DETECTION METHOD OF CIRCULATING BMP10 (BONE MORPHOGENETIC PROTEIN 10)

Abstract

The present invention relates to a method for assessing atrial fibrillation in a subject, said method comprising the steps of determining the amount of BMP10 in a sample from the subject, and comparing the amount of BMP10 to a reference amount, whereby atrial fibrillation is to be assessed. Moreover, the present invention relates to a method for diagnosing heart failure based on the determination of BMP 10 in a sample from a subject. Further, the present invention relates to a method for predicting the risk of a subject of hospitalization due to heart failure based on the determination of a BMP10-type peptide in a sample from a subject. The present invention further pertains to antibodies which bind to one or more BMP10-type peptides such as NT-proBMP10.

Claims

1. A method for assessing atrial fibrillation in a subject, comprising the steps of a) determining, in at least one sample from the subject, the amount of one or more BMP10-type peptides (Bone Morphogenetic Protein 10-type peptides), and b) comparing the amount of the one or more BMP10-type peptides to a reference amount for the one or more BMP10-type peptides, whereby atrial fibrillation is to be assessed, wherein step a) comprises contacting the sample with at least one agent that is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1.

2. The method of claim 1, wherein said agent is a monoclonal antibody, or antigen binding fragment thereof.

3. The method of claim 2, wherein the monoclonal antibody, or antigen-binding fragment thereof, is capable of binding a) to an epitope contained in amino acid region 171 to 185 of SEQ ID NO: 1 (LESKGDNEGERNMLV, SEQ ID NO: 3), wherein the epitope is an epitope contained in amino acid region 173 to 181 of SEQ ID NO: 1 (SKGDNEGER, SEQ ID NO: 4), b) to an epitope contained in amino acid region 37 to 47 of the polypeptide shown in SEQ ID NO: 1 (SLFGDVFSEQD, SEQ ID NO 2), or c) to an epitope contained in amino acid region 291 to 299 of SEQ ID NO: 1 (SSGPGEEAL, SEQ ID NO: 5).

4. The method of claim 2, wherein the monoclonal antibody, or antigen-binding fragment thereof, comprises a heavy chain variable domain that is at least 80%, 85%, 90%, 95%, 98%, 99%, or 100% identical to the heavy chain variable domain comprising a sequence shown in SEQ ID NO: 7, 8, 9, 10, 11, 12, 13, 14 or 15 (see Table A) and/or a light chain variable domain that is, in increasing order of preference at least 80%, 85%, 90%, 95%, 98%, 99% or 100% identical to the light chain variable domain comprising a sequence as shown in SEQ ID NO: 16, 17, 18, 19, 20, 21, 22, 23 or 24 (see Table B).

5. The method of claim 2, wherein the monoclonal antibody, or antigen-binding fragment thereof, comprises (a) a light chain variable domain comprising (a1) a light chain CDR1 sequence selected from SEQ ID NOs:34-42, (a2) a light chain CDR2 shown selected from SEQ ID NOs:52-60, and (a3) a light chain CDR3 selected from SEQ ID NOs:70-78, and (b) a heavy chain variable domain comprising (b1) a heavy chain CDR1 selected from SEQ ID NOs:25-33, (b2) a heavy chain CDR2 selected from SEQ ID NOs:43-51, and (b3) a heavy chain CDR3 selected from SEQ ID NOs:61-69.

6. The method of claim 1, wherein the assessment of atrial fibrillation is the diagnosis of atrial fibrillation.

7. The method of claim 1, wherein the assessment of atrial fibrillation is the prediction of the risk of an adverse event associated with atrial fibrillation, and wherein the adverse event associated with atrial fibrillation is recurrence of atrial fibrillation and/or stroke.

8. The method of claim 7, wherein an amount of the BMP10-type peptide is indicative for a subject who is at risk of suffering from an adverse event associated with atrial fibrillation and/or wherein an amount of the BMP10-type peptide below the reference amount is indicative for a subject who is not at risk of suffering from an adverse event associated with atrial fibrillation.

9. A method for diagnosing heart failure, said method comprising the steps of (a) determining, in at least one sample from the subject, the amount of one or more BMP10-type peptides (Bone Morphogenetic Protein 10-type peptide) and the amount of at least one further biomarker selected from the group consisting of a natriuretic peptide, ESM-1 (Endocan), Ang2, and FABP3 (Fatty Acid Binding Protein 3), and (b) comparing the amount of the BMP10-type peptide to a reference amount for the BMP10-type peptide and comparing the amount of the at least one further biomarker to a reference amount for said at least one further biomarker, whereby heart failure is to be diagnosed. wherein step a) comprises contacting the sample with at least one agent that is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1, wherein the at least one agent is a monoclonal antibody, or fragment thereof, as defined in claim 3.

10. An in vitro method for assessing atrial fibrillation, for diagnosing heart failure, or for predicting the risk of a subject of hospitalization due to heart failure in a subject, the method comprising contacting at least one sample from the subject with at least one agent that specifically binds to one or more BMP10-type peptides in the at least one sample, wherein the at least one agent is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1, wherein the at least one agent is a monoclonal antibody, or antigen-binding fragment thereof, as defined in claim 2.

11. A monoclonal antibody, or antigen-binding fragment thereof, which binds one or more BMP10-type peptides, wherein the antibody or fragment thereof is capable of binding a) to an epitope contained in amino acid region 171 to 185 of SEQ ID NO: 1 (LESKGDNEGERNMLV, SEQ ID NO: 3), wherein the epitope is an epitope contained in amino acid region 173 to 181 of SEQ ID NO: 1 (SKGDNEGER, SEQ ID NO: 4), b) to an epitope contained in amino acid region 37 to 47 of the polypeptide shown in SEQ ID NO: 1 (SLFGDVFSEQD, SEQ ID NO 2), or c) to an epitope contained in amino acid region 291 to 299 of SEQ ID NO: 1 (SSGPGEEAL, SEQ ID NO: 5).

12. A monoclonal antibody, or antigen-binding fragment thereof, which binds one or more BMP10-type peptides, wherein the antibody or fragment thereof comprises (a) a light chain variable domain comprising (a1) a light chain CDR1 sequence selected from SEQ ID NOs:34-42, (a2) a light chain CDR2 shown selected from SEQ ID NOs:52-60, and (a3) a light chain CDR3 selected from SEQ ID NOs:70-78, and (b) a heavy chain variable domain comprising (b1) a heavy chain CDR1 selected from SEQ ID NOs:25-33, (b2) a heavy chain CDR2 selected from SEQ ID NOs:43-51, and (b3) a heavy chain CDR3 selected from SEQ ID NOs:61-69.

13. A monoclonal antibody, or antigen-binding fragment thereof, which binds one or more BMP10-type peptides, wherein the monoclonal antibody, or antigen-binding fragment thereof, comprises a heavy chain variable domain that is at least 80%, 85%, 90%, 95%, 98%, 99%, or 100% identical to the heavy chain variable domain comprising a sequence shown in SEQ ID NO: 7, 8, 9, 10, 11, 12, 13, 14 or 15 (see Table A) and/or a light chain variable domain that is, in increasing order of preference at least 80%, 85%, 90%, 95%, 98%, 99% or 100% identical to the light chain variable domain comprising a sequence as shown in SEQ ID NO: 16, 17, 18, 19, 20, 21, 22, 23 or 24 (see Table B).

