Protein marker combinations for age prediction and use thereof

Abstract

The present application relates to a protein marker combination for age prediction and its use. The present application also relates to the use of reagents for detecting the protein marker combination and methods for age prediction and senescence assessment, wherein the protein marker combination (1) comprises human serum albumin, prothrombin, neutrophil defensin 1, vitamin D binding protein, insulin-like growth factor 1, insulin-like growth factor 2, insulin-like growth factor binding protein 5, serpin, procollagen C-endopeptidase enhancer 1 and plasminogen; or (2) comprises post-translational modifications of human serum albumin, said post-translational modifications comprising cysteine redox state, methionine oxidation, lysine glycation and C-terminal integrity of human serum albumin. The present application also relates to a method for identifying the midlife senescence turning point between 50 and 60 years of age by assessing post-translational modifications of human serum albumin.

Description

Technical Field

[0001] This application belongs to the fields of biomedical technology and in vitro diagnostics (IVD), specifically relating to a method for age prediction and biological aging assessment based on plasma proteomics and / or serum albumin post-translational modifications (PTM). This application also relates to a method for identifying the mid-aging transition period of 50-60 years of age. Background Technology

[0002] Aging is a complex biological process characterized by the progressive decline of multiple physiological functions and is the most important risk factor for many chronic diseases such as cardiovascular disease, neurodegenerative diseases, metabolic diseases and tumors.

[0003] Biological age, compared to chronological age, better reflects an individual's true physiological state and health risks. Therefore, constructing an "aging clock" based on molecular biological indicators has become a hot topic in aging research in recent years. Currently, the mainstream aging clocks mainly include: DNA methylation clocks (such as the Horvath clock, PhenoAge, and GrimAge), transcriptomic clocks, metabolomic clocks, and proteomic clocks. Among these, the plasma proteomic aging clock has received widespread attention from academia and industry due to its advantages such as easy access to blood samples, the dynamic reflection of systemic physiological states by the proteome, and good reproducibility of detection. It has also been reported to independently predict the risk of developing various chronic diseases and the risk of all-cause mortality.

[0004] M. Austin Argentieri et al. (Argentieri et al., "Proteomic aging clock predicts mortality and risk of common age-related diseases in diverse populations", Nature Medicine, 2024 Sep;30(9):2450-2460. doi: 10.1038 / s41591-024-03164-7) used the Olink Explore 3072 platform to detect 2,897 plasma proteins in the UK Biobank (n=45,441). Using Boruta feature screening and LightGBM modeling, they identified 204 age-related proteins, predicting age with a Pearsonr of 0.94. They also found that this protein clock independently predicted the incidence and all-cause mortality risk of 18 major chronic diseases. In validation studies in biobanks in China and Finland, ProtAge showed similar age prediction accuracy to the UKB. Meanwhile, the research team simplified ProtAge to obtain a version with 20 proteins (ProtAge20), which can achieve almost the same prediction performance as ProtAge.

[0005] Lehallier et al. (Benoit Lehallier et al., "Undulating changes in human plasma proteome profiles across the lifespan", Nature Medicine, 2019 Dec;25(12):1843-1850. doi: 10.1038 / s41591-019-0673-2) used the SomaScan platform to detect 2,925 plasma proteins in 4,263 plasma samples from individuals aged 8 to 95. They proposed that the human plasma proteome exhibits non-linear and fluctuating changes throughout the lifespan, and found three distinct peaks in the 4th, 7th, and 8th decades, reflecting different biological processes.

[0006] Siino et al. (Valentina Siino et al., "Plasma proteome profiling of healthy individuals across the life span in a Sicilian cohort with long-lived individuals", Aging Cell, 2022 Sep;21(9):e13684. doi: 10.1111 / acel.13684) used non-targeted mass spectrometry proteomics to study age-related changes in plasma proteins throughout the lifespan of healthy individuals (including centenarians), and identified a number of age-related proteins.

[0007] Chia-Ling Kuo et al. (Chia-Ling Kuo et al., "Proteomic aging clock (PAC) predicts age-related outcomes in middle-aged and older adults", Aging Cell. 2024;23(8):e14195. doi:10.1111 / acel.14195.) Based on the UK Biobank-PPP (n=53,021), they constructed a protein aging clock (PAC) targeting all-cause mortality risk (rather than just age-related) using 2,923 plasma proteins on the Olink Explore 3072 platform. They found that it was significantly associated with the onset and mortality of various age-related diseases.

[0008] Hamilton Se-Hwee Oh et al. ("Organ aging signatures in the plasma proteome track health and disease", Nature, 2023Dec;624(7990):164-172. doi: 10.1038 / s41586-023-06802-1) constructed organ-specific aging clocks from plasma proteins using the SomaScan platform and found that accelerated organ aging is significantly associated with multiple organ-specific diseases, dementia, and mortality risk.

