WO2013120995A1 - Exosomal biomarker for cardiovascular events in females - Google Patents

Abstract

The invention relates to methods of determining, in respect of a female subject, the risk of developing a cardiovascular event, such as coronary artery disease, by measuring SerpinF2 protein levels in a bodily fluid sample comprising exosomes from said female subject and determining whether said level indicates whether said female subject is at risk of developing a cardiovascular event.

Description

EXOSOMAL BIOMARKER FOR CARDIOVASCULAR EVENTS IN FEMALES

FIELD OF THE INVENTION

The invention relates to the field of medicine and in particular to prognostic methods of assessing the risk of cardiovascular disease and to methods of predicting whether a subject is at risk of developing a cardiovascular disease such as coronary heart disease (CHD) . The invention

furthermore relates to a specific biomarker present in exosomal samples and that is used in the methods of the invention .

BACKGROUND OF THE INVENTION

The Global Burden of Disease Study revealed that cardiovascular diseases are a leading cause of death for women and men worldwide. Ageing of the world population will push up the incidence of cardiovascular disease even higher. Disabling cardiovascular disease can be prevented by optimal treatment when patients at risk can be identified at an early stage. This urgently calls for novel biomarkers that successfully predict risk for cardiovascular disease at an early stage. These new biomarkers are also required to measure the effectiveness of newly developed treatments and to identify those patients at high risk who are eligible for more intensive treatment and cardiovascular screening programs .

Obviously, established cardiovascular risk factors, including dyslipidemia, smoking, hypertension and diabetes mellitus have been incorporated into algorithms for

cardiovascular risk assessment. However, identifying

beforehand what subjects are at risk of developing

cardiovascular disease remains difficult. Although the traditional risk factors may explain part of the risk for cardiovascular disease this will change with the elderly population since the relative risk associated with the established risk factors is diminishing with advancing age. The identification of prognostic biomarkers would be of major added value in recognizing patients who are at risk of suffering future cardiovascular events and who could then be targeted for aggressive preventive measures. For primary cardiovascular events, the prognostic value of currently known biomarkers is very limited since these biomarkers only moderately add to standard risk factors.

A recent report of the EuroHeart project (a 3-year project co-funded by the European Commission Public Health Programme, developed by the European Society of Cardiology and running from 2007-2010) has shown that women are still underrepresented in many cardiovascular clinical trials, while important gender differences are present within most areas of heart disease. As a consequence, scientific

evidence for most established biomarkers and treatment regimen is generally based on cohorts that predominantly contain male subjects. There is sufficient clinical and scientific evidence pointing to major differences in

clinical presentation and disease pathogenesis between males and females. This evidence (summarized below) will support the expressed need for biomarkers that specifically address females.

Notably, cardiovascular disease is the cause of death in just as many males as females. Heart disease is the most common cause of death in women in every major developed country and most emerging economies. The burden of

cardiovascular disease is increasing in middle-aged women relative to men. While cardiovascular death rates are declining in men, this is being stable or even increasing in women. Females and males differ in the presentation of symptoms. In addition, the traditional tests widely used to identify evidence of coronary heart disease are shown to be less sensitive and less specific in women than in men.

In order to target the rising cardiovascular death rates in women, identification and treatment of high-risk women is essential. The under representation of women has led to identification of biomarkers that are specific for men, but show inconsistent results in women, such as C- reactive protein. It is likely that these inconsistent results in the predictive value of promising biomarkers in men and women are due to differences in pathophysiology of cardiovascular disease. To identify patients at highest risk of cardiovascular disease, cardiovascular risk prediction scores, based on the prevalence of general risk factors are currently being used.

However, as mentioned, this approach stems from the knowledge that traditional risk factors for cardiovascular disease are similar in men and women, and across various regions of the world. Yet, for women, up to 20% of all coronary events occur in the absence of the traditional risk factors used in these risk scores. In addition, many women with increased prevalence of risk factors do not even experience events. Finally, it is evident that there are female related risk determinants for cardiovascular disease that are associated with pregnancy such as pre-eclampsia, premature and dysmature pregnancies. The pathophysiology is poorly understood and risk markers for these young female groups at higher risk do not exist.

Atherosclerosis develops over decades and has a

prolonged asymptomatic phase during which it is possible to modify its course. Although there is very limited data available on gender differences, it is known that women have smaller and stiffer macro- and microvasculature, with more diffuse atherosclerosis, and their microvessels appear to be more frequently dysfunctional compared with men. Moreover, coronary plaque erosions are more frequently found in women presenting with an acute coronary syndrome while in men, plaque ruptures are usually responsible for thrombus embolization. In the Athero-Express Study, women who underwent a carotid endarterectomy showed a more stable, less inflammatory plaque compared to men, although their clinical presentation was similar. These gender related differences in etiology may have impact on the predictive value of markers that are being tested and used to identify high risk individuals.

In summary, there is a general need for biomarkers predicting cardiovascular events. Cardiovascular biomarker discoveries have been mainly based on cohorts with a majority of male patients. However, it is evident that cardiovascular disease in female patients presents

differently and there is also sufficient evidence pointing to differences in pathogenic mechanisms in cardiovascular disease between males and females. There is an urgent need for biomarkers predicting cardiovascular disease that is gender specific and generally applies to females.

The ideal approach in the search for biomarkers is an unbiased approach. Novel molecular techniques such as proteomics opened new possibilities for this purpose.

Recently, in the laboratory of the inventors of the present invention this technique was successfully used to discover novel biomarkers for cardiovascular disease in

atherosclerotic plaques (WO 2009/017405). However, using atherosclerotic plaque samples has the disadvantage that it requires invasive procedures. Hence, there is a need for measuring biomarkers in easy to obtain and more patient- friendly sample types such as blood samples. SUMMARY OF THE INVENTION

The present invention relates to a method for

predicting whether a female subject is at risk of developing a cardiovascular event (in the future) , said method

comprising the steps of measuring the SerpinF2 protein level in a sample from said female subject and determining whether the measured SerpinF2 protein level indicates that said female subject is at risk of developing a cardiovascular event. An important aspect of the invention is that said protein concentration is determined in a specific sample type, namely exosomes that appears to be a unique enriched source of proteins and RNA compared to the regular plasma and serum sample type and very suitable for the rapid detection of SerpinF2 levels.

Preferably said exosome sample is a sample from a body fluid such as urine, amniotic fluid, malignant ascites, broncho-alveolar lavage fluid, synovial fluid, breast milk, saliva or blood. In a more preferred embodiment the exosome sample is a plasma sample or a serum sample. The current invention therefore provides, in a particularly preferred embodiment, a simple non-invasive method for measuring

SerpinF2 in the exosomal fraction of plasma and serum samples. The female subject is preferably a woman that is younger or equal to 58-60 years of age.

The subject is preferably an asymptomatic female and may or may not suffer from a risk factor involved in

developing a cardiovascular event such as dyslipidemia, smoking, hypertension, diabetes mellitus, and/or high cholesterol levels. Clearly, the term 'may suffer from' in relation to smoking means that the female subject is a smoker. Preferably, the female subject has not yet

experienced a cardiovascular event and said cardiovascular event is therefore a primary event. In another embodiment of the invention, the female subject may already have

experienced a cardiovascular event, or she may be under treatment for a primary- or secondary cardiovascular event.

