Clear Cell Carcinomas of the Gynecologic Tract and Thromboembolic Events: What Do We Know So Far?

The observation that cancer patients may have an increased risk for thromboembolic events was originally made by the French internist Armand Trousseau in 1865 [1]. Almost 150 years later, it is now well-recognized that cancer patients have a four- to seven-fold increase in their risk for venous thromboembolic events relative to their counterparts without cancer, and that 1–15% of cancer patients develop clinical thromboembolic events over the course of their care [2–4]. This increased risk has a multifactorial basis, the precise components of which are not entirely clear at the present time. However, in cancer patients, risk factors for increased thrombosis often abound, including the presence of medical comorbidities, invasive supportive care, chemotherapeutic and other pharmacologic agents, major surgeries, advanced stage disease, and specific malignancies that have been found to increase the risk of thromboembolic events (including gliomas, renal, gastric and pancreatic carcinomas) [2,5,6]. In ovarian cancer patients, the risk of symptomatic thromboembolic events ranges from 5.2 to 16.6% [6–8]. Furthermore, the incidence of clinically silent thromboembolic events, as identified by radiological imaging approaches or serum markers of thrombosis, is exceptionally high in patients with ovarian cancer [9]. In two studies, 15–19% of the thromboembolic events were identified during preoperative work-up, 14–33% were identified in the immediate postoperative period and 33–67% were identified during chemotherapy or long-term follow-up period [7,8]. Deep venous thrombosis significantly predominated over pulmonary emboli in one study [8], whereas these entities were approximately equally distributed in another [7]. The occurrence of thromboembolic events has been found to adversely impact survival in several reports, even in patients with localized disease [6–8]. Significant risk factors for thromboembolic events in patients with ovarian cancer largely mirror those of the general population of cancer patients [6].

One histotype of ovarian cancer that has been associated with an increased risk of thromboembolic events is clear cell carcinoma [7,10–14]. The incidence of thromboembolic events has varied significantly in various published reports on the subject, ranging from 10.8 to 42% in clear cell carcinomas as compared with 0–18% in non-clear cell cancers, but the statistical significance of the resultant differences have been a consistent finding in all reports [7,10,11,14]. Within the clear cell carcinoma subgroup of ovarian cancer patients, venous thromboembolic events have also been associated with decreased overall and progression-free survival, even after controlling for stage [10,16]. The combination of two risk factors for thromboembolic events, chemotherapy and the clear cell histotype of ovarian carcinoma, not surprisingly has a significantly negative impact on survival [10]. Interestingly, Diaz et al. recently reported that the poor prognostic significance of thromboembolic events in clear cell carcinoma is highest in early stage patients and seemed to be unrelated to either tumor burden or to mortality that is directly attributable to thrombosis [16]. This raises the intriguing possibility that the thromboembolic events are a consequence of, or are caused by, a subset of clear cell carcinomas whose biologic properties render them more aggressive clinically. Ultimately, the precise combination of factors that eventuate in the increased risk for thromboembolic events in clear cell carcinoma is unclear, but the bulk of emerging evidence indicates that a significant contributor to this risk is related to tumor cell properties. In clear cell carcinomas of the endometrium, which are much less common but display broad phenotypic overlap with their ovarian counterparts [17], a similar increased risk for thromboembolic events has been found [18,19], which bolsters the argument that the risk is directly related to the clear cell cellular phenotype. At least three procoagulant proteins have been found in clear cell carcinoma in significantly larger amounts relative to the other histotypes. IL-6, a circulating cytokine, may induce a thrombotic state not only by activating endothelial cells and platelet production, but by increasing the expression of several procoagulant proteins, including fibrinogen, tissue factor (TF), factor VIII and von Willebrand factor [20]. Using a variety of analytic approaches, Anglesio et al. demonstrated that patients with ovarian clear cell carcinomas have increased expression of IL-6 in their tumors relative to high grade serous carcinomas, as well as in their serum [21]. Another protein that has been implicated in clear cell carcinoma associated thrombosis is TF. TF, a membrane glycoprotein, has an extracellular domain that is most often expressed in cells that are not in direct contact with the blood stream. Once exposed to the bloodstream, the extracellular domain typically activates coagulation factor VII to activate the extrinsic coagulation cascade, which eventuates in the generation of thrombin and fibrin.

Ectopic factor VII expression has been identified with a higher frequency in the cells of ovarian clear cell carcinoma relative to the other histotypes. Furthermore, these tumor cells apparently secrete ‘microparticles’ with TF-factor VII activity [22]. TF has also been shown to be expressed at higher frequencies and at higher concentrations in clear cell carcinoma compared with the other histotypes in both ovarian and endometrial tumors [18,23]. It is unclear whether tumor-related TF or tumor-directed induction of host TF is the primary driver to thromboembolic events, and similarly unclear is the precise number of proteins that may activate TF. However, TF may contribute to the mortality of clear cell carcinoma patients in ways that are unrelated to the coagulation cascade, since this multifunctional protein and its associated molecular pathways have also been implicated in tumor proliferation, metastasis and angiogenesis [24,25]. Another clear cell carcinoma-related protein that may potentially activate the coagulation cascade is hepatocyte nuclear factor 1 beta (HNF1β). HNF1β is a transcription factor that is encoded by the TCF2 gene and which is significantly and more frequently upregulated in endometrial and ovarian clear cell carcinomas as compared with non-clear cell carcinomas. Using integrative bioinformatic analytic approaches, Cuff et al. identified several putative transcriptional targets for HNF1β in clear cell tumors as compared with non-clear cell tumors [15], including gene sets related to 3 proteins of the coagulation cascade (fibrinogen, prothrombin and factor XIII) 15]. Furthermore, HNF1β expression was found to be directly related to an increased risk for thromboembolic events [15].

Clear cell carcinomas of the ovary and endometrium are a subset of gynecologic carcinomas that carry an increased risk of venous thromboembolic events. This histotype represents a unique medium to study the connection between cancer-related thrombosis and general cancer-related biology and to identify factors that influence and/or promote both.