[0009] Olink Proteomics AB's Olink Explorep platform uses PEA (Proximity Extension Assay) technology, which utilizes paired antibodies each conjugated with oligonucleotides to simultaneously bind to the target protein. When two antibodies bind in close proximity, an extension reaction is triggered, and the extension product is read out by qPCR or NGS to achieve protein quantification.

[0010] SomaLogic Operating Co.'s SomaScan Assay v4.1 uses SOMAmer aptamer affinity detection, where chemically modified nucleic acid aptamers (SOMAmers) specifically bind to the target protein and are then converted into a signal that can be detected on a chip or in a fluorescence readout system.

[0011] Data-independent acquisition (DIA) mass spectrometry is a technique that uses a liquid chromatography-mass spectrometry (LC-MS) platform to continuously and systematically fragment and detect peptide ions entering the mass spectrometer according to a preset m / z window. It does not require prior reliance on antibodies, aptamers, or specific target lists. Subsequently, it combines a spectral library or the directDIA algorithm to identify and quantify peptides and their corresponding proteins.

[0012] Albumin is the main plasma protein in mammals, essential for maintaining osmotic pressure, facilitating transport, and providing antioxidant defense. Its function and stability are tightly regulated by plasma protein metabolites (PTMs) and structural integrity, including oxidation of cysteine ​​and methionine residues, lysine glycation, and C-terminal proteolysis. A key aspect of albumin's antioxidant function lies in its cysteine ​​and methionine residues. The free thiol group of Cys34 acts as the primary plasma antioxidant site, participating in complex redox cycles. This residue can undergo reversible oxidation (e.g., forming disulfides with low-molecular-weight thiols such as cysteine, homocysteine, or glutathione) or irreversible oxidation to sulfinic acid (HAS-SO2H) / sulfonic acid (HAS-SO3H). Similarly, methionine residues are readily oxidized to form methionine sulfoxide (MetO). In healthy young adults, 70% to 80% of Cys34 exists in its reduced state, but this balance is sensitive to physiological conditions. Due to its high concentration and long half-life, albumin is also a major target of glycation. This non-enzymatic modification, particularly the non-enzymatic modification of lysine residues and the N-terminus, impairs its function and is elevated in conditions such as diabetes. Furthermore, the circulating lifespan of albumin depends on the structural integrity of its C-terminus, which mediates binding to the neonatal Fc receptor (FcRn) for cell recirculation. Jeannette Nilsen et al. ("An intact C-terminal end of albumin is required for its long half-life in humans", Communications Biology, 2020 Apr 17;3(1):181. doi: 10.1038 / s42003-020-0903-7) demonstrated that the integrity of albumin's C-terminal L585 is crucial for FcRn-mediated cell recirculation and a long plasma half-life; C-terminal cleavage disrupts albumin binding to FcRn and accelerates albumin clearance.

[0013] Crucially, these modifications are closely associated with aging and age-related pathologies. The redox status of Cys34 and methionine residues changes significantly with aging and is associated with aging and conditions such as neurodegeneration and rheumatoid arthritis. Glycated albumin (GA) accumulates with aging and acts as a glucose marker and promoter of inflammatory pathways in diseases such as atherosclerosis. Age-related declines in albumin levels, along with increased C-terminal truncation observed in acute pancreatitis, further reflect systemic metabolic dysregulation during pathogenesis. Kayoko Nagumo et al. (Kayoko Nagumo et al., "Cys34-cysteinylated human serum albumin is a sensitive plasma marker inoxidative stress-related chronic diseases", PLOS ONE, 2014 Jan 8;9(1):e85216.doi: 10.1371 / journal.pone.0085216) established a quantitative method for Cys34 cysteinylation based on ESI-TOF mass spectrometry, demonstrating that the Cys-Cys34-HSA ratio is significantly elevated in chronic liver and kidney diseases and diabetes and other oxidative stress-related chronic diseases, and can serve as a sensitive marker for oxidative stress.

[0014] Previous studies have mostly focused on single PTM types or single disease contexts. No research has yet systematically and quantitatively characterized the changes in albumin PTMome with age in a unified healthy population cohort, nor has it been integrated with small protein abundance panels for age prediction and aging assessment. Existing aging clocks rely solely on protein biomarker abundance, which still has certain limitations. Summary of the Invention

[0015] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0016] Current proteomic aging clocks primarily rely on changes in protein biomarker abundance, reflecting more the quantitative changes in proteins than revealing their functional states. Protein function is also regulated at multiple levels, including post-translational modifications, turnover rates, structural conformations, subcellular localization, and interaction networks. Abundance-based models may miss crucial protein homeostasis remodeling events during aging, thus remaining insufficient in mechanistic explanation and intervention target discovery.

