and underlying mechanisms are unclear. We hypothesized that podoplanin, a sialomucin-like glycoprotein, increases the risk of VTE in primary brain tumors via its ability to induce platelet aggregation. Immunohistochemical staining against podoplanin and intratumoral platelet aggregates was performed in brain tumor specimens of 213 patients (mostly high-grade gliomas [89%]) included in the Vienna Cancer and Thrombosis Study, a prospective observational cohort study of patients with newly diagnosed cancer or progressive disease aimed at identifying patients at risk of VTE. Platelet aggregation in response to primary human glioblastoma cells was investigated in vitro. During 2-year follow-up, 29 (13.6%) patients developed VTE. One-hundred fifty-one tumor specimens stained positive for podoplanin (33 high expression, 47 medium expression, 71 low expression). Patients with podoplanin-positive tumors had lower peripheral blood platelet counts (P < .001) and higher D-dimer levels (P < .001). Podoplanin staining intensity was associated with increasing levels of intravascular platelet aggregates in tumor specimens (P < .001). High podoplanin expression was associated with an increased risk of VTE (hazard ratio for high vs no podoplanin expression: 5.71; 95% confidence interval, 1.52-21.26; P =010), independent of age, sex, and tumor type. Podoplanin-positive primary glioblastoma cells induced aggregation of human platelets in vitro, which could be abrogated by an antipodoplanin antibody. In conclusion, high podoplanin expression in primary brain tumors induces platelet aggregation, correlates with hypercoagulability, and is associated with increased risk of VTE. Our data indicate novel insights into the pathogenesis of VTE in primary brain tumors.

Figure 1. Representative samples showing podoplanin expression and intravascular platelet clusters (antiplatelet surface protein CD61) in brain tumor specimens.

(A-B) Representative examples of high podoplanin (+++)-expressing tumors. (C) Representative example of a podoplanin-negative (–) tumor. (D) The invasion zone of a podoplanin-expressing tumor is depicted. (E) A very large CD61⁺ intravascular thrombosis (+++) is shown. (F) Tumor sample with multiple CD61⁺ thrombotic vessels (++). NDP Viewing Software (NDP.view2; Hamamatsu Photonics, Japan) was used for microscopy photographs. Full scale bars represent 500 μm (A), 250 μm (B-C, E-F), and 2.5 μm (D).

Figure 2. Association between podoplanin expression levels and grade of intratumoral intravascular platelet aggregates.

Increasing levels of podoplanin expression intensity correlated with higher levels of CD61⁺ vascular platelet aggregates (χ², P < .001).

Figure 3. Cumulative incidence of venous thromboembolism (VTE) accounting for competing risk (death of any cause other than fatal VTE) according to expression levels of podoplanin.

Podoplanin expression was associated with a significantly increased risk of VTE; the probability of VTE increased with increasing levels of podoplanin staining intensity. Log-rank test comparing the cause-specific hazards of VTE (null hypothesis: all 4 podoplanin staining groups have the same VTE rate): P = .019; Gray’s test comparing the competing risk cumulative incidences (null hypothesis: all 4 podoplanin staining groups have the same VTE risk): P = .062.

Figure 4. Podoplanin expression and tumor cell–induced platelet aggregation and platelet activation in vitro.

(A) Podoplanin expression on cancer cell lines “Gli16” and “AMCH” was investigated by flow cytometry. (B) The ability of the cell lines to induce platelet aggregation was investigated by LTA: A total number of 5 × 10⁵ cancer cells were added to 225 μL of PRP,and maximal platelet aggregation (%) after 10 minutes was recorded. Gli16 cells induced significantly stronger platelet aggregation compared with control cell line AMCH (***P < .001). Platelet aggregation in response to Gli16 was abrogated by the antipodoplanin antibody NZ-1, but not by control antibody (***P < .001). ns, not significant. Box plots show data from 5 individual experiments (performed in multiple replicates; total n = 22). (C) Platelet activation and degranulation upon coincubation of 225 μL PRP with 5 × 10⁵ cancer cells was determined by a PF4 ELISA. Gli16 cells induced marked release of PF4 from platelets, which was not observed for AMCH cells. The ability of Gli16 cells to induce PF4 release from platelets could be inhibited by the anti-podoplanin antibody NZ-1, but not by control antibody. The experiment was performed in duplicates.