Metabolic marker for diagnosing diabetic secondary osteoporosis and application thereof
By screening metabolic markers such as cytosine nucleotides, ketoglutarate, and triethylamine, and combining them with exosome detection and prediction models, the diagnostic bias of diabetic secondary osteoporosis has been resolved, achieving highly accurate risk assessment and early identification.
Patent Information
- Application Number
- CN202511983034.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-26
- Publication Date
- 2026-01-27
- Estimated Expiration
- 2045-12-26
Smart Images

Figure CN121410030A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of biotechnology, and in particular to a metabolic biomarker for diagnosing diabetic secondary osteoporosis and its application. Background Technology
[0002] Prolonged hyperglycemia can impair the body's metabolic function, leading to decreased bone mass and osteoporosis, also known as diabetic osteoporosis (DOP), which increases the risk of fractures. The prevalence of osteoporosis in type 1 diabetes patients is approximately 30-50%, and their risk of hip fracture is 4.4-6 times higher than that of non-diabetic individuals. The prevalence of osteoporosis in type 2 diabetes patients is approximately 20-30%, and compared to non-diabetic individuals, the relative risk of fractures at different sites is 1.17 to 2.03.
[0003] Diagnosis of diabetic osmosis (DOP) currently relies primarily on bone mineral density (DXA) testing, but this method has significant limitations: Diabetic bone damage is centered on "deteriorating bone quality" (disrupted bone material properties and increased cortical porosity), while DXA only assesses bone mineral density, inevitably underestimating the fracture risk caused by abnormal bone quality. This deficiency is particularly pronounced in patients with type 2 diabetes, manifesting as the classic "bone mineral density paradox"—although bone mineral density values may be normal or elevated, actual bone quality is decreased, and the risk of fracture is significantly increased. Furthermore, commonly used fracture risk assessment tools neither consider the specific bone damage caused by diabetes nor treat diabetes as an independent risk factor, further leading to biased risk assessments. Therefore, developing novel biomarkers for DOP screening, establishing early diagnostic models, and identifying high-risk individuals are crucial for controlling bone metabolic disorders in diabetic patients, slowing later bone quality deterioration, and reducing fracture risk. Summary of the Invention
[0004] Therefore, the technical problem to be solved by the present invention is to overcome the problem of risk assessment bias in the existing technology of DOP diagnosis.
[0005] To address the aforementioned technical problems, this invention provides metabolic biomarkers for diagnosing diabetic secondary osteoporosis and their applications. First, this invention screens for significantly different metabolites between diabetic patients with osteoporosis (DOP) and those without osteoporosis (DM). Second, it screens for significantly different metabolites between patients with primary osteoporosis and healthy controls. After excluding the overlap of these two groups, the resulting metabolites exclude the influence of diabetic patients without osteoporosis and those with primary osteoporosis. Further screening yielded the metabolic biomarkers of this invention—cytosine nucleotide (CMP), oxoglutaric acid, and triethylamine. These three biomarkers, used individually, have an AUC > 0.8, and when used in combination, their AUC is 0.99. This provides a new diagnostic approach for the diagnosis of DOP in diabetic patients, early screening, and prevention of secondary fractures in high-risk groups, effectively overcoming the limitations of traditional bone mineral density testing in identifying osteoporosis in diabetic patients.
[0006] The first objective of this invention is to provide the application of a reagent for detecting the content of metabolic markers in the preparation of a diagnostic kit for diabetic secondary osteoporosis, wherein the metabolic markers are selected from one or more of cytosine nucleotides, ketoglutarate, and triethylamine.
[0007] Furthermore, the test sample for the diagnostic kit is an exosome.
[0008] A second objective of this invention is to provide a diagnostic kit for diabetic secondary osteoporosis, the diagnostic kit containing reagents for detecting the levels of metabolic markers, wherein the metabolic markers are those described above.
[0009] Furthermore, the diagnostic kit also includes reagents for extracting exosomes from plasma.
[0010] Furthermore, the reagents include those for detecting the concentration or content of the metabolic markers in the sample to be tested by nuclear magnetic resonance, chromatography, spectroscopy, mass spectrometry, or a combination thereof.
[0011] A third objective of this invention is to provide a reagent for detecting the concentration or content of metabolic markers in a test sample, wherein the metabolic markers are those described above, and the test sample is an exosome.