[0017] To address this deficiency, this application provides a method for age prediction and biological aging assessment based on plasma proteomics and / or serum albumin post-translational modifications.

[0018] In a first aspect, embodiments of this application provide a combination of protein biomarkers for age prediction and aging assessment, the combination of protein biomarkers comprising human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysinogen (PLG); or the combination of protein biomarkers comprises post-translational modifications of human serum albumin.

[0019] Secondly, embodiments of this application provide the use of reagents for detecting combinations of protein biomarkers in the preparation of kits for age prediction and aging assessment, wherein: the combination of protein biomarkers comprises human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysinogen (PLG); or the combination of protein biomarkers comprises post-translational modifications of human serum albumin.

[0020] Thirdly, the embodiments of this application provide a method for age prediction and aging assessment, the method comprising: The method involves obtaining a combination of protein biomarkers for the subject, wherein the combination of protein biomarkers includes human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysate (PLG); or the combination of protein biomarkers includes post-translational modifications of human serum albumin. The standardized abundance values ​​of the subjects' protein biomarker combinations were used as input features and fed into a regression model trained using the XGBoost gradient boosting regression algorithm; and The output includes the subject's predicted age.

[0021] In some embodiments, the protein biomarker combination consists of HSA, F2, HNP1, DBP, IGF1, IGF2, IGFBP5, SRGN, PCPE-1, and PLG.

[0022] In some embodiments, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycosylation, and C-terminal integrity. In some embodiments, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycosylation includes the non-enzymatic glycosylation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of the human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

[0023] In some implementations, the output results include predicted age, plasma protein age score, age deviation index, age-related stratification results, and / or a determination of whether the subject meets the predetermined characteristics of the molecular transition period of middle age.

[0024] In some embodiments, the levels of the protein biomarkers and / or post-translational modifications are assessed. In other embodiments, the levels are quantitative signals of protein abundance or peptide levels. In still other embodiments, the sample used for assessment is a whole plasma proteome sample.

[0025] Fourthly, this application provides a method for identifying the mid-aging transition period of 50-60 years of age. The method includes: z-score standardization of quantitative values ​​of multiple post-translational modifications of human serum albumin from the subject; fitting the trajectory of each post-translational modification with age using locally weighted regression; and using the common transitional changes of the multiple post-translational modifications in the 50-60 age range as a criterion to identify whether the subject is in, about to enter, or has already exceeded the mid-aging transition window of 50-60 years of age. In some embodiments, the locally weighted regression uses the LOESS (Locally Estimated Scatterplot Smoothing) algorithm.

[0026] In some embodiments, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycosylation, and C-terminal integrity. In still other embodiments, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycosylation includes the non-enzymatic glycosylation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of the human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

[0027] Specifically, the z-score transformation was performed on the modification proportion of each post-translational modification site across all samples. This involved subtracting the mean of the site and then dividing by the standard deviation, thus eliminating the influence of absolute abundance differences between different modification types or sites and ensuring comparability of their trends. Subsequently, using subject age as the independent variable and standardized modification levels as the dependent variable, locally weighted regression (e.g., the LOESS algorithm) was employed to fit the trajectory of each modification site's age-related changes. This method allows for the capture of the non-linear trend of modification levels with age without presupposing a linear relationship. The analysis revealed that different PTM types and their specific modification sites exhibit different, site-specific dynamic change patterns during aging. When different PTM types jointly exhibited a turning point within the 50-60 age range, this range was defined as the mid-age aging transition window. Furthermore, based on the subject's age and the position of the standardized levels of multiple post-translational modification sites relative to the fitted trajectory, it was determined that the subject was in, about to enter, or had already passed the mid-age aging transition window.

[0028] The implementation scheme of this application achieves at least the following technical effects: (a) The number of candidate proteins was reduced from 204-373 in existing technologies (Argentieri 2024, Lehallier 2019, etc.) to 10 core proteins, while maintaining the root mean square error (RMSE) = 6.12 years. 2 = 0.86 high performance.

[0029] (b) This invention is the first to incorporate a combination of multiple types and multiple sites of post-translational modifications of serum albumin, including Cys34 reduced state, four Met oxidation sites, four Lys glycation sites, and C-terminal integrity, totaling 10 quantitative PTM indicators, as comprehensive molecular markers of aging into the assessment of age and aging status. This combination can simultaneously reflect the body's redox state, accumulation of oxidative damage, accumulation of glycation damage, and changes in protein structural integrity, thereby overcoming the shortcomings of traditional abundance-based aging clocks, which have limited information dimensions and insufficient mechanistic explanatory power.

[0030] (c) For the first time, the 50-60 year old middle-age aging transition period is independently identified based on multiple PTM combined with LOESS trajectory, providing an early intervention window defined by molecular science.

