Soluble P-Selectin, Von Willebrand Factor, And Adamts13 Levels As Risk Factors Of Deep Vein Thrombosis In Cancer Patients Undergoing Chemotherapy

Budi Setiawan (  boedhi_smg73@yahoo.com ) Dr Kariadi General Hospital Medical Center Cecilia Oktaria Permatadewi Dr Kariadi General Hospital Medical Center Baringin de Samakto Dr Kariadi General Hospital Medical Center Ashar Bugis Dr Kariadi General Hospital Medical Center Eko Adhi Pangarsa Dr Kariadi General Hospital Medical Center Damai Santosa Dr Kariadi General Hospital Medical Center Catharina Suharti Dr Kariadi General Hospital Medical Center


Introduction
Deep vein thrombosis (DVT) and its major complication, pulmonary embolism (PE), designated together as venous thromboembolism (VTE), 1 are the leading causes of death, morbidity, late treatment, and increasing treatment costs. Moreover, DVT in cancers prompts the risk of recurring DVT, bleeding complications, long-term need of anticoagulant, and interruptions to chemotherapy programs. Cancer patients have 4-7 times higher risk of having DVT compared to those without cancers. The incidence of DVT also increases among cancer patients undergoing chemotherapy. Chemotherapy increases the risk of DVT by six times and recurrent DVT by two times. 2,3 Systemic chemotherapy may cause injuries on blood vessel walls due to its direct toxic effect on endothelial cells which prompts endothelial cells activation. The increasing activation of endothelial cells is considered as one of the causes of hemostatic activation. 4,5 Biochemically, chemotherapy-induced vascular endothelial cell activation (VECA) is indicated by the increasing the number of endothelial cells in circulation, von Willebrand Factor (VWF) in plasma and endothelial cell activation markers such as vascular cell adhesion molecule-1 (VCAM-1), soluble P-selectin (sP-selectin) and E-selectin. 5,6 An appropriate tool of DVT risk strati cations facilitates a more effective thromboprophylaxis therapy.
Khorana risk score is a score to predict the risk of cancer-related thrombosis in patients undergoing chemotherapy. 7 Various recent studies have been aimed at nding other biomarkers with high predictive value on DVT incidence. 8,9,10,11 Soluble P-selectin (CD62) is located in platelets alpha granule and Weibel-Palade body (WPB) endothelial cell. 12 Soluble P-selectin increases tissue factor (TF) expressions on monocytes and mediates TF transfer to platelets. Tissue factor is the main initiator of in vivo coagulation causing the activation of coagulation cascade. 13 Plasma levels of sP-selectin increase in DVT. High levels of sP-selectin are also related to the increasing risk of recurrent DVT, and in cancer patients, high plasma levels of sP-selectin indication of DVT incidence. 12 A platelet-adhesive blood coagulation protein, VWF is synthesized mainly in the vascular endothelial cells and megakaryocytes, stored in WPB and platelets alpha granule in the form of "ultra large" VWF (UL-VWF). Von Willebrand factor is released into plasma in the form of UL-VWF, very active in interacting with platelets and collagen. In plasma, UL-VWF is rapidly depolymerazed into small sizes ranging from 500 to 20,000 kDa. 14,15 Von Willebrand factor plays an important role in both primary and secondary hemostasis. This glycoprotein acts as a carrier for factor VIII and mediator for platelet adhesion to endothelial cells. 16 Previous studies have shown that its number would increase in cancer patients with DVT, and it is related to cancer staging and metastasis. The increasing VWF antigen and the decreasing VWF-cleaving protease will prompt thrombogenesis in cancer patients. 17 The biological role of a disintegrin-like and metalloproteinase with thrombospondin type 1, motif 13 (ADAMTS13) is to regulate thrombosis by controlling VWF activities. 9 The platelet adhesive nature of VWF depends on multimeric size and conformation in which large and long VWF multimers best react to platelets circulating in high shear stress and most thrombogenic. The VWF cleaving protease by ADAMTS13 determines the average size of circulating VWF by selectively cleaves the Tyr1605-Met1606 bond of VWF that becomes exposed under the in uence of uidic shear stress. ADAMTS13, therefore, has a signi cant systemic antithrombotic activity. 15 The previous studies have indicated lower levels of ADAMTS13 in cancer patients than those without cancers, which lead to increasing risk of DVT. They also decreased in metastasis cancer patients. 9 Considering these data, the aim of this study was to investigate the role of sP-selectin, VWF, and ADAMTS13 as predictive biomarkers for the incidence of DVT in cancer patients undergoing chemotherapy.

