Background
Acute myeloid leukemia (AML) is a highly aggressive hematological malignancy characterized by clonal expansion of transformed myeloblasts. Patients with AML are at increased risk for both hemorrhagic and thromboembolic complications [
1‐
3]. Besides disease- or treatment-related thrombocytopenia, AML patients may suffer from complex systemic coagulopathies such as overt disseminated intravascular coagulation (DIC), excessive fibrinolysis, or non-specific proteolysis. In particular, patients with acute promyelocytic leukemia (APL) are at substantial risk for a thrombohemorrhagic syndrome that, despite modern anti-leukemic therapy, still significantly contributes to early morbidity and mortality [
4].
In newly diagnosed AML, potentially large numbers of myeloblasts circulate in direct contact with the plasma compartment. It is thus very likely that the procoagulant (or fibrinolytic) phenotype of AML blasts will have a profound effect on the hemostatic system. In fact, early studies have implicated over-expression of tissue factor (TF), the principal initiator of the extrinsic coagulation protease cascade, in the pathogenesis of AML-associated DIC [
5‐
8]. More recent studies, however, have questioned the relevance of TF-driven coagulation activation to the hypercoagulable state in various hematological malignancies, including AML, and suggested alternative mechanisms such as acquired activated protein C resistance [
9,
10].
Because a thorough understanding of the cellular and molecular pathways underlying AML-associated DIC is mandatory to identify novel drug targets for the prevention and treatment of potentially fatal thrombohemorrhagic events, we conducted a prospective study to further define the correlates of systemic coagulation activation in 69 patients with newly diagnosed AML.
Discussion
In this study, we investigated PBMC-associated TF PCA in 69 patients with newly diagnosed AML and found that TF PCA expression beyond a critical level was inevitably associated with the evolution of decompensated DIC. Importantly, DIC was not restricted to APL patients, but also occurred in seven other patients, three of whom had acute (myelo)monocytic leukemia. To our knowledge, this is one of the largest prospective studies so far specifically addressing the determinants of systemic coagulation activation in newly diagnosed AML.
The increase in PBMC-associated TF PCA following cell disruption is consistent with the concept of cryptic TF that can be de-encrypted upon cellular activation or induction of apoptosis/necrosis [
24]. Previous studies have clearly shown that cryptic TF is entirely cell surface-expressed on (myelo)monocytic cells and not liberated from intracellular storage pools upon cell lysis [
25‐
27]. Instead, negatively charged phospholipids (i.e. phosphatidylserine) exposed on the outer membrane leaflet act synergistically with thiol-disulfide exchange reactions to activate surface-expressed TF [
28‐
30]. In this regard, our findings provide circumstantial clinical evidence that membrane alterations associated with apoptosis/necrosis indeed contribute to cellular TF activation in vivo, because serum LDH levels, a surrogate marker for leukemic cell turnover, significantly correlated with plasma D-dimer. Systemic coagulation activation in AML is thus, at least to a substantial extent, determined by the following three factors: specific upregulation of cellular TF PCA following leukemic transformation of hematopoietic stem cells, elevated numbers of circulating blasts, and a significant degree of spontaneous apoptosis/necrosis. In our study, none of the patients developed overt DIC after the implementation of induction chemotherapy, which may de-encrypt cellular TF through the above mechanisms [
26,
29]. Interestingly, in DIC patient no. 11 (Table
2), the systemic coagulopathy resolved during phases of AML remission, but recurred simultaneously with fulminant AML relapse, suggesting that the procoagulant phenotype of myeloblasts may be preserved despite intensive anti-leukemic therapy, including allogenic hematopoietic stem cell transplantation.
Our findings also point to a role of circulating TF-bearing MPs in the pathogenesis of AML-associated DIC, which has most recently been suggested by other experimental and clinical observations [
31‐
33]. Shedding of MPs typically results from cellular activation or apoptosis, but may occur spontaneously in solid cancers and hematological malignancies [
34]. Although MP-associated TF PCA was more frequently detected in DIC as compared to non-DIC patients, our findings do not allow the conclusion that shedding of TF-bearing MPs alone was sufficient to cause overt DIC in newly diagnosed AML. However, caution is warranted when comparing results of different studies due to a lack of standardized assays to measure MP TF.