[0031] (d) Iodoacetamide (IAA) is provided as a complementary product. 13 C-labeled iodoacetamide ( 13 A precise quantitative method for the reduced state of Cys34 differentiated by C-IAA.

[0032] (e) Using open LC-MS / MS methods such as DIA / DDA / PRM / SRM, molecular libraries are recognized without relying on antibodies or aptamers, and the platform has better standardization and transferability.

[0033] (f) This invention uses high-abundance plasma proteins as the analytical target, which has the advantages of easy sample acquisition, stable detection signal, easy internal normalization, and suitability for mass spectrometry quantification and clinical translation.

[0034] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. Other advantages of this application can be realized and obtained by means of the embodiments described in the description and the accompanying drawings. Attached Figure Description

[0035] The accompanying drawings are used to provide an understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0036] Figure 1 A scatter plot showing the correlation between the abundance of all detected plasma proteins and donor age.

[0037] Figure 2 The figure shows the results of nested cross-validation of the age prediction model based on a combination of 10 protein biomarkers in the current sample cohort. Figure 3The diagram shows a comparison between the out-of-fold predicted age and the actual age based on 10 protein features under nested cross-validation conditions (feature selection was performed within each training fold; 10-fold cross-validation was performed with 10 repetitions). Each point represents a participant, and the predicted value is the average of the results from 10 repetitions; dashed lines indicate that the predicted age is completely consistent with the actual age. Model performance is expressed as the mean ± standard deviation of the results from 10 repetitions: R0 2 = 0.821 ± 0.014, RMSE = 6.85 ± 0.27 years, MAE = 5.39 ± 0.28 years.

[0038] Figure 3 This is a secondary mass spectrum of alkylation modification of cysteine ​​at position 34 in albumin. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this application clearer, the implementation schemes of this application are described in detail below. It should be noted that, unless otherwise specified, the implementation schemes and features thereof in this application can be arbitrarily combined with each other.

[0040] This invention belongs to the fields of biomedicine and in vitro diagnostics, specifically relating to a method for age prediction and biological aging assessment based on plasma proteomics and post-translational modifications of serum albumin.

[0041] More specifically, the present invention relates to: (1) a combination of plasma protein biomarkers for age prediction (10 protein panels or multiple post-translational modifications of serum albumin); (2) related kits; and (3) the application of the above combination and kits in subject age prediction, biological age assessment and identification of the midlife (50-60 years) aging transition period.

[0042] In a first aspect, embodiments of this application provide a combination of protein biomarkers for age prediction and aging assessment, the combination of protein biomarkers comprising human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysinogen (PLG); or the combination of protein biomarkers comprises post-translational modifications of human serum albumin.

[0043] Secondly, embodiments of this application provide the use of reagents for detecting combinations of protein biomarkers in the preparation of kits for age prediction and aging assessment, wherein: the combination of protein biomarkers comprises human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysinogen (PLG); or the combination of protein biomarkers comprises post-translational modifications of human serum albumin.

[0044] Thirdly, the embodiments of this application provide a method for age prediction and aging assessment, the method comprising: The method involves obtaining a combination of protein biomarkers for the subject, wherein the combination of protein biomarkers includes human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysate (PLG); or the combination of protein biomarkers includes post-translational modifications of human serum albumin. The normalized abundance values ​​of the protein biomarker combinations of the subjects were used as input features and fed into a regression model trained based on the XGBoost algorithm; and The output includes the subject's predicted age.

[0045] In some embodiments, the protein biomarker combination consists of HSA, F2, HNP1, DBP, IGF1, IGF2, IGFBP5, SRGN, PCPE-1, and PLG.

[0046] In some embodiments, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycosylation, and C-terminal integrity. In some embodiments, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycosylation includes the non-enzymatic glycosylation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of the human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

[0047] In some implementations, the output results include predicted age, plasma protein age score, age deviation index, age-related stratification results, and / or a determination of whether the subject meets the predetermined characteristics of the molecular transition period of middle age.

[0048] In some embodiments, the levels of the protein biomarkers and / or post-translational modifications are assessed. In other embodiments, the levels are quantitative signals of protein abundance or peptide levels. In still other embodiments, the sample used for assessment is a whole plasma proteome sample.

[0049] Fourthly, the embodiments of this application provide a method for identifying the aging transition period of 50-60 years of age, the method comprising: Quantitative values ​​of multiple post-translational modifications of human serum albumin from subjects were z-score normalized. Locally weighted regression was used to fit the trajectory of each post-translational modification with age. The common turning point changes of these multiple post-translational modifications in the 50-60 age range were used as the criterion to identify whether the subject was in, about to enter, or had already passed the 50-60 age transition window of middle-age aging. In some implementations, the LOESS algorithm was used for locally weighted regression.