In another aspect the invention relates to the use of an exosome sample from a female subject, wherein the

SerpinF2 protein level in said exosome sample indicates whether said female subject is at risk of developing a cardiovascular event. In another embodiment of the

invention, the female subject may already be in the process of experiencing a cardiovascular event.

The invention also relates to a method for predicting the risk of a cardiovascular event, in particular coronary artery disease (CAD) in a female subject, based on the detection in plasma exosome samples and/or other micro- vesicles of smaller or larger size from said subject of SerpinF2.

The invention furthermore relates to SerpinF2 for use in the prognosis of the risk of a female subject developing a cardiovascular event, in particular coronary artery disease .

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows the average SerpinF2 concentration in women and men in the METEOR cohort. These values are

corrected for total exosome protein, age, systolic blood pressure, total cholesterol and high density lipoprotein.

The METEOR trial included asymptomatic females and -males.

The figure shows that the SerpinF2 levels in healthy males are lower compared to healthy females.

Figure 2 shows the frequency distribution of SerpinF2 levels in exosomes of healthy females in the PROSPECT study.

The graph shows the number of individuals for a specific value of SerpinF2. It shows that the SerpinF2 levels are not normally distributed (top panel) , many individuals have low levels of SerpinF2. For statistical analyses normally distributed data are preferred. Therefore a logarithmic transformation of the SerpinF2 data was executed (taking the log of each value) . The data then appear normally

distributed as indicated in the bottom panel. All data are therefore presented after logarithmic transformation.

Figure 3 shows the levels of SerpinF2 in females that were studied in the PROSPECT study. Data are shown for females that suffered from a primary coronary heart disease (CHD) during follow up (cases) and females that did not experience CHD during follow up (controls) . This figure shows the data for all individuals that have been studied (n=786) and also shows that the levels of SerpinF2 were lower in the cases (= patients who suffered from CHD during follow up) compared with the controls. Levels are corrected for age, blood pressure, total cholesterol, and high density lipoprotein and exosome protein concentration. This

correction is very relevant since the biomarker has to be related to CHD independent of known risk factors for CHD. The independent association indicates that the information carried by SerpinF2 has additional value in clinical

practice on top of the traditional risk factors. Thus, the observed association cannot be explained by associations with known cardiovascular risk factors alone.

Figure 4 shows a subgroup analysis of the PROSPECT study. It appears that the difference in SerpinF2 levels between cases and controls was fully explained by the women below the median age of 58 (A) . In women above the age of 58 the difference in SerpinF2 levels between those with CHD and controls was absent. This difference could not be explained by the age when women entered menopause (data not shown) . The age cut-off value (58) was based on the median value of the cohort, and is a relative arbitrary cut-off value.

Levels are corrected for age blood pressure, total

cholesterol, high density lipoprotein and total exosome protein concentration.

DETAILED DESCRIPTION

It is an object of the present invention to provide a method for predicting the risk of a female subject that may develop a (primary) cardiovascular event wherein said method largely circumvents the necessity of using invasive-, or patient unfriendly procedures.

The present invention relates to a method for

predicting (or assessing) whether a female subject is at risk of developing a cardiovascular event, said method comprising the steps of: isolating exosomes from a body sample from said female subject; measuring the SerpinF2 protein level in said exosome sample; and determining whether the measured SerpinF2 protein level indicates that said female subject is at risk of developing a

cardiovascular event. In a preferred embodiment, said body sample is urine, amniotic fluid, malignant ascites, broncho- alveolar lavage fluid, synovial fluid, breast milk, saliva or blood, preferably plasma or serum. In another preferred embodiment, said female subject is a woman that is younger or equal to 58-60 years of age. Although the cardiovascular event may be a secondary event, said cardiovascular event for which the risk is assessed is preferably a primary event. The female subject for whom the risk is assessed may be completely asymptomatic as far as cardiovascular events are concerned. However, in another preferred embodiment the methods of the invention are performed on exosome samples from a body fluid derived from a female subject that is a smoker or that is suffering from another risk factor involved in developing a cardiovascular event such as dyslipidemia, hypertension, diabetes mellitus, and/or high cholesterol levels. The female subject may also be suffering from female specific risk indicators such as polycystic ovarian syndrome or pregnancy hypertension.

The invention further relates to methods according to the invention in which the concentration (log transformed) level of SerpinF2 protein is measured in exosome samples from a female subject and wherein such concentrations will be indicative for the female subject for being at risk of developing a (primary) cardiovascular event.

The present invention also relates to the use of an exosome sample from a female subject for predicting

(assessing) whether a female subject is at risk of

developing a cardiovascular event, wherein the SerpinF2 protein level in said exosome sample indicates whether said female subject is at risk of developing of a cardiovascular event. The invention also relates to SerpinF2 for use in the prognosis of the risk of a female subject developing a

(primary) cardiovascular event, in particular coronary artery disease.

In search for gender-specific markers the inventors of the present invention at first performed histopathological analyses on human plaque samples. Initially, plaque markers were observed that are only predictive in males and not in females. This strengthened the realization that the

pathogenesis of atherosclerosis progression may differ between males and females and subsequently that biomarkers may reveal gender related differences. Prentice et al .

performed similar studies on plasma samples and determined the presence of a set of new proteins that are associated with risk of coronary heart disease among postmenopausal women (Prentice R.L. et al . (2010) Novel proteins associated with risk for coronary heart disease or stroke among postmenopausal women identified by in-depth plasma proteome profiling. Genome Med 2(7) :48) . Unfortunately, the procedure of using plasma samples is hampered by the presence of high- abundant plasma proteins such as albumin and immunoglobulins that complicate the detection of potentially interesting low-abundant proteins. The investigations on the plasma samples from which the exosomes were removed, gave mixed results and appeared non-conclusive (see WO 2012/110253).

In a next step therefore, the inventors of the present invention investigated plasma sub-fractions for the presence of proteins that could potentially have predictive value for cardiovascular events. It was then found that especially when the exosome fraction of the samples was studied, protein constitution in such samples from subjects who had suffered a cardiovascular event following the moment of sampling, in fact differed from that in patients who had not suffered such a cardiovascular event, and then it was realized that this difference could be used for prognostic research on non-symptomatic subjects that could then be identified as being at risk of developing a primary

cardiovascular event.

Protein secretion out of the cells can occur directly after production (constitutive pathway) or through a pathway in which the protein is first stored in the cell and

released after a trigger (regulatory pathway) . Secretion not only occurs with individual proteins but also occurs via vesicles containing a large number of proteins and RNA.

These vesicles are formed with a selection of lipids, protein and RNA from the secreting cell and are released as an intact vesicle, generally in the size of about 20 to about 500 nm. Vesicles in the size of about 50 to about 100 nm are referred to as ^,exosomes' and the release of exosomes has been described for various cell types, including

reticulocytes, B- and T-lymphocytes , dendritic cells, mast cells, platelets, macrophages and alveolar lung cells.

