BIOMARKER COMPOSITIONS SPECIFIC TO CORONARY HEART DISEASE PATIENTS AND USES THEREOF

Abstract

The present invention relates to a disease-specific metabolite profile, and particularly to a biomarker composition obtained by screening from blood plasma-specific profiles of coronary heart disease subjects. The present invention also relates to a use of the biomarker compositions in risk assessment, diagnosis, early diagnosis, or pathological staging of coronary heart disease, and to a method for risk assessment, diagnosis, early diagnosis, or pathological staging of coronary heart disease. The biomarker composition as provided by the present invention can be used for early diagnosis of coronary heart disease and has high sensitivity, good specificity and good application prospects.

Claims

1. A biomarker composition, comprising at least one or more selected from the following Biomarkers 1 to 6: Biomarker 1, which has a mass-to-charge ratio of 310.04±0.4 amu, and a retention time of 611.25±60 s; Biomarker 2, which has a mass-to-charge ratio of 311.05±0.4 amu, and a retention time of 611.26±60 s; Biomarker 3, which has a mass-to-charge ratio of 220.00±0.4 amu, and a retention time of 122.77±60 s; Biomarker 4, which has a mass-to-charge ratio of 247.09±0.4 amu, and a retention time of 146.37±60 s; Biomarker 5, which has a mass-to-charge ratio of 255.03±0.4 amu, and a retention time of 117.92±60 s; and Biomarker 6, which has a mass-to-charge ratio of 170.03±0.4 amu, and a retention time of 202.18±60 s.

2. The biomarker composition according to claim 1, comprising at least Biomarkers 1 to 3 and 6.

3. The biomarker composition according to claim 1, comprising Biomarkers 1 to 6.

4. A reagent composition, comprising a reagent for detecting the biomarker composition according to claim 1.

5-7. (canceled)

8. A method for risk assessment, diagnosis, early diagnosis or pathological staging of coronary heart disease, comprising a step of determining content of each biomarker of the biomarker composition according to claim 1 in a sample of a subject.

9. The method according to claim 8, wherein a liquid chromatography-mass spectrometry method is used for determining content of each biomarker of the biomarker composition of claim 1 in a sample of a subject.

10. The method according to claim 8, wherein the method further comprises a step of establishing a training set for contents of the biomarker composition in samples of a coronary heart disease subject and a normal subject.

11. The method according to claim 10, wherein the training set is established by using a multivariate statistical classification model.

12. The method according to claim 11, wherein the training set comprises data as shown in Table 2.

13. The method according to claim 8, wherein the method further comprises a step of comparing the content of each biomarker of the biomarker composition of a subject to the data of the training set, and the training set is for contents of the biomarker composition in samples of a coronary heart disease subject and a normal subject.

14. The method according to claim 13, wherein the training set is established by using a multivariate statistical classification model.

15. The method according to claim 14, wherein the training set comprises data as shown in Table 2.

16. The method according to claim 13, wherein the step of comparing the content of each biomarker is carried out by using a receiver operating characteristic curve.

17. The method according to claim 16, wherein the result from the step of comparing the content of each biomarker is interpreted by a method comprising: if a subject is assumed to be a non-coronary heart disease subject, and his probability of non-coronary heart disease diagnosed by ROC is less than 0.5 or his probability of coronary heart disease diagnosed by ROC is greater than 0.5, the subject is determined to have a high probability or a higher risk of coronary heart disease, or is diagnosed as a patent with coronary heart disease.

18-27. (canceled)

28. The method according to claim 8, wherein the sample is blood plasma or whole blood.

29. The method according to claim 11, wherein the multivariate statistical classification model is a random forest model.

30. The biomarker composition according to claim 2, further comprising Biomarker 4 and/or Biomarker 5.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0076] FIG. 1 shows total ion chromatograms of mass spectrometry for coronary heart disease group (a) and normal group (b).

[0077] FIG. 2 shows PLS-DA score plots, in which prisms (white) represent normal group, triangles (black) represent coronary heart disease group.

[0078] FIG. 3 shows a loading-plot of principal components, in which triangles (black) represent variables with VIP value greater than 1.