[0012] A fourth objective of this invention is to provide an application of a metabolic biomarker in the preparation of diagnostic products for diabetic secondary osteoporosis, wherein the metabolic biomarker is the aforementioned metabolic biomarker.
[0013] Furthermore, the diagnostic product includes a device for analyzing the levels of the metabolic markers.
[0014] The fifth objective of this invention is to provide an application of a metabolic biomarker in constructing a predictive model for diabetic secondary osteoporosis, wherein the metabolic biomarker is the aforementioned metabolic biomarker.
[0015] Furthermore, the prediction model includes the step of analyzing the concentration level data of metabolic biomarkers using one or more algorithms selected from orthogonal partial least squares discriminant analysis, partial least squares discriminant analysis, classification and logistic regression, k-nearest neighbor algorithm, Naive Bayes, support vector machine, decision tree, random forest, regression tree, gradient boosting decision tree, extreme gradient boosting algorithm, lightweight gradient boosting machine, gradient boosting machine, lasso algorithm and convolutional neural network.
[0016] Furthermore, the prediction model includes the following steps:
[0017] S1, a data receiving unit, wherein the data includes the concentration levels of metabolic markers in the sample, wherein the metabolic markers are selected from one or more of cytosine nucleotides, ketoglutarate, and triethylamine;
[0018] S2, Data Analysis Unit, wherein the data analysis unit includes analyzing the data in S1 using one or more of the following: orthogonal partial least squares discriminant analysis, partial least squares discriminant analysis, classification and logistic regression, k-nearest neighbor algorithm, Naive Bayes, support vector machine, decision tree, random forest, regression tree, gradient boosting decision tree, extreme gradient boosting algorithm, lightweight gradient boosting machine, gradient boosting machine, lasso algorithm and convolutional neural network.
[0019] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:
[0020] This invention has screened a group of metabolic biomarkers for diagnosing diabetic secondary osteoporosis, which demonstrate good diagnostic efficacy in this condition. Specifically, cytosine nucleotides, ketoglutarate, and triethylamine of this invention all exhibit an AUC greater than 0.8 when independently diagnosing diabetic secondary osteoporosis, and the combined AUC reaches as high as 0.99 when the three metabolic biomarkers are used in combination, demonstrating excellent predictive ability. The diagnostic biomarkers of this invention can eliminate the influence of metabolites in primary osteoporosis on diagnostic efficacy, significantly improving the accuracy and specificity of diagnosing diabetic secondary osteoporosis. Attached Figure Description
[0021] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.
[0022] Figure 1These are images for identifying exosome characteristics, where A is an electron micrograph of the exosome and B is a particle size distribution of the exosome.
[0023] Figure 2 This is a diagram showing the results of core metabolic biomarkers identified by SVM-RFE and RF models. In this diagram, A shows the relationship between the number of SVM-RFE features and the model error, with 20 features determined as the optimal number of features. B shows the ranking of important features selected based on the RF model. The features are arranged in descending order of the average reduction in Gini coefficient, with higher values indicating greater importance of the corresponding feature in the model's classification decision.
[0024] Figure 3 This is the intersection plot of key metabolites screened by SVM-RFE and RF models;
[0025] Figure 4 The ROC curve for N-acetyl-D-galactosamine is shown.
[0026] Figure 5 The ROC curve for N-acetyl-L-lysine is shown.
[0027] Figure 6 This is the ROC curve of cytosine nucleotides;
[0028] Figure 7 The ROC curve of docosahexaenoic acid amide is shown.
[0029] Figure 8 This is the ROC curve of pimelic acid;
[0030] Figure 9 This is the ROC curve of methylmalonic acid;
[0031] Figure 10 This is the ROC curve of pyridoxal phosphate;
[0032] Figure 11 This is the ROC curve of aldosterone;
[0033] Figure 12 This is the ROC curve of triethylamine;
[0034] Figure 13 This is the ROC curve of ketoglutarate;
[0035] Figure 14 This is the ROC curve of phycocyanin;
[0036] Figure 15 This is the ROC curve for cinconidine;
[0037] Figure 16 This is the ROC curve for oleic acid;
[0038] Figure 17This is a statistical graph showing the differential expression of three specific diagnostic marker metabolites (cytosine nucleotide, ketoglutarate, and triethylamine) between the case and control groups. Among them, AC is the statistical graph showing the differential expression in the group found in Example 1, and DF is the statistical graph showing the differential expression between the control group and the case group in Example 2.