14. A kit comprising at least one monoclonal antibody or fragment thereof, as defined in claim 11.

15. The method of claim 1, wherein the BMP10-type peptide is NT-proBMP10.

16. A method for assessing the extent of white matter lesions in a subject, said method comprising a) determining the amount of one or more BMP10-type peptides in a sample from the subject, and b) assessing the extent of white matter lesions in a subject based on the amount determined in step a), wherein step a) comprises contacting the sample with at least one agent that is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1.

17. A method for predicting dementia in a subject, said method comprising a) determining the amount of one or more BMP10-type peptides in a sample from the subject, b) comparing the amount determined in step a) to a reference, and c) predicting the risk of dementia in said subject, wherein step a) comprises contacting the sample with at least one agent that is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1.

18. A method for the assessment of whether a subject has experienced one or more silent strokes, said method comprising a) determining the amount one or more BMP10-type peptides in a sample from the subject, b) comparing the amount determined in step a) to a reference, and c) assessing whether a subject has experienced one or more silent strokes, wherein step a) comprises contacting the sample with at least one agent that is capable of binding within amino acid region 37 to 299 of the polypeptide shown in SEQ ID NO: 1.

Description

[0573] The figures show:

[0574] FIG. 1: Measurement of BMP10 ELISA in three patient groups (paroxysmal atrial fibrillation, persistent atrial fibrillation and patients in sinus rhythm)

[0575] FIG. 2: ROC curve for BMP10 in paroxysmal Afib; AUC=0.68

[0576] FIG. 3: ROC curve for BMP10 in persistent Afib; AUC=0.90 (Exploratory AFib panel: Patients with a history of atrial fibrillation covering 14 cases of paroxysmal AFib, 16 cases of persistent Afib and 30 Controls)

[0577] FIG. 4: BMP10 in differentiation of patients with Heart Failure and patients without heart failure [unit: ng/ml]

[0578] FIG. 5. BMP10 in differentiation of Heart Failure; ROC curve for BMP10; AUC=0.76

[0579] FIG. 6. Kaplan-Meier curve showing the risk for a HF hospitalization by quartiles of BMP-10 in patients with a prior history of heart failure.

[0580] FIG. 7. Kaplan-Meier curve showing the risk for a HF hospitalization by quartiles of BMP-10 in patients without a prior history of heart failure.

[0581] FIG. 8. Kaplan-Meier curve showing the risk for a stroke by dichotomized of BMP-10 (at median).

[0582] FIG. 9. Kaplan-Meier curve showing the risk for recurrence of AFib by quartiles of BMP-10 in patients undergoing pulmonary vein isolation (BEAT-PVI)

[0583] FIG. 10. Kinetic signatures for pro-peptide BMP10-binding to a selection of 8 mAbs at 37° C. obtained via SPR. Shown are the sensorgram overlays for increasing pro-peptide BMP10 concentrations, ranging between c=3.7-300 nM;

TABLE-US-00007 A) 2H8 B) 3H8 C) 8G5 D) 9E7 E) 11A5 F) 11C10 G) 14C8 H) 13G6

[0584] FIG. 11. Overlay of the normalized dissociation phase for 8 mAbs binding pro-peptide to BMP10 at 37° C.

[0585] FIG. 12. Epitopes of four monoclonal antibodies of the present invention

[0586] FIG. 13: Detection of NT-proBMP10 versus detection of mature BMP10

[0587] FIG. 14: Measurement of circulating BMP-10 in EDTA plasma samples of the SWISS AF study with Fazekas Score <2 (no) vs Fazekas Score ≥2 (yes): Detection of WMLs/Prediction of the risk of silent stroke: Circulating BMP-10 levels were assessed.

EXAMPLES

[0588] The invention will be merely illustrated by the following Examples. The said Examples shall, whatsoever, not be construed in a manner limiting the scope of the invention.

Example 1

Mapping Trial—Diagnose Patients with Atrial Fibrillation as Compared to Patients Based on Their Different Circulating NT-proBMP10 Levels

[0589] The MAPPING study related to patients undergoing open chest surgery. Samples were obtained before anesthesia and surgery. Patients were electrophysiologically characterized using high-density epicardial mapping with multi-electrode arrays (high density mapping). The trial comprised 14 patients with paroxysmal atrial fibrillation, 10 patients with persistent atrial fibrillation and 28 controls, matched to best possible (on age, gender, comorbidities). NT-proBMP10 was determined in serum samples of the MAPPING study. Elevated NT-proBMP10 levels were observed in patients with atrial fibrillation versus controls. NT-proBMP10 levels were elevated in patients with paroxysmal atrial fibrillation versus matched controls, as well as in patients with persistent atrial fibrillation versus controls.

[0590] In addition, the biomarker ESM-1 was determined in samples from the MAPPING cohort. Interestingly, it was shown that the combined determination of NT-proBMP10 with ESM-1 allowed for an increase of the AUC to 0.92 for the differentiation between persistent AF vs. SR (sinus rhythm).

[0591] In addition, the biomarker FABP-3 was determined in samples from the MAPPING cohort. Interestingly, it was shown that the combined determination of NT-proBMP10 with FABP-3 allowed for an increase of the AUC to 0.73 for the differentiation between paroxysmal AF vs. SR (sinus rhythm).

Example 2

Heart Failure Panel

[0592] The heart failure panel included 60 patients with chronic heart failure. According to the ESC guidelines criteria, heart failure was diagnosed in patients with typical signs and symptoms and objective evidence of a structural or functional abnormality of the heart at rest. Patients between 18 and 80 years with ischemic or dilated cardiomyopathy or significant valvular disease and who were able to sign the consent form were included into the study. Patients with acute myocardial infarction, pulmonary embolism or stroke in the last 6 months, further with severe pulmonary hypertension and end stage renal disease were excluded. The patients suffered predominantly from heart failure stages NYHA II-IV.

[0593] The healthy control cohort included 33 subjects. The healthy status was verified by assessing status of ECG and echocardiography results. Participants with any abnormality were excluded.

[0594] Elevated NT-proBMP10 levels were observed in serum samples of patients with heart failure versus controls.

Example 3

Biomarker Measurements

[0595] NT-proBMP10 was measured in a research grade ECLIA assay for Bone Morphogenetic Protein 10 (BMP10); ECLIA Assay from Roche Diagnostics, Germany.

[0596] For detection of NT-proBMP10 in human serum and plasma samples an antibody sandwich which specifically binds to the N-terminal prosegment of BMP10 was used. Such antibodies also bind to proBMP10 and preproBMP10. Thus, the sum of the amounts of the N-terminal prosegment of BMP10, proBMP10 and preproBMP10 was determined. Structural prediction based on findings from other BMP-type proteins as e.g. BMP9 show that BMP10 remains in a complex with proBMP10, thus detection of N-term prosegment also reflects the amount of prodomain-bound BMP10. Moreover, the homodimeric form of BMP10 can be detected, as well as heterodimeric structures, as e.g. the combination with BMP9 or other BMP-type proteins.

Example 4

The SWISS AF Study—Risk Prediction of Heart Failure Hospitalization

[0597] The data from the SWISS-AF study includes 2387 patients from which 617 have a history of heart failure (HF). BMP-10 was measured in these patients to assess its ability to predict the risk of a hospitalization due to heart failure.

[0598] As heart failure hospitalization can occur in patients with a history of heart failure and in patients without a known history heart failure the ability was assessed to predict future heart failure hospitalization was assessed in these groups independently. In total for 233 patients a hospitalization due to HF was recorded during follow-up. 125 of the 233 hospitalization occurred in patients with a prior known HF.

[0599] Prediction of HF Hospitalization in Patients with a Known History in HF

[0600] Table 1 shows the result of a cox proportional hazard model including in patients with a known history of HF. Dependent variable is the time until HF hospitalization and independent variable are log-2 the transformed NT-proBMP10 values.

