WO2017067821A1

WO2017067821A1 - A method for determining presence or risk of hemostasis disorder

Info

Publication number: WO2017067821A1
Application number: PCT/EP2016/074358
Authority: WO
Inventors: David-Alexandre TREGOUET
Original assignee: Universite Pierre Et Marie Curie (Paris 6); INSERM (Institut National de la Santé et de la Recherche Médicale); Universite D'aix-Marseille
Priority date: 2015-10-22
Filing date: 2016-10-11
Publication date: 2017-04-27

Abstract

The present invention relates to a method for determining the presence or risk of developing thrombosis or the risk of a bleeding event in a patient, which method comprises determining presence or concentration of PDGFB in a biological sample of the patient, in a biological sample of the patient.

Description

A method for determining presence or risk of hemostasis disorder

The present invention relates to a method for detecting a hemostasis disorder or abnormality, especially a thrombotic risk or a risk of bleeding event, in a patient.

Background of the invention

Hemostasis, the arrest of bleeding from an injured blood vessel, requires the combined activity of vascular, platelet, and plasma factors. Regulatory mechanisms counterbalance the tendency of clots to form. Hemostatic abnormalities (or hemostasis disorders) can lead to excessive bleeding or thrombosis.

Thrombosis is the formation of a blood clot, inside a blood vessel, obstructing the flow of blood through the circulatory system. When a blood vessel is injured, the organism uses thrombocytes and fibrin to form a blood clot to prevent blood loss. Even when a blood vessel is not injured, blood clots may form in the body in conditions such as hypercoagulability, stasis or turbulence of the blood flow.

Expensive techniques, such as echo Doppler or angiography can detect the presence of a thrombus. However no reliable and easy-to-use technique is available to determine a level of risk of thrombosis or thrombotic events.

Thrombin is an enzyme which is a central product of the response to vascular injury, displaying procoagulant, anticoagulant, antifibrino lytic and cellular effects; the magnitude and the timing of these effects are critical to normal hemostasis.

Patient with high levels of thrombin generation are at risk for thrombotic diseases such as acute ischemic stroke, venous thromboembolism and myocardial infarction while bleeding events are observed in presence of very low thrombin generation. However determining the level of thrombin generation is time-consuming and cumbersome.

In particular, venous thromboembolism (VTE) is responsible for a large public health burden in the western world, with a high incidence of (1-2/1000 person years) that exponentially increases with age. Around 10 to 20 per cent will face an impaired quality of life after a first deep vein thrombosis (DVT) and develop a post thrombotic syndrome in the lower limbs that leads to chronic swollenness and ulceration. VTE risk prediction still remains a challenge. VTE is multifactorial disease with interacting genetic, acquired and environmental risk factors. Two thirds of all VTE are provoked by i.e. trauma, surgery, immobilization and influence of estrogens, malignancy and rheumatic systemic disease. The only plasma biomarker for VTE that is in clinical use today is D-dimer, a split product from the cross-linked fibrin clot. In a diagnostic score, D-dimer has a high negative prediction value to rule out diagnosis for those with low suspicion of acute thrombosis. Still, D-dimer has very low specificity and furthermore is increased in many other conditions apart from VTE, i.e. cancer, pregnancy, inflammation and older age. A diagnostic biomarker with high positive prediction value that is comparable with Troponin for acute myocardial ischemia is still missing.

There remains a need for a rapid and reliable method for determining the presence or risk of thrombosis in a patient. Such method would make it possible to provide early treatment of the patient.

Summary of the invention

It is now provided a method for determining presence or risk of developing a hemostasis disorder in a patient, which method comprises determining presence or concentration of PDGFB, in a biological sample of the patient.

The method typically comprises comparing the PDGFB concentration in the patient with a control value, wherein a higher PDGFB concentration in the patient sample compared to the control value is indicative of presence or risk of developing thrombosis, while a lower PDGFB concentration in the patient sample compared to the control value is indicative of a risk of bleeding event in the patient.

Brief description of the drawings Figure 1:. Dot Box plots of relative levels of VWF (A) and PDGFB (B) in the discovery phase in VEBIOS and the replication phase in FARIVE. Values are in mean fluorescent intensity (MFI) as measured in the single binder bead assay on the Luminex^® system using HPA002082 antibody for VWF (A) or HPA11972 antibody for PDGFB (B); Figure 2: Protein sequence of PDGFB (SEQ ID NO:2) with target sequence of HP A antibody and IC-MS identified peptides. A protein expressed sequence tag (PREST) comprising of amino acids 27-159 was used to raise the HP A antibody HPA011972 (underlined by solid black line). The PDGFB peptides identified by IC-MS with the HPA011972 antibody in plasma are underlined by dashed black lines. Detailed description of the invention

The inventors have demonstrated that levels of PDGFB are significantly increased in patients with venous thromboembolism compared to control individuals.

On this basis, they propose to determine the presence or assess the risk of developing thrombosis in a patient, by monitoring the PDGFB concentration in a biological sample of the patient.

It is thus provided a method for determining presence or risk of developing thrombosis in a patient, which method comprises determining the PDGFB presence or concentration in a biological sample of the patient.

Such method helps the physicians to diagnose and start treatment without unnecessary delay. Moreover PDGFB measurement allows identification of patients who are likely to develop a recurrence. This is of paramount importance to avoid unnecessary prolonged anticoagulant treatment in patients at low risk of recurrence, which anticoagulant treatment may have dramatic side effects, including debilitating iatrogenic bleeding complications.

As a corollary method, it is also provided a method for determining a risk of bleeding event in a patient, which method comprises determining PDGFB concentration in a biological sample of the patient.

