Population serum proteomics uncovers a prognostic protein classifier for metabolic syndrome

Xue Cai; Zhangzhi Xue; Fang-Fang Zeng; Jun Tang; Liang Yue; Bo Wang; Weigang Ge; Yuting Xie; Zelei Miao; Wanglong Gou; Yuanqing Fu; Sainan Li; Jinlong Gao; Menglei Shuai; Ke Zhang; Fengzhe Xu; Yunyi Tian; Nan Xiang; Yan Zhou; Peng-Fei Shan; Yi Zhu; Yu-ming Chen; Ju-Sheng Zheng; Tiannan Guo

doi:10.1016/j.xcrm.2023.101172

. 2023 Aug 30;4(9):101172. doi: 10.1016/j.xcrm.2023.101172

Population serum proteomics uncovers a prognostic protein classifier for metabolic syndrome

Xue Cai ^1,⁹, Zhangzhi Xue ^1,⁹, Fang-Fang Zeng ^2,⁹, Jun Tang ^1,^3,^4,⁹, Liang Yue ^1,^3,^4,^5,⁹, Bo Wang ^6,⁹, Weigang Ge ⁶, Yuting Xie ^1,^3,^4,⁵, Zelei Miao ^1,^3,⁴, Wanglong Gou ^1,^3,⁴, Yuanqing Fu ^1,^3,⁴, Sainan Li ^1,^3,^4,⁵, Jinlong Gao ^1,^3,^4,⁵, Menglei Shuai ^1,^3,⁴, Ke Zhang ^1,^3,⁴, Fengzhe Xu ^1,^3,⁴, Yunyi Tian ^1,^3,⁴, Nan Xiang ⁶, Yan Zhou ^1,^3,^4,⁵, Peng-Fei Shan ⁷, Yi Zhu ^1,^3,^4,^5,^∗, Yu-ming Chen ^8,^∗∗, Ju-Sheng Zheng ^1,^3,^4,^∗∗∗, Tiannan Guo ^1,^3,^4,^5,^10,^∗∗∗∗

PMCID: PMC10518601 PMID: 37652016

Summary

Metabolic syndrome (MetS) is a complex metabolic disorder with a global prevalence of 20%–25%. Early identification and intervention would help minimize the global burden on healthcare systems. Here, we measured over 400 proteins from ∼20,000 proteomes using data-independent acquisition mass spectrometry for 7,890 serum samples from a longitudinal cohort of 3,840 participants with two follow-up time points over 10 years. We then built a machine-learning model for predicting the risk of developing MetS within 10 years. Our model, composed of 11 proteins and the age of the individuals, achieved an area under the curve of 0.774 in the validation cohort (n = 242). Using linear mixed models, we found that apolipoproteins, immune-related proteins, and coagulation-related proteins best correlated with MetS development. This population-scale proteomics study broadens our understanding of MetS and may guide the development of prevention and targeted therapies for MetS.

Keywords: population proteomics, metabolic syndrome, DIA-MS, machine learning, prospective prediction, blood proteomics, diabetes, apolipoproteins, triglycerides, high-density lipoproteins

Graphical abstract

Highlights

•
Over 400 proteins measured in ∼20,000 blood proteomes using DIA-MS
•
3,840 participants with two follow-up time points over 10 years
•
A machine-learning model for predicting the 10-year risk of developing MetS
•
Roles of apolipoproteins and coagulation-related proteins in MetS

Cai et al. measure over 400 proteins from ∼20,000 data-independent acquisition mass spectrometry proteomes for 7,890 serum samples from a longitudinal cohort of 3,840 participants with over 10-year follow-up. A machine-learning model is built for predicting the risk of developing MetS within 10 years, achieving an AUC of 0.774.

Introduction

Metabolic syndrome (MetS) is a complex metabolic disorder characterized by abdominal obesity, atherogenic dyslipidemia, raised blood pressure, insulin resistance, central obesity, prothrombotic and proinflammatory states.¹^,² About 20%–25% of adults worldwide suffer from MetS.³^,⁴^,⁵ In Chinese adults, the prevalence of MetS reached 24.2% (24.6% in men and 23.8% in women) in 2018.⁶ According to the National Health and Nutrition Examination Survey (NHANES), about one-third of US adults were diagnosed with MetS between 1988 and 2010.⁷ MetS has been reported to predispose to several serious diseases,⁸^,⁹^,¹⁰ including diabetes, cardiovascular disease, coronary heart disease, and several common cancers. An early prediction and diagnosis of MetS would allow faster interventions, reducing the burden on healthcare systems. There are currently diagnostic criteria of MetS, with the main components being waist circumference, glucose, blood pressure, triglycerides (TGs), high-density lipoprotein cholesterol (HDL-C).² But the pathogenesis of MetS is unclear, and there is still a lack of targeted therapy for MetS. Thus, longitudinal large cohort population studies that can effectively predict the development and explore the pathogenesis of MetS are necessary.

Plasma and serum are the predominant samples used for diagnostic analyses in clinics with low invasiveness and ease of collection and preservation. In a study involving 2,380 participants, targeted metabolomics was employed to measure serum samples, revealing a correlation between hepatic phospholipids and cardiopulmonary fitness.¹¹ Berenice and colleagues conducted an analysis of 40 metabolites collected from 602 participants using a targeted metabolomics approach. Their findings indicate that serum acylcarnitines and amino acids are associated with MetS.¹² By utilizing untargeted metabolomics, the French occupational GAZEL cohort consisting of 206 subjects achieved a successful prediction of MetS occurrence, with an area under the curve (AUC) of 0.86.¹³ Metabolomics is a potent technique for investigating metabolic disorders as it provides direct insight into an organism’s current metabolic state. However, the rapid temporal dynamics of analytes pose a significant challenge to this approach.¹⁴ In contrast to metabolomics, proteomics is a stable and dynamic reflection of both genes and the environment, making it well suited for studying large cohorts with long-term follow-up.

Plasma and serum proteomic studies based on large cohort population have used aptamer-based (SOMAscan) platform¹⁵ and proximity extension assay (Olink proteomics)¹⁶ as tools for metabolic disease.¹⁷^,¹⁸ For example, using the SOMAscan platform, Ganz et al. quantified the plasma proteins of two longitudinal cohorts of 938 and 971 samples with one follow-up point and achieved 0.70 of C statistics in the validation cohort for cardiovascular risk prediction based on 9-protein panel.¹⁹ Proteomics of 3,301 human plasma samples from two different subcohorts were also measured by the SOMAscan platform to explore the non-linear alterations in the human plasma proteome with age.²⁰ Furthermore, dozens of blood proteins were measured by the proximity extension assay (Olink proteomics) to investigate the association with BMI and waist circumference in a cross-sectional study of 3,308 participants²¹ and to predict myocardial infarction risk with a C index of 0.68 in 5,131 patients with diabetes.²² Both Olink- and SOMAscan-based proteomics are essentially targeted proteomics and measure a pre-defined list of proteins of interest based on antibody specificity. The measurement is limited to proteins with valid assays but is high throughput and transferable when the assays are established.

The rapid development of mass spectrometry (MS)-based proteomics provided the solid technical support needed by large cohort proteomics.²³ Specifically, this technology provides high-throughput and reproducible analyses with minimal sample amounts. Compared with SOMAscan and Olink, the limitation of the MS-based proteomics lies in its high cost for exploring low-abundance proteins. However, data-independent acquisition (DIA)-MS measures all flyable peptide precursor ions without prior knowledge, allowing the unbiased discovery of new biomarkers in large populations.²⁴^,²⁵ For instance, using DIA-MS, Niu et al. analyzed the plasma samples of 596 individuals and 79 liver biopsies in 3 weeks and predicted future liver-related events and all-cause fatalities.²⁶ Bruderer et al. also analyzed 1,508 plasma samples with robust capillary-flow DIA-MS to investigate potential biomarkers for weight loss.²⁷

Here, we reliably measured over 400 proteins from ∼20,000 proteomes using DIA-MS for 7,890 samples from a longitudinal cohort of 3,840 participants with two follow-up time points over 10 years. Based on our data, we built a model to predict the risk of developing MetS within 10 years. We also explored new potential biomarkers and networks associated with MetS, providing a reference for the pathogenesis and targeted therapy of MetS.

Results

Serum proteome profiling of a longitudinal cohort

Here, we included 3,840 participants from a community-based prospective cohort study, namely Guangzhou Nutrition and Health Study (GNHS; ClinicalTrials.gov: NCT03179657).²⁸^,²⁹^,³⁰^,³¹ Serum samples were collected at three of four time points: 3,479 samples were collected at baseline (between 2008 and 2010), 2,638 were collected at the second follow-up (between 2014 and 2017), and 1,773 were collected at the third follow-up (between 2018 and 2019) (Table 1). No serum samples were collected at first follow-up. On average, 5.6 years (interquartile range [IQR], Q1 5.25, Q3 5.69) passed between the baseline and the second follow-up and 3.4 years (IQR, Q1 3.19, Q3 3.53) passed between the second and the third follow-up. The statistics of the participants at the three time points are summarized in Table 1; their clinical information is provided in Table S1. In particular, the median ages for the baseline, second follow-up, and third follow-up groups were 57.4 (38.2–80.4), 63 (44.3–83.3), and 66.1 (46.8–86.2), with 69%, 68%, and 69% female participants, respectively. We next divided the serum samples into a discovery (n = 4,794) and a validation cohort (n = 3,094) randomly (Figure 1), and the sample preparation, MS data collection, and data analyses were then independently performed.

Table 1.

Statistics of the 3,840 participants from the Guangzhou Nutrition and Health Study (GNHS)

	Baseline (n = 3,480)		Third follow up (n = 1,773)
Age (years)	57.4 (38.2–80.4)	63 (44.3–83.3)	66.1 (46.8–86.2)
Gender (% of women)	69	68	69
BMI (kg/m²)	23.2 (14.8–51.3)	23.5 (14.5–57.3)	23.6(15.3–54.1)
Waist (men, cm)	86.4 (61–147)	87.8 (64.4–155.8)	88 (64.5–162)
Waist (women, cm)	81.2 (58–117.9)	86.5 (60.3–128,6)	83.6 (60.2–113.8)
Fasting serum glucose (mmol/L)	4.7 (0.2–24.6)	5 (3.2–24.6)	5.3 (3.9–21.1)
Blood triglycerides (mmol/L)	1.3 (0.01–25.4)	1.3 (0.2–15.6)	1.4 (0.45–15)
Serum HDL (mmol/L)	1.4 (0.02–3.3)	1.5 (0.6–4.6)	1.4 (0.6–4.3)
Serum LDL (mmol/L)	3.6 (0.1–10)	3.6 (0.6–11.5)	3.5 (0.9–8)
SBP	122 (73–228)	122 (63.5–241)	120 (76–205.7)
DBP	79 (46–135)	73.3 (43.7–120)	72.5 (47.7–117.3)
SBP/DBP (mmHg)	1.6 (1–2.7)	1.6 (0.9–2.8)	1.6 (1–2.5)
MetS (%)	17	21	18

Open in a new tab

Study design

GNHS study population: 7,890 samples (4,796 in the discovery cohort, 3,094 in the validation cohort) were collected from 3,840 individuals. Altogether, we generated 18,970 DIA-MS data. After data cleaning, the protein matrices (containing 4,637 samples and 438 proteins for the discovery cohort and 3,067 samples and 413 proteins for the validation cohort) were used for data mining and machine-learning modeling.