[0079] FIG. 4 shows a Volcano-plot, in which differential metabolites are located above the horizontal dotted line, wherein the materials (black triangles) on the ambilateral sides of the two vertical dashed lines are metabolites with fold-change greater than 1.2 and Q-value less than 0.05, and the materials (gray spheres) between the two vertical dashed lines are metabolites with fold-change less than 0.8 and Q-value less than 0.05.

[0080] FIG. 5 shows S-plot, in which prisms (black) represent variables with VIP greater than 1.

[0081] FIG. 6 shows score graphs for analysis of principal components, in which prisms (white) represent the normal group, triangulars (black) represent the coronary heart disease group, and the analysis of principal components is performed by analyzing 83 test set data using the disclosed markers.

[0082] FIG. 7 shows ROC diagram of random forest model (Random forest model), in which Training ROC is based on the training set, AUC=1; and Test ROC is based on the test set, AUC=1.

[0083] FIG. 8 shows ROC test set diagram, in which mass-to-charge ratios 310.04 and 311.05 are randomly removed from the training set, AUC=0.8851.

[0084] FIG. 9 shows diagram for random combinations of 6 potential markers, in which the left side of the vertical line mark gives 4 markers that need to be tested at least.

SPECIFIC MODELS FOR CARRYING OUT THE INVENTION

[0085] While the embodiments of the present invention will be described in detail with reference to the following examples, it will be understood by those skilled in the art that the following examples are intended to be illustrative of the invention and are not to be taken as limiting the scope of the invention. In the examples, when specific conditions are not given, conventional conditions or conditions recommended by the manufacturer are employed. The used reagents or instruments which manufacturers are not given are all conventional products commercially available in the markets.

[0086] The blood plasma samples of coronary heart disease and normal subjects in the present invention are from the Guangdong General Hospital.

Example 1

[0087] 1.1 Collection of samples: morning blood samples of volunteers were collected, immediately placed and stored in −80° C. low temperature refrigerator. A total of 52 blood samples were collected from the normal group and 40 blood samples were collected from the coronary heart disease group.

[0088] 1.2 Treatment of samples: frozen samples were thawed at room temperature, 500 μL of each blood plasma sample was taken and placed in 2.0 mL centrifuge tube, added with 1000 μL of methanol for dilution, centrifuged at 10000 rpm for 5 min, for standby.

[0089] 1.3 Analysis by Liquid Chromatography-Mass Spectrometry

[0090] Instrument and Equipment

[0091] HPLC-MS-LTQ Orbitrap Discovery (Thermo, Germany)

[0092] Chromatographic Conditions

[0093] Column: C18 column (150 mm×2.1 mm, 5 μm); Solvent A was 0.1% (v/v) formic acid/water, and solvent B was 0.1% (v/v) formic acid/methanol; gradient elution program: 0˜3 min, 5% B, 3˜36 min, 5%˜80% B, 36˜40 min, 80%˜100% B, 40˜45 min, 100% B, 45˜50 min, 100%˜5% B, 50˜60 min, 5% B; flow rate: 0.2 mL/min; injection volume: 20 μL.

[0094] Mass Spectrometry Conditions

[0095] ESI ion source, positive ion mode for data acquisition, the mass scanning range was 50˜1000 mass-to-charge (m/z). Ion source parameters ESI: sheath gas was 10, auxiliary air was 5, capillary temperature was 350° C., spray voltage was 4.5 KV.

[0096] 1.4 Data Processing

[0097] XCMS software (e.g., http://metlin.scripps.edu/xcms/) was used for peak detection and peak matching of raw data; and R software using PLS-DA (partial least squares-discriminant analysis) was used for pattern recognition analysis of differential variables of the metabolite profile of coronary heart disease group (FIG. 1a) and the metabolite profile of normal group (FIG. 1b), so as to establish PLS-DA mathematical model.

[0098] 1.5 Comparison and Determination of Characteristic Metabolite Profiles

[0099] The blood plasma metabolite profile of coronary heart disease patients (FIG. 1) was established by comparing the blood plasma metabolite profiles of the normal group and the coronary heart disease group. The results showed that there were significant differences in the blood plasma metabolite profiles between the normal group and the coronary heart disease group.

Example 2

[0100] 2.1 Sample collection: morning blood plasma samples of volunteers were collected, immediately placed and stored in −80° C. low temperature refrigerator. A total of 52 blood plasma samples were collected from the normal group and 40 blood plasma samples were collected from the coronary heart disease group.