In several cell types, including T cells, platelets, dendritic cells and mast cells, secretion of exosomes is regulated by specific stimuli. While early studies focused on their secretion from diverse cell types in vitro,

exosomes have now been identified in body fluids such as urine, amniotic fluid, malignant ascites, broncho-alveolar lavage fluid, synovial fluid, breast milk, saliva and blood.

Exosomes appear to have a wide range of biological functions in immune response, antigen presentation,

intracellular communication and the transfer of RNA and proteins. The inventors of the present invention realized that since exosomes express an array of proteins that reflect the originating host cell, that exosomes contain valuable information regarding ongoing (patho-) physiologic processes in the human body, including information of future cardiovascular events and that isolating exosomes and use these as the preferred sample in a prognosis method would enable one to gain more specific knowledge than when a non- exosome enriched sample was used.

An ischemic event immediately activates endothelial cells that attract platelets that also become activated. This activation of endothelial cells and platelets was found to be accompanied by the release of microvesicles that are secreted into the blood. At the point in time that an ischemic coronary event occurs, microvesicles are secreted from several cells and tissues, including the myocardium. Apoptosis of cardiomyocytes occurs almost instantly after occluding the coronary artery and subsequent ischemia. Such apoptotic cardiomyocytes secrete vesicles in the blood, generally referred to as apoptotic bodies. The inventors contemplated that since secretion not only occurs with individual proteins such as troponin but also via such vesicles containing a large number of proteins and RNA, the components found in, on or attached to the vesicles could also be used as biomarkers for cardiovascular disease.

The present invention is based on the finding that particular proteins associated with such microvesicles , and in particular exosomes, appeared as highly suitable

biomarkers to assess the risk of developing a cardiovascular event, in particular coronary artery disease in females.

The invention thus provides a method of predicting the risk of a female subject developing a cardiovascular event, in particular CHD, comprising detecting a biomarker in an exosome sample or other micro-vesicles of smaller or larger size from said subject, wherein said biomarker is SerpinF2 (IPI : IPI00879231, SWISSPROT : A2AP_HUMAN) . The IPI numbers refer to the International Protein Index

(www.ebi.ac.uk/IPI), as indexed on December 4, 2010 followed by SWISSPROT database entry name as indexed on November 30, 2010. SerpinF2 refers to the name of the gene, from which the protein product is being identified by the methods as outlined in the Example section. The skilled person is aware of the fact that the protein name (from the SerpinF2 gene) comes in different versions: alpha 2 antiplasmin, A2AP, AAP, ALPHA-2-PI, API and Alpha-2 plasmin inhibitor. For

consistency herein, the name of the gene is used. Notably, the protein itself (or a fragment thereof that can be detected) serves as the biomarker. The referenced index number (database accessions) as used herein includes

reference to fragments, isoforms and modifications thereof, hence the present invention foresees the use of fragments of the protein as well as modifications and derivatives of the protein disclosed herein as a biomarker in the context of the various aspects of the present invention.

The present invention therefore relates to a method for assessing whether a female subject is at risk of developing a cardiovascular event, said method comprising the steps of: isolating exosomes from a body sample from said female subject; measuring the SerpinF2 protein level in said exosome sample; and determining whether the measured

SerpinF2 protein level indicates that said female subject is at risk of developing a cardiovascular event. The assessment whether a female subject is at risk of developing a

cardiovascular event is performed in vitro: the isolation of exosomes from a body sample and the measurement of the

SerpinF2 protein level in said exosome sample is performed in vitro.

The invention is particularly useful for predicting the risk of a female subject developing coronary artery disease (CAD) . CAD is the result of the accumulation of atheromatous plaques within the walls of the coronary arteries. The formation of plaque in the lumen of an artery causes

narrowing of the lumen, which may end up in a total

occlusion. CAD is sometimes also referred to as coronary heart disease (CHD) .

The term ^,exosome sample' preferably refers to a sample of isolated exosomes but may also refer to a general sample of any body fluid, preferably serum or plasma, wherein said sample comprises exosomes (= body fluid samples that contain higher concentrations of exosomes than they would normally contain) , and that optionally further comprises

microvesicles that are smaller or larger than the defined exosome size (about 50 to about 100 nm) . Generally, exosomes are isolated as outlined in the example section below. Exosomes are a subgroup of the more general term microvesicles . Exosomes are vesicles with a particular size range. In this application, the term ^,exosome sample' refers to a sample that contains exosomes that have been purified to a certain extent from the originating body fluid sample. Such exosome samples may contain intact exosomes and/or lysed exosomes wherein the lysis is preferably induced by external process steps (such as the process as outlined in the example section) . As stated above, the exosome sample may contain microvesicles and vesicles that have a

(slightly) different size as compared to the exosomes.

A biomarker according to the present invention is also generally identified in the art as a ^''profile' or as a

''protein profile' . A (protein) profile according to the present invention preferably relates to SerpinF2.

In a highly preferred embodiment of the invention, the biomarker protein (or a peptide fragment thereof) is detected in an exosome sample (wherein said sample may comprise other vesicles that may be somewhat larger or smaller in size than a regular exosome) . The exosome samples may be derived from different types of body fluids such as serum, plasma or blood. Alternatively, exosomes samples from other body fluids such as urine, amniotic fluid, malignant ascites, broncho-alveolar lavage fluid, synovial fluid, breast milk and saliva can be used. The fact that exosomes from other body fluids than blood, plasma or serum could have strong predictive value for cardiovascular disease is known to the skilled person (Cheng et al . (2012) A

translation study of urine miRNAs in acute myocardial infarction. J Mol Cell Cardiol 53: 668-676). Also the company Exosome Diagnostics Inc. demonstrated the presence of two known prostate cancer biomarkers, PCA-3 and TMPRSS2- ERG in exosomes isolated from urine of patients, showing the potential for diagnosis of prostate cancer patients (Nilsson J. et al . (2009) Prostate cancer-derived urine exosomes : a novel approach to biomarkers for prostate cancer. Br J

Cancer 100:1603-1607).

In a preferred embodiment according to the present invention, the biomarker protein or a peptide fragment thereof is detected in serum- or plasma exosome samples. The biomarker that is detected may be freely present in the exosome, or attached to-, anchored in- or adhered to the exosomes. When the biomarker is extracted from the exosomes for measurement, the proteins may have been present freely or attached to-, anchored in- or adhered to the exosomes before extraction took place. In all embodiments of the invention the biomarker protein is SerpinF2.

In the methods and uses of the different embodiments of the present invention, the cardiovascular event to be predicted is preferably selected from, but is not limited to the following conditions: vascular death or sudden death, fatal or non fatal stroke, fatal or non fatal myocardial infarction, fatal or non fatal rupture of an abdominal aortic aneurysm (which is considered an atherosclerosis equivalent) , rupture of abdominal aortic aneurysm confirmed by laparatomy, cardiovascular intervention, coronary artery disease, transient ischemic attack (TIA) , peripheral

arterial disease, acute coronary syndrome, heart failure and re-stenosis of carotid, coronary, femoral or other arteries. The cardiovascular event is most preferably CAD.