"PDGFB " (Platelet-derived growth factor subunit B) is a protein that in humans is encoded by the PDGFB gene. Human PDGFB consists in a 241 amino acid sequence available on GenBank (Access Number CAG30424.1). The protein is a member of the platelet-derived growth factor family. It is functional either as a homodimer (PDGF-BB) or as a heterodimer with the platelet- derived growth factor alpha (PDGFA) polypeptide (PDGF-AB), where the dimers are connected by disulfide bonds. However non- functional monodimer chains B exist.

In the context of the present invention, PDGFB may be detected in the form of

a PDGF B monomer;

a PDGF B chain covalently, or not covalently, bound to another protein which can be o another chain B of PDGF

o chain A of PDGF

o a distinct protein The term "patient" or "subject" means any human being or non-human mammal. Especially it is a man or woman, at any age. In a particular embodiment, the patient is an individual at risk of thrombosis or thrombotic event. In particular, the subject may be at increased risk of developing a thrombus due to a medical condition which disrupts hemostasis, or to a medical or surgical intervention. In another embodiment, the patient is an individual at risk of bleeding, e.g. the subject may be at increased risk of bleeding due to a medical or surgical intervention.

The "biological sample" may be any biological sample of the patient which might contain said PDGFB, especially a liquid sample such as blood, plasma, serum, urine, saliva. In a preferred embodiment, the biological sample is blood, plasma or serum, preferably plasma.

Advantageously the biological sample, such as plasma sample, may be subject to denaturing treatments to expose hidden epitopes of the protein. For instance it may be heat treated before analysis, or subject to depletion of high abundance plasma proteins prior to analysis, as in mass spectrometry based analysis.

The PDGFB concentration in the biological sample is determined using any method known by a person skilled in the art, including immunoassays or mass spectrometry. In a preferred embodiment, the PDGFB concentration in the sample is determined using an immunoassay. Monoclonal or polyclonal raised against PDGFB, such as antibodies commercially available, may be used for that purpose.

The anti-PDGFB antibodies may be advantageously specific for PDGF chain B, in that they do not substantially bind to PDGF chain A.

Immunoassays include Enzyme-linked immunosorbent assay (ELISA), lateral flow test, latex agglutination, other forms of immunochromatography, western blot, and/or magnetic immunoassay. For instance, the immunoassay may be radial immunodiffusion, nephelemetry, or a turbidimetric assay, which are particularly convenient when assaying the PDGFB protein in a blood or plasma sample.

Advantageously the assay does not substantially cross-react with PDGF-AA, nor with epidermal growth factor, basic fibroblast growth factor, nor transforming growth factor-β. In a particular example, PDGFB can be measured by a double antibody sandwich enzyme- linked immunosorbent assay (ELISA), e.g. as described in Leitzel et al, 1991 or Rossi et al, 1998. The method advantageously comprises comparing the PDGFB concentration in the patient with a control value.

The "control value" refers to a standard value of PDGFB concentration in healthy or normal subjects or a population of healthy or normal subjects. For instance, when the biological sample is plasma, physiological plasma PDGFB concentration is known to the person skilled in the art to range from 0.1 to 0.6 ng/ml (Leitzel et al, 1991; Zehetner et al, 2014).

A higher PDGFB concentration in the patient sample compared to the control value is indicative of risk of developing thrombosis. The term "higher concentration" refers to a significantly higher concentration, e.g. of more than 10%, preferably more than 20%, 30%, 40%, or 50%. When the biological sample is plasma, a PDGFB concentration superior to 0.6 ng/ml is indicative a presence or risk of developing thrombosis.

Determining the PDGFB concentration in the sample may further allow to assess the level of risk: the greater the PDGFB concentration is, the greater the risk is. The level of thrombotic risk increases when the PDGFB concentration increases.

Conversely, a lower PDGFB concentration in the patient sample compared to the control value is indicative of risk of a bleeding event. The term "lower concentration" refers to a significantly lower concentration, e.g. of less than 10%>, preferably less than 20%>, 30%>, 40%>, or 50%>. When the biological sample is plasma, a PDGFB concentration inferior to 0. lng/ml is indicative a presence or risk of a bleeding event.

Determining the PDGFB protein concentration in the sample may further allow to assess the level of risk: the lower the PDGFB concentration is, the greater the risk of bleeding event is, and the more severe the prognostic is. The level of risk of bleeding event increases when the PDGFB concentration decreases.

The invention allows to assess the presence of a thrombus, including a developing thrombosis, or to assess the risk of thrombosis in venous or arterial circulation. The term "risk of thrombosis" or "risk of developing thrombosis" refers to predisposition or likelihood for the patient or subject to have or develop a thrombosis or suffer a thrombotic event.

In a particular embodiment, the thrombosis is venous thrombosis. In particular, it may be selected from the group consisting of deep vein thrombosis, portal vein thrombosis, renal vein thrombosis, jugular vein thrombosis, Budd-Chiari syndrome, Paget- Schroetter disease and cerebral venous sinus thrombosis.

Deep vein thrombosis (DVT) is the formation of a blood clot within a deep vein. It most commonly affects leg veins, such as the femoral vein. Classical signs of DVT include swelling, pain and redness of the affected area.

In a preferred embodiment, the venous thrombosis is phlebitis or pulmonary embolism.

In another embodiment, the thrombosis is arterial thrombosis.

In particular, the method of the invention allows to detect or assess the risk of an arterial thrombosis which may cause a stroke or a myocardial infarction.

In a particular embodiment, the subject is at increased risk of developing a thrombus due to a medical condition which disrupts hemostasis, wherein said medical condition is preferably coronary artery disease or atherosclerosis.

The method of the invention further allows to assess the risk of thrombotic events in patients with coronary artery disease, particularly acute myocardial infarction, stroke, unstable angina, stable angina, or restenosis.

In another embodiment, the subject is at increased risk of developing a thrombus due to a medical procedure, including cardiac surgery, atherectomy, cardiopulmonary bypass, or catheterization, in particular cardiac catheterization.