[0101] 2.2 Sample treatment: frozen samples were thawed at room temperature, 500 μL of each blood plasma sample was taken and placed in 2.0 mL centrifuge tube, added with 1000 μL of methanol for dilution, centrifuged at 10000 rpm for 5 min, for standby.

[0102] 2.3 Analysis by Liquid Chromatography-Mass Spectrometry

[0103] Instrument and Equipment

[0104] HPLC-MS-LTQ Orbitrap Discovery (Thermo, Germany)

[0105] Chromatographic Conditions

[0106] Column: C18 column (150 mm×2.1 mm, 5 μm); mobile phase A: 0.1% formic acid aqueous solution, mobile phase B: 0.1% formic acid in acetonitrile solution; gradient elution program: 0˜3 min, 5% B, 3˜36 min, 5%˜80% B, 36˜40 min, 80%˜100% B, 40˜45 min, 100% B, 45˜50 min, 100% 5% B, 50˜60 min, 5% B; flow rate: 0.2 mL/min; injection volume: 20 μL.

[0107] Mass Spectrometry Conditions

[0108] ESI ion source, positive ion mode for data acquisition, scanning mass m/z 50˜1000. Ion source parameters ESI: sheath gas was 10, auxiliary air was 5, capillary temperature was 350° C., cone hole voltage was 4.5 KV.

[0109] 2.4 Data Processing

[0110] XCMS software was used for relevant pretreatment of raw data to obtain a two-dimensional matrix data, and wilcox-test was used to statistically determine significant differences of peaks of metabolites; and PLS-DA (partial least squares-discriminant analysis) was used for pattern recognition analysis of differential variables of the metabolite profile of coronary heart disease group (FIG. 1a) and the metabolite profile of normal group (FIG. 1b), and potential biomarkers were screened out by VIP, Volcano-plot and S-plot in combination.

[0111] 2.5 Metabolic Profile Analysis and Potential Biomarkers

[0112] 2.5.1 Orthogonal Partial Least Squares Discriminant Analysis (PLS-DA)

[0113] PLS-DA method was used to distinguish the normal group and the coronary heart disease group, and potential markers were further screened by VIP values (Loading-plot for principal component analysis) (FIG. 3), Volcano-plot (FIG. 4) and S-plot (FIG. 5). It was shown in FIG. 3 and FIG. 4 that there were significant different metabolites in the normal group and coronary heart disease group. As shown in FIG. 5, each point in the S-plot represented a variable, and the S-plot graph showed the relevance of the variable to the model. The black prism-tagged variable was a variable with VIP greater than 1, which had a large deviation and a good correlation with the model (see FIG. 2 and FIG. 5).

[0114] 2.5.2 Potential Biomarkers

[0115] The potential markers were screened according to the VIP values of the PLS-DA model for pattern cognition. The variables with VIP values greater than 1 were extracted from the PLS-DA model, and variables with large deviation and relevance were further selected according to Loading-plot, Volcano-plot and S-plot, and 6 potential biomarkers were obtained by further combining variables with P value of less than 0.05 and Q value of less than 0.05, which were shown in Table 1.

TABLE-US-00001 TABLE 1 Potential biomarkers Ratio (normal Mass-to- Retention group/coronary charge time, Rt heart disease P Q VIP ratio (amu) (sec) group) value value value 310.04 611.25 0.0001 2.17E−15 1.92E−12 1.16 311.05 611.26 0.0025 1.35E−12 8.20E−11 1.16 220.00 122.77 0.6414 3.86E−02 2.75E−02 1.01 247.09 146.37 0.3984 5.61E−08 3.37E−07 1.24 255.03 117.92 0.3484 3.83E−06 1.21E−05 1.26 170.03 202.18 0.0156 3.79E−03 4.08E−03 4.54

[0116] 2.5.3 Principal Component Analysis (PCA)

[0117] PCA is a non-supervised pattern recognition method that can visually describe differences between samples in multidimensional space. PCA analysis was performed on 83 samples of the obese group and control group using the resultant six differential markers. It can be seen from FIG. 6, in the PCA model, the two groups were substantially divided in the first principal component orientation, indicating there were significant differences in blood plasma metabolic profiles between the normal group and the coronary heart disease group.