Based on age and gender, we randomly divided the 4,796 serum samples of our discovery cohort into 178 batches, each containing 29 unique serum samples, two biological replicates, and one quality control (QC) sample (pooled serum samples) to ensure the stability of the entire workflow (Figure 1; Table S2A). 526 randomly selected serum samples were processed twice as biological replicates. The 178 batches were acquired by MS in random order. An MS-QC sample (pooled serum peptides) was injected before each batch to ensure that the instrument was in good condition. Next, batches were acquired using a 20-min DIA-MS,²⁴ and some randomly selected batches were acquired 2–3 times as technical replicates, summing up to 11,646 MS data, including 301 QC samples, 526 biological replicates, 288 MS-QC samples, and 5,735 technical replicates (Figure 1). Finally, we used the software tool DIA-NN³² to generate a matrix of 583 proteins and 11,646 MS data with 45% missing values.

To obtain high-quality proteomics data for our subsequent analyses, we cleaned and sorted the protein matrix returned by DIA-NN. First, we excluded 445 MS data with protein identifications below 75% of the median protein identification (<245), which may have been affected by an imperfect sample preparation or the instrument status. We also excluded proteins with missing values exceeding 80% of the protein matrix (n = 145) (Figure 1). The median Pearson correlation coefficient (r) of the MS-QC samples achieved 0.94, proving the high stability of the MS instruments (Figure 2A). Next, we checked the MS-QC samples of each batch and excluded 167 MS data with no MS-QC samples and no biological or technical replicates, as it would have been hard to assess the quality of their MS data. Finally, we corrected the batch effects using ProteomeExpert³³ and observed a significant improvement in our data (Figure 2B). We marked workflow-QC samples from red to green according to the date order of data acquisition by MS, showing minimal batch effects (Figure 2B). The r values of 512 biological replicates and 5,317 technical replicates were calculated, and the median r values of the two sets of replicates were 0.97 and 0.96 (Figure 2C), respectively. These results prove the high consistency and reproducibility of our data. For our subsequent data analyses, we combined the quantitative results of those replicates with r >0.8.

QC analyses and dataset pre-processing

(A) Pearson correlation of the MS-QC samples from the discovery and validation cohorts.

(B) Batch effect evaluation using principal-component analysis (PCA) and the QC samples from the discovery and validation cohorts.

(C) Pearson correlation of the technical and biological replicates from the discovery and validation cohorts. Samples were processed twice as biological replicates and were acquired twice by MS as technical replicates.

The experimental design of our validation cohort was the same as that of our discovery cohort. We divided 3,094 serum samples into 132 batches (Table S2A). A total of 7,324 MS data (containing 247 QC samples, 338 biological replicates, 213 MS-QC samples, and 3,432 technical replicates) were measured using 20-min DIA-MS. Next, we excluded 194 MS data with protein identifications below 75% of the median protein identification (<238). Finally, we focused on 413 proteins that were identified in both the discovery and the validation cohorts. The median r values of the MS-QC samples, the biological replicates, and the technical replicates were 0.93 (Figure 2A), 0.96, and 0.95 (Figure 2C), respectively. We also corrected the batch effect using ProteomeExpert and observed a significant improvement (Figure 2B).

After these steps, from the discovery cohort, we obtained a protein matrix containing 4,637 samples and 438 proteins, with 19% missing values (Figure 1; Table S2B), and from the validation cohort, we obtained a protein matrix containing 3,067 samples and 413 proteins, with 18.4% missing values (Figure 1; Table S3B).

A protein-based classifier for predicting MetS insurgence

For the discovery cohort, we collected a set of samples at the baseline time point, including 267 non-MetS samples from individuals who were diagnosed with MetS at the second or third follow-up and 588 non-MetS samples from individuals who were not diagnosed with MetS at any of our time points (Figure 3A). Using these data, we built a machine-learning model to evaluate the serum proteins’ ability to predict the risk of developing MetS within 10 years.

Machine-learning model for predicting the risk of developing MetS

(A) Case and control samples for the machine-learning model.

(B) Workflow of the machine-learning model built with quantitative proteomics data and clinical features.

(C) Shapley importance of variables in the model.

(D) Receiver operating characteristic (ROC) plot of the machine-learning model.

We used LightGBM to generate a machine-learning model using the discovery dataset of 855 samples containing 288 serum proteins as well as age and gender information (Figure 3B; Table S4A). First, we randomly split the dataset into a training dataset (n = 684) and an internal validation dataset (n = 171). Then, a 2-step feature selection was used for choosing protein features from the 288-serum proteome (Figure 3B). Specifically, the final model features were apolipoprotein A-I (APOA1), sex hormone-binding globulin (SHBG), APOD, age, vitronectin (VTN), attractin (ATRN), clusterin (CLU), heparin cofactor 2 (SERPIND1, HCII), alpha-1B-glycoprotein (A1BG), APOC2, immunoglobulin heavy variable 4-39 (IGHV4-39), and APOB-100 (APOB) (Figure 3C). We optimized our model using a 10-fold cross-validation, that is, internal validation was performed ten times. Next, we tested the model using the randomly selected internal validation dataset and achieved an AUC of 0.784, indicating that our model can efficiently predict the risk of developing MetS within 10 years (Figure 3D). Finally, we tested the model using 242 samples from an independent validation cohort and achieved an AUC of 0.774 (Figure 3D; Table S4B). The confusion matrices are provided in Figure S1. Due to the relatively smaller sample size of the validation cohort compared with the discovery cohort, the significance ranking of selected features in the model may fluctuate when applied to the validation cohort. However, the primary features, such as APOA1, age, APOD, and VTN, remain relatively stable (Figure 3C). In order to investigate the model’s applicability across different age ranges, we tested it using 133 samples (119 controls and 14 cases) from post-menopausal women, resulting in an AUC of 0.799 (Figure S1). These results show that the selected features are stable and can be well generalized to independently collected samples and that the protein used in the model may present as promising biomarker candidates.

Serum proteins regulated in MetS samples

To find which serum proteins are potentially affected by MetS, we used two linear mixed models to examine (1) the relationship between protein level and MetS at the baseline and (2) the dynamic relative changes along the three time points. For the baseline analysis, we used 1,736 baseline samples (309 MetS, 1,427 non-MetS), and for the dynamic analysis, we used 4,561 samples from all three time points (986 MetS, 3,575 non-MetS). With the baseline analysis, we found that 175 proteins were significantly dysregulated (B-H adjusted p < 0.05) in the MetS compared with the non-MetS group (Table S4C), while 194 proteins (B-H adjusted p < 0.05) were associated with the dynamic development of MetS over 10 years (Table S4D). A total of 143 proteins were dysregulated according to both analyses (Figure 4A).

Performance of new potential protein biomarkers of MetS and their related pathways

(A) Dysregulated proteins highly associated with MetS were selected using linear mixed models. The inner circle represents dysregulated proteins of the baseline model, while the outer circle represents dysregulated proteins of the dynamic model. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05.

(B) Enriched pathways analyzed by IPA of 175 and 194 significantly dysregulated proteins. The size of circle represents the −log10(p value) and the color represents the ratio, which is calculated by dividing the number of analysis-ready molecules by the total number of molecules within the pathway.

(C) Top (|β| > 0.2) significantly dysregulated proteins associated with MetS using the linear mixed model.

(D) Expression of typical dysregulated proteins in discovery cohort and validation cohort. B-H adjusted p values of these proteins are < 0.001.

To find which pathways may be affected by MetS, we analyzed the 175 and 194 dysregulated proteins using Ingenuity Pathway Analysis (IPA). We found that 11 pathways were significantly dysregulated according to our baseline analysis (Figure 4B; Table S4E) and 12 were significantly dysregulated according to our dynamic analysis (the threshold for both analyses was −log(p) >1.3) (Figure 4B; Table S4F). In particular, ten pathways were significantly dysregulated according to both analyses, including farnesoid X receptor/retinoid X receptor (FXR/RXR) activation, liver X receptor/retinoid X receptor (LXR/RXR) activation, atherosclerosis signaling, maturity-onset diabetes of young (MODY) signaling, and acute-phase response signaling (Figure 4B). Seven pathways with the most significant p values were mainly enriched in apolipoproteins (Tables S4E and S4F); among these seven, FXR/RXR activation and LXR/RXR activation are involved in metabolism and are associated with RXR activation.³⁴ RXR activation is known to correlate highly with MetS, as drugs targeting RXR and its heterodimeric partners are already in clinical use against MetS.³⁵ Additionally, MetS has been reported to increase the risk of atherosclerosis³⁶ and is a clinical characteristic of MODY.³⁷ Acute-phase response signaling has also been associated with MetS.³⁸ The growth hormone signaling pathway was significantly dysregulated only in the baseline analysis (Figure 4B), indicating that this pathway was dysregulated in patients with MetS but possibly not linked with its development. In particular, the increase in growth hormone can cause insulin resistance in men.³⁹ Type II diabetes mellitus and the AMPK signaling pathways were significantly dysregulated in the dynamic analysis only (Figure 4B), indicating these pathways may be primarily involved in the development of MetS. Diabetes is a component of MetS,⁴⁰ and evidence suggests that the dysregulation of the AMPK signaling pathway may lead to MetS.⁴¹

The expression of 32 significantly dysregulated proteins (|β| > 0.2; see STAR Methods), among the 175 proteins and 194 proteins, showed the same performance in the discovery and the validation cohorts (Figure 4C). These proteins were dysregulated in MetS with respect to the non-MetS patients and were linked to the development of MetS. The elevated expression of APOC2 and VTN (Figure 4D) and the lower expression of AOPA1, APOD, A1BG, CLU, and APOC2 (Figure 4D) in patients with MetS were consistent with the results of above analysis (Figure 3D). In addition, two apolipoproteins, APOC3 and APOF, and several other reported proteins, pigment epithelium-derived factor (PEDF), alpha-1-antitrypsin (A1AT), and afamin (AFM), were also dysregulated in patients with MetS (Figure 4D). These results show that apolipoproteins and their related pathways are closely related to the occurrence and development of MetS.

Molecular characterization of MetS subtypes

MetS can be diagnosed when at least three of five diagnostic indicators are met: waist circumference, fasting blood glucose, blood pressure (systolic blood pressure/ diastolic blood pressure [SBP/DBP]), fasting TG, and fasting HDL-C (see STAR Methods).⁴⁰ Subtypes of MetS are based on different indicators. Here, we compared the different types of MetS and non-MetS and explored the molecular mechanism of MetS based on different indicators.

We selected 444 MetS samples that met the indicators of abnormal TG (Tables S5A and S5B) and 444 non-MetS control samples with normal TG and matching age, gender, and other MetS indicators the same as MetS. Similarly, we selected 169, 235, 384, and 354 paired samples to analyze abnormal HDL, glucose, waist, and SBP/DBP, respectively (Tables S5A and S5B). No dysregulated proteins were discovered in the waist- and SBP/DBP-based MetS samples, and only one up-regulated protein was deregulated in the glucose-based MetS (Figure S2). On the other hand, six up-regulated and three down-regulated proteins were observed in the TG-based MetS samples (Figure 5A, B-H adjusted p value < 0.05, |log2(fold change)| > 0.2), while one up-regulated and six down-regulated proteins were observed in the HDL-based MetS samples (Figure 5B, B-H adjusted p value < 0.05, |log2(fold change)| > 0.2). The expressions of these dysregulated proteins were consistent in the discovery and the validation cohorts (Figures 5C and 5D).

Dysregulated proteins in different types of MetS

(A) Volcano plot of the dysregulated proteins in TG-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2.

(B) Volcano plot of the dysregulated proteins in HDL-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2.

(C) Expression of the dysregulated proteins in TG-based MetS and non-MetS.

(D) Expression of the dysregulated proteins in HDL-based MetS and non-MetS.

See also Figure S2.

Next, we analyzed the dysregulated proteins in TG- and HDL-based MetS samples using the IPA. The network enriched in the dysregulated proteins from the TG-based MetS samples (Figure S3) showed that the increased abundance of APOE, APOC1, APOC2, and APOC3 led to the activation of HDL and low-density lipoprotein (LDL) as well as inhibition of very-low-density lipoprotein cholesterol (VLDL-C) and insulin. In contrast, in the HDL-based MetS, the decrease in APOD and APOF led to the inhibition of HDL and predicted the inhibition of LDL (Figure S4).