When the patient is diagnosed with a thrombus, or risk of developing a thrombus according to the method of the invention, appropriate treatment may be prescribed. Such treatment may include administration of anticoagulants, such as Vitamin K antagonists, and/or heparin.

In a particular embodiment, the method of the invention thus comprises an additional step of treating the subject diagnosed with a thrombosis or a risk of developing a thrombosis.

As used herein, the term "treatment" or "therapy" includes curative and/or prophylactic treatment. More particularly, curative treatment refers to any of the alleviation, amelioration and/or elimination, reduction and/or stabilization (e.g., failure to progress to more advanced stages) of a symptom, as well as delay in progression of a symptom of a thrombosis.

Prophylactic treatment or "prevention" refers to any of: halting the onset, reducing the risk of development, reducing the incidence, delaying the onset, reducing the development, as well as increasing the time to onset of symptoms of a thrombosis. In the context of the present invention, the term "preventing" more particularly applies to a subject who is at risk of developing a thrombosis.

In another embodiment, the invention allows to assess the risk of bleeding event. The term "risk of bleeding event" refers to predisposition or likelihood for the patient or subject to suffer a bleeding event. The term "bleeding event", also referred to haemorrhagic episode, includes severe or life-threatening bleeding events as well as moderate bleeding events. Severe or life- threatening bleeding events cause haemodynamic compromise requiring intervention (e.g. blood or fluid replacement, inotropic support, surgical repair). Moderate bleeding events can require blood transfusion.

In a particular embodiment, the patient is affected with a systemic inflammatory disorder, which may be caused by an infection (sepsis) by a pathogen, such as a bacteria, but not necessarily so. In another embodiment, the patient is affected with a viral haemorrhagic fever.

When the patient is diagnosed with an increased risk of bleeding or haemorrhagia according to the method of the invention, extra care should be taken in case of medical or surgical intervention.

In a particular embodiment, the method of the invention thus comprises an additional step of treating the subject diagnosed with a risk of bleeding event. As used herein, the term "treatment" or "therapy" includes curative and/or prophylactic treatment. Treatment may involve administration of pro-coagulant agents or agents that modulate hemostasis.

The Example illustrates the invention without limiting its scope.

EXAMPLE: Identification of a novel plasma biomarker for venous thromboembolism Based on a high-throughput technology, the inventors aimed to identify novel biomarkers that may improve current clinical diagnosis and prediction tools for VTE risk. They conducted a high-throughput affinity plasma proteomic screening targeting about 400 proteins in relation to VTE risk in a case-control study composed of 88 VTE patients and 85 controls from Sweden. Main findings were replicated in an independent French sample collection enrolling 580 cases and 589 controls.

Material and methods

Discovery cohort -Venous thromboembolism biomarker study (VEBIOS) VEBIOS was initiated in 2010 at Karolinska University hospital in collaboration with Karolinska Institutet and Science for Life Laboratory, (KTH Royal Institute of Technology, Stockholm, Sweden) with the aim of identifying and translating biomarkers for VTE using resources available through the HPA-Project (Uhlen et al, 2015). VEBIOS is an on-going case- control study since January 2011, conducted by the Coagulation Unit at Karolinska University Hospital with recruitment of cases from three regional hospitals. Eligible cases are patients from 18 to 70 years of age with first VTE, which were referred from the Emergency clinics after diagnosis for a follow up at the respective outpatient clinics for Thrombosis and Hemostasis. VTE was confirmed by diagnostic imaging, venous ultrasonography in patients with DVT of the lower limbers, and computed tomography pulmonary angiography (CTPA) or ventilation perfusion scintigraphy (V/Q lung scan) in patients with pulmonary embolism (PE). Cases have been treated with anticoagulants during six to twelve months with no restriction in the choice of anticoagulant medication (i.e. vitamin K antagonist (VAK), direct oral anticoagulant (DOAC) or low molecular weight heparin (LMWH). Cases were identified by the physicians at the out clinics at respective hospital and asked to participate and donate blood after discontinuation of anticoagulant treatment. Cases found with severe thrombophilia after inclusion i.e. Antithrombin -, Protein S and Protein C deficiency, antiphospho lipid syndrome, homozygosity for either Factor V Leiden (G1691A) or the G20210A polymorphism in Prothrombin gene, or a combined heterozygosity were excluded from the study. Controls were continuously recruited from the population living in Stockholm County, based on a randomized selection using the Swedish Tax Agency register (Skatteverket: https://www.skatteverket.se/). Controls were contacted over phone by a research nurse at the Karolinska University Hospital Coagulation Unit and asked if willing to participate and donate blood, and in the case of a positive response followed by information sent by mail. Controls were matched to cases for age (± 2 years) and gender and included within one year after the matched case index date. Exclusion criteria for all eligible participants were personal history of VTE, pregnancy during the last three months before the index date (women only) or an active cancer within the last five years. All participants were free from anticoagulant medication for at least one month before time of blood sampling. Only persons who (a) could understand and read Swedish, (b) were in a mental condition to be capable to give an informed consent and (c) did not have short life expectancy were asked to participate. If an eligible participant fulfilled the inclusion criteria and accepted participation, the date of blood sampling was defined as the index date in the study. Data collection: All participants were asked to fill in a questionnaire regarding (a) demography, (b) provoking factors within three months preceding the VTE diagnosis, or the time of sampling for the controls, (c) co-existing cardiovascular risk factors and disease, autoimmune -, systemic - or chronic disease, (d) current health situation; alcohol consumption, smoking habits, physical activity (e) family history of VTE and other cardiovascular disease (f) ongoing medication (g) and for women only; reproductive history and hormonal use.