[0601] As visible by the hazard ratio and the low p-value NT-proBMP10 is able to predict the risk for HF hospitalization significantly in patients with a known history of HF. As NT-proBMP10 values were log-2 transformed before they were entered into the model the hazard ratio can interpreted that risk increase by 3.43 for a patient if the value of NT-proBMP10 doubles

TABLE-US-00008 TABLE 1 Summary of cox proportional hazard model for NT-proBMP10 (log-2 trans-formed) predicting the risk of HF hospitalization in patients with a known history of HF. Hazard Ratio 95% Confidence Interval P-Value 3.43 2.23-5.27 <0.001

[0602] FIG. 6 shows a Kaplan-Meier curve which displays the risk of HF hospitalization by quartiles of NT-proBMP10. It is visible that the risk increases constantly with increasing NT-proBMP10 values and the highest risk is observed for patients with NT-proBMP10 levels within the highest quartile.

[0603] Prediction of HF Hospitalization in Patients without a Known History in HF

[0604] Table 2 shows the result of a cox proportional hazard model including in patients without a known history of HF. Dependent variable is the time until HF hospitalization and independent variable are the log-2 transformed NT-proBMP10 values.

[0605] As visible by the hazard ratio and the low p-value NT-proBMP10 is able to predict the risk for HF hospitalization significantly in patients without a known history of HF. As NT-proBMP10 values were log-2 transformed before they were entered into the model the hazard ratio can interpreted that risk increase by 3.43 for a patient if the value of NT-proBMP10 doubles

TABLE-US-00009 TABLE 2 Summary of cox proportional hazard model for NT-proBMP10 (log-2 trans-formed) predicting the risk of HF hospitalization in patients without a known history of HF. Hazard Ratio 95% Confidence Interval P-Value 4.24 2.52-7.15 <0.001

[0606] FIG. 7 shows a Kaplan-Meier curve which displays the risk of HF hospitalization by quartiles of NT-proBMP10. It is visible that the risk increases with increasing NT-proBMP10 values and the risk is highest for patients with NT-proBMP10 levels within the two highest quartiles.

Example 5

The SWISS AF Study—Risk Prediction of Stroke

[0607] The ability of circulating NT-proBMP10 to predict the risk for the occurrence of stroke was verified (in reference to example 3) in a prospective, multicentric registry of patients with documented atrial fibrillation (Conen D., Swiss Med Wkly. 2017 Jul. 10; 147:w14467). NT-proBMP10, results were available for 65 patients with an event and 2269 patients without an event.

[0608] In order to quantify the univariate prognostic value of NT-proBMP10 proportional hazard models were used with the outcome stroke.

[0609] The univariate prognostic performance of NT-proBMP10 was assessed by two different incorporations of the prognostic information given by NT-proBMP10.

[0610] The first proportional hazard model included NT-proBMP10 binarized at the median (2.2 ng/mL) and therefore comparing the risk of patients with NT-proBMP10 below or equal to the median versus patient with NT-proBMP10 above the median.

[0611] The second proportional hazard model included the original NT-proBMP10 levels but transformed to a log 2 scale. The log 2 transformation was performed in order to enable a better model calibration.

[0612] In order to get estimates for the absolute survival rates in the two groups based on the dichotomized baseline NT-proBMP10 measurement (<=2.2 ng/mL vs >2.2 ng/mL) a Kaplan-Meier plot was created.

[0613] In order to assess if the prognostic value of NT-proBMP10 is independent from known clinical and demographic risk factors a weighted proportional cox model including in addition the variables age, and history of Stroke/TIA/Thromboembolism was calculated. These were the only significant clinical risk predictors on the whole cohort (including all controls).

[0614] In order to assess the ability of NT-proBMP10 to improve existing risk scores for the prognosis of stroke the CHADS.sub.2 the CHA.sub.2DS.sub.2-VASc and the ABC score were extended by NT-proBMP10 (log 2 transformed). Extension was done by creating a portioned hazard model including NT-proBMP10 and the respective risk score as independent variables.

[0615] The c-indices of the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc and ABC score were compared to the c-indices of these extended models.

[0616] Results

[0617] Table 1 shows the results of the two univariate weighted proportional hazard models including the binarized or the log 2 transformed NT-proBMP10. The association between the risk for experiencing a stroke with the baseline value of NT-proBMP10 is not significant in the model using log 2-transformed NT-proBMP10 as a risk predictor but close to the significance level of 0.05.

[0618] For the model using the binarized NT-proBMP10 the p-value is slightly higher. It could be argued however with a higher number of events the effect could be statistically significant. The hazard ration for the binarized NT-proBMP10 implies a 1.5-fold higher risk for a stroke in the patient group with baseline NT-proBMP10>2.2 ng/mL versus the patient group with baseline NT-proBMP10<=2.2 ng/mL. This can be seen also in FIG. 8 displaying the Kaplan Meier curves for the two groups.

[0619] The results of the proportional hazard model including NT-proBMP10 as log 2 transformed linear risk predictor suggest the log 2 transformed values NT-proBMP10 are proportional to the risk for experiencing a stroke. The hazard ratio of 2.038 can be interpreted in a way that a 2-fold decrease of NT-proBMP10 is associated with 2.038 increase of risk for a stroke.

TABLE-US-00010 TABLE 1 Results result of the univariate weighted proportional hazard model including the binarized and log2 transformed NT-proBMP20. Hazard Ratio (HR) 95%-CI HR P-Value NT-proBMP10 log2 1.523 0.930-2.495 0.095 Baseline BMP10 <=2.2 ng/mL 2.038 0.994 4.179 0.052 vs NT-proBMP10 >2.2 ng/mL

[0620] Table 2 shows the results of a proportional hazard model including NT-proBMP10 (log 2 transformed) in the combination with clinical and demographic variables. It is visible that the prognostic value of NT-proBMP10 diminishes to some extend but this could be partially also being explained by low statistical power of the model.

TABLE-US-00011 TABLE 2 Multivariate proportional hazard model including NT-proBMP10 and relevant clinical and demographic variables. Hazard Ratio (HR) 95%-CI HR P-Value Age 1.0615 1.0256-1.0987 0.0007 History 1.9186 1.1451-3.2145 0.0133 Stroke/TIA/embolism NT-proBMP10 (log2 1.2253 0.5545-2.7076 0.6155 transformed)

[0621] Table 3 shows the results of the weighted proportional hazard model combining the CHADS.sub.2 score with NT-proBMP10 (log 2 transformed). In this model NT-proBMP10 can add prognostic information to the CHADS.sub.2 score but with a p-value above 0.05 which can however be tolerated with respect to the low sample size.

TABLE-US-00012 TABLE 3 Weighted proportional hazard model combining the CHADS.sub.2 score with NT-proBMP10 (log2 transformed) Hazard Ratio (HR) 95%-CI HR P-Value CHADS.sub.2 score 1.3792 1.1590-1.6413 0.0003 NT-proBMP10 (log2 1.5046 0.7094-3.1911 0.2869 transformed)

[0622] Table 4 shows the results of the weighted proportional hazard model combining the CHA.sub.2DS.sub.2-VASc score with NT-proBMP10 (log 2 transformed). Also in this model NT-proBMP10 can add prognostic information to the CHA.sub.2DS.sub.2-VASc score but with a p-value above 0.05 which can however be tolerated with respect to the low sample size.