[0050] In some embodiments, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycosylation, and C-terminal integrity. In still other embodiments, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycosylation includes the non-enzymatic glycosylation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of the human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

[0051] Example 1: Plasma proteomics analysis of 133 healthy subjects and construction of a 10-protein age prediction model 1.1 Subject Cohort A total of 133 healthy subjects from Shandong Provincial Hospital (affiliated with Shandong First Medical University) were included in the study and divided into four age groups: <40 years old (youth), 40-60 years old (middle-aged), 60-70 years old (early old age), and ≥70 years old (late old age). The male-to-female ratio and exact number of participants in each age group were as follows: <40 years old: 16 males / 15 females; 40-60 years old: 22 males / 14 females; 60-70 years old: 19 males / 11 females; ≥70 years old: 18 males / 18 females.

[0052] Ethics Approval Number: NSFC: NO.2025-502, Biomedical Research Ethics Committee Approval Number of Shandong Provincial Hospital Affiliated to Shandong First Medical University. All participants signed informed consent forms.

[0053] 1.2 Preparation of plasma proteomic samples for DIA analysis Prepare whole plasma proteome samples according to the following steps, and input 1 μL of plasma from each sample: (1) After peripheral blood was collected at room temperature using EDTA anticoagulant tubes, plasma was separated by centrifugation at 4°C and 1000×g for 15 min, and then aliquoted and stored at -80°C. (2) Take about 1 μL of plasma, add sodium deoxycholate (SDC) at a final concentration of 1% (w / v) to denature it, heat at 95℃ for 5-10 min, and obtain the denatured plasma sample; (3) After the denatured plasma sample (90 μL) has cooled, add 10 μL of a chloroacetamide / tris(2-carboxyethyl)phosphine hydrochloride (CAA / TCEP) mixture with a molar ratio of 4:1 and carry out the reduction and alkylation reaction at 95 °C for 10 min (or at room temperature for 30 min). (4) Add trypsin (Promega, catalog number V5113) at a protein to trypsin mass ratio of 50:1 and enzymatically hydrolyze at 37℃ for 12-16 h; (5) Desalting with SDB-RPS StageTip (Empore™, catalog number: 760030C): Elution is performed as follows: Add 2 volumes (200 μL) of buffer 1 to the enzymatically digested sample and transfer the sample to SDB-RPS StageTip. Centrifuge at 3000×g for 5 min and repeat this process once. Replace the centrifuge tube and add 200 μL of buffer 1 to SDB-RPS StageTip. Centrifuge at 3000×g for 2 min. Add another 200 μL of buffer 2 to SDB-RPS StageTip and centrifuge at 3000×g for 2 min. Replace the centrifuge tube and add 200 μL of buffer 3 to SDB-RPS StageTip. Centrifuge at 3000×g for 2 min, collect the eluent, and concentrate the eluent under vacuum. Buffer 1 is an isopropanol solution containing 10% TFA; Buffer 2 contains 95% water, 5% acetonitrile and 0.05% TFA; Buffer 3 is 60% acetonitrile containing 10 μL of ammonia per mL.

[0054] (6) Redissolve in 20 μL of 0.1% TFA aqueous solution for LC-MS / MS analysis.

[0055] 1.3 LC-MS / MS DIA Data Acquisition Chromatography: Vanquish Neo UHPLC, IonOpticks Aurora 25 cm UHPLC C18 column (1.6 μm), 24 min gradient.

[0056] Mass spectrometry: Orbitrap Astral mass spectrometer, DIA mode: MS1 resolution 240,000, scan range 380-980 m / z; MS2 rapid acquisition in the Astral analyzer, using two sets of sequential isolation windows to ensure precursor coverage.

[0057] 1.4 Data Analysis (Spectronaut directDIA workflow) Spectronaut version 18.7.240325.55695, directDIA workflow, both precursor and protein FDR set to 1% (dynamic quality tolerance), cross-sample normalization (Automatic) enabled, interference correction and MS2 level quantification filtering enabled, other parameters kept at default.

[0058] 1.5 Protein Feature Preprocessing (based on scikit-learn ColumnTransformer) (i) Expand the semicolon-separated multiple accessions and merge them by protein, then transpose them into a sample × protein matrix; (ii) Remove proteins with a missing rate > 30%; (iii) Perform log2(x+1) transformation on strictly positive values; (iv) Imput the median of numerical features + z-score standardization; (v) Imput the mode of categorical variables (gender) + one-hot encoding.

[0059] 1.6 Construction of XGBoost Age Prediction Model All analyses were performed using Python 3 + scikit-learn + XGBoost, with random_state = 430.

