Abstract
Background: Acute promyelocytic leukemia (APL), characterized by the t(15;17) translocation, represents a favorable-risk subgroup of AML patients when treated with ATRA-based regimens. The leading cause of early death and the major unmet medical need in this disease are the bleeding complications, which are mainly attributed to aberrant expression of Tissue Factor (TF) and Annexin A2 (ANXA2) on leukemic promyelocytes, leading to disseminated intravascular coagulation and to hyperfibrinolysis, respectively. APL patients are also at increased risk of venous and arterial thromboses. The mechanisms underlying APL-associated hemostatic complications are not fully elucidated.
Aims and Methods: We analyzed the transcriptome of 30 APL specimens comprised in the Leucegene 430 AML cohort, as well as normal hematopoietic cell populations (n=63) aiming to better understand the hemostasis-related transcriptomic landscape of this subgroup.
Results: We identified podoplanin (PDPN) as the most significantly and differentially overexpressed gene in APL (median 2.6 vs 0 RPKM for APL and AML, respectively, q-value = 7.3 x 10-29) and determined that PDPN gene expression correlates with protein surface expression as assessed by flow cytometry on primary APL cells (r=0.89). TF (6.6 vs 1.6, p=1.8x10-9) and ANXA2 (142.8 vs 69.1, p=1.9x10-5) are also more expressed in APL, but less specifically than PDPN. Podoplanin is a glycoprotein that physiologically binds to its receptor, CLEC2, on platelets to induce platelet aggregation causing the separation of blood and lymphatic vessels during embryogenesis (Uhrin et al Blood 2010). As anticipated, we found that PDPN is not expressed in all studied sorted cell subpopulations from normal blood or bone marrow specimens, including in promyelocytes (median of 0 RPKM for all), indicating that platelets are most likely not exposed to PDPN in the adult vasculature and that this protein is ectopically expressed on APL promyelocytes.
Next, we used lentiviral gene transfer to engineer OCI-AML5 cells, which do not express PDPN, to ectopically express this protein (AML5PDPN). When mixed with platelet-rich plasma (PRP), we found that AML5PDPN cells have much greater platelet-binding and activating capacity than AML5CTRL cells. Using light transmission aggregometry, only AML5PDPN cells, not AML5CTRL cells, could induce platelet aggregation at tested concentrations. These results indicate that PDPN expression on leukemia cells is sufficient to induce platelet binding, activation and aggregation. Using patient-derived cells, we similarly found that PDPN-expressing primary APL cells (n=3 samples), but not PDPN-negative APL (n=2) or AML (n=5) cells, have the capacity to bind and activate platelets and to induce platelet aggregation. Our results suggest that PDPN could contribute to APL-related thromboses, as was recently reported in brain tumors (Riedl et al Blood 2017).
We next tested the hypothesis that PDPN expression on human myeloid blasts leads to platelet consumption, thrombocytopenia and bleeding in vivo . NSG mice were transplanted with either AML5CTRL or AML5PDPN cells and were monitored until they showed clinical signs of leukemia. Platelet counts rapidly and significantly dropped in the AML5PDPN-cohort at day 25 and beyond when compared to control animals (median 343 vs 1005 x 109/L, p < 0.0001). Mice platelet counts were inversely correlated with circulating PDPN load (r=-0.88) and animals with PDPN+ leukemias had markedly prolonged bleeding time (median > 15:00 vs 7:45, p = 0.0003).
Finally, we hypothesized that ATRA reduces bleeding complications in patients, at least partially, by decreasing PDPN expression. We found that PDPN surface protein on primary APL promyelocytes is markedly decreased within the first 24 hours of ATRA treatment in vitro, and levels reached less than 10% on day 4 when compared to untreated cells (DMSO). Arsenic trioxide or dexamethasone treatment led to a modest or no reduction in PDPN expression, respectively.
Conclusion: PDPN expression is specific to APL and, by inducing platelet binding, activation and aggregation, it contributes to platelet consumption and to bleeding complications. Our findings may also expose a new promising angle for the treatment of the early hemostatic complications found in APL patients.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.