[0039] Figure 18 These are receiver operating characteristic (ROC) curves for identifying DOP patients in a diabetic patient population using three specific diagnostic marker metabolites (cytosine nucleotide, ketoglutarate, and triethylamine). In Example 1, A is the ROC curve of using the three metabolic markers alone in the discovery group; B is the ROC curve of using the three metabolic markers in combination in the discovery group; C is the ROC curve of using the three metabolic markers alone in Example 2; and D is the ROC curve of using the three metabolic markers in combination in Example 2. Detailed Implementation
[0040] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.
[0041] Example 1: Screening of differentially metabolized plasma exosomes between patients with diabetic secondary osteoporosis and diabetic non-osteoporosis patients.
[0042] 1. Research subjects and research design
[0043] All participants obtained written informed consent before participating in the study. Venous blood was collected from participants in the morning on an empty stomach, plasma was separated, and stored at -80°C.
[0044] The subjects in this study included two groups: (a) the discovery group (diabetic patients group): 20 diabetic non-osteoporosis (DM) subjects formed the control group, and 20 diabetic secondary osteoporosis (DOP) subjects formed the case group; (b) the identification group (non-diabetic patients group): including 50 healthy subjects and 50 primary osteoporosis subjects.
[0045] The inclusion and exclusion criteria for participants are as follows:
[0046] (1) Women aged ≥65 years were included;
[0047] (2) The inclusion criteria for diabetes mellitus are typical symptoms of diabetes mellitus (polydipsia, polyphagia, polyuria, unexplained weight loss) + fasting plasma glucose (FPG) ≥7mmol / L;
[0048] (3) The inclusion criteria for osteoporosis were based on bone mineral density measured by dual-energy X-ray absorptiometry (T-value of femoral neck bone mineral density less than -2.5).
[0049] (4) Exclude patients with a history of fractures, liver and kidney dysfunction, thyroid dysfunction, chronic inflammation and infection, malignant tumors, etc.
[0050] 2. Isolation and identification of plasma exosomes
[0051] Exosomes were isolated from plasma samples using a plasma exosome affinity extraction kit. Particle size was calculated using nanoparticle tracking analysis, and exosome morphology was detected by transmission electron microscopy. The exosomes exhibited a typical "tea saucer" structure, with sizes mostly ranging from 60 nm to 80 nm (see...). Figure 1 ).
[0052] 3. Extraction of exosome metabolites
[0053] (1) Transfer the sample to a 2mL centrifuge tube and add 100mg of glass beads;
[0054] (2) Add 1.0 mL of a mixed solution of acetonitrile:methanol:H2O (2:2:1) and vortex for 30 s;
[0055] (3) Place the centrifuge tube in the instrument’s 2mL adapter, immerse it in liquid nitrogen for 5 minutes to freeze quickly, then thaw it at room temperature. Place the centrifuge tube in the 2mL adapter again, install it in the tissue homogenizer, and homogenize at 60Hz for 2 minutes. Repeat twice.
[0056] (4) Take out the centrifuge tube, process the sample at 12000 rpm and 4℃ for 10 min, collect the supernatant and then concentrate and dry it;
[0057] (5) The sample was reconstituted by adding 300µL of acetonitrile: 0.1% formic acid (1:9) to a 2-chloro-L-phenylalanine solution (2.5μg / mL), filtered through a 0.22μm membrane, and then detected by liquid chromatography-mass spectrometry (LC-MS).
[0058] 4. Metabolite LC-MS analysis
[0059] LC chromatographic conditions: An ACQUITYUPLC® HSST3 (2.1 × 100 mm, 1.8 µm) column was used, with a flow rate of 0.3 mL / min, a column temperature of 40 °C, and an injection volume of 2.0 μL. Positive ion mode was used, with a mobile phase of 0.1% formic acid acetonitrile (B1) and 0.1% formic acid water (A1). The gradient elution program was: 0–1 min, 8% B1; 1–8 min, 8%–98% B1; 8–10 min, 98% B1; 10–10.1 min, 98%–8% B1; 10.1–12 min, 8% B1. In negative ion mode, the mobile phase consisted of acetonitrile (B2) and 5 mM ammonium formate water (A2), with the gradient elution program as follows: 0-1 min, 8% B2; 1-8 min, 8%-98% B2; 8-10 min, 98% B2; 10-10.1 min, 98%~8% B2; 10.1-12 min, 8% B2.