The finding that plasma TF antigen was significantly increased in DIC as compared to non-DIC patients and healthy controls may be interpreted in favor of our conclusion that TF is an important determinant of DIC evolution in newly diagnosed AML. It is important to note, however, that solid-phase immunoassays may preferentially capture soluble (i.e. truncated) TF, which is functionally inactive, rather than MP-associated TF. Furthermore, there is concern about a lack of specificity when using human plasma in commercial TF ELISAs, which may result in inappropriately high background signals [
35,
36]. Nevertheless, in our study, plasma TF antigen was specifically upregulated in patients with DIC (Fig.
4c). Because the commercial TF ELISA used in our study does not detect alternatively spliced TF (unpublished data), future studies should therefore investigate, if proteolytic degradation of full-length TF by enzymes liberated from dying myeloblasts represents a novel mechanism of controlling the extrinsic coagulation pathway in newly diagnosed AML. Taken together, functional assays are more sensitive and specific than ELISAs in detecting minute amounts of biologically active TF, but TF immunoassays may provide useful complementary information in complex disease states.
An important finding of our study is that significant systemic coagulation activation, as indicated by excessively elevated plasma D-dimer, may occur in the absence of TF PCA overexpression by PBMCs. In fact, three of the eleven DIC patients (27%) had TF PCA levels below the threshold values of 2,500 AU/10
6 cells for intact and 8,000 AU/10
6 cells for disrupted PBMCs (Fig.
2). All of these patients had quite high numbers of circulating blasts, but two of them had only moderately increased serum LDH levels (Table
2). These observations point to alternative cellular and molecular pathways contributing to systemic coagulation in newly diagnosed AML. In this regard, negatively charged polymers such as polyphosphates or nucleic acids (i.e. RNA and DNA) have recently been identified as key mediators of intrinsic contact pathway activation in vivo [
37]. Consistent with a highly aggressive hematological malignancy leading to substantial spontaneous cell death, we found significantly increased levels of cell-free plasma DNA in the AML patient cohort. Furthermore, DNA levels correlated with plasma D-dimer and serum LDH. Even when the seven DIC patients with exceedingly high PBMC-associated TF PCA were excluded, levels of cell-free plasma DNA were associated with the degree of systemic coagulation activation (Fig.
5e). These findings may be regarded as circumstantial clinical evidence that liberation of DNA from apoptotic/necrotic AML blasts indeed contributes to intravascular thrombin generation through contact activation of factor XII [
38] and may thus stimulate further research into the role of the intrinsic coagulation pathway in AML-associated coagulopathies. It has to be noted, however, that we did not specifically consider the contribution of cytoplasmic RNA [
23], which may be liberated from dying cells at an earlier stage than nuclear DNA and which may be more closely correlated with serum LDH levels. Nevertheless, DNA-mediated activation of factor XII could explain why other studies did not find a significant contribution of the TF-dependent coagulation pathway to systemic coagulation activation in various hematological malignancies, including AML [
9,
10]. Another potential mediator of PCA in AML is cancer procoagulant (CP), a cysteine proteinase that activates factor X independently of factor VIIa (FVIIa) [
39], although more recent studies have questioned a predominant role of CP in paraneoplastic coagulation activation [
40].
In solid malignancies, TF expression has been linked to tumor angiogenesis. In particular, TF has been shown to regulate VEGF production via its cytoplasmic domain [
41], and TF-FVIIa-mediated cell signaling through cleavage of protease-activated receptor-2 (PAR-2) is crucial to neoplastic blood vessel formation [
42]. Although such a relationship has also been postulated for hematological malignancies [
43‐
45], our findings do not support an obvious association between TF PCA expression and VEGF secretion by isolated PBMCs in newly diagnosed AML. It has to be noted, however, that TF PCA, as measured in our study, is not directly coupled to TF signaling functions, because cryptic (i.e. non-coagulant) TF still binds FVIIa and thus retains its ability to activate PAR-2 [
46]. Furthermore, measurement of ex vivo VEGF production by isolated PBMCs may not adequately reflect bone marrow microvessel density, which is also likely dependent on additional angiogenic growth factors (e.g. bFGF, HGF) not addressed in our study [
16].