TABLE-US-00013 TABLE 4 Weighted proportional hazard model combining the CHA.sub.2DS.sub.2-VASc score with NT-proBMP10 (log2 transformed) Hazard Ratio (HR) 95%-CI HR P-Value CHA.sub.2DS.sub.2-VASc 1.2756 1.0992-1.4803 0.1281 score NT-proBMP10 (log2 1.4308 0.6645-3.0806 0.3600 transformed)

[0623] Table 5 shows the results of the weighted proportional hazard model combining the ABC score with NT-proBMP10 (log 2 transformed). In this model the estimated hazard ratio diminishes and NT-proBMP10 likely can't add any prognostic performance

TABLE-US-00014 TABLE 5 Weighted proportional hazard model combining the ABC score with NT-proBMP10 (log2 transformed) Hazard Ratio (HR) 95%-CI HR P-Value ABC score 1.1839 1.1046-1.2688 <0.0001 NT-proBMP10 (log2 0.7321 0.3123-1.7161 0.4731 transformed)

[0624] Table 6 shows the estimated c-indexes of NT-proBMP10 alone, of the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc, the ABC score and of the weighted proportional hazard model combining the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc, the ABC score with NT-proBMP10 (log 2) on the case cohort selection. It can be seen that the addition of NT-proBMP10 improves the c-index of the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc score but not the ABC score.

[0625] The differences in c-index are 0.019, 0.015 and −0.002 for the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc, the ABC score respectively.

TABLE-US-00015 TABLE 6 C-indexes of NT-proBMP10, the CHA.sub.2DS.sub.2-VASc score and the combination of the CHA.sub.2DS.sub.2-VASc score with NT-proBMP10 and C-indexes of the CHADS.sub.2 and ABC score and their combination with NT-proBMP10. C-Index NT-proBMP10 univariate 0.577 CHADS.sub.2 0.629 CHADS.sub.2 + NT-proBMP10 0.643 CHA.sub.2DS.sub.2-VASc 0.616 CHA.sub.2DS.sub.2-VASc + NT-proBMP10 0.627 ABC score 0.692 ABC score + NT-proBMP10 0.690

Example 6

The BEAT-AF-PVI Study—Prediction of Risk of Recurrent AFib After Pulmonary Vein Isolation and Catheter Ablation

[0626] The ability of NT-proBMP10 to predict the risk of future recurrent atrial fibrillation episodes was assessed in the BEAT-AF-PVI study. The BEAT-AF-PVI study (Knecht S, International Journal of Cardiology, Volume 176, Issue 3, 2014, Pages 645-650) is a prospective cohort study including patients with atrial fibrillation who underwent a pulmonary vein isolation. One of the collected study endpoints was the time until the first recurrence of atrial fibrillation. Thus, the ability of circulating NT-proBMP10 to predict the risk for the reoccurrence of AFib after PVI and catheter ablation was verified in a prospective, multicentric registry of patients with documented atrial fibrillation (Zeljkovic I., Biochem Med. 2019; 29: 020902).

[0627] NT-proBMP10 measurements and information about atrial fibrillation recurrence were available in 719 patients. For 310 out of 719 patients an atrial fibrillation recurrence was observed. NT-proBMP10 was measured in a research grade ECLIA assay for Bone Morphogentic Protein 10 (NT-proBMP10), from Roche Diagnostics, Germany.

[0628] The ability of NT-proBMP10 to predict the risk of recurrent atrial fibrillation was assessed by a Cox (proportional hazard) regression model. Table 1 shows the results of the proportional hazard model. The results show that with increasing values of NT-proBMP10 the risk of recurrent atrial fibrillation increases significantly. As the NT-proBMP10 was included in the model log 2 transformed to improve model calibration one can interpret the hazard ration as 1.91 risk increase if NT-proBMP10 increases 2-fold.

TABLE-US-00016 TABLE 7 Summary of the Cox regression model predicting the risk of recurrent AFIB by the log2 transformed values of NT-proBMP10. Hazard Hazard Ratio Ratio 95% CI P-Value NT-proBMP10 (log2) 1.91 1.24-2.93 0.003

[0629] Alternatively NT-proBMP10 can be used also in a binarized (e.g. split at median of 1.7 ng/mL) form for risk prediction of atrial fibrillation. Table 8 indicates that the risk in the patients group with NT-proBMP10 levels above the median is elevated by 32%. This risk difference is again statistically significant.

TABLE-US-00017 TABLE 8 Summary of the Cox regression model predicting the risk of recurrent AFIB by binarized NT-proBMP10 at the observed median value. Hazard Hazard Ratio Ratio 95% CI P-Value NT-proBMP10 >1.7 ng/mL 1.32 1.05-2.65 0.018

Example 7

Assessment of Recurrent AF with Circulating NT-proBMP10

[0630] The GISSI AF study relates to patients in sinus rhythm (SR) with a history of atrial fibrillation (AF), but without significant left ventricular dysfunction or heart failure. All patients underwent biochemistry NT-proBMP10 assessment and electrocardiography at 3 times during 1 year follow up.

[0631] Circulating NT-proBMP10 levels have been determined in samples of n=281 patients with blood sampled and biomarkers assayed, in SR at 6-month visit and in samples of n=33 patients with blood sampled and biomarkers assayed, with ongoing AF at 6-month visit. Antibodies against pro-peptide BMP10 were used.

TABLE-US-00018 TABLE 9 Measurement of circulating NT-proBMP10 in GISSI AF patients at 6-month visit in SR and with ongoing AF GISSI AF SR (n = 281) AF (n = 33) p-value NT-proBMP10 1.97 2.31 <0.0001 (median; ng/mL) [1.75-2.33] [2.04-2.67]

[0632] As shown in table 9, NT-proBMP10 was observed in patients with ongoing AF at sampling versus patients in SR at the time of blood sampling at the 6—month visit of the GISSI AF study.

[0633] It is obvious, that for NT-proBMP10 small, but highly significant delta changes of marker elevations in AF pts vs SR pts could be detected. In the 33 patients with ongoing AF at the time of sampling median NT-proBMP10 values of 2,31 [2,04-2,67] vs 1,97[1,75 — 2,33] ng/mL were observed versus the 281 patients in SR at the time of blood sampling.

[0634] Samples of n=105 patients were in SR at 6-month visit, but had experienced more than one recurrence of AF from randomization to the 6—month visit. All 105 patients spontaneously converted to SR.

TABLE-US-00019 TABLE 10 Measurement of circulating NT-proBMP10 in GISSI AF patients in SR; Case control of patients experiencing recurrent AF several days before sampling Days after AF 1-7 (n = 11) 8-30 (n = 17) >30 (n = 77) NT-proBMP10 2.30 1.90 1.90 (median; ng/mL) [1.65-2.45] [1.67-2.50] [1.75-2.25] 0-7 days n = 11 patients 8-30 days n = 17 patients >30 days n = 77 patients

[0635] As shown in table 10, NT-proBMP10 titers were observed to be elevated for up to 7 days after AF in patients, that spontaneously converted to SR. In the 11 patients with preceding AF up to 7 days before sampling median values of 2,30 [1,65 — 2,45] versus 1,90 [1,75-2,25] ng/mL were observed versus the 77 patients with preceding AF more than 30 days before blood sampling. It is striking, that the very same median values for NT-proBMP10 were observed up to 7 days after preceding AF before sampling as in patients with ongoing AF. In the 11 patients in SR after AF up to 7 days ago at the time of sampling median NT-proBMP10 values of 2,30 [1,65 — 2,45] ng/mL were observed. As shown in table 9 in the 33 patients with ongoing AF at the time of in SR at the time of blood sampling median NT-proBMP-10 values of 2,31 [2,04 — 2,67] ng/ml were observed.