Proteins associated with MetS initiation

To investigate the development mechanisms of MetS, we next focused on participants that were not diagnosed with MetS at the baseline but developed MetS within 10 years. By looking at the five diagnostic indicators of MetS mentioned above, some of these people met 1–2 indicators, and others met none. Therefore, we explored the proteins that significantly changed in the serum of participants that developed MetS based on the five abnormal indicators.

We selected 146 individuals from the discovery cohort that only met the waist indicator for MetS at baseline (Tables S6A and S6B). Of these individuals, only 19 of them developed MetS. As the remaining 127 participants had an abnormal waist indicator also at the two follow-up visits, they acted as controls. To minimize the influence of sample size, we employed oversampling in our analysis (see STAR Methods). In the MetS patients, 12 proteins were significantly up-regulated (B-H adjusted p value < 0.05, log2(fold change) > 0.2), while 24 were significantly down-regulated (B-H adjusted p value < 0.05, log2(fold change) < -0.2), at the baseline compared with the control group (Figure 6A; Table S6C). These proteins included angiotensinogen (AGT), APOD, and thrombospondin-1 (THBS1).

Dysregulated proteins in the development of MetS

(A) Volcano plot and boxplots of the dysregulated proteins in the development of waist-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2.

(B) Volcano plot and boxplots of the dysregulated proteins in the development of TG-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2.

(C) Volcano plot and boxplots of the dysregulated proteins in the development of SBP-/DBP-based MetS. The cutoff of dysregulated proteins has been set at B-H adjusted p value < 0.05, |log2(fold change)| > 0.2.

(D) Boxplots of the overlapping core dysregulated proteins in the development of waist-based MetS, TG-based MetS, and SBP-/DBP-based MetS. The numbers between the paired box plots represent the B-H adjusted p values.

We next focused on TGs. A total of 24 cases and 112 controls were selected similarly to the previous analysis on the waist indicator. Only five proteins were significantly up-regulated (adjusted p value < 0.05, log2(fold change) > 0.2), and six were significantly down-regulated (adjusted p value < 0.05, log2(fold change) < -0.2, Figure 6B; Tables S6A, and S6D), including Xaa-Pro aminopeptidase 1 (XPNPEP1) and APOF. Using the SBP/DBP indicator, 45 cases and 246 controls were selected; eight proteins were significantly up-regulated (adjusted p value < 0.05, log2(fold change) > 0.2) and 19 were significantly down-regulated (adjusted p value < 0.05, log2(fold change) < -0.2, Figure 6C; Tables S6A and S6E), including APOC1 and APOC4. Finally, as less than ten cases satisfied the selection criteria based on the blood glucose and HDL indicators, we did not analyze them.

Among the dysregulated proteins we identified, adiponectin (ADIPOQ) and taste receptor type 1 member 3 (TAS1R3) were dysregulated both in the TG-based and the waist-based MetS analyses (Figure 6D). This finding suggests that these two proteins may play a key role in the development of MetS. Additionally, SHBG was dysregulated both in the TG-based and the SBP/DBP-based MetS analyses and was also selected in our machine-learning model for predicting MetS risk.

Discussion

We generated a population proteomics resource based on 7,890 serum samples collected from a prospective cohort of 3,840 participants over 10 years, with a total of 18,970 DIA-MS data. We then generated a machine-learning model to predict the risk for an individual of developing MetS over the next 10 years. We also used this dataset to explore the molecular bases of MetS onset and development.

To our knowledge, our study provides the largest DIA-based serum proteome from a prospective population cohort with two follow-up time points to date. Niu et al. recently built a dataset of 311 plasma proteins from 596 individuals using a 21-min DIA-MS and Orbitrap Exploris 480 mass spectrometer.²⁶ Also, Orwoll et al. identified 224 proteins in 1,196 serum samples using liquid chromatography-ion mobility-MS (LC-IMS-MS) with a 58-min LC gradient.⁴² In our study, using a 20-min DIA-MS, 438 and 413 proteins were identified, with 19% and 18% missing values, in the independent protein matrices of the discovery and the validation cohorts, respectively.

A combination of blood proteomics and physiological status in a longitudinal population study can be used for disease risk prediction, as has been shown for various metabolic diseases with high incidence.⁴³^,⁴⁴ We used this to study MetS. In particular, we built a prospective machine-learning model to predict the risk of developing MetS within 10 years.

Elhadad et al. generated a LASSO model for the prospective analysis of incident MetS (mean follow-up time = 6.5) in a German population based on the data of 8 proteins.¹⁸ Their training cohort of 623 participants achieved an AUC of 0.74, but no validation cohort was used. Our study used 11 serum proteins plus age as model features. The AUC of our model was 0.784 in the discovery cohort (855 individuals) and 0.774 in the validation cohort (242 individuals). Also, our resource has an additional follow-up time point of 10 years, providing more accurate data for the long-term prediction of MetS. Age and gender were included as features during model generation, but the model did not select gender as a final feature. Therefore, we conclude that gender was not a significant factor in the model.

Our proteomics analysis provides essential insights into MetS, highlighting the apolipoproteins that play a crucial role in its pathophysiology. Apolipoproteins changes are associated with insulin resistance.⁸ In particular, APOC3, a component of the TG-rich very LPLs (VLDLs) and HDLs in plasma,⁴⁵ had been reported as a MetS risk factor in the Canadian⁴⁶ and Turkish populations.⁴⁷ Our results showed that the apolipoproteins APOC2, APOC3, APOA1, APOD, and APOF are not only associated with MetS onset but also with the development of MetS. Also, the expression of different apolipoproteins showed important variations in our data: APOC2 and APOC3 had high expression levels, while APOA1, APOD, and APOF had low expression levels. The higher expressions of APOC1, APOC2, APOC3, and APOE were already reported in MetS.⁴⁸ Therefore, exploring the biology of different apolipoproteins can provide more precise knowledge of the pathophysiology of MetS. In our data, APOC1, APOC2, APOC3, and APOE were up-regulated in the TG-based MetS, while APOD and APOF were down-regulated in HDL-based MetS. These observations suggest that apolipoproteins may have different roles and influences in MetS of different subtypes. Indeed, APOCs play an essential role in the etiology of human hyperlipidemias,⁴⁹ especially APOC2⁵⁰ and APOC3.⁵¹ Our data showed that these proteins may be related to MetS primarily by affecting TGs. Circulating APOD and APOF are mainly present in HDL⁵²^,⁵³ and may thus be related to MetS by affecting HDL. When we compared the serum proteomes of patients with and without MetS (Figure 4), two pathways, namely FXR/RXR activation (Figure S5) and LXR/RXR activation (Figure S6), were significantly enriched. According to IPA, apolipoproteins including APOA1, APOC1, APOC2, APOC4, APOE, and APOB modulate the hydrolysis of TGs, C metabolism, C secretion, and C efflux (Figures S5 and S6). Multiple metabolites involved in these metabolic pathways have been reported to be related to MetS. Examples of such metabolites include oleic acid,¹² epinephrine,⁵⁴ phosphatidylcholine,⁵⁵ C,⁵⁶ and homocysteine⁵⁷ (Figures S5 and S6).

In addition to apolipoproteins, we found that several other proteins are significantly associated with MetS development. Some of these proteins were already linked to MetS by other studies. SHBG can regulate the plasma metabolic clearance rate of steroid hormones and, in agreement with our findings, has been reported to predict the development of MetS in a population-based cohort study of middle-aged Finnish men.⁵⁸ VTN is a multifunctional adhesive glycoprotein that exerts regulatory functions in coagulation, fibrinolysis, or the plasminogen activation system.⁵⁹ In a prospective study, baseline plasma VTN was found as a marker of incident MetS at 9 years,⁶⁰ consistent with our findings. Also, the serum levels of PEDF encoded by SERPINF1 (a glycoprotein that belongs to the superfamily of serine protease inhibitors) have been reported associated with MetS in a cross-sectional study of the Japanese population.⁶¹ Furthermore, in a 10-year prospective study of Chinese men, PEDF was also associated with the development of MetS.⁶² A1AT encoded by SERPINA1 is an acute-phase inflammation marker that is associated with the development of the MetS.⁶³ AFM, a human plasma vitamin E-binding glycoprotein, was found to be strongly associated with the prevalence in a cross-sectional manner and development of MetS in a prospectively epidemiological study.⁶⁴ We also identified proteins that have not been reported to be directly associated with MetS. A1BG, which belongs to the Ig family, was reported to be elevated in the urine of patients with diabetes.⁶⁵^,⁶⁶ Interestingly, the lower expression was observed in the serum of patients with MetS or those who develop MetS within 10 years, indicating that the protein may be related to the humoral regulatory system of MetS. Another immune-related protein, ATRN, shows a higher level in circulating blood monocytes of human subjects because of obesity.⁶⁷ IGHV4-39, an immunoglobulin produced by B lymphocytes, has been reported to promote insulin resistance.⁶⁸ These results suggest that the immune system and coagulation-related systems may also be important regulatory systems in the development of MetS.

The core mechanism of metabolic diseases such as hypertension, hyperglycemia, hyperlipidemia, and obesity, which often appear together and develop into MetS, is still unclear. Here, we explored the dysregulated proteins linked to the development of MetS. We found that specific dysregulated proteins are connected with various metabolic abnormal states and MetS development, providing a new basis for understanding this mechanism. Among these proteins, we found AGT, THBS1, XPNPEP1, ADIPOQ, and TAS1R3. AGT is an essential component of the renin-angiotensin system (RAS), and increased insulin stimulates AGT production.⁶⁹ We found a higher expression of AGT in the population that later developed MetS, suggesting that AGT positively correlates with the development of MetS, consistent with other studies.⁷⁰ THBS1, an adipose-derived matricellular protein, has been reported as a biomarker of MetS in Japanese subjects.⁷¹ ADIPOQ is an anti-inflammatory cytokine and a biomarker for the development of MetS,⁸^,⁷² and it was down-regulated in our data. A lower ADIPOQ level was observed in Chinese adolescents,⁷³ peri- and post-menopausal women,⁷⁴ young Polish subjects⁷⁵ with MetS, and in a prospective study of the Korean general population with MetS.⁷⁶ Finally, TAS1R3 is a putative taste receptor reported to negatively correlate with glucose levels, TGs, and MetS.⁷⁷ XPNPEP1 contributes to the degradation of bradykinin, and the bradykinin system has beneficial effects on hypertension⁷⁸^,⁷⁹ and type 2 diabetes.⁸⁰ This agrees with our data showing higher levels of XPNPEP1 in the populations that later developed MetS. These findings are again consistent with our results and corroborate the robustness of our findings, which include other proteins that have not yet been linked to MetS development.

The GNHS community-based prospective cohort study was initiated in 2008. Comprehensive phenotypic data for the individuals involved in this study have been recorded over time, including but not limited to questionnaires, physical examinations, dual-energy X-ray absorptiometry, ultrasonography, and blood/urine/faces tests.²⁸^,²⁹^,³⁰^,³¹ The proteomic data acquired in this study could be potentially applied in understanding other health status and diseases such as diabetes.

In conclusion, we generated a resource of serum proteomics using DIA-MS based on a prospective population cohort with a 10-year follow-up. Using this data, we built a model for predicting the risk of developing MetS within 10 years. We also found new potential protein biomarkers directly related to MetS, including multiple immune-related proteins and pathways. This finding provides new insights into the pathophysiology of MetS. Our data also showed that apolipoproteins, SHBG, VTN, PEDF, A1AT, and AFM are not only related to the current status of MetS but are also related to its development over time. These findings broaden our understanding of MetS and provide information for future research and treatment.