Blood sampling: Blood sampling was carried out at the Coagulation unit by a research nurse. All participants had to be fasting overnight and sampled the following morning 7-10 a.m. Each participant donated to the biobank; citrate and EDTA plasma, and whole blood for genetic analysis. Plasma samples were centrifuged at 2000g for 15 min at room temperature and aliquots were snap frozen and stored at-80 °C. Blood and plasma samples were sent to the Karolinska University Laboratory for the analysis of high sensitive CRP (hsCRP), D-dimer, blood glucose, creatinine, albumin, blood count, lipid profile, and liver enzymes. The cases were sampled one to six months after discontinuation of anticoagulant treatment. Biometry: At day of blood sampling (index date), the research nurse measured weight with clothes on (kg), length (m), blood pressure and pulse in left arm (mmHg) in sitting position, and circumference of waist and hip (cm)

Ethical permission: Informed written consent was obtained from all participants in accordance with the Declaration of Helsinki. VEBIOS was approved by the regional research ethics committee in Stockholm, Sweden (KI 2010/636-31/4).

Replication study -The FARIVE study

FAPJVE is a French multicentre case-control study carried out between 2003-2009 with the aim to study the interaction of environmental, genetic and biologic risk factor for first VTE and risk of recurrence. The study has previously been well described and involves consecutive inpatients or outpatients from 18 years of age, confirmed by diagnostic imaging and treated for a first episode of DVT and/or PE (Zhu et al, 2009).The controls were age- and sex matched and consisted of in- and out- patients, free of personal history of venous and arterial thrombotic disease. Exclusion criteria for all eligible participants were diagnosis of cancer, short life expectancy owing to other causes, renal - or liver failure. The replication study was based on a subset from FARIVE with genotype data (n=1200) (Tregouet et al, 2009). According to protocol, blood sampling was carried out at respective center and each participant donated citrate plasma and whole blood to the biobank, centralized to Hopital Europeen Georges Pompidou (HEGP) in Paris. The plasma was prepared by centrifugation for 20 minutes at 2000 g at room temperature and aliquots were snap frozen and stored at-80 °C. Cases were sampled at the study entry within the first weeks after diagnosis and if possible one month after anticoagulant withdrawal. Informed written consent was obtained from all participants in accordance with the Declaration of Helsinki. The study was approved by the Paris Broussais- HEGP ethics committee in Paris (2002-034).

Selection of candidate proteins and antibody reagents for Discovery analysis in VEBIOS

The selection of candidate proteins, 'targets', was based on evidence grade of association with VTE or intermediate traits, HPA antibodies availability and their pre-set quality criteria.

Evidence grade was divided into four groups, 'robust', 'supported', 'plausible' and 'hypothesised':

- 'Robust' was regarded as targets with a well-established association with VTE from the literature i.e. the protein von Willebrand factor (VWF), or candidate genes i.e. KNG1 supported by expression analysis and/or functional analysis (Morange et al 2011).

- 'Supported' was defined as (a) genetic data such as single nucleotide polymorphism (SNP) in gene/locus associated with VTE i.e. BAB but with no other type of supportive evidence (Antoni et al, 2010) and (b) targets i.e. Macrophage metalloelastase (MMP12) associated with arterial thrombosis and/or cardiovascular events based on genetic and functional data (Traylor et al, 2014).

- 'Plausible 'was defined as (a) proteins involved in intermediate traits of thrombosis (i.e. Protein disulphide isomerase A3 (PDIA3) (Wang et al, 2013) and (b) gene expression data linked to expression in the endothelium from in-house or public datasets, and (c) suggested plasma proteins, previously identified with significant association to cardiovascular events in an in-house proteomics study performed within the Protein Affinity Plasma Profiling group at SciLifeLab.

- Finally, the 'Hypothesis' group included proteins without supportive evidence indicating a direct role in thrombosis for the particular gene, but where tentative candidates were known to be involved in pathways of interest in thrombosis or intermediate traits, such as ORMl-like protein 3 (ORMDL3) (Nails et al, 2011). Antibodies corresponding to the proposed targets were obtained from the protein specific affinity reagents resource generated by the Human Proteome Atlas program (HP A). In the latest release of the Atlas (version 13), there are now 24 000 antibodies corresponding to 17 000 protein-coding genes in human (Uhlen et al, 2015). Selected targets were checked for protein epitope signature tags (PrEST), antibodies availability on the publicly available portal 'www.proteinatlas.org' and then assessed according to quality criteria. The latter were (a) concentration higher than 0.05mg/ml and (b) clear indication of the specific binding to the candidate target protein or protein fragment. This was based on confirmation of protein binding by mass spectrometry, control of cross-reactivity by protein arrays with PrEST fragments, determination of antibody specificity with western blot and finally immunohistochemical or immunofluorescence stained tissue arrays that had been annotated by trained pathologist (Uhlen et al, 2015).

From the 586 proteins initially selected as candidate targets, 408 candidates tagged by 755 HPA antibodies remained after the above filtering (Table 1). For 235 of these candidates, more than one HPA antibody was included. These correspond to different antibodies targeting either the same ('twins') or a different ('siblings') recombinant expressed region of the target protein (Table 1).

Table 1 : Summary of the selection on candidate proteins and corresponding HPA antibodies Candidates Corresponding HPA Ab per candidate

1 HPA

VEREMA Suggested Qualified* Ab 2 HPA Abs 3 HPA Abs 4 HPA Abs n HPA Abs

Robust 38 20 4 12 3 1 41

Supportive 142 99 47 29 13 8 176

Plausible 326 237 100 86 36 15 440

Hypothesis 80 54 22 23 6 3 98

N 586 410 173 150 58 27 755

Where: * is qualified according to HPA availability of corresponding Ab

The polyclonal antibody reagent HPA011972 is raised against a 133 amino acid sequence o the PDGFB protein:

EELYEMLSDHSIRSFDDLQRLLHGDPGEEDGAELDLNMTRSHSGGELESLARGRRSL GSLTIAEPAMIAECKTRTEVFEISRRLIDRTNANFLVWPPCVEVQRCSGCCNNRNVQC RPTQVQLRPVQVRKIEIV (SEQ ID NO: 1). The antigen is recombinantly expressed in E. coli, and injected into rabbits with Freund's adjuvant, so as to produce polyclonal antibodies, including HPA011972 reagent.