[0636] Data evaluation showed that patients with NT-proBMP10 levels (independent from other biomarkers) above a reference value (in the study >2,0 ng/mL) are suspicious to have recurrent atrial fibrillation after therapeutic interventions, e.g. after cardioversion. The differentiation of poor and good therapy responders supports decision making, which patient may not profit from a therapy in order to avoid costly therapies and associated burden, but poor outcome to the patient.

[0637] It is even shown, that elevated NT-proBMP10 levels may be detected in patients presenting in sinus rhythm up to 7 days after a preceding AF episode.

[0638] In summary, the diagnosis of paroxysmal atrial fibrillation in patients presenting within 7 days later in sinus rhythm may be achieved with detection of enhanced levels of NT-proBMP10 alone or in combination with a marker of cardiac injury (eg. cTNThs) and/or a marker of heart failure (NT-ProBNP).

Example 8

Detection of NT-proBMP10 Versus Detection of Mature BMP 10

[0639] For comparison of detecting mature BMP10 and NT-proBMP10 the Elecsys prototype detection method as described in example 14 was measured in a head to head analysis with BMP-10 ELISA (R&D systems DuoSet DY2926-05) detecting mature BMP-10 homodimer (aa 317-424). Samples that were measured are deriving from the mapping cohort as described in example 1. Diagnosed patients with atrial fibrillation as compared to patients based on their different circulating BMP10 levels applying novel method for detection of NT-proBMP10 in comparison to commercial immunoassays for detection of mature BMP10 from R&D systems. Serum samples from patients were electrophysiologically characterized using high-density epicardial mapping with multi-electrode arrays (MAPPING study) (FIG. 13, or Table below).

TABLE-US-00020 sample Mature BMP10 [pg/ml] Diagnosis Patient 1 52.4 Sinus rhythm Patient 2 605 Sinus rhythm Patient 3 552 Sinus rhythm Patient 4 217 Persistent AF Patient 5 2374 Sinus rhythm Patient 6 74.5 Paroxysmal AF

[0640] For only six out of 52 samples mature BMP10 levels could be detected. These are reflecting 4 in sinus rhythm, 1 paroxysmal and 1 persistent AF patient. Whereas NT-proBMP10 levels could be detected for all samples as described in example 1.

[0641] This finding of only 11.5% of samples having detectable levels of mature BMP10 suggests that physiologically the mature form is underrepresented in circulation due to internalization upon receptor binding. Thus, NT-proBMP10 is circulating in a more stable form and higher detectable concentration levels. Allowing clinical decision making based on circulating NT-proBMP10 levels.

Example 9

Immunization of Rabbits for Generation of Antibodies Against BMP-10

[0642] Here we describe the development antibodies with the ability to bind Bone Morphogenetic Protein-10 (NT-proBMP10). For the generation of such antibodies, we immunized 12 weeks old NZW rabbits with rec. NT-proBMP10, a polypeptide comprising the first 312 amino acids of preproBMP10. All rabbits were subjected to repeated immunizations. In the first month the animals were immunized weekly. From the second month onward the animals were immunized once per month. For the first immunization, we dissolved 500 μg of the immunogen in 1 mL 140 mM NaCl and we emulsified the solution in 1 ml CFA. For all following immunizations, CFA was replaced by IFA.

Example 10

Development of Antibodies Binding NT-proBMP10

[0643] For the development of antibodies binding to BMP-10, B-cell cloning as described in Seeber et al. (2014), PLoS One. 2014 Feb. 4; 9(2) was used. Firstly, the PBMC pool of cells was prepared from whole blood of the immunized animals by ficoll gradient centrifugation. For the enrichment of antigen reactive B-cells from the PBMC pool randomly biotinylated NT-proBMP10 was immobilized on streptavidin coated magnetic beads (Miltenyi). For the coating of the beads the protein was used at a concentration of 1 μg/ml. Therefore, we incubated the prepared PBMC pool of the immunized animals with the NT-proBMP10-coated beads for 1 h. For the enrichment of antigen-reactive B-cells MACS columns (Miltenyi) were used. B-cell sorting and incubation was done as described in Seeber et al. (2014), PLoS One. 2014 Feb. 4; 9(2). To identify NT-proBMP10 reactive clones by ELISA we immobilized NT-proBMP10 on the surface of 96we11 plates and the concentration used for immobilization was 250 ng/ml. After washing the plates were blocked with 5% BSA to reduce background signals. The plates were washed again and 30 μl of the primary rabbit B-cell supernatant were transferred to the 96 well plates and incubated for 1 h at room temperature. For the detection of antibodies bound to the screening peptides, HRP-labeled F(ab′)2 goat anti-rabbit Fcγ (Dianova) and ABTS (Roche) as a substrate were added. Clones binding to plate bound NT-proBMP10 were selected for subsequent molecular cloning as described in Seeber et al. (2014), PLoS One. 2014 Feb. 4; 9(2).

Example 11

Kinetic Screening

[0644] Information antigen: Pro-peptide BMP10 (R&D-Systems) and in-house construct “312” (pre-pro-peptide BMP10), both represent the prodomain of BMP10±19 aa N-terminal leader peptide. Recombinant human BMP-10, R&D-Systems, Cat. No 2926-BP/CF, Lot: Qual0518031, disulfide-linked homodimer, MW 24.40 kDa.

[0645] Kinetic Screening

[0646] The kinetic screening was performed at 37° C. on a GE Healthcare Biacore 4000 instrument. A Biacore CM5 Series S sensor was mounted to the instrument, hydrodynamical addressed and preconditioned according to the manufacturer's instructions. The system buffer was HBS-EP (10 mM HEPES, 150 mM NaCl, 1 mM EDTA, 0.05% (w/v) P20). The system buffer supplemented with 1 mg/mL CMD (Carboxymethyldextran, Fluka) was used as sample buffer.

[0647] A rabbit antibody capture system was immobilized on the sensor surface. A polyclonal goat anti-rabbit IgG Fc capture antibody GARbFcγ (Code-No. 111-005-046; Jackson Immuno Research) was amine coupled using the EDC/NHS-chemistry according to the to the manufacturer's instructions.

[0648] 30 μg/mL GARbFcγ in 10 mM sodium acetate buffer (pH 4.5) were preconcentrated to the spots 1, 2, 4 and 5 in the flow cells 1, 2, 3 and 4 and covalently bound to the CMD-surface with densities of approximately 10000 RU. Free activated carboxyl groups were subsequently saturated with 1 M ethanolamine pH 8.5.

[0649] The spots 1 and 5 were used for the interaction measurements and spots 2 and 4 served as references. Each rabbit antibody supernatant suspension was diluted 1:5 in sample buffer and was injected at a flow rate of 10 μL/min for 2 minutes. The rabbit antibody Capture Level (CL) in resonance units RU was monitored.

[0650] Construct “312” was injected with a single concentration c=150 nM to the respective surface displayed anti-pro-peptide BMP10 rabbit mAb at 30 μL/min. The association and dissociation phase was monitored for 5 minutes each. After each cycle of kinetics determination the rabbit clones were completely washed from the sensor surface by an injection of 10 mM Glycine pH 1.5 at 20 μL/min for 30 seconds. Report points Binding Late (BL), shortly before the end of the pro-peptide BMP10 injection and Stability Late (SL), shortly before the end of the dissociation were extracted from obtained sensorgram. They are used to characterize the antibody/antigen binding stability. Furthermore, the dissociation rate constant k.sub.d [s.sup.−1] was calculated according to a Langmuir 1:1 model. The antigen/antibody complex stability halftime (minutes) was calculated according to the formula ln(2)/60*k.sub.d. The Molar Ratio, representing the binding stoichiometry, was calculated with the formula:

MW (antibody)/MW (antigen)*BL (antigen)/CL (antibody)

[0651] 280 rabbit antibodies were tested using this approach. 18 Abs were identified with suitable kinetic properties meeting the criteria for the Elecsys-platform.