Study setting
This prospective cohort study was performed in Dr. Kariadi hospital, the university hospital of Diponegoro University, Semarang, Indonesia. The study protocol was approved by the Ethical Committee of the Dr. Kariadi Hospital, and written informed consent was obtained from all study participants.

Patients and data collection
During November 2016 and February 2017, 40 consecutive patients with active cancers undergoing chemotherapy were enrolled in the study followed consecutively for 3 months. The inclusion criterias for the study were patients with newly diagnosed of cancer with histological con rmation, age over 18 years, willingness to participate, and undersigned a written informed consent. Exclusion criteria were overt bacterial or viral infection within the last two weeks, venous or arterial thromboembolism within the last three months and continuous anticoagulation with vitamin K-antagonists or low molecular weight heparin (LMWH), and surgery or radiotherapy within the last two weeks.
Before this study was conducted, all patients had been informed about study's details in an individual interview. Then, anamnesis on patients' cancer history, tumor site, tumor histology and tumor stage were documented. The patients who met the inclusion criteria were selected as the subjects of this study. Samples were examined twice, before and after chemotherapy, to see the plasma levels of sP-selectin, VWF and ADAMTS13. The patients were monitored for 3 months. In the rst, second and third month, evaluations on DVT occurrence were conducted with Wells pretest probability model to assess DVT. If the Wells score was ≥ 2, color duplex sonography was performed. If there were no symptoms of DVT, and Wells score was < 2, color duplex sonography examination was conducted to determine DVT incidence at the end of the third month of chemotherapy.

Color duplex sonography
Color duplex sonography was performed at the Division of Radiology of Dr. Kariadi Hospital, Semarang, Indonesia. Patients with clinically suspected DVT and Wells score ≥ 2 were performed color duplex ultrasonography. To avoid investigator-related variations of the results, color duplex sonography was performed in each patient by the same investigator at all time points.

Measurement of sP-selectin, VWF and ADAMTS13
Venous blood specimens were collected by sterile and atraumatic antecubital venipuncture, and were collected in citrate vacutainer tubes SST 5 ml, containing 0.5 ml of liquid anticoagulant. To obtain platelet-poor plasma, the citrated blood was centrifuged at 1000 g (3000 rpm) for 15 minutes. Plasma aliquots were stored at-20 °C until they were assayed to determine sP-selectin, VWF, and ADAMTS13 plasma levels in series. Samples were coded prior to laboratory analysis.
Measurements of sP-selectin levels were carried out using a recombinant human P-selectin Immunoassay/CD62P catalog number ADP3 (R&D Systems, Inc. Blood samples were scheduled to be drawn at the following time-points: (i) baseline, before initial chemotherapy, and (ii) the day after initial chemotherapy was completely administrated.

Statistical analysis
Continuous variables were summarized with mean (standard deviation [SD]) or median (25th-75th percentile) whereas categorical data were described by absolute frequencies and percentages. The correlation between two continuous variables was evaluated with Spearman rank correlation coe cient. Independent t-tests were done to compare numeric variables between DVT and non DVT patients. Test for normality of the data with Shapiro-Wilk test. Independent t-test was done when the data was normally distributed. Mann-Whitney was done when the data was not normally distributed. Receiver operating characteristic (ROC) curve analysis was performed to determine the cut-off point levels of sP-selectin, VWF, and ADAMTS13 for DVT incidence of cancer patients undergoing chemotherapy. The cut-off-point of signi cance was p = 0.05 with 95% con dence interval. Relative risk (RR) was determined with chisquare test between two comparative groups. All potential predictors were entered simultaneously in a multivariate logistic regression model that was reduced using a backward selection method. All test was two sided with p < 0.05 considered statistically signi cant. The Statistical Package for Social Science (IBM v 21; SPSS Inc., USA) was used for data analysis.