[0060] MS mass spectrometry conditions: positive ion spray voltage 3.50 kV, negative ion spray voltage -2.50 kV, sheath gas 40 arb, auxiliary gas 10 arb. Capillary temperature 325℃. First-stage full scan at a resolution of 60,000 m / z, first-stage ion scan range 100-1000 m / z. Second-stage fragmentation was performed using high-energy collisional dissociation (HCD) at 30% collision energy, second-stage resolution 15,000 m / z. The first four ions acquired were fragmented, and unnecessary tandem mass spectrometry (MS / MS) information was removed using dynamic exclusion.
[0061] 5. Data Preprocessing
[0062] (1) The raw LC-MS data was converted to mzXML format by the MSConvert tool in Proteo Wizard software, and peak identification, peak extraction, peak alignment and integration were performed by the XCMS toolkit based on R language. Mass spectrometry peaks with relative standard deviation (RSD) <30% were removed by Compound Discover 3.1 software.
[0063] (2) Metabolites in the analyzed samples were identified and annotated by comparing secondary mass spectra with multiple databases such as KEGG, LIPIDMaps, HMDB, MassBank, MzClou and a self-built standard library.
[0064] (3) Multivariate statistical analysis: First, unsupervised principal component analysis (PCA) was used to observe the natural clustering trend and outliers among the samples to assess the overall data quality. Then, supervised partial least squares discriminant analysis (PLS-DA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were used to maximize the differences between groups and find the metabolite biomarkers that best distinguish each group. To effectively avoid overfitting problems that may occur during the modeling process, a permutation test was used, specifically performing 100 permutation operations (n=100) to enhance the stability and reliability of the model.
[0065] 6. Screening of metabolic biomarkers
[0066] (1) Perform a t-test on the data and calculate the variable importance projection (VIP) value and the adjusted significance probability (P) value of the independent variables between each group. FDR Based on P FDR The study used a threshold of <0.05 and VIP>1 to screen for significantly differentially expressed metabolites between diabetic patients with osteoporosis (DOP) and those without osteoporosis (DM). Secondly, it screened for significantly differentially expressed metabolites between patients with primary osteoporosis and healthy controls. After excluding the overlap between these two groups, 30 differentially expressed metabolites unique to the discovery group were identified. These potential DOP-specific metabolic biomarkers were included in subsequent analyses and further screening.
[0067] (2) To reduce data dimensionality, prevent model overfitting, and identify the most biologically significant biomarkers, this study employed two feature selection methods: Support Vector Machine-Recursive Feature Elimination (SVM-RFE) and Random Forest (RF). Cross-validation of these two methods was used to ensure the robustness of the selected feature set. Feature selection first employed SVM-RFE, recursively removing features that contributed least to the model to obtain the optimal feature subset. We used a linear kernel SVM as the base learner and utilized a cross-validation recursive feature elimination method (RFECV) with 10x cross-validation to automatically determine the optimal number of features. The CV error was minimized to 0.00158 when 20 features were included. Secondly, we used the RF algorithm to evaluate feature importance, setting the number of decision trees (ntree) to 500, with all other parameters using default values. The optimal feature subset determined by SVM-RFE cross-validation and the 18 features selected by Random Forest that significantly contributed to model accuracy were retained. Ultimately, 13 characteristic metabolites shared by the two sets were identified as key biomarkers for DOP (see [link to relevant documentation]). Figure 2-3 ).
[0068] 7. ROC Analysis
[0069] To further verify the specificity of the above 13 metabolites in the diagnosis of DOP, univariate receiver operating characteristic (ROC) curve analysis was performed on the biomarkers. The results are as follows: Figures 4-16 As shown, the area under the curve (AUC) of 12 out of the 13 metabolites was greater than 0.80 (excluding N-acetyl-L-lysine, whose AUC was less than 0.8), indicating diagnostic value. Cytosine nucleotides (AUC=0.98), ketoglutarate (AUC=0.91), and triethylamine (AUC=0.81) effectively distinguished the case group from the control group, demonstrating high accuracy and discriminative power.