[0652] Kinetic Characterization

[0653] Detailed kinetic investigations were performed using the BIAcore 8K instrument from GE Healthcare. The rabbit mAbs<propeptid BMP10>clones 2H8, 3H8, 8G5, 9F7, 11A5, 11C10, 14C8 and 13G6—identified via the kinetic screening—were kinetically characterized in detail for binding pro-peptide BMP10 at 37° C.

[0654] A Biacore CM5 Series S sensor (Lot #10281824/10276998) was mounted to the instrument.

[0655] Amine Coupling of Capture Molecules

[0656] A rabbit antibody capture system was immobilized on the sensor surface. A polyclonal goat anti-rabbit IgG Fc capture antibody GARbFcγ (Code-No. 111-005-046, Lot #131053; Jackson Immuno Research) was amine coupled using the EDC/NHS-chemistry according to the manufacturer's instructions: running buffer: HBS-N buffer (10 mM HEPES, 150 mM NaCl, pH 7.4), activation by mixture of EDC/NHS, the capture-Ab was diluted in coupling buffer NaAc, pH 5.0, c=30 μg/mL; finally remaining activated carboxyl groups were blocked by injection of 1 M Ethanolamime pH 8.5; Ab densities reached between 11200-12700 RU

[0657] Kinetic Characterization for Pro-Peptide BMP10 Binding to a Selection of mAbs at 37° C.

[0658] The system and sample buffer was HBS-EP (10 mM HEPES, 150 mM NaCl, 1 mM EDTA, 0.05% (w/v) P20, pH 7.4).

[0659] Flow cells 2 of channels 1, 2, 3, 4, 5, 6, 7 and 8 were used for the interaction measurements and flow cells 1 of each channel served as references. Each rabbit antibody was diluted to 3 nM in sample buffer and was injected at a flow rate of 5 μL/min for 2 minutes. The rabbit antibody Capture Level (CL) in resonance units RU was monitored.

[0660] A series of increasing concentrations c=3.7-300 nM pro-peptide BMP10 “BMP10(312)” was injected to the respective surface displayed anti-pro-peptide BMP10 rabbit mAb at 60 μL/min with replicate for the concentration c=33.3 nM. The association phase was monitored for 3 minutes; dissociation phase was monitored for 10 minutes. After each cycle of kinetics determination, the rabbit clones were eluted from the capture system by an injection of 10 mM Glycine pH 2.0 for 1 minute, followed by two consecutive injections of 10 mM Glycine pH 2.25 at 20 μL/min for 1 minute.

[0661] The dissociation rate constant k.sub.d was evaluated using a Langmuir 1:1 fit model according to the BIAcore™ evaluation software Insight SW V 2.0 from GE Healthcare. The antigen/antibody complex stability half-life time (minutes) was calculated according to the formula ln(2)/60*k.sub.d.

[0662] The Molar Ratio, representing the binding stoichiometry, was calculated with the formula:

MW (antibody)/MW (antigen)*BL (antigen)/CL (antibody)

[0663] Since the pro-peptide BMP10 is a dimeric molecule, the obtained affinities are avidity-burdened, therefore, representing apparent data.

[0664] Results

[0665] Kinetic Screening

[0666] 280 rabbit antibodies were tested using the kinetic screening approach. 18 Abs with suitable kinetic properties were identified and further characterized.

[0667] Detailed Kinetic Characterization

[0668] From 280 kinetically screened rabbit monoclonal antibodies 18 antibodies were selected.

[0669] Detailed concentration-dependent kinetic investigations showed, that the pro-peptide BMP 10 interaction does not behave according to a Langmuir 1:1 interaction.

[0670] The pro-peptide BMP10 is a dimeric molecule, the interactions are probably avidity burdened. The kinetic data represents apparent data, but can be characterized by visual inspection of the sensorgrams and by a quantification of the antibody linear dissociation phases. The complex half-lifes vary between t/2 diss=6 and >115 minutes. The binding stoichiometries are between 1.2 to 1.4 and indicate a 2:1 binding stoichiometry.

[0671] All Abs with exception of 14C8 and 8G5 show suitable kinetic profiles: comparable fast complex formation velocity and complex half-lifes t/2 diss>15 minutes. The Molar Ratios for all Abs indicate a 2:1 binding stoichiometry. Clone 8G5 and 14C8 show a slightly slower association and less complex stability than other clones. Both cover the same epitope region. Clones 2H8 and 3H8, each covers a unique epitope region.

[0672] Clone 13G6 shows the slowest dissociation (k.sub.d<1.0E−04 s.sup.−1), resulting in the complex-half-life t/2-diss>115 minutes. Clone 11A5 shows a suitable kinetic signature with fast complex formation velocity and a complex half-life of 30 minutes. The Molar Ratio indicates fully functional Ab with a 2:1 binding stoichiometry.

[0673] Results are summarized in the Table 11, Sensorgram overlays are shown in FIG. 10. The normalized antibody dissociation phases for the 8 clones are shown in FIG. 11

TABLE-US-00021 TABLE 11 Kinetic constants and affinity data for the pro-peptide BMP10-binding to a selection of mAbs<pro-peptide BMP10> measured by SPR (BIACORE ® 8K) at 37° C. CL k.sub.d t/2 diss R.sub.max MR mAb [RU] [s.sup.−1] [min] [RU] [ ] 2H8 208 3.0E−04 ± <0.1% 38 65 1.2 3H8 162 7.8E−04 ± <0.1% 15 53 1.3 8G5 164 2.1E−03 ± <0.1% 6 52 1.2 9E7 189 5.4E−04 ± <0.1% 21 70 1.4 11A5 187 3.9E−04 ± <0.1% 30 68 1.4 11C10 187 4.5E−04 ± <0.1% 26 64 1.3 14C8 155 1.2E−03 ± <0.1% 10 53 1.3 13G6 219 <1.0E−04 ± 1.1%  >115 74 1.3 k.sub.d dissociation rate constant [s.sup.−1] t.sub./2-diss antigen/antibody complex stability half-life [min], calculated according to the formula ln(2)/60*k.sub.d R.sub.max maximum analyte response [RU] MR Molar Ratio: Ratio R.sub.max experimental vs. theoretical, bound analyte per mAb

Example 12

Epitope Mapping Using Peptide Microarrays

[0674] The epitope mapping of antibody clones was carried out by means of a library of overlapping immobilized peptide fragments (length: 15 amino acids, 14 amino acids overlap) corresponding to the sequence of human bone morphogenetic protein 10. Peptides were synthesized with an automated synthesizer (Intavis MultiPep RS) on modified cellulose disks which were dissolved after synthesis. The solutions of the individual peptides were then spotted onto coated microscope slides. The synthesis was carried out stepwise utilizing 9-fluorenylmethoxycarbonyl (Fmoc) chemistry on amino-modified cellulose disks in a 384-well synthesis plate. In each coupling cycle, the corresponding amino acids were activated with a solution of DIC/HOBt in DMF. Between coupling steps, un-reacted amino groups were capped with a mixture of acetic anhydride, diisopropylethyl amine and 1-hydroxybenzotriazole. Upon completion of the synthesis, the cellulose disks were transferred to a 96-well plate and treated with a mixture of trifluoroacetic acid (TFA), dichloromethane, triisoproylsilane (TIS) and water for side chain deprotection. After removal of the cleavage solution, the cellulose bound peptides were dissolved with a mixture of TFA, TFMSA, TIS and water, precipitated with diisopropyl ether and re-suspended in DMSO. These peptide solutions were subsequently spotted onto Intavis CelluSpots™ slides using an Intavis slide spotting robot.