Results
During November 2016 and February 2017, 40 patients with newly diagnosed cancer were enrolled in the study. One patient died before given chemotherapy on 6th weeks. The median age of the patients was 49 (20-71) years old with male patients more dominant (55% vs 45%). The most frequent cancer types were colorectal cancer (45%) and cervical cancer (15%) ( Table 1).  (25) Colon, n (%) 8 (20) Cervical, n (%) 6 (15) Pancreas, n (%) 3 (7. The cancer patients' levels of sP-selectin were relatively high (normal value levels of sP-selectin 0.99-47.7 ng/mL) both before and after chemotherapy. The mean levels of sP-Selectin pre-chemotherapy and post-chemotherapy were 87.61 ± 39.03 ng/mL and 98.52 ± 54.76 ng/mL, respectively (Table 2). In patients with DVT, the levels of sP-selectin pre-chemotherapy and post-chemotherapy were higher than patients without DVT. The median levels of sP-selectin pre-chemotherapy and post-chemotherapy in the group of patients with DVT were 121.0, IQR 107.5-230.6 ng/mL and 204.4, IQR 110.9-278.3 ng/mL, respectively. In cancer patients with DVT, the levels of sP-selectin were signi cantly higher than those without DVT in both pre-chemotherapy and post-chemotherapy (p = 0.041; p = 0.000, respectively) ( Table 3 and Fig. 1, A). The levels of VWF in the cancer patients were relatively high (normal value levels of VWF 0.5-1.5 U/mL) pre-chemotherapy and post-chemotherapy. The mean levels of VWF pre-chemotherapy and postchemotherapy were 1.57 ± 0.87 U/mL and 1.83 ± 0.89 U/mL, respectively (Table 2). In patients with DVT, the levels of VWF pre-chemotherapy and post-chemotherapy were higher than those without DVT. The median levels of VWF pre-chemotherapy and post-chemotherapy in the patients with DVT were 3.03, IQR 2.97-3.75 U/mL and 3.43, IQR 3-3.97 U/mL, respectively. In cancer patients with DVT, the levels of VWF were signi cantly higher than those without DVT in both pre-chemotherapy and post-chemotherapy (p = 0.001; p = 0.002, respectively) ( Table 3 and Fig. 1,B).
The levels of ADAMTS13 in the cancer patients were relatively normal (normal value levels of ADAMTS13 0.72-1.17 U/mL) pre-chemotherapy and post-chemotherapy. The mean levels of ADAMTS13 prechemotherapy and post-chemotherapy were 1.06 ± 0.50 U/mL and 1.13 ± 0.55 U/mL, respectively ( Table 2). In patients with DVT, the levels of ADAMTS13 pre-chemotherapy and post-chemotherapy were lower than those of the patients without DVT. The median levels of ADAMTS13 pre-chemotherapy and post-chemotherapy in the patients with DVT were 0,74, IQR 0,45 − 0,81 U/mL and 0,42, IQR 0.31-0.52 U/mL, respectively. In cancer patients with DVT, the levels of ADAMTS13 were signi cantly lower than patients without DVT in both pre-chemotherapy and post-chemotherapy (p = 0.007 and p = 0.001, respectively) ( Table 3 and Fig. 1,C).   Table 4).
Finally, in a stepwise manner, multiple logistic regression analysis was performed using DVT incidence as dependent variable vs. three possible biomarkers (sP-selectine, vWF, and ADAMTS pre-and postchemotherapy) starting with a full model and removing non-signi cant variable one by one. The potential independent variable were categorized as dichotomy using pre-determined cut off value. The nal result showed that VWF levels > 2.99 U/mL pre-chemotherapy was the only predictor of DVT incidence, with a RR of 11.1 (95% CI, 1.95-62.74, p = 0,007)