The method of the present invention may suitably be used for risk stratification and/or patient selection (such as for clinical trials) , for monitoring disease and the development thereof and for monitoring the treatment of disease and/or the effect of (possibly different types of) treatment. The biomarker may be used as clinical marker for safety and efficacy studies (e.g. as surrogate endpoint marker) . The invention also relates to SerpinF2 for use in prognosis and assessing the risk of a (human) subject, in particular a female subject, which may or may not yet have been identified as a patient, in developing a cardiovascular event, in particular CAD. In a further embodiment, the

(protein) profile relates to two or more biomarkers, in which on top of SerpinF2, an additional second, or third, or further protein is selected from the group consisting of Vitronectin, CD14, Cystatin C, Plasminogen, Nidogen 2 and SerpinGl .

The cardiovascular event is preferably a primary event in a (human) female subject that has not yet suffered from a cardiovascular event. In another embodiment, the event may be a secondary or further event occurring in a (human) female subject already having suffered from one or more earlier events. In yet another aspect, the invention relates to a method in discriminating between patients that already had a cardiovascular event and are at risk of suffering an additional event and patients who had such an event and do not have an increased risk of suffering a further event. The event is preferably CAD.

In methods of the present invention the prognosis is performed by using exosomes as the sample type and by measuring the concentration level of SerpinF2 in such samples. Surprisingly, the inventors of the present

invention found that the presence of, and especially the specific level of SerpinF2 in exosome samples from (human) subjects appeared predictive of the risk of developing cardiovascular disease, specifically in female subjects.

The present invention relates to a method for

predicting the risk of a cardiovascular event, in particular coronary artery disease, in a female (human) subject, in which the method is based on the detection of SerpinF2 levels in exosome samples from said subject. In one

embodiment, the invention provides a method of predicting the risk of a female subject developing a cardiovascular event, in particular coronary artery disease, comprising detecting the biomarker SerpinF2 in an exosome sample from said female subject. The method according to the present invention can be used for primary prevention in females.

The skilled person understands that instead of

detecting the complete biomarker protein, one may also detect peptide fragments of said biomarker proteins which are derived from the biomarker proteins by fragmentation thereof. The term peptide fragment as used herein refers to peptides having between and including 5 and 50 amino acids. These peptide fragments preferably provide a unique amino acid sequence of the protein, and are associated with the cardiovascular events as disclosed herein, in particular CAD. The proteins and/or peptide fragment may optionally be detected as chemically modified proteins and/or peptides. Such chemical modification may for instance be selected from the group consisting of glycosylation, oxidation,

(permanent) phosphorylation, reduction, myristylation, sulfation, acylation, acetylation, ADP-ribosylation, amidation, hydroxylation, iodination, and methylation. A large number of possible protein modifications are described in the RESID database at http://www.ebi.ac.uk/RESID (release 2 December 2010; Garavelli J.S. (2004) The RESID Database of Protein Modifications as a resource and annotation tool. Proteomics 4:1527-1533; Farriol-Mathis N. et al . (2004) Annotation of post-translational modifications in the Swiss- Prot knowledge base. Proteomics 4 ( 6) : 1537-1550 ) . The skilled person is well aware of such protein modifications. In general, biomarkers are found inside exosomes but they may also be physically connected or linked to exosomes, which means both in and on their surface. When on their surface the biomarker can be either attached to the

membrane, e.g. expressed on or in the membrane surface or anchored therein, or be in loose connection therewith, i.e. adhered to the exosome without being physically attached to or in the membrane. The biomarkers may also be part of the membrane. The biomarker SerpinF2 has been associated with the membrane but it is unclear with what mechanism it is attached .

The inventors found that biomarkers that are attached, anchored or adhered to the exosome can also be detected in samples of body fluid, in particular in serum or plasma. In a preferred embodiment of the invention, SerpinF2 or a peptide fragment thereof is detected in, on or attached to exosomes from serum, plasma or blood.

The invention further relates to a kit for performing any one of the methods disclosed herein, wherein the kit comprises means for detecting the presence of the biomarker as defined above, in particular SerpinF2. The means for detecting the presence of the biomarker may be multifold and are preferably antibodies, antibody fragments or antibody derivates or the biomarker (s) can be detected via mass spectrometry and flow cytometry. The antibody-based

detection means optionally comprise a detectable label. The person skilled in the art of protein detection is well aware of the wide array of possible applications using antibodies, antibody fragments or antibody derivatives to detect

proteins and determine and measure protein levels in a sample .

The kit according to the invention is intended for use in a method of predicting the risk of a female subject developing a cardiovascular disease, in particular coronary artery disease, by determining the presence of the biomarker SerpinF2 in or on exosomes of the female subject. In

addition, means for detecting other biomarkers may be included. The kit may further comprise reagents and/or instructions for using the means for detecting the biomarker in any such method.

The methods of the present invention may be applied in the risk assessment for primary- as well as for secondary cardiovascular events. In one embodiment, the invention is used to do a precautionary risk assessment in females that did not previously have a cardiovascular event and do not suffer from risk factors. In another embodiment, the methods of the present invention relate to a female subject that either: a) suffers from a risk factor involved in developing a cardiovascular event such as dyslipidemia, smoking (= said female subject is a smoker), hypertension, diabetes

mellitus, and/or high cholesterol levels or female specific risk indicators such as polycystic ovarian syndrome or pregnancy hypertension (including pre-eclampsia) but has not yet experienced a cardiovascular event (which would then be indicative for the risk of a primary event) , b) that does or does not suffer from a risk factor involved in developing a cardiovascular event and/or has already experienced a cardiovascular event (which would then be indicative for a secondary or further event) , or c) that is being treated for a cardiovascular event.

The levels of the SerpinF2 protein in the exosome samples from the female subjects are indicative for

assessing the risk of developing a cardiovascular event, in particular CAD. At a younger age (<60 years old), a log transformed level of SerpinF2 between 11.2 and 11.8 pg/ml, preferably between 11.3 and 11.7 pg/ml, more preferably between 11.4 and 11.6 pg/ml and most preferably about 11.5 pg/ml is found to be predictive. All values given are after total protein correction. Figure 4 shows that a log

transformed level of SerpinF2 that is lower than 11.6 pg/ml indicates that said female subject is at risk of developing a cardiovascular event. Hence, in another preferred

embodiment, the invention provides a method according to the invention wherein a SerpinF2 concentration in an exosome sample of a woman younger than approximately 60 years (= younger than 58-60 years) is indicative for an upcoming primary cardiovascular event (whenever that might occur in the rest of the life of that individual) when it is lower than 11.8 pg/ml, preferably lower than 11.7 pg/ml, more preferably lower than 11.6 pg/ml, and most preferably lower than 11.5 pg/ml. It should be noted - and it will be

appreciated by the skilled person - that the concentration values as provided herein were obtained using a particular detection method (LUMINEX) and that by using a different method (for example ELISA or Mass Spectrometry) for

measuring protein levels other values could possibly be the outcome. Hence, the concentration level of 11.6 pg/ml that is the approximate and most preferred cut-off value for determining whether a risk of CHD in women exists or not (as disclosed by the present invention) might depend on the detection method that is applied. In a preferred embodiment the LUMINEX method as outlined in the Examples section is used to determine the SerpinF2 protein level in exosome samples. This potential difference of detection method vis^¬ a-vis the SerpinF2 level does not influence the strong predictive value of the protein concentration as provided by the present invention.