Limitations of the study

The findings of this study have to be seen in the light of some limitations. First, we used a short-gradient, high-throughput proteomics workflow, which is more suitable for analyzing large cohorts. Consequently, the serum proteins we identified are relatively high in abundance. Also, our discovery and validation cohorts are from the same city. Our results, therefore, require a follow-up study using independent populations from different areas.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples

Serum samples	community-based prospective cohort GNHS	This paper (Table S1)

Chemicals, peptides, and recombinant proteins

Ammonium bicarbonate (ABB)	General-Reagent	Cat #G12990A
Urea	Sigma-Aldrich	Cat #U1250
Tris (2-carboxyethyl) phosphine (TCEP)	Adamas-beta	Cat # 61820E
Iodoacetamide (IAA)	Sigma-Aldrich	Cat #I6125
Trypsin	Hualishi Tech	Cat # HLS TRY001C
Trifluoroacetic acid (TFA)	Fisher Chemical	Cat # 85183
Water	Fisher Chemical	Cat #W6-4
Acetonitrile	Fisher Chemical	Cat # A955-4
Formic acid (FA)	Fisher Chemical	Cat # A117-50
Methanol	Sigma-Aldrich	Cat # HLS TRY001C

Critical commercial assays

96-well SOLAμ™ SPE plate	Thermo Fisher Scientific	Cat # 62209-001

Deposited data

Mass spectrometry data	This paper	http://proteomecentral.proteomexchange.org(PXD039236, PXD039231, PXD038253)

Software and algorithms
Analyst TF 1.7.1	AB Sciex	Cat # OPTON-30965
DIA-NN (1.8)	Demichev et al.³²	https://github.com/vdemichev/DiaNN
R version 3.6.1	R Project	https://www.r-project.org
Ingenuine pathway analysis (version 51963813)	Kramer et al.⁸³	https://www.qiagen.com/cn/

Other

Eksigent NanoLC 400 System	Eksigent	https://sciex.com/products/hplc-products/nanolc-products/nanolc-400
TripleTOF 5600 system	SCIEX	https://sciex.com/br/products/mass-spectrometers/qtof-systems/tripletof-systems/tripletof-5600-system

Open in a new tab

Resource availability

Lead contact

Further information should be directed to and will be fulfilled by the Lead Contact Tiannan Guo (guotiannan@westlake.edu.cn).

Materials availability

This study did not generate new unique reagents.

Experimental model and subject details

Ethics approval and consent to participate

The Guangzhou Nutrition and Health study protocol received approval from both the Ethics Committee of the School of Public Health at Sun Yat-sen University and the Ethics Committee of Westlake University. Prior to participating in the study, all participants provided written informed consent, as required by ethical standards.

Patients and samples

This study was based on the previously reported community-based prospective cohort Guangzhou Nutrition and Health Study (GNHS, ClinicalTrials.gov identifier: NCT03179657).²⁸^,²⁹^,³⁰^,³¹ Briefly, 3840 of 4048 participants aged 40–83 were enrolled in this study. The participants were recruited between 2008 and 2013 and followed up until 2019 (Figure 1, Tables 1, and S1A). In particular, participants were evaluated and blood samples were taken from them at up to three time points: baseline (between 2008 and 2013), second follow-up (between 2014 and 2017), and third follow-up (between 2018 and 2019). Venous whole blood samples were collected from all participants early in the morning before having food using serum separation tubes, and then centrifuged at 1811 G for 10 min for serum collection. The serum samples were frozen at −80°C before the analysis.

The diagnostic criteria for MetS in this study refer to the 2016 Chinese guidelines for managing dyslipidemia in adults.⁴⁰ Specifically, they included three or more of the following: central obesity/abdominal obesity: a waist circumference of at least 90 cm for men and at least 85 cm for women; hyperglycemia: fasting blood glucose of at least 6.10 mmol/L (110 mg/dL) or 2-h blood glucose after glycemic load of at least 7.80 mmol/L (140 mg/dL) or confirmed diabetes for patients that received treatment; hypertension: blood pressure of at least 130/85 mmHg or patients with hypertension that received treatment; fasting TG of at least 1.7 mmol/L (150 mg/dL); fasting HDL-C lower than 1.0 mmol/L (40 mg/dL).

Method details

Protein digestion

Peptides were extracted from the serum samples as previously described.⁸⁴^,⁸⁵ Briefly, 1 μL of serum sample was lysed using 20 μL of lysis buffer (8 M urea (Sigma, Catalog #U1230) in 100 mM ammonium bicarbonate, ABB) at 32°C for 30 min, then reduced and alkylated using 10 mM tris (2-carboxyethyl) phosphine (Sigma Catalog #T4708) and 40 mM iodoacetamide (Sigma, Catalog # SLCD4031), respectively. Before the enzymatic digestion, 70 μL of 100 mM ABB were added to the samples to dilute the urea. The protein extracts were then digested with a two-step overnight tryptic digestion (Hualishi Tech. Ltd, Beijing, China), using an enzyme-to-substrate ratio of 1:60 (final ratio 1:30) at 32°C for 4 h and then another 12 h. The digestion was then stopped by adjusting the pH to 2–3 using 1% trifluoroacetic acid (Thermo Fisher Scientific, Catalog #T/3258/PB05). Before the MS analysis, peptides were cleaned with HRP SOLAu columns (Thermo Fisher Scientific, San Jose, USA).

Mass spectrometric analysis

The peptide samples were injected into an Eksigent NanoLC 400 System (Eksigent, Dublin, CA, USA) and analyzed using SWATH-MS on a TripleTOF 5600 system (SCIEX, CA, USA) equipped with an electrospray ionization (ESI) source, as previously described.⁸⁵ Briefly, 0.5 μg of peptides were enriched in a trap column (5 μm, ChromXP C18CL, 120 Å, 10 × 0.3 mm) at a flow rate 10 μL/min for 3 min, and then separated on an analytical column (3 μm, ChromXP C18CL, 120 Å, 150 × 0.3 mm) with a 20-min LC gradient of 5–30% buffer B. Buffer A contained 2% acetonitrile (ACN) and 0.1% formic acid (FA) in HPLC water, while buffer B contained 98% ACN and 0.1% FA in HPLC water. Before analysis, the trap column was conditioned using buffer A for 2.5 min, and then the analytical column was conditioned for 3 min. For SWATH-MS, a 55 variable Q1 isolation window scheme was set as described in a previous study.⁸⁵ Bovine serum albumin (BSA) digests for mass and RT calibration were performed every day. Blank sample (buffer A) was performed every five samples. The batches were acquired by MS in random order. The samples in the same batch were injected into MS according to their serial numbers. An MS-QC sample (pooled serum peptides) was injected before each batch for quality control. Multiple maintenance operations, including MS calibration and column replacement, are performed between batches. Samples from the same batch are collected for MS analysis within one day to ensure comparability.

Proteome data analysis

The MS files were analyzed using DIA-NN (1.8)³² against a plasma spectral library⁸⁵ containing 5102 peptides and 819 unique proteins from the Swiss-Prot database of Homo sapiens. In the DIA-NN settings, the software automatically sets the retention time extraction window, while the m/z extraction window for MS1 and MS2 was set to 20 ppm and 50 ppm, respectively. Protein and peptide false discovery rates were set not to exceed 1%. Protein inference was set to the protein names (from the FASTA file), and the cross-run normalization was set as ‘RT-dependent’.

Machine learning

In the machine learning analysis, we used LightGBM Library (Ver. 3.3.2) in Python (Ver. 3.9) to generate a machine learning model using the discovery dataset of 855 samples containing 288 serum proteins, as well as age and gender information. First, we randomly split the dataset into a training dataset (n = 684) and an internal validation dataset (n = 171). Then, a 2-step feature selection was used for choosing protein features from the discovery dataset and other features from the patient’s clinical information (i.e., age, gender) to be included in the final classifier. The features were firstly selected based on proteome data and clinical data independently. In the second selection step, preselected features of the two categories were combined and any unrelated proteome was further eliminated, especially when the proteome was correlated with age and gender. The purpose of splitting the feature selection into two steps was to preserve proteomes that may have a weak contribution to classification, as including clinical information in the first place may result in the elimination of these proteomes due to their stronger contributions. During the first step, we used a Shapley value-based recursive feature elimination (RFE) algorithm,⁸⁶ followed by a Shapely value-based Boruta algorithm for serum proteome feature selection, and the feature importance-based RFE for age and gender selection. In the second step, we combined preselected proteomes and clinical information from the first step and performed another Shapley value-based Boruta algorithm to obtain the final feature set. Through feature selection, 11 proteins and age were selected as the final features. We then optimized our model using the training dataset with 10-fold cross-validation and tested the model with the internal validation dataset and the independent validation cohort.

Quantification and statistical analysis

The linear mixed model for the baseline analysis used the data collected at the baseline visit, while the linear mixed model for the dynamic analysis used the data collected from all three time points. For the two linear mixed models, MetS and non-MetS were set as outcome variables; the protein expression, age, and gender were set as fixed effects; the patient ID was set as a random effect in the linear mixed model for dynamic analysis. The value β represents the regression coefficient, which indicates the influence of independent variable effect on dependent outcome variable in the regression equation. The larger the regression coefficient is, the greater the influence of effect on outcome variable is. The positive regression coefficient means that outcome variable increases with the increase of effect, while the negative regression coefficient means that outcome variable decreases with the increase of effect.

Acknowledgments

This work is supported by grants from the National Key R&D Program of China (nos. 2022YF0608403 and 2021YFA1301600), the National Natural Science Foundation of China (nos. 82073546, 81773416, and 82103826), and the Zhejiang Provincial Natural Science Foundation of China (LQ21H260002). We thank Zhiheng Rao, Youqi Liu, Hao Chen, Zhengyang Luo, Bingjie Wang, and Huidi Ma for assistance in data generation and analysis. We thank the Westlake University Supercomputer Center for assistance in data generation and storage. We thank all study participants of the Guangzhou Nutrition and Health Study and all other team members involved in the cohort study.

Author contributions

T.G., J.-S.Z., Y.-m.C., and Y. Zhu designed and supervised the project. Y.-m.C. and F.-F.Z. established the cohort and collected the phenotype data and the samples. X.C., J.T., L.Y., Y.X., Z.M., W.G., Y.F., S.L., M.S., K.Z., F.X., Y.T., N.X., and Y. Zhou generated the data. X.C., Z.X., B.W., W.G., and J.G. analyzed the data. X.C., L.Y., P.-F.S., Y. Zhu, and T.G. drafted the manuscript with input from all co-authors. X.C., Z.X., F.Z., J.T., L.Y., and B.W. contributed equally to this work.

Declaration of interests

Y. Zhu and T.G. are shareholders of Westlake Omics, Inc. B.W., W.G., and N.X. are employees of Westlake Omics, Inc. A patent related to this study has been submitted to the China National Intellectual Property Administration.

Inclusion and diversity

We support inclusive, diverse, and equitable conduct of research.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT in order to improve language and readability. After using this service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Published: August 30, 2023

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.xcrm.2023.101172.

Contributor Information

Yi Zhu, Email: zhuyi@westlake.edu.cn.

Yu-ming Chen, Email: chenyum@mail.sysu.edu.cn.

Ju-Sheng Zheng, Email: zhengjusheng@westlake.edu.cn.

Tiannan Guo, Email: guotiannan@westlake.edu.cn.