[0675] For epitope analysis, the prepared slides were washed with ethanol and then Tris-buffered saline (TBS; 50 mM Tris, 137 mM NaCl, 2.7 mM KCl, pH 8) before a blocking step was carried out for 1 h at 37° C. with 5 mL 10× Western Blocking Reagent (Roche Applied Science), 2.5 g sucrose in TBS, 0.1% Tween 20. After washing (TBS+0.1% Tween 20), the slides were incubated with a solution (1 μg/mL) of antibody clones in TBS+0.1% Tween 20 at 37° C. 1 h. After washing, the slides were incubated for detection with an anti-rabbit secondary HRP-antibody (1:20000 in TBS-T) followed by incubation with DAB substrate. Positive SPOTs were assigned to the corresponding peptide sequences.

TABLE-US-00022 TABLE 12 Epitopes Position in the protein (UniProtKB O95393, Analyzed Antibodies SEQ ID NO: 1 Detected Epitope MAB<BMP10>rRb-3H8 HC3/LC3 37 to 47 S-L-F-G-D-V-F-S-E- (p10qx/p10qy)-IgG(SPA) Q-D (SEQ ID NO: 2) MAB<BMP10>rRb-8G5 HC3/LC5 — no linear epitope (p10qz/p10ra)-IgG(SPA) detected MAB<BMP10>rRb-9E7 HC12/LC7 no linear epitope (p10rw/p10rx)-IgG(SPA) detected MAB<BMP10>rRb-11A5 HC6/LC3.2 171 to 185 L-E-S-K-G-D-N-E- (p10se/p10sf)-IgG(SPA) G-E-R-N-M-L-V (SEQ ID NO: 3 MAB<BMP10>rRb-2H8 HC7/LC6 291 to 299 S-S-G-P-G-E-E-A-L (p10qv/p10qw)-IgG(SPA) (SEQ ID NO: 5) MAB<BMP10>rRb-14C8 HC2/LC3 no linear epitope (p10th/p10ti)-IgG(SPA) detected MAB<BMP10>rRb-13G6 173 to 181 S-K-G-D-N-E-G-E-R (p10tf/p10tg)-IgG(SPA) (SEQ ID NO: 4)

Example 13

Selection of Antibodies for Sandwich Assays

[0676] Biotinylated and Ruthenylated clones 2H8, 3H8, 8G5, 9E7 11A5, 11C10, 14C8, 13G6 were tested in multiple sandwiches for identification of sandwich partners for further Elecsys immunoassay development. Relation of recognition of recombinant NT-proBMP10 (22-312) at 0,01 ng/ml with blank values across sandwich combinations (Table 13) reflects a signal-to-noise ratio of 1,16 and 1,22 for the combination of clones 11A2 and 11C2 either on the biotinylated or ruthenylated side. The combination 13G6 and 11C10 reflects ratios of 1,22 and 1,23, indicating high grade of performance independent of the orientation. A similar observation is made for the combination of 3H8 and 9E7 with ratios of 1,22 and 1,23 respectively but higher blank values compared to the combination of 11C10 and 13G6. The other combinations described achieve good ratios, yet only for one sandwich orientation.

TABLE-US-00023 TABLE 13 Antibody sandwich characteristics regarding maximum signal readout on El- ecsys ECLIA measuring cell for concentration levels of recombinant NT-proBMP10 (22- 312) at 0.01 ng/ml and the signal-to-noise ratio of 0.01 ng/ml divided by the blank value without spiked-in recombinant protein (S/N). Clone-IgG-Ru/Clone-IgG-Bi [ng/ml] 11A5/ 11A5/ 13G6/ 11C10/ 3H8/ 13G6/ 11C10/ 9E7/ 3H8/ 11C10/ BMP10 (312) 11C10 3H8 11C10 11A5 9E7 3H8 13G6 3H8 11C10 3H8 0.01 1147 1193 865,5 1012 1668 922 1005 1087.5 1756.5 1114.5 0 987.5 1015 712 827 1363 748 815 884.5 1430 859.5 S/N 1.16 1.18 1.22 1.22 1.22 1.23 1.23 1.23 1.23 1.30

[0677] Recognition of 20 native samples derived from healthy donors (Table 14) shows the range of detection for NT-proBMP10 in native serum samples, allowing to determine baseline values for clinical sample measurement. These are in a comparable range for both Ru- and Bi-orientations of 11C10 and 13G6. A similar range is detected by 11C10 and 11A5 also in both orientations.

TABLE-US-00024 TABLE 14 Antibody sandwich characteristics regarding signal readout on Elecsys ECLIA measuring cell for concentration levels of 20 native healthy human serum samples depicted are minimum, maximum, and median values. Clone-IgG-Ru/Clone-IgG-Bi 11C10/13G6 11C10/11A5 3G6/11C10 11A5/11C10 median 35546 42507 21968 31176.5 min 22175 20474 15786 23059 max 52515 63268 29712 44245

Example 14

Biomarker Measurement (Exemplary Method for the Detection of NT-proBMP10)

[0678] Serum and plasma concentrations of the biomarker NT-proBMP10 were measured with the commercialized Elecsys® reagents from Roche Diagnostics (Mannheim, Germany). The biomarker NT-proBMP10 was measured with prototype Elecsys® reagents from Roche Diagnostics (Mannheim, Germany).

[0679] NT-proBMP10-Assay using biotinylated rabbit MAB<NT-proBMP10>F(ab′)2-Bi and ruthenylated MAB<NT-proBMP10>-Ru:

[0680] An electrochemiluminescence immunoassay (ECLIA) for the specific measurement of NT-proBMP10 in particular in human serum or plasma samples was developed using the Elecsys® cobas analyzer e601. The Elecsys NT-proBMP10 immunoassay is an electrochemiluminescence im-munoassay (ECLIA) that functions via the sandwich principle. There are two antibodies included in the assay, namely a biotinylated monoclonal antibody F(ab′)2-fragment MAB<BMP10> (MAB<BMP10_22-312>Bi; capture antibody, such as 11C10) and a ruthenylated monoclonal anti-BMP10 antibody MAB<BMP10> (MAB<BMP10_22-312>-Ru; detection antibody, such as 13G6), which form sandwich immunoassay complexes with NT-proBMP10 in the sample. The complexes are then bound to solid-phase streptavidin-coated microparticles. These are captured magnetically onto the electrode surface, leading to chemiluminescence emission upon application of voltage to the electrode, which is measured by a photomultiplier. Results are determined via an instrument-specific calibration curve determined by series of 6 calibrators with different concentrations of NT-proBMP10 across the measuring range. Samples are measured applying assay protocol 2 with pipetting volumes of 20 ul of the sample, 75 ul of reagent 1 (R1), 75 μl of reagent 2 (R2) and 30 μl of magnetic beads. R1 is containing MAB<BMP10>F(ab′)2-Bi in phosphate reaction buffer and reagent 2 (R2) is containing MAB<BMP10>-Ru in the same reaction buffer.