It is concluded that SerpinF2 is a very useful and strong new marker that can now be used in the prognosis of cardiovascular events next to the generally known risk factors such as age, blood pressure, HDL cholesterol, total cholesterol, current smoking status and use of blood

pressure lowering medication. Preferably, the SerpinF2 levels should be corrected for protein amount that is derived during the exosome isolations from a fixed amount of plasma. As expected, a higher amount of protein derived from a fixed amount of plasma will result, on average, in a higher amount of SerpinF2 per unit plasma. Therefore

correction for exosome derived protein concentration in the fixed amount of plasma is preferred. The values of lower than 11.6 and 11.5 pg/ml are after total protein correction.

The methods of the present invention may be performed using different means of measuring the level of SerpinF2 in an exosome sample from the female subject. A preferred method for measuring the SerpinF2 level is by applying

SerpinF2 binding moieties, preferably antibodies. The person skilled in the art is aware of the power of using antibodies in assays such as ELISAs in determining protein

concentrations. Antibodies may come in a wide variety of forms, such antibody fragments, and derivatives. Antibodies may be chimeric, humanized, human, or be derived from other suitable sources such as mice, rats and rabbits. Preferably, said SerpinF2 protein level in said exosome sample is determined by using an antibody, or fragment or derivative thereof that interacts with said SerpinF2 protein and wherein said interaction is indicative of the concentration of SerpinF2 in said sample. Other concentration detection methods known to the skilled person may be used, such as mass spectrometry or a different antibody-based assay such as an ELI SA kit.

The inventors of the present invention found earlier that a certain set of exosomal proteins are useful in predicting the risk of cardiovascular events (disclosed in WO 2011/083145 and WO 2012/110253) in both male and female subjects. Although not specified for gender, it was found that the level of Vitronectin, CD14, SerpinF2, Cystatin C, Plasminogen, Nidogen 2 and/or SerpinGl in exosome samples could provide further indications of developing (primary) cardiovascular events. Hence, in a preferred embodiment of the present invention the level of one or more of these additional proteins (besides SerpinF2) is determined and used in concert with the SerpinF2 level to address the risk of developing a (primary or secondary) cardiovascular event, in particular CAD.

The present invention also relates to methods wherein an additional step relates to obtaining a blood sample from said female subject and more preferably obtaining an exosome sample from said blood sample.

The present invention also relates to the use of an exosome sample from a female subject, wherein the SerpinF2 protein level in said exosome sample indicates whether said female subject is at risk of developing a cardiovascular event. The present invention also relates to methods for predicting the risk of a cardiovascular event, in particular coronary artery disease, in a female subject, based on the detection of SerpinF2 in plasma exosome samples and/or other micro-vesicles of smaller or larger size from said subject.

As shown in Figure 4, the predictive value of SerpinF2 for the occurrence of events is particularly suitable for testing females (compared with males) below the median age of the PROSPECT cohort (58 years) that was used in the experiment. Thus, this invention is typically applicable in females at an age where females are still at intermediate or low risk to develop cardiovascular disease in the next 10 years of their lives. The invention therefore is applicable for females at intermediate and low risk following the mostly used risk charts such as the Framingham Risk Score.

Females that according to the Framingham Risk Score are at high risk are not the right target population since these women are already optimally treated. It is however relevant to have a means to predict the risk of developing

cardiovascular disease in females that are classified to be at intermediate and low risk. The mostly used charts such as the Framingham Risk Score show that the risk typically shifts from intermediate and low to high between ages 58-60. As shown herein, the inventors found that the predictive value of the SerpinF2 concentration in the exosomes samples taken from female subjects was strongest in females that were younger than the age of 58-60 years. Hence, the methods of the present invention are preferably performed on exosome samples that are derived from female subjects that are younger than- or equal to the age of 58, 59 or 60.

According to another aspect thereof the invention provides a method for predicting whether a female subject is at risk of developing a cardiovascular event, said method comprising the steps of: obtaining an exosome sample from a female subject; measuring the SerpinF2 protein level in said exosome sample from said female subject; and determining whether the measured SerpinF2 protein level indicates that said female subject is at risk of developing a

cardiovascular event.

The present invention will be further illustrated in the Examples that follow and that are not intended to limit the invention in any way. EXAMPLES

EXAMPLE 1 SerpinF2 levels differ between asymptomatic men and women

In earlier experiments it was found that the level or certain proteins might be indicative for the future

appearance of cardiovascular disease (WO 2011/083145 and WO 2012/110253). Notably, the use of exosome samples appeared to be particularly useful because SerpinF2 levels in general plasma samples gave mixed results whereas the SerpinF2 levels in the supernatant of plasma samples from which exosomes and other microvesicles were removed demonstrated a loss of predictive value.

SerpinF2 levels were initially studied in a group of healthy individuals that were enrolled in the so-called

'METEOR' trial (individuals from the US and the EU) . This trial was set up to study the effect of statin therapy on atherosclerosis via monitoring intima-media thickness (IMT). The individuals were middle-aged with moderately elevated cholesterol and low risk of Cardiovascular Disease (CVD) .

Exosome sample preparation

Exosomes were isolated from the plasma samples by centrifuging the plasma samples at 3000 x g for 15 min at RT prior to use. A pre-treated 0.45 ym filter with 100 μΐ of pre-heated MQ water at 37 °C was prepared, centrifuged at 10000 x g for 2 min at RT, and transferred into a new empty filter tube. The plasma was added into the pre-treated filter followed by centrifugation at 12000 x g for 10 min at RT . 250 μΐ of filtered plasma was taken and mixed thoroughly with 63 μΐ of ExoQuick solution for an overnight incubation at 4°C. The next morning, the precipitated exosomes were collected as a pellet using centrifugation at 1500 x g for 30 min at R . After removing the supernatant, the pellet was centrifugated again at 1500 x g for 5 min at RT to remove the remaining supernatant.

To isolate protein, the pellet containing exosomes was resuspended in Roche Complete Lysis-M buffer containing protease inhibitors (EDTA-free) and incubated for 30 min at RT . To lyse a pellet derived from 250 μΐ of plasma, 100 μΐ of Roche Complete Lysis-M buffer was used. To help with the lysis, the pellet was pipetted up and down for a couple of times and continued to incubate for another 10 min at RT . A pre-treated 0.22 μιη filter with 50 μΐ of Roche Complete Lysis-M buffer was prepared, centrifuged at 10000 x g for 2 min at RT, and transferred into a new empty filter tube. The suspension was added to this pre-treated filter followed by centrifugation at 15000 x g for 10 min at RT . The solution was collected after filtration and stored in 20 μΐ aliquots at -80°C.