Supplemental information

Document S1. Figures S1–S6 and Tables S1–S6

mmc1.pdf^{(1.2MB, pdf)}

Document S2. Article plus supplemental information

mmc2.pdf^{(8.2MB, pdf)}

Date and code availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the iProX partner repository⁸¹^,⁸² with the dataset identifier PXD039236, PXD039231, and PXD038253. The phenotype data can be requested by e-mail from the corresponding author Yu-ming Chen (chenyum@mail.sysu.edu.cn). The code used in this study has been deposited at GitHub under https://github.com/guomics-lab/GNHS. Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

References

1.Expert Panel on Detection, E., and Treatment of High Blood Cholesterol in, Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) JAMA. 2001;285:2486–2497. doi: 10.1001/jama.285.19.2486. [DOI] [PubMed] [Google Scholar]
2.Alberti K.G.M.M., Eckel R.H., Grundy S.M., Zimmet P.Z., Cleeman J.I., Donato K.A., Fruchart J.-C., James W.P.T., Loria C.M., Smith S.C., et al. Harmonizing the Metabolic Syndrome. Circulation. 2009;120:1640–1645. doi: 10.1161/CIRCULATIONAHA.109.192644. [DOI] [PubMed] [Google Scholar]
3.International Diabetes Federation: The IDF consensus worldwide definition of the metabolic syndrome.
4.Ranasinghe P., Mathangasinghe Y., Jayawardena R., Hills A.P., Misra A. Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review. BMC Publ. Health. 2017;17:101. doi: 10.1186/s12889-017-4041-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.do Vale Moreira N.C., Hussain A., Bhowmik B., Mdala I., Siddiquee T., Fernandes V.O., Montenegro Júnior R.M., Meyer H.E. Prevalence of Metabolic Syndrome by different definitions, and its association with type 2 diabetes, pre-diabetes, and cardiovascular disease risk in Brazil. Diabetes Metabol. Syndr. 2020;14:1217–1224. doi: 10.1016/j.dsx.2020.05.043. [DOI] [PubMed] [Google Scholar]
6.Li Y., Zhao L., Yu D., Wang Z., Ding G. Metabolic syndrome prevalence and its risk factors among adults in China: A nationally representative cross-sectional study. PLoS One. 2018;13:e0199293. doi: 10.1371/journal.pone.0199293. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.National Center for Health Statistics, D.o.H.I.S . National Center for Chronic Disease Prevention and Health Promotion. Centers for Disease Control and Prevention, Division of Diabetes Translation; 2012. Crude and age-adjusted percentage of civilian, noninstitutionalized adults with diagnosed diabetes, United States, 1980–2010. [Google Scholar]
8.Eckel R.H., Grundy S.M., Zimmet P.Z. The metabolic syndrome. Lancet. 2005;365:1415–1428. doi: 10.1016/S0140-6736(05)66378-7. [DOI] [PubMed] [Google Scholar]
9.Alshehri A.M. Metabolic syndrome and cardiovascular risk. J. Family Community Med. 2010;17:73–78. doi: 10.4103/1319-1683.71987. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Esposito K., Chiodini P., Colao A., Lenzi A., Giugliano D. Metabolic syndrome and risk of cancer: a systematic review and meta-analysis. Diabetes Care. 2012;35:2402–2411. doi: 10.2337/dc12-0336. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Floegel A., Wientzek A., Bachlechner U., Jacobs S., Drogan D., Prehn C., Adamski J., Krumsiek J., Schulze M.B., Pischon T., Boeing H. Linking diet, physical activity, cardiorespiratory fitness and obesity to serum metabolite networks: findings from a population-based study. Int. J. Obes. 2014;38:1388–1396. doi: 10.1038/ijo.2014.39. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Palacios-González B., León-Reyes G., Rivera-Paredez B., Ibarra-González I., Vela-Amieva M., Flores Y.N., Canizales-Quinteros S., Salmerón J., Velázquez-Cruz R. Targeted Metabolomics Revealed a Sex-Dependent Signature for Metabolic Syndrome in the Mexican Population. Nutrients. 2022;14:3678. doi: 10.3390/nu14183678. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Pujos-Guillot E., Brandolini M., Pétéra M., Grissa D., Joly C., Lyan B., Herquelot É., Czernichow S., Zins M., Goldberg M., Comte B. Systems Metabolomics for Prediction of Metabolic Syndrome. J. Proteome Res. 2017;16:2262–2272. doi: 10.1021/acs.jproteome.7b00116. [DOI] [PubMed] [Google Scholar]
14.Schrimpe-Rutledge A.C., Codreanu S.G., Sherrod S.D., McLean J.A. Untargeted Metabolomics Strategies-Challenges and Emerging Directions. J. Am. Soc. Mass Spectrom. 2016;27:1897–1905. doi: 10.1007/s13361-016-1469-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Gold L., Ayers D., Bertino J., Bock C., Bock A., Brody E.N., Carter J., Dalby A.B., Eaton B.E., Fitzwater T., et al. Aptamer-based multiplexed proteomic technology for biomarker discovery. PLoS One. 2010;5:e15004. doi: 10.1371/journal.pone.0015004. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Assarsson E., Lundberg M., Holmquist G., Björkesten J., Thorsen S.B., Ekman D., Eriksson A., Rennel Dickens E., Ohlsson S., Edfeldt G., et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One. 2014;9:e95192. doi: 10.1371/journal.pone.0095192. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Ngo D., Sinha S., Shen D., Kuhn E.W., Keyes M.J., Shi X., Benson M.D., O'Sullivan J.F., Keshishian H., Farrell L.A., et al. Aptamer-Based Proteomic Profiling Reveals Novel Candidate Biomarkers and Pathways in Cardiovascular Disease. Circulation. 2016;134:270–285. doi: 10.1161/CIRCULATIONAHA.116.021803. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Elhadad M.A., Wilson R., Zaghlool S.B., Huth C., Gieger C., Grallert H., Graumann J., Rathmann W., Koenig W., Sinner M.F., et al. Metabolic syndrome and the plasma proteome: from association to causation. Cardiovasc. Diabetol. 2021;20:111. doi: 10.1186/s12933-021-01299-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Ganz P., Heidecker B., Hveem K., Jonasson C., Kato S., Segal M.R., Sterling D.G., Williams S.A. Development and Validation of a Protein-Based Risk Score for Cardiovascular Outcomes Among Patients With Stable Coronary Heart Disease. JAMA. 2016;315:2532–2541. doi: 10.1001/jama.2016.5951. [DOI] [PubMed] [Google Scholar]
20.Lehallier B., Gate D., Schaum N., Nanasi T., Lee S.E., Yousef H., Moran Losada P., Berdnik D., Keller A., Verghese J., et al. Undulating changes in human plasma proteome profiles across the lifespan. Nat. Med. 2019;25:1843–1850. doi: 10.1038/s41591-019-0673-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Ponce-de-Leon M., Linseisen J., Peters A., Linkohr B., Heier M., Grallert H., Schöttker B., Trares K., Bhardwaj M., Gào X., et al. Novel associations between inflammation-related proteins and adiposity: A targeted proteomics approach across four population-based studies. Transl. Res. 2022;242:93–104. doi: 10.1016/j.trsl.2021.11.004. [DOI] [PubMed] [Google Scholar]
22.Ferreira J.P., Sharma A., Mehta C., Bakris G., Rossignol P., White W.B., Zannad F. Multi-proteomic approach to predict specific cardiovascular events in patients with diabetes and myocardial infarction: findings from the EXAMINE trial. Clin. Res. Cardiol. 2021;110:1006–1019. doi: 10.1007/s00392-020-01729-3. [DOI] [PubMed] [Google Scholar]
23.Zhu Y., Aebersold R., Mann M., Guo T. SnapShot: Clinical proteomics. Cell. 2021;184:4840–4840.e1. doi: 10.1016/j.cell.2021.08.015. [DOI] [PubMed] [Google Scholar]
24.Gillet L.C., Navarro P., Tate S., Röst H., Selevsek N., Reiter L., Bonner R., Aebersold R. Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis. Mol. Cell. Proteomics. 2012;11 doi: 10.1074/mcp.O111.016717. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Guo T., Kouvonen P., Koh C.C., Gillet L.C., Wolski W.E., Röst H.L., Rosenberger G., Collins B.C., Blum L.C., Gillessen S., et al. Rapid mass spectrometric conversion of tissue biopsy samples into permanent quantitative digital proteome maps. Nat. Med. 2015;21:407–413. doi: 10.1038/nm.3807. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Niu L., Thiele M., Geyer P.E., Rasmussen D.N., Webel H.E., Santos A., Gupta R., Meier F., Strauss M., Kjaergaard M., et al. Noninvasive proteomic biomarkers for alcohol-related liver disease. Nat. Med. 2022;28:1277–1287. doi: 10.1038/s41591-022-01850-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Bruderer R., Muntel J., Müller S., Bernhardt O.M., Gandhi T., Cominetti O., Macron C., Carayol J., Rinner O., Astrup A., et al. Analysis of 1508 Plasma Samples by Capillary-Flow Data-Independent Acquisition Profiles Proteomics of Weight Loss and Maintenance. Mol. Cell. Proteomics. 2019;18:1242–1254. doi: 10.1074/mcp.RA118.001288. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Jiang Z., Sun T.Y., He Y., Gou W., Zuo L.S.Y., Fu Y., Miao Z., Shuai M., Xu F., Xiao C., et al. Dietary fruit and vegetable intake, gut microbiota, and type 2 diabetes: results from two large human cohort studies. BMC Med. 2020;18:371. doi: 10.1186/s12916-020-01842-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Miao Z., Lin J.S., Mao Y., Chen G.D., Zeng F.F., Dong H.L., Jiang Z., Wang J., Xiao C., Shuai M., et al. Erythrocyte n-6 Polyunsaturated Fatty Acids, Gut Microbiota, and Incident Type 2 Diabetes: A Prospective Cohort Study. Diabetes Care. 2020;43:2435–2443. doi: 10.2337/dc20-0631. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Gou W., Ling C.W., He Y., Jiang Z., Fu Y., Xu F., Miao Z., Sun T.Y., Lin J.S., Zhu H.L., et al. Interpretable Machine Learning Framework Reveals Robust Gut Microbiome Features Associated With Type 2 Diabetes. Diabetes Care. 2021;44:358–366. doi: 10.2337/dc20-1536. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Gou W., Fu Y., Yue L., Chen G.D., Cai X., Shuai M., Xu F., Yi X., Chen H., Zhu Y., et al. Gut microbiota, inflammation, and molecular signatures of host response to infection. J Genet Genomics. 2021;48:792–802. doi: 10.1016/j.jgg.2021.04.002. [DOI] [PubMed] [Google Scholar]
32.Demichev V., Messner C.B., Vernardis S.I., Lilley K.S., Ralser M. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. s. 2020;17:41–44. doi: 10.1038/s41592-019-0638-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Zhu T., Chen H., Yan X., Wu Z., Zhou X., Xiao Q., Ge W., Zhang Q., Xu C., Xu L., et al. ProteomeExpert: a docker image based web-server for exploring, modeling, visualizing, and mining quantitative proteomic data sets. Bioinformatics. 2021;37:273–275. doi: 10.1093/bioinformatics/btaa1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Desvergne B. RXR: from partnership to leadership in metabolic regulations. Vitam. Horm. 2007;75:1–32. doi: 10.1016/S0083-6729(06)75001-4. [DOI] [PubMed] [Google Scholar]
35.Shulman A.I., Mangelsdorf D.J. Retinoid x receptor heterodimers in the metabolic syndrome. N. Engl. J. Med. 2005;353:604–615. doi: 10.1056/NEJMra043590. [DOI] [PubMed] [Google Scholar]
36.McNeill A.M., Rosamond W.D., Girman C.J., Golden S.H., Schmidt M.I., East H.E., Ballantyne C.M., Heiss G. The metabolic syndrome and 11-year risk of incident cardiovascular disease in the atherosclerosis risk in communities study. Diabetes Care. 2005;28:385–390. doi: 10.2337/diacare.28.2.385. [DOI] [PubMed] [Google Scholar]
37.Thanabalasingham G., Pal A., Selwood M.P., Dudley C., Fisher K., Bingley P.J., Ellard S., Farmer A.J., McCarthy M.I., Owen K.R. Systematic assessment of etiology in adults with a clinical diagnosis of young-onset type 2 diabetes is a successful strategy for identifying maturity-onset diabetes of the young. Diabetes Care. 2012;35:1206–1212. doi: 10.2337/dc11-1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Pickup J.C., Mattock M.B., Chusney G.D., Burt D. NIDDM as a disease of the innate immune system: association of acute-phase reactants and interleukin-6 with metabolic syndrome X. Diabetologia. 1997;40:1286–1292. doi: 10.1007/s001250050822. [DOI] [PubMed] [Google Scholar]
39.Rizza R.A., Mandarino L.J., Gerich J.E. Effects of growth hormone on insulin action in man. Mechanisms of insulin resistance, impaired suppression of glucose production, and impaired stimulation of glucose utilization. Diabetes. 1982;31:663–669. doi: 10.2337/diab.31.8.663. [DOI] [PubMed] [Google Scholar]
40.Joint committee for guideline revision 2016 Chinese guidelines for the management of dyslipidemia in adults. J. Geriatr. Cardiol. 2018;15:1–29. doi: 10.11909/j.issn.1671-5411.2018.01.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Ruderman N., Prentki M. AMP kinase and malonyl-CoA: targets for therapy of the metabolic syndrome. Nat. Rev. Drug Discov. 2004;3:340–351. doi: 10.1038/nrd1344. [DOI] [PubMed] [Google Scholar]