[0060] XGBoost regression hyperparameters: n_estimators = 400, max_depth = 3, learning_rate = 0.1, reg_lambda(L2λ) = 1.0; 10-fold cross-validation was used, and all models used the same fold assignment. Performance is measured in terms of out-of-fold predicted RMSE / MAE / R. 2 Evaluation: The optimal model is selected based on the minimum RMSE; feature importance is extracted using impurity-based importance.

[0061] 1.7 Model Performance and 10-Protein Core Panel (a) Performance of the whole-protein feature XGBoost model OOF: RMSE = 6.24 years, R 2 = 0.85; (b) After screening by Top-K importance (K = 3, 5, 10, 15) and then training, the OOF performance of the Top-10 protein model was: RMSE = 6.12 years, R 2 = 0.86, with almost no loss in performance and even slightly better than the full protein model, proving that the 10-protein small panel can retain most of the age characterization information; (c) The importance of the 10 core proteins and their characteristics are detailed in Table 1; the performance evaluation results of the age prediction model built based on these 10 core proteins are detailed in Table 1. Figure 1 and Figure 2 middle; (d) HSA abundance decreased significantly with age: the average level in the youth group (<40 years) was 1.19 times higher than that in the middle-aged group (40-60 years) and 1.79 times higher than that in the elderly group (significant trend), indicating that HSA is the most important single age-related protein.

[0062] Table 1: Information on 10 core plasma protein biomarkers and their sources Research has found that human serum albumin is the most important age-related protein, so the inventors further studied the differential post-translational modifications of human serum albumin in different age groups.

[0063] Example 2: Quantitative method for reduced IAA + 13C-IAA dialkylation of albumin Cys34 2.1 Sample Preparation (1) Plasma proteins were denatured with 6 M guanidine hydrochloride; (2) Add 20 mM IAA to the final concentration, react overnight in the dark at room temperature, and alkylate the originally free Cys34 mercapto groups in the body; (3) The sample was transferred to Vivacon. 500 ultrafiltration tubes were washed twice by centrifugation at 12,000×g, room temperature, and 15 min with 50 mM ammonium bicarbonate as washing solution to remove residual IAA. (4) Add 10 mM DTT (dithiothreitol), keep at 37℃ for 30-60 min, to reduce the remaining disulfide bonds and the Cys34 that has been reversibly oxidized; (5) Add 13 C-IAA, reacted in the dark at room temperature for 3 h, to perform isotopic differential alkylation of the newly released thiol groups in step 4; (6) Remove by ultrafiltration washing twice with 50 mM ammonium bicarbonate. 13 C-IAA; (7) Add trypsin at a mass ratio of 50:1 (protein:enzyme) and incubate overnight at 37°C; (8) Collect peptides by centrifugation at 12,000×g for 15 min, concentrate under vacuum, reconstitute with 20 μL 0.1% TFA, and prepare for LC-MS / MS analysis.

[0064] 2.2 LC-MS / MS DDA Data Acquisition Chromatography: Thermo-Dionex Ultimate 3000+ self-packed 100 μm × 300 mm, 1.9 μm C18 column (Dr. Maisch GmbH, Germa), 120 min program, 0.30 μL / min constant flow elution. Phase A was 0.1% formic acid in water, and phase B was 80% acetonitrile-0.1% formic acid in water. Mass spectrometry: Orbitrap Exploris 480, DDA mode; MS1 full scan 300-1800 m / z, resolution 70,000; HCD fragmentation Top 20 strongest ions.

[0065] 2.3 Quantitative Formula for Reduced State Proportion The characteristic peptide ALVLIAFAQYLQQC(IAA)PFEDHVK was detected, and the mass spectrum is shown below. Figure 3 As shown: ● Reduced form (IAA notation): m / z = 831.09 (z = 3); ● 13 C isotope oxidized form ( 13 C-IAA label): m / z = 831.76 (z = 3); ● Intermediate isotope peak (used to subtract natural abundance interference): m / z = 831.42.

[0066] Formula for calculating the proportion of reduced state: Free Thiol Ratio = I(831.09) / {I(831.09) + [I(831.76) - 0.79 × I(831.42)]} Where I represents abundance, I(831.09) represents the abundance of the reduced characteristic peptide ALVLIAFAQYLQQC(IAA)PFEDHVK, and I(831.76) represents... 13 The abundance of the C isotope oxidized characteristic peptide ALVLIAFAQYLQQC(IAA)PFEDHVK, I(831.42) represents the abundance of the intermediate isotope peak.

[0067] 2.4 Typical Results ● <40 years old group: The proportion of restored individuals is about 40%, with little difference between individuals; ● In the three groups of 40-60 years old, 60-70 years old, and ≥70 years old: the proportion of reduced state was about 35%, showing a significant decrease and then remaining stable; ● There was no significant difference between men and women (p > 0.05); ● Note that the antioxidant capacity of Cys34 has already undergone an irreversible decline in middle age (40-60 years old), serving as a marker of early aging.