Protein plasma profiling

Protein plasma profiles were generated using two suspension bead arrays in which HPA antibodies were coupled to magnetic color-coded beads as previously described (Drobin et al, 2013; Neiman et al, 2013). This allows for protein profiling of biotinylated and heat-treated plasma proteins with a multiplexed high throughput approach that does not involve any purification, fractionation or depletion steps (Bystrom et al 2014). Each suspension bead array was composed of 384 antibodies and the method holds the possibility to generate protein profiles in 384 samples in parallel (Drobin et al, 2013; Neiman et al, 2013). Protein profiling in VEBIOS was performed in citrate and EDTA plasma from the same individuals, using identical conditions. Plasma sample selection

From the original 190 VEBIOS participants, 177 individuals were used for the discovery stage. Nine individuals were excluded because of they could not be properly matched for age (1 case and 8 controls) to avoid any bias. Three women were excluded due to provoked VTE in combination with oestrogen containing hormonal contraceptives and menopausal replacement treatment were omitted, and one sample had to be removed before the experiment after been identified with blood borne infection). The plasma samples were randomised to 96 well plates. In FARIVE, serving as a replication for VEBIOS results, citrate plasma samples were available from 603 cases and 597 controls from 13 centers. While 587 cases in FARIVE were sampled within the first weeks after diagnosis of VTE and initiation of anticoagulant treatment, 16 cases were sampled post-thrombosis. The latter were then excluded from the analysis. The cases were paired to controls from the same center, and matched to gender and age. The pairs were randomly placed on the same 384 well plates within a 96-well area. Clinical characteristic of the VEBIOS and FARIVE study samples are shown in Table 2 and Table 3, respectively. Table 2: Clinical characteristics of VEBIOS - Discovery study

VEBIOS

Variables 89 cases 89 controls

freq. n freq.

Localization of the thrombosis

Deep vein thrombosis, lower limbs 47 0.53

- Proximal 36 0.77

Pulmonary embolism 45 0.51

Gender and biometry

Gender; women 35 0.39 35 0.39

Age (years) (mean ± SD and range) 51,1 ±10.8 (20-70) 51,6 ±10.8 (21-70)

BMI (kg/m2) (mean ± SD and range) 26.7 ±4.6 (17-46) 25.6 ±4 (19-38)

Obese, BMI > 30 kg/m2 20 0.22 16 0.19

Cardiovascular risk factors

Hypertension* 19 0.21 10 0.11

Hyperlipidemia* 5 0.06 5 0.06

Diabetes Mellitus* 2 0.02 5 0.06

Current smokingt 15 0.17 7 0.08

Family history

VTE, First degree relative < 60 years old 15 0.17 1 0.01

Provoked risk factors† 39 0.44 17 0.19

Trauma 22 0.22 7 0.08

Cast, orthosis, etc. 15 0.17 1 0.01

Surgery 16 0.18 2 0.02

Pregnancy, post-partum 3 months 0 0 0 0

Contraceptives** 11 0.31 9 0.26

Menopausal replacement therapy†† 5 0.14 1 0.03 Where: n, numbers; freq., frequency; SD; standard deviation; * , medical treatment; †, within three month from diagnose or index date; J, daily within the last year; **, estrogen containing (oral, patch, vaginal devices);††, estrogen containing (oral only)

Table 3: Clinical characteristics of FARIVE - verification study

FARIVE

Variables 603 cases 597 controls

n freq. n freq.

Localisation of the thrombosis

Deep vein thrombosis (DVT), lower

limbs 102 0.28

Pulmonary embolism (PE) 147 0.41

DVT and PE 105 0.3

Gender and biometry

Gender; women 363 0.60 340 0.57

Age (years) (mean ± SD and range) 53.1±19.7 (17-91) 51.2±18.4 (18-89) Height (m) (mean ± SD and range) 1.68±0.9 (1.48-2) 1.68±0.9 (1.48-2.100)

BMI (kg/m2) (mean ± SD and range) 26.5±5.6 (15-55) 25.9±6 (15-47

Obese, BMI > 30 kg/m2 135 0.22 114 0.19

Cardiovascular risk factors

Hypertension* 194 0.32 247 0.42

Hyperlipidaemia* 156 0.26 134 0.22

Diabetes Mellitus* 40 0.07 103 0.17

Current smoking† 100 0.17 138 0.24

Family history

VTE, First degree relative 145 0.24 NA

Provoked risk factor s†

Trauma 24 0.04 NA

Cast, orthosis 22 0.04 NA

Surgery 109 0.18 NA

Pregnancy, post-partum 3 months 21 0.03 NA

Hormonal use (women only)

Contraceptives** 146 0.40 NA

Menopausal replacement therapy** 28 0.08 NA

Where: n, numbers; freq., frequency; SD; standard deviation; *, according to medical records or treatment;†, within three month from diagnose or index date; % 4 participants smoked less than 1 cigarettes, per day, 10 missing information; **, not defined