[0070] Example 2: Validation of differential plasma exosome metabolites between patients with diabetic secondary osteoporosis and diabetic non-osteoporosis patients.
[0071] 1. Validation of DOP-related exosome metabolites
[0072] This example included 17 patients with diabetes without osteoporosis (control group) and 9 patients with diabetes secondary to osteoporosis (case group). Using the aforementioned detection and statistical analysis methods, a stable characteristic profile containing three metabolites (cytosine nucleotide, ketoglutarate, and triethylamine) was identified and validated in plasma exosomes. The three metabolites reproduced significant differences between the case and control groups and showed a trend consistent with the case-control group variation in Example 1. Figure 17 The remaining nine metabolites showed different trends between the case and control groups compared to the trends in the case and control groups in Example 1, or the trends were similar but the differences between the groups were not significant. Therefore, the remaining nine metabolites were excluded.
[0073] 2. ROC analysis
[0074] ROC analysis results showed that cytosine nucleotides, ketoglutarate, and triethylamine exhibited excellent specificity and sensitivity in predicting DOP. For example... Figure 18 As shown, the area under the curve (AUC) of each individual metabolite is greater than 0.8, and the AUC of the diagnostic model when used in combination reaches as high as 0.99. This indicates that the above three metabolites can serve as specific metabolic markers for DOP and can be used for the diagnosis of DOP.
[0075] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.
Claims
1. The application of reagents for detecting the content of metabolic markers in the preparation of diagnostic kits for diabetic secondary osteoporosis, characterized in that, The metabolic markers are selected from one or more of cytosine nucleotides, ketoglutarate, and triethylamine.
2. The application according to claim 1, characterized in that, The diagnostic kit is used to detect exosomes.
3. A diagnostic kit for diabetic secondary osteoporosis, characterized in that, The diagnostic kit contains reagents for detecting the levels of metabolic markers, wherein the metabolic markers are those described in claim 1.
4. The diagnostic kit according to claim 3, characterized in that, The diagnostic kit also includes reagents for extracting exosomes from plasma.
5. The diagnostic kit according to claim 3, characterized in that, The reagents include those for detecting the concentration or content of the metabolic markers in the sample to be tested by nuclear magnetic resonance, chromatography, spectroscopy, mass spectrometry, or a combination thereof.
6. A reagent for detecting the concentration or content of metabolic markers in a sample, characterized in that, The metabolic biomarker is the metabolic biomarker described in claim 1, and the sample to be tested is an exosome.
7. The application of metabolic markers in the preparation of diagnostic products for diabetic secondary osteoporosis, characterized in that, The metabolic biomarker is the metabolic biomarker described in claim 1.
8. The application according to claim 7, characterized in that, The diagnostic product includes equipment for analyzing the levels of the metabolic markers.
9. The application of metabolic biomarkers in constructing a predictive model for diabetic secondary osteoporosis, characterized in that, The metabolic biomarker is the metabolic biomarker described in claim 1.
10. The application according to claim 9, characterized in that, The prediction model includes the steps of analyzing the concentration levels of metabolic biomarkers using one or more algorithms selected from orthogonal partial least squares discriminant analysis, partial least squares discriminant analysis, classification and logistic regression, k-nearest neighbor algorithm, Naive Bayes, support vector machine, decision tree, random forest, regression tree, gradient boosting decision tree, extreme gradient boosting algorithm, lightweight gradient boosting machine, gradient boosting machine, lasso algorithm and convolutional neural network.
Citation Information
Patent Citations
Bone dysbolism disease marker and use thereof
JP2003125769A
Method for predicting and / or predicting osteoporotic fractures, method for outputting fracture scores, learning model generation method, learning model, method for predicting risk factors for osteoporotic fractures, graph creation method, program, information processing device, and learning dataset creation method
JP7408000B1
Cited By
Postmenopausal osteoporosis risk prediction and traditional Chinese and western medicine intervention management system
CN121617562A
Protein marker for diagnosing postmenopausal osteoporosis and application thereof
CN121899419A
Protein markers for diagnosing postmenopausal osteoporosis and use thereof
CN121899419B