42.Orwoll E.S., Wiedrick J., Nielson C.M., Jacobs J., Baker E.S., Piehowski P., Petyuk V., Gao Y., Shi T., Smith R.D., et al. Proteomic assessment of serum biomarkers of longevity in older men. Aging Cell. 2020;19:e13253. doi: 10.1111/acel.13253. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Schüssler-Fiorenza Rose S.M., Contrepois K., Moneghetti K.J., Zhou W., Mishra T., Mataraso S., Dagan-Rosenfeld O., Ganz A.B., Dunn J., Hornburg D., et al. A longitudinal big data approach for precision health. Nat. Med. 2019;25:792–804. doi: 10.1038/s41591-019-0414-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Robbins J.M., Peterson B., Schranner D., Tahir U.A., Rienmüller T., Deng S., Keyes M.J., Katz D.H., Beltran P.M.J., Barber J.L., et al. Human plasma proteomic profiles indicative of cardiorespiratory fitness. Nat. Metab. 2021;3:786–797. doi: 10.1038/s42255-021-00400-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Yao Z., Wang Y. Apolipoprotein C-III and hepatic triglyceride-rich lipoprotein production. Curr. Opin. Lipidol. 2012;23:206–212. doi: 10.1097/MOL.0b013e328352dc70. [DOI] [PubMed] [Google Scholar]
46.Pollex R.L., Hanley A.J.G., Zinman B., Harris S.B., Khan H.M.R., Hegele R.A. Metabolic syndrome in aboriginal Canadians: prevalence and genetic associations. Atherosclerosis. 2006;184:121–129. doi: 10.1016/j.atherosclerosis.2005.03.024. [DOI] [PubMed] [Google Scholar]
47.Onat A., Hergenç G., Sansoy V., Fobker M., Ceyhan K., Toprak S., Assmann G. Apolipoprotein C-III, a strong discriminant of coronary risk in men and a determinant of the metabolic syndrome in both genders. Atherosclerosis. 2003;168:81–89. doi: 10.1016/s0021-9150(03)00025-x. [DOI] [PubMed] [Google Scholar]
48.Savinova O.V., Fillaus K., Jing L., Harris W.S., Shearer G.C. Reduced apolipoprotein glycosylation in patients with the metabolic syndrome. PLoS One. 2014;9:e104833. doi: 10.1371/journal.pone.0104833. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Jong M.C., Hofker M.H., Havekes L.M. Role of ApoCs in lipoprotein metabolism: functional differences between ApoC1, ApoC2, and ApoC3. Arterioscler. Thromb. Vasc. Biol. 1999;19:472–484. doi: 10.1161/01.atv.19.3.472. [DOI] [PubMed] [Google Scholar]
50.Breckenridge W.C., Little J.A., Steiner G., Chow A., Poapst M. Hypertriglyceridemia associated with deficiency of apolipoprotein C-II. N. Engl. J. Med. 1978;298:1265–1273. doi: 10.1056/NEJM197806082982301. [DOI] [PubMed] [Google Scholar]
51.Carlson L.A., Ballantyne D. Changing relative proportions of apolipoproteins CII and CIII of very low density lipoproteins in hypertriglyceridaemia. Atherosclerosis. 1976;23:563–568. doi: 10.1016/0021-9150(76)90016-2. [DOI] [PubMed] [Google Scholar]
52.McConathy W.J., Alaupovic P. Isolation and partial characterization of apolipoprotein D: a new protein moiety of the human plasma lipoprotein system. FEBS Lett. 1973;37:178–182. doi: 10.1016/0014-5793(73)80453-3. [DOI] [PubMed] [Google Scholar]
53.Olofsson S.O., McConathy W.J., Alaupovic P. Isolation and partial characterization of a new acidic apolipoprotein (apolipoprotein F) from high density lipoproteins of human plasma. Biochemistry. 1978;17:1032–1036. doi: 10.1021/bi00599a014. [DOI] [PubMed] [Google Scholar]
54.Seematter G., Binnert C., Tappy L. Stress and metabolism. Metab. Syndr. Relat. Disord. 2005;3:8–13. doi: 10.1089/met.2005.3.8. [DOI] [PubMed] [Google Scholar]
55.Kulkarni H., Meikle P.J., Mamtani M., Weir J.M., Barlow C.K., Jowett J.B., Bellis C., Dyer T.D., Johnson M.P., Rainwater D.L., et al. Variability in associations of phosphatidylcholine molecular species with metabolic syndrome in Mexican-American families. Lipids. 2013;48:497–503. doi: 10.1007/s11745-013-3781-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Wang S. Association between serum low-density lipoprotein cholesterol and metabolic syndrome in a working population. Lipids Health Dis. 2021;20:73. doi: 10.1186/s12944-021-01500-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Vuksan-Ćusa B., Jakovljević M., Sagud M., Mihaljević Peleš A., Marčinko D., Topić R., Mihaljević S., Sertić J. Metabolic syndrome and serum homocysteine in patients with bipolar disorder and schizophrenia treated with second generation antipsychotics. Psychiatr. Res. 2011;189:21–25. doi: 10.1016/j.psychres.2010.11.021. [DOI] [PubMed] [Google Scholar]
58.Laaksonen D.E., Niskanen L., Punnonen K., Nyyssönen K., Tuomainen T.P., Valkonen V.P., Salonen R., Salonen J.T. Testosterone and sex hormone-binding globulin predict the metabolic syndrome and diabetes in middle-aged men. Diabetes Care. 2004;27:1036–1041. doi: 10.2337/diacare.27.5.1036. [DOI] [PubMed] [Google Scholar]
59.Preissner K.T. Structure and biological role of vitronectin. Annu. Rev. Cell Biol. 1991;7:275–310. doi: 10.1146/annurev.cb.07.110191.001423. [DOI] [PubMed] [Google Scholar]
60.Alessi M.C., Nicaud V., Scroyen I., Lange C., Saut N., Fumeron F., Marre M., Lantieri O., Fontaine-Bisson B., Juhan-Vague I., et al. Association of vitronectin and plasminogen activator inhibitor-1 levels with the risk of metabolic syndrome and type 2 diabetes mellitus. Results from the D.E.S.I.R. prospective cohort. Thromb. Haemostasis. 2011;106:416–422. doi: 10.1160/TH11-03-0179. [DOI] [PubMed] [Google Scholar]
61.Yamagishi S.I., Adachi H., Abe A., Yashiro T., Enomoto M., Furuki K., Hino A., Jinnouchi Y., Takenaka K., Matsui T., et al. Elevated serum levels of pigment epithelium-derived factor in the metabolic syndrome. J. Clin. Endocrinol. Metab. 2006;91:2447–2450. doi: 10.1210/jc.2005-2654. [DOI] [PubMed] [Google Scholar]
62.Chen C., Tso A.W.K., Law L.S.C., Cheung B.M.Y., Ong K.L., Wat N.M.S., Janus E.D., Xu A., Lam K.S.L. Plasma level of pigment epithelium-derived factor is independently associated with the development of the metabolic syndrome in Chinese men: a 10-year prospective study. J. Clin. Endocrinol. Metab. 2010;95:5074–5081. doi: 10.1210/jc.2010-0727. [DOI] [PubMed] [Google Scholar]
63.Setoh K., Terao C., Muro S., Kawaguchi T., Tabara Y., Takahashi M., Nakayama T., Kosugi S., Sekine A., Yamada R., et al. Three missense variants of metabolic syndrome-related genes are associated with alpha-1 antitrypsin levels. Nat. Commun. 2015;6:7754. doi: 10.1038/ncomms8754. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Kronenberg F., Kollerits B., Kiechl S., Lamina C., Kedenko L., Meisinger C., Willeit J., Huth C., Wietzorrek G., Altmann M.E., et al. Plasma concentrations of afamin are associated with the prevalence and development of metabolic syndrome. Circ. Cardiovasc. Genet. 2014;7:822–829. doi: 10.1161/CIRCGENETICS.113.000654. [DOI] [PubMed] [Google Scholar]
65.Soggiu A., Piras C., Bonizzi L., Hussein H.A., Pisanu S., Roncada P. A discovery-phase urine proteomics investigation in type 1 diabetes. Acta Diabetol. 2012;49:453–464. doi: 10.1007/s00592-012-0407-0. [DOI] [PubMed] [Google Scholar]
66.Gourgari E., Ma J., Playford M.P., Mehta N.N., Goldman R., Remaley A.T., Gordon S.M. Proteomic alterations of HDL in youth with type 1 diabetes and their associations with glycemic control: a case-control study. Cardiovasc. Diabetol. 2019;18:43. doi: 10.1186/s12933-019-0846-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Laudes M., Oberhauser F., Schulte D.M., Schilbach K., Freude S., Bilkovski R., Schulz O., Faust M., Krone W. Dipeptidyl-peptidase 4 and attractin expression is increased in circulating blood monocytes of obese human subjects. Exp. Clin. Endocrinol. Diabetes. 2010;118:473–477. doi: 10.1055/s-0030-1249014. [DOI] [PubMed] [Google Scholar]
68.Winer D.A., Winer S., Shen L., Wadia P.P., Yantha J., Paltser G., Tsui H., Wu P., Davidson M.G., Alonso M.N., et al. B cells promote insulin resistance through modulation of T cells and production of pathogenic IgG antibodies. Nat. Med. 2011;17:610–617. doi: 10.1038/nm.2353. [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Harte A.L., McTernan P.G., McTernan C.L., Crocker J., Starcynski J., Barnett A.H., Matyka K., Kumar S. Insulin increases angiotensinogen expression in human abdominal subcutaneous adipocytes. Diabetes Obes. Metabol. 2003;5:462–467. doi: 10.1046/j.1463-1326.2003.00274.x. [DOI] [PubMed] [Google Scholar]
70.Aubert J., Safonova I., Negrel R., Ailhaud G. Insulin down-regulates angiotensinogen gene expression and angiotensinogen secretion in cultured adipose cells. Biochem. Biophys. Res. Commun. 1998;250:77–82. doi: 10.1006/bbrc.1998.9185. [DOI] [PubMed] [Google Scholar]
71.Matsuo Y., Tanaka M., Yamakage H., Sasaki Y., Muranaka K., Hata H., Ikai I., Shimatsu A., Inoue M., Chun T.H., Satoh-Asahara N. Thrombospondin 1 as a novel biological marker of obesity and metabolic syndrome. Metabolism. 2015;64:1490–1499. doi: 10.1016/j.metabol.2015.07.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Matsuzawa Y., Funahashi T., Kihara S., Shimomura I. Adiponectin and metabolic syndrome. Arterioscler. Thromb. Vasc. Biol. 2004;24:29–33. doi: 10.1161/01.ATV.0000099786.99623.EF. [DOI] [PubMed] [Google Scholar]
73.Li Q., Xu L.Z., Zou T., Ai P., Huang G.H., Li P., Zheng A.P. Correlation of serum adiponectin and adiponectin gene polymorphism with metabolic syndrome in Chinese adolescents. Eur. J. Clin. Nutr. 2015;10:62–67. doi: 10.1038/ejcn.2014.152. [DOI] [PubMed] [Google Scholar]
74.Wattanapol P., Vichinsartvichai P., Sakoonwatanyoo P. Serum adiponectin is a potential biomarker for metabolic syndrome in peri-and postmenopausal women. Gynecol. Endocrinol. 2020;36:620–625. doi: 10.1080/09513590.2020.1742688. [DOI] [PubMed] [Google Scholar]
75.Kowalska I., Straczkowski M., Nikolajuk A., Adamska A., Karczewska-Kupczewska M., Otziomek E., Kinalska I., Gorska M. Insulin resistance, serum adiponectin, and proinflammatory markers in young subjects with the metabolic syndrome. Metabolism. 2008;57:1539–1544. doi: 10.1016/j.metabol.2008.06.008. [DOI] [PubMed] [Google Scholar]
76.Kim J.Y., Ahn S.V., Yoon J.H., Koh S.B., Yoon J., Yoo B.S., Lee S.H., Park J.K., Choe K.H., Guallar E. Prospective study of serum adiponectin and incident metabolic syndrome: the ARIRANG study. Diabetes Care. 2013;36:1547–1553. doi: 10.2337/dc12-0223. [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Bertran L., Portillo-Carrasquer M., Martínez S., Aguilar C., Lopez-Dupla M., Riesco D., Binetti J., Vives M., Sabench F., Del Castillo D., et al. Expression of Jejunal Taste Receptors in Women with Morbid Obesity. Nutrients. 2021;13:2437. doi: 10.3390/nu13072437. [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Sharma J.N. Hypertension and the bradykinin system. Curr. Hypertens. Rep. 2009;11:178–181. doi: 10.1007/s11906-009-0032-7. [DOI] [PubMed] [Google Scholar]
79.Agarwal R. Bradykinin and inhibition of angiotensin-converting enzyme in hypertension. N. Engl. J. Med. 1999;340:967–968. [PubMed] [Google Scholar]
80.Mandle R.J., Colman R.W., Kaplan A.P. Identification of prekallikrein and high-molecular-weight kininogen as a complex in human plasma. Proc. Natl. Acad. Sci. USA. 1976;73:4179–4183. doi: 10.1073/pnas.73.11.4179. [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Ma J., Chen T., Wu S., Yang C., Bai M., Shu K., Li K., Zhang G., Jin Z., He F., et al. iProX: an integrated proteome resource. Nucleic Acids Res. 2019;47:D1211–D1217. doi: 10.1093/nar/gky869. [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Chen T., Ma J., Liu Y., Chen Z., Xiao N., Lu Y., Fu Y., Yang C., Li M., Wu S., et al. iProX in 2021: connecting proteomics data sharing with big data. Nucleic Acids Res. 2022;50:D1522–D1527. doi: 10.1093/nar/gkab1081. [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Shen B., Yi X., Sun Y., Bi X., Du J., Zhang C., Quan S., Zhang F., Sun R., Qian L., et al. Proteomic and Metabolomic Characterization of COVID-19 Patient Sera. Cell. 2020;182:59–72.e15. doi: 10.1016/j.cell.2020.05.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Zhang Y., Cai X., Ge W., Wang D., Zhu G., Qian L., Xiang N., Yue L., Liang S., Zhang F., et al. Potential Use of Serum Proteomics for Monitoring COVID-19 Progression to Complement RT-PCR Detection. J. Proteome Res. 2022;21:90–100. doi: 10.1021/acs.jproteome.1c00525. [DOI] [PubMed] [Google Scholar]
86.Scott M., Lundberg S.-I.L. 2017. A unified approach to interpreting model predictions. (Advances in Neural Information Processing Systems). [Google Scholar]
83.Krämer J., Green J., Pollard J., Jr, Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30:523–530. doi: 10.1093/bioinformatics/btt703. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S6 and Tables S1–S6