Statistical analyses

Median fluorescent intensity (MFI) values were obtained from the suspension bead array assays readout from the Luminex instrument. In the preprocessing of the proteomic raw data, samples with too low bead counts were excluded; 4 samples in VEBIOS and 4 in FARIVE. Probabilistic quotient method (PQM) was used to normalise the data for sample-by-sample variation (Dieterle et al, 2006). Robust PCA method was applied to identify study-specific outliers leading to the additional exclusion of 14 individuals in FARIVE but none in VEBIO (Hubert et al, 2005). No detectable heterogeneity in the data across the 13 FARIVE centres were observed. Finally, 88 cases and 85 controls were available for analysis in the discovery VEBIOS study and 580 cases and 589 controls at the replication stage. At both steps, the association of proteomic markers with VTE was tested using linear regression analysis while adjusting for age, gender and additionally by center in FARIVE. Log-transformation was applied to proteomic biomarkers to remove any skewness in the proteomic data distribution. The analyses were performed using the R statistical computing software (Ihaka et al, 1996). Immunocapture mass spectrometry (IC-MS) for verification ofHPA011972 antibody target

IC-MS was performed as described by Neiman et al. with some modifications (Neiman M et al, 2013). In brief, 100 μΐ, of pooled citrate plasma from the VEBIOS sample set was diluted, heated at 56°C and incubated overnight with beads covalently coupled to the HPA011972 antibody, or with normal rabbit IgG as a negative control (rlgG). Beads washes and protein digestion were carried using a Trypsin/Lys-C mixture (Promega) at an enzyme-to-protein ratio of 1 :50. Each IC experiment was performed in duplicate. After digestion samples were analyzed using a Q-Exactive HF (Thermo), equipped with an Ultimate 3000 RSLC nanosystem (Dionex). The column used was a 15cm x 75 μιη ID Easy spray column packed with 3μιη CI 8 (#E.800, Thermo) installed on to the nano-electrospray ionization (NSI) source. Samples were re- suspended in 10 μΕ of solvent A (90% water, 5% DMSO, 5% acetonitrile (ACN), 0.1% formic acid (FA)). The injection volume was 5 μΐ, and the flow was set at 0.25 μΕ/ηιίη. Peptide elution was performed with a linear gradient from 3 % to 30% of solvent B (90% ACN, 5% DMSO, 0.1% FA) in 50 min and a subsequent increase to 99% of solvent B in 7 min. The MS analysis was performed in full scan mode (mass-to-charge ratio (m/z) of 300-1600 with a resolution of 60,000, and the top 5 most abundant ions were selected for higher energy collision (HCD). The MS/MS analysis method resolution was 30,000, with an isolation width of 2 m/z, and a dynamic exclusion time of 60 sec. The MS2 spectra were queried against Uniprot complete human proteome database updated 20150521 , and using the target-decoy search strategy, by the engine Sequest under the platform Proteome Discoverer (PD, vl .4.0.339, Thermo Scientific). The following settings were applied: precursor mass tolerance of 10 ppm; product mass tolerances of 0.02 Da for HCD-FTMS; trypsin as selected enzyme, with an allowance of 2 missed cleavages; carbamidomethylation on cysteine (57.021 Da) as fixed modification; and as variable modifications oxidation of methionine (15.995 Da), methylation of lysine (14.016 Da), acetylation of lysine (42.011 Da), and phosphorylation of serine, threonine and tyrosine (79.966 Da). Peptides found at 1% false discovery rate (FDR) were used to infer protein identities. For each IC, proteins identified were considered potential antibody targets if they fulfilled the following criteria: a peptide-spectrum match (PSM) >2; identified in both duplicates; frequency of identification lower than 50% in an internal database containing the most frequent proteins identified by IC-MS in control plasma due to non-specific interaction. This database allowed narrowing the list of proteins for target verification. This database was also used to assign a z- score to each protein. A z-score of >4 was considered as cut-off to claim a protein as specifically enriched. Data were plotted using R Studio (Version 0.99.441, Boston, MA). Verification of PDGFB as the HPAOl 1972 antibody target

From the VEBIOS discovery screen, 11 samples were selected with different PDGFB levels, as measured in median fluorescent intensity (MFI) by the HPAOl 1972 single binder Luminex assay. The samples were measured in parallel on the Luminex^® system with single binder assay (SBA) and sandwich assay, and in a commercially available immunoassay kit system (ELLA by ProteinSimple; SPCKA-PS-000179), which allows for absolute quantification in pg/ml. For the bead based sandwich assay the HPAOl 1972 capture antibody was coupled to magnetic beads and a biotinylated goat anti-human PDGF-BB was used as a detection antibody (RnD DY220, part 840926). For the ELLA assay, all reagents were provided by the manufacturer and compartmentalized in the cartridge. In brief, samples were thawed on ice and spun down for 2 min at 2000 rpm. Plasma samples were diluted 1 : 1 or 1 :10 with SD13 buffer and applied into the ELLA cartridge and incubated with Luminex beads, respectively.