Example 15

Prediction of Silent Brain Infarcts (LNCCI and SNCI) Based on Circulating BMP-10 Levels

[0681] BMP-10 in the assessment of silent brain infarcts provides a method to [0682] 1. Predicting the risk of silent brain infarcts in patients with atrial fibrillation based on circulating BMP-10 levels in serum/plasma (SWISSAF study, Table 15+16) [0683] 2. Improving the prediction of clinical accuracy of clinical stroke risk scores for silent brain infarcts based on circulating BMP-10 levels in serum/plasma (e.g. CHA.sub.2DS.sub.2-VASc, CHADS2 score) (SWISS AF study, Table 17)

[0684] The ability of circulating BMP-10 to predict the risk for the occurrence of silent infarcts was assessed in the SWISS AF study (Conen D., Forum Med Suisse 2012; 12:860-862; Conen et al., Swiss Med Wkly. 2017; 147). Patients of the SWISS AF cohort have a median age of 74 years, a rate of prior clinical strokes or TIA of 20%, a rate of vascular diseases of 34% and a history of diabetes of 17%.

[0685] BMP-10 was measured in the complete SWISS AF study with a pre-commercial assays were used for bone morphogenetic protein 10 (BMP-10) (high-throughput Elecsys® immunoassays; Roche Diagnostics, Mannheim, Germany). For detection of BMP-10 a sandwich-immunoassay was developed for the cobas Elecsys® ECLIA platform.

[0686] Because the estimates from a naïve proportional hazard model on the case control cohort would be biased (due to the altered proportion of cases to controls), a weighted proportional hazard model was used. Weights are based on the inverse probability for each patient to be selected for the case control cohort. In order to get estimates for the absolute survival rates in the two groups based on the dichotomized baseline BMP-10 measurement (<=median vs>median) a weighted version of the Kaplan-Meier plot was created.

[0687] In order to assess the ability of BMP-10 to improve existing risk scores for the prognosis of stroke the CHADS.sub.2 the CHA.sub.2DS.sub.2-VASc and the ABC score were extended by BMP-10 (log 2 transformed). Extension was done by creating a portioned hazard model including BMP-10 and the respective risk score as independent variables.

[0688] The c-indices of the CHADS.sub.2, the CHA.sub.2DS.sub.2-VASc and ABC score were compared to the c-indices of these extended models. For the calculation of the c-index in the case-cohort setting a weighted version of the c-index was used as proposed in Ganna (2011).

[0689] Results

TABLE-US-00025 TABLE 15 Significant altered circulating levels of BMP-10 in patients with brain lesions (SWISS AF study). Brain lesions include LNCCI and SNCI. Values are median (1.sup.st; 3.sup.rd Quartile). Any brain No Yes All lesions on bMRI n = 1078 n = 631 p n = 1709 BMP-10 ng/ml 2.17 2.32 <0.001 2.23 (1.87; 2.51) (1.98; 2.73) (1.91; 2.60)

[0690] Patients with LNCCI or SNCI on the bMRI were older (75.0 vs 68.1 years, p<0.0001), had more often permanent AF (28.4 vs 17.8%, p=0.0002), higher systolic BP levels (136.7 vs 131.3 mmHg, p<0.0001) and a higher CHA2DS2-VASc score (3.2 vs 2.1 points, p<0.0001), but showed no difference in the rate of oral anticoagulation (90.3 vs 88.5%, p=0.32). As shown in Table 15, BMP-10 has significantly higher levels in patients with brain lesions.

[0691] As demonstrated in Table 15, the risk of silent brain infarcts in patients with atrial fibrillation can be assessed based on circulating BMP-10 levels in serum/plasma.

TABLE-US-00026 TABLE 16 Significant multivariable-adjusted hazard ratios (HR) (95% confidence intervals (CI) for BMP-10 associated with silent infarcts (presence of LNCCI and SNCI) Biomarkers Presence of LNCCI Bone morphogenetic protein 10 OR (95% CI) p-value Model 1 1.73 (1.17; 2.56) 0.006 Model 2 1.57 (1.05; 2.34) 0.028

[0692] Model 1 was adjusted for age and sex.

[0693] Model 2 was additionally adjusted for systolic blood pressure, prior major bleeding, diabetes, peripheral vascular disease, BMI, smoking status, use of oral anticoagulation and antiplatelet medication.

[0694] Biomarkers were logarithmized.

[0695] As shown in Table 16, BMP-10 was significantly associated with LNCCI after multivariable adjustments for age and sex (Model 1) or for age, sex, systolic blood pressure, prior major bleeding, diabetes, peripheral vascular disease, BMI, smoking status, use of oral anticoagulation and antiplatelet medication.

[0696] Therefore the risk of silent brain infarcts in patients with atrial fibrillation can be assessed based on circulating BMP-10 levels in serum/plasma.

TABLE-US-00027 TABLE 17 Significant improvement of the CHAD2DS2-VASc score with BMP-10 for the relation to large non-cortical infarcts. Predictor variables were logarithmized biomarkers in addition to CHADS2-VA2SC score, the outcome variable was presence/ absence of large non-cortical and cortical infarcts. Presence of LNCCI or SNCI Biomarkers AUC CHA2DS2-VASc 0.696 (0.67-0.721)  CHA2DS2-VASc + 0.699 (0.673-0.724) Bone morphogenetic protein 10

[0697] When we added individual biomarkers to the CHA2DS2-VASc score, the AUC (95% CI) was improved by BMP-10 0.699 (0.673-0.724) as demonstrated in Table 17.

[0698] The combination of BMP-10 with clinical parameters of the CHA2DS2-VASC score well predicted clinically silent brain infarcts and outperformed the CHA2DS2-VASc score. Early clinical identification of patients at risk of cognitive decline might allow for better diagnostic and preventive measures.

Example 16

Prediction of White Matter Lesions Based on Circulating BMP-10 Levels

[0699] Data in the SWISS-AF data shows that BMP-10 correlates with existence of large non-cortical and cortical infarcts (LNCCI) in patients.

[0700] The extent of matter lesions can be expressed by the Fazekas score (Fazekas, J B Chawluk, A Alavi, H I Hurtig, and R A Zimmerman American Journal of Roentgenology 1987 149:2, 351-356). The Fazekas score is ranging from 0 to 3. 0 indicates no WML, 1 mild WML, 2 moderate WML and 3 severe WML.

[0701] In order to compare the association of BMP-10 with large non-cortical and cortical infarcts (LNCCI) patients were classified in two groups, Fazekas Score <2 (no) vs Fazekas Score ≥2 (yes). FIG. 14 shows that BMP-10 is increased in patients with moderate or severe WMLs versus patients with mild or no WMLs.

[0702] WML extent can be caused by clinical silent strokes (Wang Y, Liu G, Hong D, Chen F, Ji X, Cao G. White matter injury in ischemic stroke. Prog Neurobiol. 2016; 141:45-60. doi:10.1016/j.pneurobio.2016.04.005). This further advocates the usefulness of BMP-10 to predict the risk for clinical stroke.

[0703] The ability of circulating BMP-10 to discriminate between patients with Fazekas Score <2 (no) versus Fazekas Score ≥2 (yes) is indicated by the AUC of 0.62. White matter changes in the brain of dementia patients. Advanced age and changes in WML scores have been described to be associated with severity of dementia in Alzheimers disease patients (Kao et al., 2019).

[0704] Age is also an important predictor of clinical stroke. Therefore, it is plausible that data of significantly increased BMP-10 levels in the circulation indicate not only moderate or severe large non-cortical and cortical infarcts (LNCCI), but also indicate age related brain diseases, e.g. vascular dementia.