[0068] Example 3: Quantitative analysis of albumin methionine oxidation (Met123 / 298 / 446 / 548) 3.1 Retrieving data from the DDA database using Proteome Discoverer (v2.5) Database: Albumin FASTA; Enzyme: Trypsin, maximum 2 missed cleavages; Variable modification: amidation (C), 13 C-amidation, oxidation (M), saccharification (K); MS1 mass tolerance 20 ppm; MS2 tolerance 0.02 Da; fixed value PSM Validator.

[0069] 3.2 By comparing the intensity ratio of oxidized (+16 Da) and unoxidized peptides at each site, the site-specific oxidation ratio can be obtained.

[0070] 3.3 Typical Results (Oxidation Ratio of 4 Sites with Age): ● Met123: <40 years old, approximately 0.10-0.15; ≥70 years old, significantly increases to approximately 0.22 (p<0.001), located on Domain I, the highly solvent-exposed ring, and the earliest responder; ● Met298: <40 years old, approximately 0.08; ≥70 years old, approximately 0.20 (p<0.001), located near the Domain II hydrophobic ligand binding interface; ● Met446: <40 years old, approximately 0.08; ≥70 years old, approximately 0.18 (p<0.001), located in the hydrophobic core of Domain III; ● Met548: <40 years old, approximately 0.10; ≥70 years old, approximately 0.22 (p<0.001), located on exposed surfaces of Domain III; ● No significant differences were found between men and women; all four sites showed a trend of accelerated oxidation after age 60.

[0071] Example 4: Quantitative analysis of albumin lysine glycosylation (K64 / K137 / K262 / K525) 4.1 The search variable modification increases saccharification (K) (+162 Da or corresponding Amadori product mass), and other parameters are the same as in Example 3.

[0072] 4.2 By comparing the intensity ratio of glycosylated and unglycosylated characteristic peptides at each site, the glycation ratio at each site can be obtained.

[0073] 4.3 Typical Results: ● K64, K262, K525: Gradually increase with age, significantly increasing from <40 years to >70 years (p<0.05), with K525 being the most abundant site for glycation (approximately 27%→32%). ● K137: It exhibits a biphasic trajectory of first decreasing and then increasing, and then rebounds after about age 55; ● No significant sex differences were observed at any of the four loci; they could serve as candidate biomarkers for biological aging and metabolic drift.

[0074] Example 5: Quantification of albumin C-terminal integrity (LVAASQAALGL / LVAASQAALG) 5.1 Two C-terminal peptides were detected: the intact LVAASQAALGL (containing L585, which is essential for binding to FcRn) and the truncated LVAASQAALG (lacking L585).

[0075] 5.2 Calculate the intact / total C-terminal peptide ratio (i.e., the integral ratio).

[0076] 5.3 Typical Results: ● Percentage of intact / total C-terminal peptides: <40 years old group: approximately 85%; 40-60 years old group: approximately 88%; 60-70 years old group: approximately 88%; ≥70 years old group: approximately 90%; ● The proportion of intact / total C-terminal peptides gradually increased with age (p<0.05), indicating that C-terminal proteolytic activity decreased with age, and albumin C-terminus was more intact in older individuals; ● There was no significant difference between men and women. While this phenomenon prolongs the circulating half-life of albumin, it also exposes it to oxidative and glycation environments for a longer period, leading to increased cumulative damage.

[0077] Example 6: Integrated Age Prediction and Biological Age Assessment 6.1 Input characteristics: Quantitative values ​​of the 10 protein biomarkers described in this paper (log2 transformation + z-score normalization) or quantitative values ​​of the 10 PTM sites of human serum albumin described in this paper (z-score normalization).

[0078] 6.2 Output: (a) Actual age prediction (XGBoost regression); (b) Biological age assessment value (based on the aging score of the PTM combination); (c) Age Gap = Predicted Age - Actual Age; (d) Stratification of aging rate: When |Age Gap| < 5 years → normal for the same age; Age Gap > +5 years → accelerated aging; Age Gap < -5 years → delayed aging.

[0079] Example 7: Identification of the aging transition period in middle age (50-60 years old) (multiple PTM combined with LOESS trajectory analysis) 7.1 The quantitative values ​​of the 10 post-translational modifications of albumin were normalized by z-score.

[0080] 7.2 Fit the LOESS trajectory with age as x and z-score as y, and assign colors to the four types of PTM (Cys34=red, Met=blue, K-glycation=green, C-terminus=purple).