mmc1.pdf^{(1.2MB, pdf)}

Document S2. Article plus supplemental information

mmc2.pdf^{(8.2MB, pdf)}

Data Availability Statement

[bib1] 1.Expert Panel on Detection, E., and Treatment of High Blood Cholesterol in, Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) JAMA. 2001;285:2486–2497. doi: 10.1001/jama.285.19.2486. [DOI] [PubMed] [Google Scholar]

[bib2] 2.Alberti K.G.M.M., Eckel R.H., Grundy S.M., Zimmet P.Z., Cleeman J.I., Donato K.A., Fruchart J.-C., James W.P.T., Loria C.M., Smith S.C., et al. Harmonizing the Metabolic Syndrome. Circulation. 2009;120:1640–1645. doi: 10.1161/CIRCULATIONAHA.109.192644. [DOI] [PubMed] [Google Scholar]

[bib3] 3.International Diabetes Federation: The IDF consensus worldwide definition of the metabolic syndrome.

[bib4] 4.Ranasinghe P., Mathangasinghe Y., Jayawardena R., Hills A.P., Misra A. Prevalence and trends of metabolic syndrome among adults in the asia-pacific region: a systematic review. BMC Publ. Health. 2017;17:101. doi: 10.1186/s12889-017-4041-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib5] 5.do Vale Moreira N.C., Hussain A., Bhowmik B., Mdala I., Siddiquee T., Fernandes V.O., Montenegro Júnior R.M., Meyer H.E. Prevalence of Metabolic Syndrome by different definitions, and its association with type 2 diabetes, pre-diabetes, and cardiovascular disease risk in Brazil. Diabetes Metabol. Syndr. 2020;14:1217–1224. doi: 10.1016/j.dsx.2020.05.043. [DOI] [PubMed] [Google Scholar]

[bib6] 6.Li Y., Zhao L., Yu D., Wang Z., Ding G. Metabolic syndrome prevalence and its risk factors among adults in China: A nationally representative cross-sectional study. PLoS One. 2018;13:e0199293. doi: 10.1371/journal.pone.0199293. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.National Center for Health Statistics, D.o.H.I.S . National Center for Chronic Disease Prevention and Health Promotion. Centers for Disease Control and Prevention, Division of Diabetes Translation; 2012. Crude and age-adjusted percentage of civilian, noninstitutionalized adults with diagnosed diabetes, United States, 1980–2010. [Google Scholar]

[bib8] 8.Eckel R.H., Grundy S.M., Zimmet P.Z. The metabolic syndrome. Lancet. 2005;365:1415–1428. doi: 10.1016/S0140-6736(05)66378-7. [DOI] [PubMed] [Google Scholar]

[bib9] 9.Alshehri A.M. Metabolic syndrome and cardiovascular risk. J. Family Community Med. 2010;17:73–78. doi: 10.4103/1319-1683.71987. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Esposito K., Chiodini P., Colao A., Lenzi A., Giugliano D. Metabolic syndrome and risk of cancer: a systematic review and meta-analysis. Diabetes Care. 2012;35:2402–2411. doi: 10.2337/dc12-0336. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Floegel A., Wientzek A., Bachlechner U., Jacobs S., Drogan D., Prehn C., Adamski J., Krumsiek J., Schulze M.B., Pischon T., Boeing H. Linking diet, physical activity, cardiorespiratory fitness and obesity to serum metabolite networks: findings from a population-based study. Int. J. Obes. 2014;38:1388–1396. doi: 10.1038/ijo.2014.39. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Palacios-González B., León-Reyes G., Rivera-Paredez B., Ibarra-González I., Vela-Amieva M., Flores Y.N., Canizales-Quinteros S., Salmerón J., Velázquez-Cruz R. Targeted Metabolomics Revealed a Sex-Dependent Signature for Metabolic Syndrome in the Mexican Population. Nutrients. 2022;14:3678. doi: 10.3390/nu14183678. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Pujos-Guillot E., Brandolini M., Pétéra M., Grissa D., Joly C., Lyan B., Herquelot É., Czernichow S., Zins M., Goldberg M., Comte B. Systems Metabolomics for Prediction of Metabolic Syndrome. J. Proteome Res. 2017;16:2262–2272. doi: 10.1021/acs.jproteome.7b00116. [DOI] [PubMed] [Google Scholar]

[bib14] 14.Schrimpe-Rutledge A.C., Codreanu S.G., Sherrod S.D., McLean J.A. Untargeted Metabolomics Strategies-Challenges and Emerging Directions. J. Am. Soc. Mass Spectrom. 2016;27:1897–1905. doi: 10.1007/s13361-016-1469-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] 15.Gold L., Ayers D., Bertino J., Bock C., Bock A., Brody E.N., Carter J., Dalby A.B., Eaton B.E., Fitzwater T., et al. Aptamer-based multiplexed proteomic technology for biomarker discovery. PLoS One. 2010;5:e15004. doi: 10.1371/journal.pone.0015004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] 16.Assarsson E., Lundberg M., Holmquist G., Björkesten J., Thorsen S.B., Ekman D., Eriksson A., Rennel Dickens E., Ohlsson S., Edfeldt G., et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One. 2014;9:e95192. doi: 10.1371/journal.pone.0095192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Ngo D., Sinha S., Shen D., Kuhn E.W., Keyes M.J., Shi X., Benson M.D., O'Sullivan J.F., Keshishian H., Farrell L.A., et al. Aptamer-Based Proteomic Profiling Reveals Novel Candidate Biomarkers and Pathways in Cardiovascular Disease. Circulation. 2016;134:270–285. doi: 10.1161/CIRCULATIONAHA.116.021803. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Elhadad M.A., Wilson R., Zaghlool S.B., Huth C., Gieger C., Grallert H., Graumann J., Rathmann W., Koenig W., Sinner M.F., et al. Metabolic syndrome and the plasma proteome: from association to causation. Cardiovasc. Diabetol. 2021;20:111. doi: 10.1186/s12933-021-01299-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Ganz P., Heidecker B., Hveem K., Jonasson C., Kato S., Segal M.R., Sterling D.G., Williams S.A. Development and Validation of a Protein-Based Risk Score for Cardiovascular Outcomes Among Patients With Stable Coronary Heart Disease. JAMA. 2016;315:2532–2541. doi: 10.1001/jama.2016.5951. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Lehallier B., Gate D., Schaum N., Nanasi T., Lee S.E., Yousef H., Moran Losada P., Berdnik D., Keller A., Verghese J., et al. Undulating changes in human plasma proteome profiles across the lifespan. Nat. Med. 2019;25:1843–1850. doi: 10.1038/s41591-019-0673-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Ponce-de-Leon M., Linseisen J., Peters A., Linkohr B., Heier M., Grallert H., Schöttker B., Trares K., Bhardwaj M., Gào X., et al. Novel associations between inflammation-related proteins and adiposity: A targeted proteomics approach across four population-based studies. Transl. Res. 2022;242:93–104. doi: 10.1016/j.trsl.2021.11.004. [DOI] [PubMed] [Google Scholar]

[bib22] 22.Ferreira J.P., Sharma A., Mehta C., Bakris G., Rossignol P., White W.B., Zannad F. Multi-proteomic approach to predict specific cardiovascular events in patients with diabetes and myocardial infarction: findings from the EXAMINE trial. Clin. Res. Cardiol. 2021;110:1006–1019. doi: 10.1007/s00392-020-01729-3. [DOI] [PubMed] [Google Scholar]