[0118] 2.5.4 Receiver Operating Characteristic Curve (ROC)

[0119] The six potential markers were discriminated in the normal group and the coronary heart disease group by using a random forest model (Random Forest).sup.[7] and receiver operating characteristic curve (ROC).sup.[8]. The data of peak areas of 92 metabolite profiles of the normal group and the coronary heart disease group were selected and used as training set via ROC modeling (see references [7] and [8]) (Table 2). In addition, 83 test samples (including 38 coronary heart disease samples and 45 normal control samples) were selected as test set. The test results showed AUC=1, FN (false negative)=0, FP (false positive)=0 (FIG. 7). Thus, the present invention has high accuracy and specificity, and has good prospects to be developed as a diagnosis method to provide a basis for diagnosis of coronary heart disease.

TABLE-US-00002 TABLE 2 Peak area data of training set metabolite profiles Group (1: Coronary heart disease group; Mass-to-charge ratio Sample No. 0: normal group) 310.0474 311.0511 220.0088 247.0927 255.0378 170.0328 CD5751_1 1 0.035971 0.028941 2.0698 1.847557 3.447373 1.838497 CD5767_1 1 0.518048 0.528351 1.67918 1.057417 1.427185 0.642363 CD5778_1 1 0.118419 0.114578 5.274362 1.636753 5.136574 0.856236 CD5779_1 1 0.830761 0.825451 1.981177 1.820087 0.678725 3.258685 CD5782_1 1 2.273 2.266954 2.183817 1.795206 2.589921 0.378109 CD5783_1 1 0.142664 0.143573 1.950855 1.546727 1.105065 1.232417 CD5788_1 1 0.297865 0.296303 0.80925 0.357977 0 0.163506 CD5795_1 1 5.468021 5.71178 1.796937 1.627936 2.73744 0.247029 CD5796_1 1 0.218944 0.217782 3.1631 1.449651 2.229824 0.525126 CD5797_1 1 1.4614 1.436343 3.226244 0.713308 2.545898 0.692436 CD5805_1 1 0.79711 0.813508 2.906224 1.529678 0.3218 1.169494 CD5814_1 1 2.094091 2.036124 1.027309 0.67968 0.706373 0.526758 CD5816_1 1 2.73073 2.79561 1.510684 1.580599 1.310362 0.566224 CD5819_1 1 0.001538 0.01335 1.591497 1.396263 0.941163 0 CD5822_1 1 3.744198 3.809473 2.267795 1.429436 2.422802 0.086586 CD5831_1 1 3.260472 3.348123 2.334711 1.218704 3.769648 1.490568 CD5832_1 1 0.028251 0.024733 1.883654 1.928723 4.108324 0.871313 CD5833_1 1 3.207548 3.22654 1.264351 0.833532 5.375295 0.26837 CD5838_1 1 0.672543 0.669593 2.269154 1.465057 1.48128 1.196006 CD5851_1 1 2.225022 2.232207 2.231923 1.564476 1.626777 1.065598 CD5860_1 1 3.588629 3.619216 1.623603 0.349931 1.466329 2.37164 CD5863_1 1 0.01132 0.011593 2.806604 1.334728 1.423662 0.469683 CD5867_1 1 0.096234 0.095499 1.864931 1.113036 4.784302 0.359933 CD5871_1 1 3.275621 3.38236 1.351764 0.405 2.682353 0.089361 CD5877_2 1 2.900862 2.903243 6.918055 0.863341 1.525894 0.248852 CD5881_2 1 4.546949 4.59075 6.97697 0.302554 1.321564 1.111718 CD5884_2 1 0.457603 0.461464 9.928549 2.066072 12.84567 0.449829 CD5891_2 1 0.098234 0.106791 0.84521 1.595202 2.156983 0.800175 CD5892_2 1 0 0.000933 5.960872 1.560398 3.367482 0.560264 CD5898_2 1 0.000153 0.004361 0.817571 1.463604 1.986249 0.713979 CD5900_2 1 2.296194 2.291767 6.686202 0.650392 1.617971 0.69235 CD5916_2 1 0.229847 0.112529 7.224221 1.473773 1.725251 0.841159 CD5923_2 1 1.470656 1.481594 11.02546 1.473311 1.955908 1.359043 CD5925_2 1 0.000292 0 5.76178 1.351079 1.62615 0.127491 CD5926_2 1 0.000984 0.002923 0.668754 1.208468 1.655 0.554825 CD5931_2 1 3.045312 3.090104 0.555586 0.246728 3.868446 