Discussion
This study found that the levels of sP-selectin and VWF in cancer patients with DVT pre-chemotherapy and post-chemotherapy were higher than patients without DVT. Furthermore, the levels of sP-selectin and VWF were getting higher after chemotherapy. In DVT cancer patients, the levels of sP-selectin and VWF were signi cantly higher than patients without DVT in both pre-chemotherapy and post-chemotherapy (p = 0.041; p = 0.000 and p = 0,001; p = 0,002, respectively).
In the patients with DVT, the levels of ADAMTS13 pre-chemotherapy and post-chemotherapy were lower than patients without DVT. The levels of ADAMTS13 were signi cantly lower than patients without DVT in both pre-chemotherapy and post-chemotherapy (p = 0.007 and p = 0.001, respectively).
It is known that sP selectin and VWF are stored in platelets alpha granules and WPB bodies endothelial cells. 12,14 The study conducted by Payne H et al. on mice showed that endothelial activation and WPB exocytosis were immunity mechanism which initiated DVT. In this process, WPB fused with plasma membrane and released its constituents such as VWF and sP-selectin, in membrane, which allowed cell recruitments. It corresponded with the increasing levels of sP-selectin and VWF in the patients with DVT. 18 Soluble P-selectin and VWF levels in the cancer patients with DVT, which were already high before chemotherapy and increased higher after chemotherapy showed that immune systems and in ammations, played signi cant roles in DVT pathogenesis. 1 Cancers and chemotherapy treatments can cause in ammatory conditions. 19 In cancers, there are in ammatory stimulates in the tumor microenvironment. 20,21 In ammatory stimulates in cancers and chemotherapy activate intracellular signal pathways which activate the production of in ammatory mediators. 22,23 Blood vessel invasions by cancer cells, diagnostic and therapeutic interventions, such as central venous catheter insertion, also cause injuries in cancer patients' blood vessels. This leads to endothelial cells damage which prompts endothelial cells activation resulting in in ammatory responses. 19 Systemic chemotherapy may cause injuries on blood vessel walls due to its direct toxic effect on endothelial cells which prompts endothelial cell activation. The increasing activation of endothelial cells is considered as one of the causes of increasing hemostatic activation. 4,5 Biochemically, chemotherapyinduced VECA is indicated by increasing endothelial cells in circulation, VWF in plasma and endothelial cell activation markers such as VCAM-1, sP-selectin and E-selectin. 5,6 The rising endothelial activities as a result of in ammatory induce various changes in endothelial cells and leukocytes and platelets which encourage procoagulant and prothrombotic surfaces in blood vessel walls that increase the risk of DVT. 6,24 The adhesion process of leukocytes to blood vessel endothelial cell is a feature of in ammatory proces. 6 There are leukocyte rolling, adhesion, and movement to the injured spots controlled by selectin, integrin, and other adhesive molecules. Some adhesive molecules such as P-selectin, involved in in ammation, contribute in thrombosis. 15 The interaction between P-selectin and its main receptor in leukocytes, Pselectin glycoprotein ligand 1 (PSGL-1), causes neutrophil and macrophage recruitment; together with other mediators, they induce leucocytes to produce procoagulant microparticles. Subsequently, P-selectin prompts rising TF expressions in monocytes and mediates TF transfer to platelets. Tissue factor is the main initiator of in vivo coagulant, causing the activation of coagulant cascade extrinsic pathways. 13 The study conducted by Ay et al. has showed that an elevated levels of sPselectin is a useful parameter to identify patients who are at risk of thrombotic events, showing an odds ratio of 2.6 for patients with serum levels of sPselectin > 53.1 ng/dl after adjusted for age, sex, surgery, chemotherapy and radiotherapy. 25 Von Willebrand factor has a role in both primary and secondary hemostasis. This glycoprotein acts as a carrier for factor VIII and a platelet adhesion mediator to endothelial cells. 16 Von Willebrand factor is an adhesive protein capable of attaching to tumor cells and platelets, which can cause microthrombi formations. Platelet-tumor cell aggregation also facilitate the metastasis process by making it easier for tumor cells to adhere and migrate through vessel walls. 26 The previous study has shown that the numbers will increase in cancer patients with DVT and are related to cancer stage and metastasis. The increasing VWF antigen and the decreasing VWF-cleaving protease will prompt thrombogenesis in cancer patients. 