Results

In this study, the inventors employed an affinity proteomics approach to identify proteins associated with VTE risk. From the 408 proteins tested for association in the VEBIOS study, they identified plasma levels for four proteins as significantly associated with the disease after Bonferroni correction (i.e. p <6.6xl0^"5) (Table 4). The strongest association was observed for HIVEP1 targeted by the HPA050724 antibody, showing significantly increased protein levels in VEBIOS patients compared to controls (mean MFI level (log); 6.820 vs 6.734, p= 9.5xl0^~6). Similarly, protein levels of VWF (HPA002082; 6.845 vs 6.678, p= 2.0x l0^"5) and of PDGFB recognized by HPAOl 1972 (8.363 vs 8.285, p = 3. Ox 10^"5) were also observed to be significantly increased in patients compared to controls. Conversely, GPX3 levels were significantly (p = 1.4x l0⁵) lower in patients than in controls (HPA059686; 5.788 vs 5.872). In order to provide additional support for these findings obtained from citrated plasma, we analysed the target protein levels for the four suggested VTE-associated HPA antibodies in EDTA plasma of the same VEBIOS individuals. Strong concordance was observed between citrate and EDTA based measurements with Pearson's correlations of 0.65, 0.69, 0.83, and 0.64 for HIVEP1, GPX3, VWF and PDGFB, respectively. VTE-association p-values derived from EDTA measurements were 5.9 10^"6, 1.9x l0^"3, 2.2x l0^"7 and l .lxlO^"3 for HIVEP1, GPX3, VWF and PDGFB, respectively (Table 4). The inventors then sought to replicate VEBIOS results in citrate plasma samples of 580 VTE cases and 589 controls from the FARIVE study where the four HPA antibodies targeting the 4 proteins were measured using the same affinity proteomics technology. At the p = 0.0125 statistical threshold that corresponds to the Bonferroni threshold correcting for the number of tests performed (i.e. n = 4), two HPA antibodies were replicated (Table 4). The levels of VWF and PDGFB were significantly increased in patients compared to controls, 7.529 vs 7.396 (p = 5.2xl0^"10) and 8.317 vs 8.280 (p = 1.7xl0^"7) as observed in VEBIOS. In Figure 1, results as measured in MFI for PDGFB in the discovery and the replication phase respectively is plotted. A marginal trend (p = 0.036) for a decrease of GPX3 levels in patients, consistent with VEBIOS findings, was also observed but failed to reach the pre-specified statistical threshold. Finally, similarly to what was observed in VEBIOS, HIVEP1 levels tended to increase in cases compared to controls (6.384 vs 6.370, p = 0.160) but the association did not reach significance (Table 4).

Table 4: Main findings¹ at proteins that associated² with VTE risk in the VEBIOS study and their replication in FA IVE

VEBIOS-Citrate VEBIOS-EDTA FARIVE-Citrate

Controls Cases Controls Cases Controls Cases

Binder name protein

N = 85 N = 88 N = 86 N = 87 N = 589 N = 580

6.734 (0.012) 6.820 (0.015) 7.066 (0.015) 7.172 (0.018) 6.370 (0.006) 6.384 (0.006)

HPA050724 HIVEP1

β = +0.089 (0.019); p = 9.5x l0^"6 β = +0.107 (0.023); p = 5.9x l0^"6 β = +0.012 (0.009) ; p = 0.160

5.872 (0.016) 5.788 (0.012) 5.791 (0.021) 5.713 (0.016) 6.699 (0.006) 6.679 (0.007)

HPA059686 GPX3

β = -0.086 (0.019); p = 1.4X 10^"5 β = -0.080 (0.025); p = 1.9x l0^"3 β = -0.018 (0.009); p = 0.036

6.678 (0.025) 6.845 (0.033) 6.586 (0.030) 6.841 (0.043) 7.396 (0.015) 7.529 (0.016)

HPA002082 VWF

β = +0.170 (0.040); p = 2.0x l0^"5 β = -0.260 (0.048); ρ = 2.2^χ 10^"7 β = +0.130 (0.021); p = 5.2x l0-¹⁰

8.285 (0.012) 8.363 (0.016) 8.418 (0.013) 8.480 (0.015) 8.280 (0.005) 8.317 (0.006)

HPA011972 PDGFB

β = +0.082 (0.019); p = 3.0x l0^"5 β = +0.063 (0.019); p = 1.1 x lO^"3 β = +0.037 (0.007); p = 1.7X 10^"7

Where: ¹ Shown values correspond to mean of fluorescence intensity relative to levels of the protein targeted by the HPA antibodies (log) MFI separately in controls and cases with the standard error. The β coefficient correspond to the log levels increase in cases compared to controls, adjusted for age, sex and additionally centre in FARIVE with standard error and ² is the Bonferroni threshold corrected for the number of teste HPA antibodies in VEBIOS (6.6x 10-5 = 0.05/755).

We performed IC-MS to verify that PDGFB was specifically bound by HPAOl 1972. Three peptides from PDGFB were identified. 244 proteins were identified and data were filtered as described in the methods section, using an internal database of the proteins most frequent found in plasma by ICMS, and z-scores were calculated. The highest z-score of 4.9 was obtained for PDGFB confirming that the HPAOl 1972 antibody specifically binds PDGFB in plasma.

To further confirm that PDGFB was specifically bound by HPAOl 1972, we performed a comparative analysis of a subset of VEBIOS plasma using ELLA micro fluidics immunoassay system and MFI measurements from the bead-based Luminex® assays obtained using the HPAOl 1972 as capture antibody either in the single binder assay or in the sandwich assay. Both assays correlated well (p = 0.61; p=0.04 and p = 0.70; p=0.01, respectively) with absolute pg/mL levels of PDGFB in the plasma samples obtained by the microfluidic immunoassay system, which were in agreement with previously reported PDGFB levels in plasma (Polanski M et al. 2007).

To conclude, plasma levels of two protein, VWF and PDGFB, were found robustly associated with VTE in VEBIOS and FARIVE. These two markers were moderately correlated, p = 0.42 and p = 0.26 in VEBIOS and FARIVE, respectively, and remained significantly associated with VTE when both were entered in a joint logistic regression model. For instance, in such a joint model adjusted for age, sex and centre in FARIVE, the significance for association of VWF and PDGFB protein with VTE risk was p < 0.001 and p = 0.002. Further, there was no influence of VKA treatment in subgroup analysis of PDGFB and INR (data not shown).

