The hemostatic system in cancer

The hemostatic system forms an integral part of the innate immune system and contributes to tumor growth and metastasis. Many malignancies are associated with a hyper-thrombotic state that derives foremost from tissue factor (TF), the initiator of the coagulation cascade, expressed by tumor cells and innate immune cells and stromal cells under the lingering chronic inflammatory conditions in the tumor microenvironment (TME). Thrombosis can also be promoted by changes in platelet production and thrombocytosis, and endothelial activation that is frequently associated with cancer therapy. The clinical management and the prevention of cancer-associated thrombosis now include direct oral anticoagulants (DOACs), and they may influence the tumor-promoting effects of hemostasis beyond the prevention of thrombosis.

Coagulation proteases and cell signaling

Coagulation activation of TF is not only relevant for hemostasis and thrombosis but also induces cell signaling. FVIIa, FXa and thrombin are the relevant coagulation factors that cleave protease-activated receptors (PARs) and thereby influence the function of the tumor and immune cells. Whereas liver-derived FVII and FX contribute to intravascular coagulation and thrombosis, they may also reach the perivascular but not the deeper tissue in the tumor to induce coagulation signaling. Here, the locally produced coagulation proteases are indispensable. In this context, proteases synthesized by immune and tumor cells can become activators of PARs and thereby alter the TME.

Coagulation and immune evasion in cancer

Tumors acquire multiple mechanisms to evade the host’s immune response for enhanced tumor growth and metastasis. Tumor antigens taken up by tumor-associated macrophages (TAMs) and dendritic cells are presented to CD8⁺ T cells that play pivotal roles in killing tumor cells. The TME can instruct TAMs to suppress tumor-killing T cells and lead to an immune-suppressive TME that eventually facilitates tumor growth and invasion. We recently demonstrated that coagulation factors interacting with TF, specifically FVII and FX, are synthesized by TAMs in the TME. The cell-autonomous synthesis of specifically FX promoted tumor progression and FXa inhibition by DOACs, but not heparins restricted to the intravascular space prevented the tumor-promoting activities of TAM-synthesized FXa. Additional studies elucidate potential coagulation-related targets that may contribute to the beneficial effects of DOACs in tumor suppression.

Thrombin–PAR1 interaction in cancer

Thrombin is the major protease that activates PAR1. Ubiquitous PAR1 knockout mice in spontaneous cancer models indicated a role for PAR1 in the prostate (TRAMP) and intestinal (APC^Min/+) tumor progression. In these mice, PAR1 deletion affects not only tumor cells but also bystander cells in the TME, e.g., endothelial cells and immune cells. Deletion of PAR1 resulted in more aggressive tumor growth, which was linked to apoptosis induction in transformed epithelia. Another study analyzed the effect of specific deletion of PAR1 in a pancreatic adenocarcinoma cell line KPC (Kras^G12D/+; TP53^R172H/+; Elas-cre^ER/+) and implanted the tumor cells into mice with very low circulating prothrombin levels or mice with ubiquitous PAR1^KO. This study precisely localized an evasive tumor mechanism to thrombin-PAR1 signaling in cancer cells since tumors lacking PAR1 expression grew much slower compared to those with unaltered PAR1 levels. The reduced tumor growth was associated with increased immune cell infiltration in PAR1-deficient tumors. Because immune cell infiltrated, ‘hot tumors’ have better clinical outcomes, these preclinical data indicated an important role for coagulation signaling in anti-tumor immune responses.

Thrombin-mediated TGF-β release from platelets drives cancer progression

Another evasive immune mechanism for thrombin involves platelets. Thrombin can activate TGF-β and TGF-β, in turn upregulates PAR1, suggesting a potential amplification loop involving PAR1. TGF-β is a key factor inducing fibrosis, limiting immune cell infiltration and inducing immune-suppressive regulatory T cells. Thrombin also cleaves GARP from the platelet surface, activating TGF-β, and subsequently fostering fibrosis and expanding Tregs for enhanced tumor growth. This signaling pathway is blocked by the thrombin inhibitor dabigatran, indicating a mechanism by which DOACs targeting thrombin can improve anti-tumor immunity.

Macrophage-derived FX fosters tumor growth via PAR2

Our data indicated a relevant more upstream coagulation signaling pathway that is not influenced by thrombin inhibition. In the preclinical spontaneous breast cancer model PyMT, as well as several transplantable tumor models in immunocompetent mice lacking either F10 expression in the myeloid compartment (F10^flLysM-cre; F10^flCx3Cr1-cre) or being resistant to FXa-mediated PAR2 cleavage (in the PAR2 G37I mutant mouse model), we showed that monocyte/macrophage-specific expression of FX promotes tumor immune evasion through cell-autonomous PAR2 signaling. This pathway converges with other innate immune signaling pathways to control dendritic cell phenotypes and to impair cytotoxic T-cell responses. Disrupting this tumor-specific pathogenic signaling by pharmacological inhibition of FXa with rivaroxaban led to an increase in antigen-presenting cells within the tumor-draining lymph nodes and the expansion of cytotoxic T cells in the TME, and reduced tumor progression. These experiments identified a novel tumor-specific pathogenic pathway that should be particularly sensitive to DOACs targeting FXa specifically but not other anticoagulants in clinical use.

DOACs synergize with checkpoint inhibitors in mice

Checkpoint inhibitor therapy with anti-PD-1 or anti-PD-L1 targets the immune suppressive effects of macrophages and tumor cells in the TME on tumor-infiltrating cytotoxic T cells. The coagulation- and platelet-related immune suppressive effects can be remarkably reversed by pharmacological inhibition of FXa with rivaroxaban or thrombin with dabigatran to enhance the efficacy of checkpoint inhibitor therapy with α-PD-L1 or α-PD1 in preclinical models. This synergy in the case of FXa specific DOACs can be explained by the DOAC-induced stimulation of antigen-presenting and immune-priming dendritic cells, which expand the pool of CTL susceptible for checkpoint inhibitor therapy-induced activation. The clinical relevance of these findings, nevertheless, remained uncertain.

FXa inhibitors synergize with checkpoint inhibitors in melanoma patients

A recent retrospective study of metastatic melanoma patients receiving immune checkpoint inhibitors revealed new insights into the relevant coagulation proteases that can be targeted with specific DOACS for improved clinical outcomes. In this cohort stratified for concomitant anticoagulant used during the initial course of checkpoint inhibitor therapy, similar numbers of patients received either low-molecular-weight heparins (LMWH), vitamin K antagonists (VKA), or FXa DOACs. Only patients receiving FXa inhibitors showed a significantly prolonged progression-free survival, whereas patients treated with LMWH or VKA did not differ from patients without anticoagulation. As the outcome for LMWH was expected because of its limited tissue penetrance, the inefficiency of VKA is surprising and may indicate either inefficient inhibition of the synthesis of FX in the TME or additional important roles of other vitamin K-dependent coagulation proteases in anti-tumor immunity. The remarkable and strong synergistic effect observed for FXa DOACs on checkpoint inhibitor therapy warrants confirmation in prospective trials.