0.107213 CD5934_2 1 2.761015 2.807205 0.932432 1.439827 2.482806 0.332513 CD5935_2 1 2.59472 2.570944 6.481325 1.600715 1.829884 0.653901 CD5988_2 1 0.074743 0.076354 8.927902 0.287215 4.092279 0.856305 CD5990_2 1 0.528431 0.531329 7.455305 2.119379 8.704949 0.402132 N165E_2 0 0 0.001198 2.273735 0.219128 1.404477 0.468192 N167E_2 0 0.000153 0.003134 0.330128 1.427848 2.519408 0.947302 N168E_2 0 0 0.004276 0.532362 1.193247 2.307861 0.635855 N170E_2 0 0.000123 0.006678 1.865563 0.62836 1.752023 0.876193 N171E_2 0 0.000108 0.002391 0.69803 1.43365 1.397692 0.506712 N185E_2 0 0 0.000993 0.556966 1.31429 1.097642 0.570415 N186E_2 0 0 0.000913 0.315125 1.586638 0.800838 0.930858 N187E_2 0 0 0.004406 2.299563 0.286376 1.737154 1.048589 N190E_2 0 0.0043 0.002773 1.338285 1.36053 1.854572 1.10703 N191E_2 0 0 0.002541 4.508579 1.504058 4.033288 0.69268 N195E_2 0 0 0.004779 0.385074 0.251441 0.94138 1.314871 N197E_2 0 0 0.002084 2.091143 1.331258 1.345158 0.77953 N198gan_huruilian_2 0 0.000116 0.003454 4.933436 0.018222 0 1.500137 N199gan_linrufang_2 0 0 0.001958 2.272987 0.022897 0 1.087806 N200E_2 0 0 0.000896 0.427487 0.433903 2.398669 0.340811 N201gan_lvhuiX_2 0 0 0.002775 2.366315 0.013556 0.035438 0.362495 N203gan_1 0 0 0.003349 0.887285 0.078262 0 3.139258 N204gan_wangmiaorong_1 0 0.000167 0.00143 0.909609 0.06722 0.104947 2.562048 N205gan_liuqifang_1 0 0 0 0.853212 0.049627 0.049099 3.671745 N206gan_liuguoying_1 0 0.000268 0.002434 0.860768 0.071575 0 3.307561 N207E_2 0 0.001081 0.002212 0.662516 0.635003 2.114432 0.689146 N208gan_zhengshuX_2 0 0 0.00256 4.877083 0.023979 0 0.575836 N209gan_wangxin_2 0 9.43E−05 0.00151 1.282423 0.013061 0 1.529018 N212E_1 0 0.000208 0.006068 1.451385 1.210609 0.883391 2.824106 N213E_1 0 0 0.003607 0.213409 1.329888 0.812546 0.567079 N214E_1 0 0.000194 0.002669 0.345 1.358322 0.37904 2.964935 N215E_1 0 0 0.001307 0.616592 2.579593 1.53582 0.878895 N217E_1 0 0.000261 0.002067 1.393546 1.469455 0.781215 2.802722 N218E_1 0 0.000154 0 1.002086 0.994796 1.424977 0.515163 N220E_1 0 0 0.002992 0.836052 1.463508 0.756875 3.049891 N222E_1 0 0 0.00519 1.476197 1.434959 0.853225 1.892389 N223E_1 0 0 0.002726 0.994038 1.259697 0.732209 2.262991 N226E_1 0 0 0.033521 0.245588 1.383716 1.351605 0.71867 N227E_1 0 0 0.006596 1.112362 1.38151 1.48167 0.675359 N228E_1 0 0.001776 0.001104 2.128 0.720528 0.740941 1.334337 N229E_1 0 0 0.006393 1.382065 0.957774 0.537195 2.750898 N231E_1 0 0 0.002535 0.864953 1.816202 2.984069 2.654364 N232E_1 0 0 0.002986 0.396396 0.411945 0.42109 4.548935 N233E_1 0 0 0.006421 1.02865 0.464872 2.490481 1.610518 N234E_1 0 0 0.004642 1.071868 1.526404 1.642752 0.63371 N235E_1 0 0 0 1.634843 0.480648 1.674944 0.985628 N236E_1 0 0 0.004328 1.166732 0.867571 1.2451 1.58132 N237E_1 0 0.000284 0.002743 1.382313 0.273098 0.619799 5.461519 N238E_1 0 0 0.002365 0.138054 1.218771 0.687922 1.281134 N239E_1 0 0.000128 0.005445 3.55656 1.314373 0.736699 0.753425 N241E_1 0 0.000128 0.00495 1.320156 0.331016 0.915875 1.332363 N242E_1 0 0.000169 0.003374 1.959169 0.166331 0.319434 1.013391 N243E_1 0 0 0.001174 1.286338 1.603114 1.067129 2.545529 N244E_1 0 0 0.020157 0.885433 1.110168 1.142692 1.858838 N245E_1 0 0.000156 0.001082 0.237614 0.360275 0.671792 2.873411 N247E_2 0 0.000107 0.004216 0.247196 1.321152 0.972867 0.926548 N248E_2 0 0 0.001756 0.264963 1.250571 0.716003 1.141921