[0081] 7.3 Common Observations: ● Cys34 reduced state: It continues to decline before the age of 50, and tends to stabilize after the age of 50 (the irreversible decline has reached its limit). ● Oxidation of Met123 / 298 / 446 / 548: Accelerates significantly after age 60; ● K-glycation: K64, K262, and K525 gradually increase across adulthood; K137 first decreases and then increases (inflection point around age 55); C-end integrity: increases before age 50, plateaus between 50 and 60 years old, and continues to increase after age 60; ●In summary, multiple PTM types in the 50-60 age range exhibit transformative changes, constituting a molecular window for the transition of middle-aged aging.

[0082] 7.4 Output: The subject's assessment results in the three categories of "before the transition period / during the transition period / after the transition period" can be combined with the Age Gap in Example 6 to provide a comprehensive "early / mid / late aging risk warning".

[0083] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A combination of protein biomarkers for age prediction and aging assessment, said combination comprising human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D-binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysinogen (PLG); preferably, said combination of protein biomarkers consists of HSA, F2, HNP1, DBP, IGF1, IGF2, IGFBP5, SRGN, PCPE-1, and PLG; or The protein biomarker combination comprises post-translational modifications of human serum albumin; preferably, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycation, and C-terminal integrity; more preferably, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycation includes the non-enzymatic glycation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

2. The use of reagents for detecting combinations of protein biomarkers in the preparation of kits for age prediction and aging assessment, wherein: The protein biomarker combination comprises human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D-binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysate (PLG); preferably, the protein biomarker combination consists of HSA, F2, HNP1, DBP, IGF1, IGF2, IGFBP5, SRGN, PCPE-1, and PLG; or The protein biomarker combination comprises post-translational modifications of human serum albumin; preferably, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycation, and C-terminal integrity; more preferably, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycation includes the non-enzymatic glycation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

3. The use according to claim 2, wherein the level of the protein biomarker and / or post-translational modification is evaluated; preferably, wherein the level is a quantitative signal of protein abundance or a quantitative signal of peptide.

4. The use according to claim 2 or 3, wherein the sample used for evaluation is a whole plasma proteome sample.

5. A method for age prediction and aging assessment, the method comprising: The method involves obtaining indicators of a combination of protein biomarkers for the subject, wherein the combination of protein biomarkers comprises human serum albumin (HSA), prothrombin (F2), neutrophil defensin 1 (HNP1), vitamin D-binding protein (DBP), insulin-like growth factor 1 (IGF1), insulin-like growth factor 2 (IGF2), insulin-like growth factor binding protein 5 (IGFBP5), serine proteoglycan (SRGN), procollagen C-endopeptidase enhancer 1 (PCPE-1), and plasminogen lysate (PLG); preferably, the combination of protein biomarkers consists of HSA, F2, HNP1, DBP, IGF1, IGF2, IGFBP5, SRGN, PCPE-1, and PLG; or the combination of protein biomarkers comprises a fraction of human serum albumin. Post-translational modifications; preferably, the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycosylation, and C-terminal integrity; more preferably, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycosylation includes the non-enzymatic glycosylation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG. The standardized abundance values ​​of the subjects' protein biomarker combinations were used as input features and fed into a regression model trained using the XGBoost gradient boosting regression algorithm; and The output includes the subject's predicted age.

6. The method according to claim 5, wherein, The output results include predicted age, plasma protein age score, age deviation index, age-related stratification results, and / or the determination results of whether the subject meets the predetermined characteristics of the molecular transition period in middle age.

7. The method according to claim 5 or 6, wherein the level of the protein marker and / or post-translational modification is evaluated; preferably, wherein the level is a quantitative signal of protein abundance or a quantitative signal of peptide; preferably, the sample used for evaluation is a whole plasma proteome sample.

8. A method for identifying the aging transition period of 50-60 years old, the method comprising: The quantitative values ​​of multiple post-translational modifications of human serum albumin from the subjects were standardized by z-score, and the trajectory of each PTM with age was fitted using local weighted regression. Based on the common turning point changes of the multiple post-translational modifications in the 50-60 age range, the subjects were identified as being in, about to enter, or have exceeded the 50-60 age transition window of middle-age aging.

9. The method according to claim 8, wherein the local weighted regression is performed using the LOESS algorithm; the post-translational modifications include cysteine ​​(Cys) redox state, methionine (Met) oxidation, lysine (Lys) glycation, and C-terminal integrity; more preferably, the cysteine ​​(Cys) redox state includes the reduced proportion of human serum albumin Cys34; the methionine (Met) oxidation includes the methionine sulfoxide (MetO) levels at human serum albumin Met123, Met298, Met446, and Met548 sites; the lysine (Lys) glycation includes the non-enzymatic glycation levels at human serum albumin K64, K137, K262, and K525 sites; and the C-terminal integrity includes the ratio of human serum albumin C-terminal peptide LVAASQAALGL to LVAASQAALG.

Images