Patients presenting with DIC had a higher probability of early death (Fig.
6f). In this regard, our observation that TF PCA of both intact and disrupted PBMCs was associated with the presence of DIC may point to a prognostic role of TF in newly diagnosed AML. However, using either the upper limit of normal or the median TF PCA level of intact PBMCs as an arbitrary cut-off value, we did not observe an effect of increased TF PCA expression on median or OS in our patient cohort.
Our study has several limitations. First, although the total patient cohort is one of the largest so far, the generalizability of conclusions is still limited by the relatively small sample size of DIC patients. Likewise, much larger studies may be required to define the true impact of leukemic cell TF PCA expression on clinical outcome in newly diagnosed AML. Second, we used a qualitative TGA to measure MP-associated TF PCA. Although this assay may adequately mirror TF-driven coagulation activation in a plasma environment, quantitative assays such as a previously described chromogenic FXa generation assay [
36] may provide more subtle information on the role of TF-bearing MPs in AML-associated DIC. Third, the impact of induction chemotherapy or treatment with all-trans retinoic acid on TF PCA expression and MP shedding was not addressed in our study. Finally, albeit novel, our findings on the association of cell-free DNA with plasma D-Dimer are rather descriptive, and mechanistic studies with functional readouts and different methods of DNA isolation and quantification are thus required to further define the role of intrinsic contact pathway activation in newly diagnosed AML.
Methods
Patients
The study protocol was approved by the ethics committee of the city of Hamburg, Germany (no. OB-059/05), and the study was conducted in accordance with the second Declaration of Helsinki. All patients and controls provided written informed consent.
Patients were eligible if they had been admitted to the University Medical Center Hamburg-Eppendorf, Germany, for diagnostic work-up and subsequent treatment of presumed AML, which was confirmed by standard cytological, cytogenetic, and immunophenotypic criteria in 69 patients (Table
1). In five and seven patients, however, the final diagnoses were ALL or MDS, respectively. These patients served as additional reference groups for the analysis of PBMC-associated TF PCA (see below). The ALL patient cohort comprised four females and one male (mean age, 37 ± 14 years), and the MDS patient cohort comprised three females and four males (65 ± 9 years). The group of healthy controls (n = 10) included four females and six males (54 ± 18 years). Although consecutive patient enrollment was not feasible for logistic or consenting reasons, patients were enrolled on a random basis between August 2005 and November 2014, thus grossly representing the population of AML patients treated at our institution.
Patients had to be at least 18 years of age with no clinical and laboratory evidence for acute bacterial or viral infections or severe hepatic or renal dysfunction. Major surgical procedures within the preceding 4 weeks and any systemic anticoagulation prior to study enrollment were exclusion criteria.
Blood sampling and processing
All blood samples were taken before the initiation of specific anti-leukemic therapy. Venous blood was drawn by puncture of an antecubital vein or through a central venous catheter into plastic tubes prefilled with either 3.2% (0.109 M) sodium citrate or lithium heparin. Citrate-anticoagulated whole blood was centrifuged for 18 min at 4,000 rpm (1,200×g) to obtain platelet-poor plasma, which was snap-frozen in liquid nitrogen and stored in aliquots at −80°C until analysis. Heparin-anticoagulated whole blood was used to isolate peripheral blood mononuclear cells (PBMCs) as described below.
Measurement of plasma D-dimer and thrombin–antithrombin (TAT) complexes
Plasma D-dimer and TAT complexes were quantified using the Innovance™ D-dimer test on a BCS™ coagulation analyzer and the Enzygnost™ TAT micro kit, respectively (Siemens Healthcare).
Isolation of PBMCs
PBMCs were isolated from heparinized whole blood by density gradient centrifugation on Ficoll-Hypaque (Pharmacia) as previously described [
26]. Cells were washed in phosphate-buffered saline (PBS), adjusted to 1 × 10
6/mL, and immediately analyzed for TF procoagulant activity (PCA).