[0120] Using the random forest model to calculate the classification ability of the six potential biomarkers for the obese group and the normal group, the results of the classification ability (arranged from high to low) were shown in Table 3. The markers in the table should be tested using at least above 4 markers (FIG. 9), so as to maintain high sensitivity and specificity.

TABLE-US-00003 TABLE 3 Classification ability of potential biomarkers Metabolite Interpreting Interpreting Mean Mean (mass-to-charge value of value of Decrease Decrease ratio) (amu) normal group obese group Accuracy Gini 310.04 0.168808 0.203038 0.180087 17.56405 311.05 0.111137 0.137121 0.119923 14.38203 220.00 0.023173 0.020618 0.021529 3.634677 170.03 0.017363 0.015979 0.016477 3.425181 255.03 0.010633 0.005938 0.008497 3.596366 247.09 0.007744 0.006723 0.007192 2.095741

[0121] If mass-to-charge ratios, such as 310.04 and 311.05, were randomly removed from the training set, the resultant ROC test set (the above 83 test set samples) had AUC=0.8851, AUC decreased significantly, FN=0.184 and FP=0.200, FN and FP significantly increased (FIG. 8), which indicated the ability for diagnosis of coronary heart disease decreased.

REFERENCES

[0122] [1] Finegold, J A; Asaria, P; Francis, D P. Mortality from ischaemic heart disease by country, region, and age: Statistics from World Health Organisation and United Nations. International journal of cardiology. 4 Dec. 2012, 168 (2): 934-45. [0123] [2] World Health Organization Department of Health Statistics and Informatics in the Information, Evidence and Research Cluster. The global burden of disease 2004 update. Geneva: WHO. 2004. ISBN 92-4-156371-0. [0124] [3] Elizabeth Barrett-Connor. Gender differences and disparities in all-cause and coronary heart disease mortality: epidemiological aspects. Best Pract Res Clin Endocrinol Metab. 2013 August; 27(4):481-500. [0125] [4] Madjid M, Willerson J T. Inflammatory markers in coronary heart disease. Br Med Bull. 2011; 100:23-38. doi: 10.1093/bmb/ldr043. Epub 2011 Oct. 18. [0126] [5] Spoletini Il, Vitale C, Rosano G M. Biomarkers for predicting postmenopausal coronary heart disease. Biomark Med. 2011 August; 5(4):485-95. doi: 10.2217/bmm.11.51. [0127] [6] Kishore Kumar Pasikanti, Kesavan Esuvaranathan, Paul C. Ho, et al. Noninvasive urinary metabonomic diagnosis of human bladder cancer. Journal of Proteome Research, 2010, 9, 2988-2995. [0128] [7] Liaw, Andy & Wiener, Matthew. Classification and Regression by randomForest, R News (2002), Vol. 2/3 p. 18. [0129] [8] Jianguo Xia, David I. Broadhurst, Michael Wilson, David S. Wishart. Translational biomarker discovery in clinical metabolomics: an introductory tutorial. Metabolomics (2013) 9:280-299.