This present work is the first large scale comprehensive screening for plasma proteomic biomarkers for VTE risk, and includes both a discovery and a replication phase in independent samples. From 755 HPA antibodies assessed in the VEBIOS study, four antibodies targeting WVF, PDGFB, HIVEP1 and GPX3, were selected for replication in the FARIVE study according to stringent statistical criterion. Two antibodies statistically replicated, VWF and PDGFB, and both proteins exhibited increased plasma levels in cases compared to controls. Von Willebrand factor (VWF) was included in our list of candidates because of its well established association with thrombosis. The observed association for VWF with VTE in our study served as a proof of principle and positive control that PDGFB is a bio marker for VTE. REFERENCES

Antoni G, Morange PE, Luo Y, et al. A multi-stage multi-design strategy provides strong evidence that the BAB locus is associated with early-onset venous thromboembolism. J Thromb Haemost. 2010;8(12):2671-2679.

Bystrom S, Ayoglu B, Haggmark A, et al. Affinity proteomic profiling of plasma, cerebrospinal fluid, and brain tissue within multiple sclerosis. Journal of proteome research. 2014;13(11):4607-4619.

Dieterle F, Ross A, Schlotterbeck G, Senn H. Probabilistic quotient normalization as robust method to account for dilution of complex biological mixtures. Application in 1H NMR metabonomics. Analytical chemistry. 2006;78(13):4281-4290.

Drobin K, Nilsson P, Schwenk JM. Highly multiplexed antibody suspension bead arrays for plasma protein profiling. Methods in molecular biology. 2013;1023: 137-145.

Hubert M, Rousseeuw PJ, Vanden Branden K. ROBPCA: a new approach to robust principal components analysis. Techno metrics. 2005 (47): 64-79.

Ihaka R, Gentleman R. R: A Language for Data Analysis and Graphics Journal of Computational and Graphical Statistics. 1996;5 (3):299-314

Leitzel et al, 1991, Cancer Research 51 :4149-4154

Morange PE, Oudot-Mellakh T, Cohen W, et al. KNG1 Ile581Thr and susceptibility to venous thrombosis. Blood. 2011.

Nails MA, Couper DJ, Tanaka T, et al. Multiple loci are associated with white blood cell phenotypes. PLoS genetics. 201 l;7(6):el002113. Neiman M, Fredolini C, Johansson H, et al. Selectivity analysis of single binder assays used in plasma protein profiling. Proteomics. 2013;13(23-24):3406-3410.

Polanski M, Anderson NL. A list of candidate cancer biomarkers for targeted proteomics. Biomarker insights. 2007;1 : 1-48.

Rossi E, Casali B, Regolisti G, Davoli S, Perazzoli F, Negro A, Sani C, Tumiati B, Nicoli D. Increased plasma levels of platelet-derived growth factor (PDGF-BB + PDGF-AB) in patients with never-treated mild essential hypertension. Am J Hypertens. 1998 Oct;l l(10): 1239-43.

Traylor M, Makela KM, Kilarski LL, et al. A novel MMP12 locus is associated with large artery atherosclerotic stroke using a genome-wide age-at-onset informed approach. PLoS genetics. 2014;10(7):el004469.

Tregouet DA, Heath S, Saut N, et al. Common susceptibility alleles are unlikely to contribute as strongly as the FV and ABO loci to VTE risk: results from a GWAS approach. Blood. 2009;113(21):5298-5303.

Uhlen M, Fagerberg L, Hallstrom BM, et al. Proteomics. Tissue-based map of the human proteome. Science. 2015;347(6220): 1260419.

Wang L, Wu Y, Zhou J, et al. Platelet-derived ERp57 mediates platelet incorporation into a growing thrombus by regulation of the alphallbbeta3 integrin. Blood. 2013;122(22):3642-3650.

Zehetner C, Kirchmair R, Neururer SB, Kralinger MT, Bechrakis NE, Kieselbach GF. Systemic upregulation of PDGF-B in patients with neo vascular AMD. Invest Ophthalmol Vis Sci. 2014 Jan 20;55(l):337-44.

Zhu T, Carcaillon L, Martinez I, et al. Association of influenza vaccination with reduced risk of venous thromboembolism. Thrombosis and Haemostasis. 2009;102(6): 1259- 1264.

Claims

1. A method for determining presence or risk of developing a hemostasis disorder in a patient, which method comprises determining presence or concentration of PDGFB in a biological sample of the patient.

2. The method of claim 1, wherein PDGFB is detected as a monomer or in a form bound to another protein.

3. The method of claim 1 or 2, wherein the hemostasis disorder is thrombosis.

4. The method of claim 3, further comprising comparing the PDGFB concentration in the patient with a control value, wherein a higher PDGFB concentration in the patient sample compared to the control value is indicative of presence or risk of developing thrombosis.

5. The method of claims 3 or 4, wherein the thrombosis is venous thrombosis.

6. The method of claim 5, wherein the venous thrombosis is selected from the group consisting of deep vein thrombosis, portal vein thrombosis, renal vein thrombosis, jugular vein thrombosis, Budd-Chiari syndrome, Paget-Schroetter disease and cerebral venous sinus thrombosis.

7. The method of claim 5, wherein the venous thrombosis is phlebitis or pulmonary embolism.

8. The method of claims 3 or 4, wherein the thrombosis is arterial thrombosis, which optionally causes a stroke or a myocardial infarction.

9. The method of claim 8, wherein said patient is at increased risk of developing a thrombus due to a medical condition which disrupts hemostasis, wherein said medical condition is preferably coronary artery disease or atherosclerosis.

10. The method claim 8, wherein said patient is at increased risk of developing a thrombus due to a medical procedure, including cardiac surgery, atherectomy, cardiopulmonary bypass, or catheterization, in particular cardiac catheterization.

11. The method of claim 1 or 2, for determining a risk of bleeding event in a patient.

12. The method of claim 11, further comprising comparing the PDGFB concentration in the patient with a control value, wherein a lower PDGFB concentration in the patient sample compared to the control value is indicative of a risk of bleeding.

13. The method of any of claims 1 to 12, wherein the biological sample is a blood, plasma, or serum sample.

14. The method of any of claims 1 to 13, wherein the patient is human.