Measurement of cellular TF PCA
A previously described single-stage clotting assay was used to measure TF-specific PCA in PBMCs that were either left intact or disrupted by repeated freeze-thawing [
47]. Briefly, 100 µL of cell suspension (1 × 10
6/mL) were incubated with 100 µL of normal human plasma (NHP) (HemosIL™; Instrumentation Laboratory) in the cuvettes of a KC10 coagulation instrument (Amelung) for 2 min at 37°C. Following the addition of CaCl
2 (final concentration, 5 mM), times until fibrin clot formation were recorded. All measurements were carried out in triplicates in the presence of 30 µg/mL inhibitory TF monoclonal antibody (#4509; American Diagnostica) or control IgG (Sigma). Clotting times were converted into arbitrary PCA units (AU) by reference to a standard curve obtained by serial dilutions of lipidated recombinant human TF (Innovin™; Dade Behring).
Measurement of microparticle-associated TF PCA
Plasma microparticles (MPs) were isolated by centrifugation at 100,000×
g for 90 min at 10°C. Pellets were resuspended in the original volume with PBS. MP-associated TF activity was measured by a TGA according to the method developed by Hemker [
48,
49]. Briefly, 20 µL of MP suspension were added, along with 20 µL of a mixture containing CaCl
2 and the fluorogenic substrate Z-GGA-AMC-HCl (Bachem BioSciences), to 80 µL NHP in the presence of 25 µg/mL corn trypsin inhibitor (Enzyme Research Laboratories). Thrombin generation was measured over a period of 90 min using a Synergy 2 fluorometer (BioTek). MP suspensions were added as the only source of TF and phospholipids to initiate coagulation. Inhibitory TF antibody H36 (50 µg/mL) (Sunol Molecular) was employed to verify TF specificity.
Measurement of plasma TF antigen
Plasma TF antigen was measured using the Quantikine™ enzyme-linked immunosorbent assay (ELISA) (R&D Systems) according to the manufacturer’s instructions.
Measurement of VEGF antigen in plasma and culture supernatant
The Quantikine™ ELISA (R&D Systems) was used to measure vascular endothelial growth factor (VEGF) antigen levels in plasma samples and culture supernatants. Supernatants were prepared by incubating isolated PBMCs (1 × 106/mL) for 24 h at 37°C and 5% CO2 in RPMI culture medium in the presence of 10% fetal calf serum. Cells were removed by centrifugation and supernatants stored at −80°C until analysis.
Quantification of cell-free plasma DNA
Plasma DNA was quantified based on a previously described method [
22]. Briefly, 2 µL of plasma were diluted in 98 µL of PBS containing 0.1% bovine serum albumin. Diluted plasma was then mixed with 100 µL PBS containing Sytox™ Green nucleic acid stain (Invitrogen) at a final concentration of 2 µM to label DNA fluorescently. Fluorescence was recorded using MTP reader (Tecan) with a 485 nm excitation and 535 nm emission filter set. Autofluorescence was considered as background and determined in samples mixed with PBS without Sytox™ Green. DNA concentrations were calculated based on a standard curve obtained by known concentrations of DNA.
Statistical analysis
Normally and non-normally distributed data were presented as mean ± standard deviation and median with (interquartile) range, respectively. Differences between two groups were analyzed using the two-sided Student’s t test or the Mann–Whitney U test. Differences between multiple groups were analyzed using the Kruskal–Wallis test. Categorical data were analyzed using the Fisher’s exact test. Correlation coefficients were according the method of Spearman. Kaplan–Meier survival curves were compared using the log-rank (Mantel–Cox) test. All analyses were carried out using GraphPad Prism™ software. A P value of <0.05 was considered statistically significant.
Authors’ contributions
CD recruited patients, collected and interpreted clinical data, and critically revised the manuscript. AA designed experiments, interpreted data, and co-wrote the manuscript. BS and MJA performed experiments and analyzed data. TF designed experiments, analyzed data, and critically revised the manuscript. MD performed experiments, analyzed data, and critically revised the manuscript. JLF and CB oversaw the study, contributed to data interpretation, and critically revised the manuscript. FL designed the study, analyzed data, and wrote the first draft of the manuscript. All authors read and approved the final manuscript.