SerpinF2 concentration measurement

SerpinF2 protein concentrations in the exosome samples were determined on Luminex, an assay known to the person skilled in the art (Baker H.N. et al . (2012) Conversion of a capture ELISA to a Luminex xMAP assay using a multiplex antibody screening method. J Visualized Experiments:

www . j ove . com/video/4084 / ; Wang L-S . et al . (2012) Comparison of xMAP and ELISA assays for detecting cerebrospinal fluid biomarkers of Alzheimer's Disease. J Alzheimer's Disease 31:439-445). For this, purified recombinant SerpinF2 protein (R&D Systems cat . # 1470-PI) and the following two antisera were used: a mouse monoclonal IgG2A antibody (R&D Systems cat . # MAB1470) and a biotinylated polyclonal goat IgG antibody (R&D Systems cat . # BAF1470). For the Luminex assay, SerpinF2 capture antibody was coupled to MapgPlex Microspheres (beads) using EDC and

Sulfo-NHS solutions from Pierce. In a 96-well plate, 50 μΐ sample and approximately 2500 SerpinF2 antibody-coated beads were added to each well. The mixture was further incubated at RT for 45 min in order for the analyte (SerpinF2 present in the sample) to bind specifically to the beads. The unbound materials were removed by washing the plate with Bio-Plex wash buffer from Bio-Rad on a 96-well magnetic washer. In the next step, biotinylated-detection antibodies were added into each well. These specifically bind the target analytes that were captured on the beads. Another wash step was performed to remove unbound biotinylated antibodies on a 96-well magnetic washer using Bio-Rad' s Bio- Plex wash buffer. Next, Streptavidin-Phycoerythrin (PE) (BD PharMingen) was used to detect and bind to the biotin- labeled antibodies. The excess Streptavidin-PE was removed by washing the plate on a 96-well magnetic washer using the same wash buffer from Bio-Rad. Finally, Strep-PE signals per well was measured on the Luminex system.

In the final exosome sample total protein content was measured using a standard bicinchoninic acid assay (BCA) method. The BCA is a biochemical procedure widely used to determine the total protein concentration in a solution. The total protein concentration is shown by a color change of sample solution from green to purple in proportion to protein concentration that can be measured using a

spectrophotometer .

Results

The levels in exosome samples from all individuals baseline (when drug therapy was not started yet) were compared to check for gender-specific differences. This gender-specific analysis showed that in this group of healthy asymptomatic people SerpinF2 concentrations in exosome samples were consistently lower in males compared with females. Figure 1 shows the difference between the SerpinF2 levels in these METEOR samples between women and men. The conclusion is that in healthy males SerpinF2 levels in exosome samples are lower than in healthy females. Since the high value of SerpinF2 is specifically predictive for a beneficial outcome in females at low or intermediate risk (see below) , this observation is relevant since it shows that the high levels in younger females are not specific for individuals included in the cohort.

EXAMPLE 2 SerpinF2 protein levels in exosome samples from male and female subjects in relation to intima media thickness (IMT) .

Notably, intima media thickness (IMT) of the carotid artery (measured with ultrasound) has been widely used as a surrogate for the presence of systemic atherosclerotic disease. Average values of the thickness of the intima-media complex strongly depend on age and risk factors, yet a recent report shows that the IMT in the general population of 58 years is 0.73 mm (standard deviation 0.16) . With every increase in IMT of 0.10 mm, the risk for stroke and

myocardial infarction increased with -10% (Den Ruijter H.M. et al . (2012) Common carotid intima-media thickness

measurements in cardiovascular risk prediction : a metaanalysis. JAMA 308(8) :796- 803).

The inventors of the present invention wondered whether the SerpinF2 level in exosome samples of patients who suffered from any manifestation of cardiovascular disease, diabetes or hypercholesterolemia would show some relation to the IMT of these patients. The hypothesis was that the value of SerpinF2 as a biomarker would be supported when a

correlation existed with an established imaging marker for atherosclerotic disease, in this case being IM .

For this, an experiment was set up with a total of 1031 samples (820 men and 211 women) from the Second

Manifestations of ARTerial disease (SMART) Study. The SMART study had the objective to investigate brain changes on MRI in patients with symptomatic atherosclerotic disease.

Between May 2001 and December 2005, 1309 patients newly referred to the University Medical Center Utrecht (UMCU, the Netherlands) with manifest coronary artery disease, cerebral vascular disease, peripheral arterial disease or an

abdominal aortic aneurysm, and without magnetic resonance (MR) contraindications were included. During a single day visit to the UMCU, an MRI of the brain was performed in addition to a physical examination. Also, the IMT of the carotid artery was determined by ultrasound using methods known to the person skilled in the art. Blood and urine samples were taken. As part of the SMART study, risk

factors, medical history, and functioning were assessed with questionnaires that the patients filled in before their visit to the medical center. All cohort members were

followed for clinical cardiovascular events for a minimum of three years. At the time of the present experiment, plasma was still available of 1060 patients for biomarker analysis. The SMART study and SMART-MR study were approved by the ethics committee of the UMCU and written informed consent was obtained from all participants.

Exosome samples were prepared and SerpinF2 protein concentrations were determined as outlined in Example 1. Results

The SerpinF2 levels were categorized in quartiles (Ql, Q2, Q3 and Q4, wherein Ql is the group with the lowest concentration of SerpinF2 and Q4 has the highest

concentration of SerpinF2) for each stratum (on the one hand females and males younger than 60 years (<60); on the other hand females and males older or equal to 60 years (=> 60)) and then studied in respect of its relation with IMT. 60 years might appear as an arbitrary cut off, but 60 is an age where many females enter a high risk category (more than 20% risk suffering from a cardiovascular event the coming 10 years) .

The IMT as measured in the male subjects did not have any relation to the SerpinF2 levels in the exosome samples in this cohort (data not shown) . Surprisingly however, SerpinF2 levels in exosome samples did appear to be

indicative for the occurrence of cardiovascular disease in female subjects: The IMT was significantly lower in females younger than 60 years when SerpinF2 levels were higher than the median value (in Q4) . This observation was not observed in the other 3 groups. This indicates that specifically in the group of females below 60, atherosclerotic disease is less in those with higher SerpinF2 levels. This provides a strong indication that SerpinF2 levels are associated with cardiovascular disease, especially in women younger than 60 years. Table I shows the average IMT measurements in the female patients younger than 60 years (left four quartiles Q1-Q4) and older or equal to 60 years (right four quartiles Q1-Q4) . Age <60 Age =>60

Table I. Values of the data from the women in the SMART cohort experiment. Number = number of female subjects in each quartile; IMT = intima media thickness (in mm); * p=0.023, indicating a significant difference for IMT between the groups Ql, Q2, Q3 and Q4 in the females younger than 60.

The overall conclusion from this experiment is that higher values of SerpinF2 in the exosome samples (quartile Q4) are indeed associated with a lower IMT in women younger than 60 years, which means: a lower risk for Cardiovascular Disease at a later stage in life.

EXAMPLE 3 SerpinF2 as the hiomarker of choice for predicting cardiovascular events in asymptomatic women

In a next experiment, it was questioned whether the association between SerpinF2 levels in exosome samples and atherosclerosis could be translated to coronary heart disease (CHD) and whether lower SerpinF2 levels predict long-term CHD risk in asymptomatic women in which risk prediction urgently needs improvement.