17 High VWF levels were signi cantly associated with the risk of developing DVT in cancer patients. 27 The rise of VWF in circulation causes increasing ADAMTS13 activities, which results in decreasing levels of ADAMTS13. Mechanistically, this association is hypothesized that DVT related to cancers can be based on platelet aggregations mediated by VWF. Bauer et al. have recently showed that melanoma cells could activate vascular endothelial cells and push them to release ULVWF followed by platelets aggregation in vitro. They further explained that a combination of VWF release and decreased ADAMTS13 in tumor tissues tend to prompt procoagulant environment. 28 The decrease in ADAMTS-13 activities, which regulate size and, therefore, affect VWF cleansing from circulation, may result in a rise in plasma VWF. Although the mechanism prompting a decrease in plasma ADAMTS-13 in cancer patients has not been fully understood, various oncogenes have been found to regulate the expression of extracellular proteinases, including matrix-degrading metalloproteinases, which can directly disrupt ADAMTS13 activities. 26 The result of this study also showed that the cut-off points of sP-selectin and VWF levels were high both pre-chemotherapy and post-chemotherapy and the low levels of ADAMTS13 pre-chemotherapy and post-chemotherapy posed high relative risk for DVT incidence. The levels of sP-selectin, VWF and ADAMTS13 pre-chemotherapy with cut-off point > 106.7 ng/mL, > 2.99 U/mL and < 0,80 U/mL, respectively had RR for DVT incidence being 16  The data above indicated that the cut-off points of sP-selectin, VWF, and ADAMTS13 levels both prechemotherapy and post-chemotherapy could be used to predict the incidence of DVT in the cancer patients undergoing chemotherapy. Multiple logistic regression analysis is performed to determine the strongest biomarker as a predictor of DVT incidence pre-and post-chemotherapy.
Multiple logistic regression analysis showed that VWF levels pre-chemotherapy was the strongest biomarker as a predictor of DVT incidence pre-chemoterapy. Cancer patients whose initially have VWF levels > 2.99 U/mL will have 11-fold increased risk of developing DVT (RR 11,1;95% CI 1.95-62.74, p = 0.007) independently to other coagulation marker such as sP-selectin and ADAMTS13. The data above indicated that the cut-off points of VWF levels > 2.99 U/mL effectively predict the incidence of DVT in cancer patients undergoing chemotherapy.
Deep vein thrombosis has long been considered as a blood clotting disorder. The new understanding of DVT mechanism with some evidence and research data above shows the important role of immune systems and in ammations in the pathogenesis of DVT. Deep vein thrombosis is a process related to immunity and in ammation, not just the coagulation process that causes thrombosis. Local endothelial activations and the release of WPB present in endothelial cells have a very important role in initiating DVT. In addition to vascular systems, the immune system is emerging as a pivotal player in the pathophysiology of DVT. 1,29 The paradigm above opens the way to new thoughts on studies about new potential thromboprophylaxis strategies, or suggests a role for anti-in ammatory, which could be used for DVT prevention with a lower risk of bleeding complications than conventional therapeutic approaches. 1,29 The limitation of this study is that sP-Selectin, VWF, ADAMTS13 and color duplex sonography serial examinations were not conducted monthly. They can be used to see the changes in the levels of sPselectin, VWF and ADAMTS13 every chemotherapy cycle, so that the changes in biomarker levels and the incidence of DVT can be observed regularly. It will allow us to see the changes in biomarker levels affecting DVT incidence more accurately.

Conclusion
Our study demonstrated that (i) plasma levels of sP-selectin and VWF were high and ADAMTS13 were low in cancer patients with DVT (ii) plasma levels of VWF more than 2.99 U/mL is an independent risk factor for the incidence of DVT in cancer patients undergoing chemotherapy, which indicated VWF could be an early and helpful marker for thromboprophylaxis therapy in cancer patients undergoing chemotherapy.
Further research is needed to understand the important role of immune system and in ammation in the pathogenesis of DVT and thus opens the way to new thoughts on new potential thromboprophylaxis strategies, or suggest a role for anti-in ammatory, which could be used for DVT prevention with a lower risk of bleeding complications in comparison to conventional therapeutic approaches. 1