To study this, samples were used of a primary

prevention cohort: ^Λ PROSPECT" . This is a cohort study among 17,357 women, aged 49-70 (mean age 57 years) that were recruited during a national breast cancer screening between 1992 and 1997 in the Utrecht area (the Netherlands) . The study followed the women in time for cardiovascular disease. Importantly, at recruitment, these women were free from cardiovascular disease. All participants consented to participate in the study and the follow-up. Each participant filled out a general questionnaire. Participants were also physically examined. 97.5% of the participants donated a blood sample. The women were followed at regular intervals for the occurrence of adverse clinical events such as breast cancer but also cardiovascular events.

During the follow-up period approximately 2,000

participants experienced one or more CHD events. From these CHD cases, 230 subjects were selected to determine the level of biomarker proteins in exosome samples derived from their donated blood. 600 women from the complete cohort at

baseline were randomly selected as controls. This particular study is referred to as a 'case-cohort study' . In a case cohort study a random sample is selected from the entire source population (that is including all controls and cases) . If that sample is unbiased (sampling done

independently from exposure status and disregarding sampling variation) the distribution of exposed and unexposed persons in the sample is expected to reflect the exposure

distribution in the source population at the beginning of the cohort. This is an important aspect of case-cohort studies. The sample should be representative of the

population giving rise to cases (the source population) regarding exposure.

Non-fasting blood was used to determine the expression level of four proteins: CD14, SerpinF2, SerpinGl and

Cystatin C. Exosomes were isolated out of plasma samples as outlined in Example 1. Similar as what was outlined in examples 1 and 2, the concentration of protein was measured in these exosome samples to study the relation of that concentration to cardiovascular events in asymptomatic females. Total protein content was used to correct the protein data of the identical samples.

Results

After correction for total protein content, SerpinF2 showed a significant difference (p<0.05) between patients with CHD and controls. Unlike CD14, Cystatin C, and

SerpinGl, SerpinF2 held its significant difference in women between events and controls after correction for risk factors. CD14, SerpinGl and Cystatin C appeared not to be significantly associated with the occurrence of coronary heart disease events during follow up in the women who were included (data not shown) . For baseline data see Table II. Baseline SerpinF2 levels were measured using multiplex antibody based assays, as outlined above.

Table II Baseline values of PROSPECT cohort. CHD = Coronary Heart Disease; sd = standard deviation; BP = Blood Pressure; BMI = Body Mass Index . The distribution of data is shown in Figure 2. Because the levels were not normally distributed, a logarithmic transformation of the exosomal SerpinF2 data was performed.

Logistic regression models were used to determine the associations between the markers and the occurrence of CHD. Logarithmic transformation is executed because SerpinF2 data are not normally distributed in the cohort while the

logarithmic levels are. Due to the logarithmic scale the differences in numbers seem small.

In this PROSPECT cohort, it was found that the SerpinF2 protein levels were significantly associated with coronary heart disease (β=-0.183, p=0.038). This association

indicates that in the event that a female has an increased SerpinF2 concentration, the chance of getting CHD during follow-up is reduced and that low levels of SerpinF2 are associated with increased risk.

This association remained present after correction for traditional risk factors such as cholesterol levels, HDL levels, blood pressure and diabetes. Values were also corrected for the amount of protein that was harvested from the exosomes per fixed unit of plasma. This was necessary since protein amounts differed per unit plasma.

It was concluded that SerpinF2 carries information on CHD risk, independent of traditional risk factors for CHD. This clearly shows that SerpinF2 levels can be used as a useful biomarker for the risk prediction for coronary heart disease in asymptomatic women, on top of known

cardiovascular risk factors.

In yet another experiment, the exosome sample SerpinF2 protein expression levels of cases who suffered from CHD during follow up were compared with those of healthy control females in the PROSPECT cohort. Figure 3 shows the average SerpinF2 levels, corrected for total protein and for the traditional risk factors age, blood pressure, total

cholesterol, HDL cholesterol, current smoking and diabetes. A significant negative correlation between SerpinF2 levels and the incidence of CHD cases was observed. The higher the level of SerpinF2 in the exosome samples of these female subjects, the lower the risk of developing a myocardial infarction. After correction for the risk factors age, blood pressure, total cholesterol, HDL cholesterol, current smoking and diabetes, this difference was no longer

significant but the trend was still evident. Table III

displays the relation between SerpinF2 levels in exosomes and the incidence of CHD in females. The data show that with a log increase in SerpinF2, the chance of getting a CHD event is reduced with -20% (odds ratio of 0.81) . An odds ratio of 1 indicates that there is no difference between groups, whereas an odds ratio that is lower than 1 indicates that the risk is lower when the SerpinF2 levels increase. The numbers given between brackets indicate the confidence interval for the odds ratio. When the numbers do not include the value 1, then the difference is significant (p<0.05) : a higher level is associated with a lower risk for myocardial infarction. The odds ratio for SerpinF2 alone and with total protein (+) was calculated.

In general, the significance of a comparison is

influenced by the sample size, the variability and the observed difference. When more people are included (a higher sample size) , it is more easy to show that a small

difference is ^''significant' (e.g. p=0.05 indicates that there is a less than 5% chance that the conclusion ^Λ there is a difference' is false) . The standard deviation (SD) is a measure of variability. The SD is influenced by the

variability of a specific measurement and also by the sample size. The more patients are included the more accurate the provided mean value will be with a smaller confidence interval (also a measure of variability) . Finally, the observed difference is relevant since the bigger the

difference the lower the p-value. Univariate Multivariate: adjusted for cardiovascular risk factors

p-value Odds ratio p-value Odds ratio

SerpinF2 -p=0.002 0.81 (0.72- p=0.501 0.95

0.93) (0.81- 1.11)

SerpinF2 -p=0.121 0.89 (0.76- p=0.175 0.89

+ 1.03) (0.74- protein 1.06)

Table III

EXAMPLE 4 SerpinF2 protein concentrations in exosome samples are particularly predictive for cardiovascular events when measured in younger women

From the results in Examples 1 to 3 it can be concluded that the correlation between events and SerpinF2 levels in exosome samples are different when looked at in an elderly cohort with existing cardiovascular disease, as compared to the levels in younger women without cardiovascular disease. Therefore, it was tested whether the predictive value of SerpinF2 was driven by age.

Indeed, the interaction term (SerpinF2 * age) was significant in the full model, adjusted for traditional risk factors (p=0.045). This indicated that in the PROSPECT cohort age influenced SerpinF2 and events. To study this effect of age in more depth, the studied cohort was split in two parts by the median age of the total population, see

Table IV. The mean age of the women in PROSPECT is 58 years. Therefore, the cohort was split in a group of woman younger or equal to 58 years of age (<=58) or older than 58 years (>58) . PROSPECT <=58 years PROSPECT > 58 years

Controls (n) 311 267

CHD Cases (n) 65 143

Table IV Number of cases in the stratified groups

The average SerpinF2 levels, stratified for age in controls and in CHD cases are depicted in Figure 4: (A) shows the levels in the group <= 58 years and (B) shows the levels in the group > 58 years. Clearly the SerpinF2 levels appear much higher in the younger group that did not

experience a myocardial infarction, than in women of the same age group that, after sampling, did experience a primary event. In the older group this effect has

disappeared in that older women not experiencing a

myocardial infarction had similar SerpinF2 levels in their exosomes as the women of that age group that did experience a primary event.

The odds ratio of SerpinF2 levels and the occurrence of a CHD event were calculated with and without adjusting for the cardiovascular risk factors age, blood pressure, total cholesterol, HDL cholesterol, current smoking and diabetes. The results are shown in Table V (female subjects <= 58 years) and Table VI (female subject > 58 years) .

For women below the age of 58 the analysis shows that for every log increase in SerpinF2 levels, the chance of getting a CHD the coming years is lower (-approximately 35% lower, odds of 0.65) .

However, for women above 58, this is not the case as the odds ratio is equal to ~1 which indicates that there are no differences in the chance of getting CHD when SerpinF2 are either lower or higher.

The odds ratio is lower than 1 indicating that a higher level of SerpinF2 is associated with a lower chance of CHD events. When the confidence interval of the odds ratio includes the value 1, then the difference is not

significant. For this reason the confidence interval of the odds ratio is provided between brackets. The more the value "1" is diverging from the confidence interval, the more significant the difference.

Table V Odds ratio between SerpinF2 levels and CHD in women <= 58 years

Table VI Odds ratio between SerpinF2 levels and CHD in women > 58 years

It is evident that the predictive value of exosomal SerpinF2 concentrations in the PROSPECT cohort is stronger in female subjects with an age below or equal to the median value of 58 years in the group tested. Notably, even after correcting for risk factors, it was found that when exosome samples were isolated, SerpinF2 concentrations therein appeared to be predictive for primary events. Importantly, this indicates that the predictive value could not be fully explained by any of the risk factors, and that therefore SerpinF2 appears to be a very strong prognostic marker for primary cardiovascular events next to the generally accepted and known risk factors.

It was further checked whether the values of SerpinF2 were in any way related to hormone levels, for instance in women that went through menopause with or without Hormone Replacement Therapy (HRT) , that is often provided for relief purposes during menopause. It is generally held that HRT women have higher estrogen and progesterone levels in their blood. However, any correlation between high or low hormone levels (due to HRT or in relation to menopause) with the SerpinF2 levels in the exosome samples from these females was absent (data not shown) .

It should be noted that in the older women the

predictive value decreased. Moreover, it appeared that with older age the SerpinF2 values in the control group was lower than the SerpinF2 levels in the control group at younger age. Thus, the higher levels that seemed beneficial for outcome were mostly observed in the females that were younger. The cut-off of 58 as found in this example is relatively arbitrary, as outlined above. As explained before, the median age used in this example more or less correlates with the age at which the risk level of females changes from intermediate/low to high in the presently used risk charts (58-60 years) . The present invention is thus particularly useful for younger females that would not be identified on the basis of the current risk charts but that can be identified as running the risk of developing a primary cardiovascular event with the method of the invention, as long as the Se pinF2 level is measured in exosome samples.

Claims

1. Method for predicting whether a female subject is at risk of developing a cardiovascular event, said method comprising the steps of:

- isolating exosomes from a body sample from said female subject;

- measuring the SerpinF2 protein level in said exosome sample; and

- determining whether the measured SerpinF2 protein level indicates that said female subject is at risk of developing a cardiovascular event.

2. Method according to claim 1, wherein said body sample is urine, amniotic fluid, malignant ascites, broncho- alveolar lavage fluid, synovial fluid, breast milk, saliva or blood, preferably plasma or serum.

3. Method according to claim 1 or 2 wherein said female subject is a woman that is younger or equal to 58-60 years of age.

4. Method according to any one of claims 1 to 3, wherein said cardiovascular event is a primary event.

5. Method according to any one of claims 1 to 4, wherein said cardiovascular event is selected from the group consisting of: myocardial infarction, rupture of an

abdominal aortic aneurysm, rupture of abdominal aortic aneurysm, cardiovascular intervention, coronary artery disease, transient ischemic attack (TIA) , peripheral

arterial disease, acute coronary syndrome, heart failure and re-stenosis of a coronary artery, and wherein the cardiovascular event is in particular coronary artery disease (CAD) .

6. Method according to any one of claims 1 to 5, wherein said female subject is suffering from a risk factor involved in developing a cardiovascular event, said risk factor being dyslipidemia, smoking, hypertension, diabetes mellitus, and/or high cholesterol levels, and/or wherein said female subject is suffering from female specific risk indicators such as polycystic ovarian syndrome or pregnancy hypertension .

7. Method according to any one of claims 1 to 6, wherein a log transformed level between 11.2 and 11.8 pg/ml, preferably between 11.3 and 11.7 pg/ml, more preferably between 11.4 and 11.6 pg/ml and most preferably about 11.5 pg/ml SerpinF2 after correction for total protein, indicates that said female subject is at risk of developing a

cardiovascular event, wherein said levels are preferably measured with a Luminex assay.

8. Method according to any one of claims 1 to 7, wherein said SerpinF2 protein level is determined by using an antibody, or fragment or derivative thereof, that

interacts with said SerpinF2 protein and wherein said interaction is indicative of the concentration of SerpinF2 in said sample.

9. Method according to any one of claims 1 to 8, further comprising the additional step of extracting

proteins from said isolated exosomes before determining the SerpinF2 protein level.

10. Use of an exosome sample from a female subject for predicting whether a female subject is at risk of developing a cardiovascular event, wherein the SerpinF2 protein level in said exosome sample indicates whether said female subject is at risk of developing of a cardiovascular event.

11. Use according to claim 10, wherein said female subject is a woman that is younger or equal to 58-60 years of age.

12. Use according to claim 10 or 11, wherein a log transformed level of between 11.2 and 11.8 pg/ml, preferably between 11.3 and 11.7 pg/ml, more preferably between 11.4 and 11.6 pg/ml and most preferably about 11.5 pg/ml

SerpinF2, after correction for total protein, indicates that said female subject is at risk of developing a

cardiovascular event.

13. Method for predicting the risk of a cardiovascular event, in particular coronary artery disease, in a female subject, based on the detection of SerpinF2 in an exosome samples from said female subject.

14. SerpinF2 for use in the prognosis of the risk of a female subject developing a cardiovascular event, in

particular coronary artery disease.

15. SerpinF2 for use according to claim 14, wherein said cardiovascular event is a primary event.

16. SerpinF2 for use according to claim 14 or 15, wherein said cardiovascular event wherein said

cardiovascular event is selected from the group consisting of: myocardial infarction, rupture of an abdominal aortic aneurysm, rupture of abdominal aortic aneurysm,

cardiovascular intervention, coronary artery disease, transient ischemic attack, peripheral arterial disease, acute coronary syndrome, heart failure and re-stenosis of a coronary artery, and wherein the cardiovascular event is in particular coronary artery disease.