[bib23] 23.Zhu Y., Aebersold R., Mann M., Guo T. SnapShot: Clinical proteomics. Cell. 2021;184:4840–4840.e1. doi: 10.1016/j.cell.2021.08.015. [DOI] [PubMed] [Google Scholar]

[bib24] 24.Gillet L.C., Navarro P., Tate S., Röst H., Selevsek N., Reiter L., Bonner R., Aebersold R. Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis. Mol. Cell. Proteomics. 2012;11 doi: 10.1074/mcp.O111.016717. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] 25.Guo T., Kouvonen P., Koh C.C., Gillet L.C., Wolski W.E., Röst H.L., Rosenberger G., Collins B.C., Blum L.C., Gillessen S., et al. Rapid mass spectrometric conversion of tissue biopsy samples into permanent quantitative digital proteome maps. Nat. Med. 2015;21:407–413. doi: 10.1038/nm.3807. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Niu L., Thiele M., Geyer P.E., Rasmussen D.N., Webel H.E., Santos A., Gupta R., Meier F., Strauss M., Kjaergaard M., et al. Noninvasive proteomic biomarkers for alcohol-related liver disease. Nat. Med. 2022;28:1277–1287. doi: 10.1038/s41591-022-01850-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Bruderer R., Muntel J., Müller S., Bernhardt O.M., Gandhi T., Cominetti O., Macron C., Carayol J., Rinner O., Astrup A., et al. Analysis of 1508 Plasma Samples by Capillary-Flow Data-Independent Acquisition Profiles Proteomics of Weight Loss and Maintenance. Mol. Cell. Proteomics. 2019;18:1242–1254. doi: 10.1074/mcp.RA118.001288. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Jiang Z., Sun T.Y., He Y., Gou W., Zuo L.S.Y., Fu Y., Miao Z., Shuai M., Xu F., Xiao C., et al. Dietary fruit and vegetable intake, gut microbiota, and type 2 diabetes: results from two large human cohort studies. BMC Med. 2020;18:371. doi: 10.1186/s12916-020-01842-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Miao Z., Lin J.S., Mao Y., Chen G.D., Zeng F.F., Dong H.L., Jiang Z., Wang J., Xiao C., Shuai M., et al. Erythrocyte n-6 Polyunsaturated Fatty Acids, Gut Microbiota, and Incident Type 2 Diabetes: A Prospective Cohort Study. Diabetes Care. 2020;43:2435–2443. doi: 10.2337/dc20-0631. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Gou W., Ling C.W., He Y., Jiang Z., Fu Y., Xu F., Miao Z., Sun T.Y., Lin J.S., Zhu H.L., et al. Interpretable Machine Learning Framework Reveals Robust Gut Microbiome Features Associated With Type 2 Diabetes. Diabetes Care. 2021;44:358–366. doi: 10.2337/dc20-1536. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Gou W., Fu Y., Yue L., Chen G.D., Cai X., Shuai M., Xu F., Yi X., Chen H., Zhu Y., et al. Gut microbiota, inflammation, and molecular signatures of host response to infection. J Genet Genomics. 2021;48:792–802. doi: 10.1016/j.jgg.2021.04.002. [DOI] [PubMed] [Google Scholar]

[bib32] 32.Demichev V., Messner C.B., Vernardis S.I., Lilley K.S., Ralser M. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. s. 2020;17:41–44. doi: 10.1038/s41592-019-0638-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Zhu T., Chen H., Yan X., Wu Z., Zhou X., Xiao Q., Ge W., Zhang Q., Xu C., Xu L., et al. ProteomeExpert: a docker image based web-server for exploring, modeling, visualizing, and mining quantitative proteomic data sets. Bioinformatics. 2021;37:273–275. doi: 10.1093/bioinformatics/btaa1088. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Desvergne B. RXR: from partnership to leadership in metabolic regulations. Vitam. Horm. 2007;75:1–32. doi: 10.1016/S0083-6729(06)75001-4. [DOI] [PubMed] [Google Scholar]

[bib35] 35.Shulman A.I., Mangelsdorf D.J. Retinoid x receptor heterodimers in the metabolic syndrome. N. Engl. J. Med. 2005;353:604–615. doi: 10.1056/NEJMra043590. [DOI] [PubMed] [Google Scholar]

[bib36] 36.McNeill A.M., Rosamond W.D., Girman C.J., Golden S.H., Schmidt M.I., East H.E., Ballantyne C.M., Heiss G. The metabolic syndrome and 11-year risk of incident cardiovascular disease in the atherosclerosis risk in communities study. Diabetes Care. 2005;28:385–390. doi: 10.2337/diacare.28.2.385. [DOI] [PubMed] [Google Scholar]

[bib37] 37.Thanabalasingham G., Pal A., Selwood M.P., Dudley C., Fisher K., Bingley P.J., Ellard S., Farmer A.J., McCarthy M.I., Owen K.R. Systematic assessment of etiology in adults with a clinical diagnosis of young-onset type 2 diabetes is a successful strategy for identifying maturity-onset diabetes of the young. Diabetes Care. 2012;35:1206–1212. doi: 10.2337/dc11-1243. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Pickup J.C., Mattock M.B., Chusney G.D., Burt D. NIDDM as a disease of the innate immune system: association of acute-phase reactants and interleukin-6 with metabolic syndrome X. Diabetologia. 1997;40:1286–1292. doi: 10.1007/s001250050822. [DOI] [PubMed] [Google Scholar]

[bib39] 39.Rizza R.A., Mandarino L.J., Gerich J.E. Effects of growth hormone on insulin action in man. Mechanisms of insulin resistance, impaired suppression of glucose production, and impaired stimulation of glucose utilization. Diabetes. 1982;31:663–669. doi: 10.2337/diab.31.8.663. [DOI] [PubMed] [Google Scholar]

[bib40] 40.Joint committee for guideline revision 2016 Chinese guidelines for the management of dyslipidemia in adults. J. Geriatr. Cardiol. 2018;15:1–29. doi: 10.11909/j.issn.1671-5411.2018.01.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib41] 41.Ruderman N., Prentki M. AMP kinase and malonyl-CoA: targets for therapy of the metabolic syndrome. Nat. Rev. Drug Discov. 2004;3:340–351. doi: 10.1038/nrd1344. [DOI] [PubMed] [Google Scholar]

[bib42] 42.Orwoll E.S., Wiedrick J., Nielson C.M., Jacobs J., Baker E.S., Piehowski P., Petyuk V., Gao Y., Shi T., Smith R.D., et al. Proteomic assessment of serum biomarkers of longevity in older men. Aging Cell. 2020;19:e13253. doi: 10.1111/acel.13253. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib43] 43.Schüssler-Fiorenza Rose S.M., Contrepois K., Moneghetti K.J., Zhou W., Mishra T., Mataraso S., Dagan-Rosenfeld O., Ganz A.B., Dunn J., Hornburg D., et al. A longitudinal big data approach for precision health. Nat. Med. 2019;25:792–804. doi: 10.1038/s41591-019-0414-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] 44.Robbins J.M., Peterson B., Schranner D., Tahir U.A., Rienmüller T., Deng S., Keyes M.J., Katz D.H., Beltran P.M.J., Barber J.L., et al. Human plasma proteomic profiles indicative of cardiorespiratory fitness. Nat. Metab. 2021;3:786–797. doi: 10.1038/s42255-021-00400-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] 45.Yao Z., Wang Y. Apolipoprotein C-III and hepatic triglyceride-rich lipoprotein production. Curr. Opin. Lipidol. 2012;23:206–212. doi: 10.1097/MOL.0b013e328352dc70. [DOI] [PubMed] [Google Scholar]

[bib46] 46.Pollex R.L., Hanley A.J.G., Zinman B., Harris S.B., Khan H.M.R., Hegele R.A. Metabolic syndrome in aboriginal Canadians: prevalence and genetic associations. Atherosclerosis. 2006;184:121–129. doi: 10.1016/j.atherosclerosis.2005.03.024. [DOI] [PubMed] [Google Scholar]

[bib47] 47.Onat A., Hergenç G., Sansoy V., Fobker M., Ceyhan K., Toprak S., Assmann G. Apolipoprotein C-III, a strong discriminant of coronary risk in men and a determinant of the metabolic syndrome in both genders. Atherosclerosis. 2003;168:81–89. doi: 10.1016/s0021-9150(03)00025-x. [DOI] [PubMed] [Google Scholar]

[bib48] 48.Savinova O.V., Fillaus K., Jing L., Harris W.S., Shearer G.C. Reduced apolipoprotein glycosylation in patients with the metabolic syndrome. PLoS One. 2014;9:e104833. doi: 10.1371/journal.pone.0104833. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] 49.Jong M.C., Hofker M.H., Havekes L.M. Role of ApoCs in lipoprotein metabolism: functional differences between ApoC1, ApoC2, and ApoC3. Arterioscler. Thromb. Vasc. Biol. 1999;19:472–484. doi: 10.1161/01.atv.19.3.472. [DOI] [PubMed] [Google Scholar]

[bib50] 50.Breckenridge W.C., Little J.A., Steiner G., Chow A., Poapst M. Hypertriglyceridemia associated with deficiency of apolipoprotein C-II. N. Engl. J. Med. 1978;298:1265–1273. doi: 10.1056/NEJM197806082982301. [DOI] [PubMed] [Google Scholar]

[bib51] 51.Carlson L.A., Ballantyne D. Changing relative proportions of apolipoproteins CII and CIII of very low density lipoproteins in hypertriglyceridaemia. Atherosclerosis. 1976;23:563–568. doi: 10.1016/0021-9150(76)90016-2. [DOI] [PubMed] [Google Scholar]

[bib52] 52.McConathy W.J., Alaupovic P. Isolation and partial characterization of apolipoprotein D: a new protein moiety of the human plasma lipoprotein system. FEBS Lett. 1973;37:178–182. doi: 10.1016/0014-5793(73)80453-3. [DOI] [PubMed] [Google Scholar]

[bib53] 53.Olofsson S.O., McConathy W.J., Alaupovic P. Isolation and partial characterization of a new acidic apolipoprotein (apolipoprotein F) from high density lipoproteins of human plasma. Biochemistry. 1978;17:1032–1036. doi: 10.1021/bi00599a014. [DOI] [PubMed] [Google Scholar]

[bib54] 54.Seematter G., Binnert C., Tappy L. Stress and metabolism. Metab. Syndr. Relat. Disord. 2005;3:8–13. doi: 10.1089/met.2005.3.8. [DOI] [PubMed] [Google Scholar]

[bib55] 55.Kulkarni H., Meikle P.J., Mamtani M., Weir J.M., Barlow C.K., Jowett J.B., Bellis C., Dyer T.D., Johnson M.P., Rainwater D.L., et al. Variability in associations of phosphatidylcholine molecular species with metabolic syndrome in Mexican-American families. Lipids. 2013;48:497–503. doi: 10.1007/s11745-013-3781-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib56] 56.Wang S. Association between serum low-density lipoprotein cholesterol and metabolic syndrome in a working population. Lipids Health Dis. 2021;20:73. doi: 10.1186/s12944-021-01500-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib57] 57.Vuksan-Ćusa B., Jakovljević M., Sagud M., Mihaljević Peleš A., Marčinko D., Topić R., Mihaljević S., Sertić J. Metabolic syndrome and serum homocysteine in patients with bipolar disorder and schizophrenia treated with second generation antipsychotics. Psychiatr. Res. 2011;189:21–25. doi: 10.1016/j.psychres.2010.11.021. [DOI] [PubMed] [Google Scholar]

[bib58] 58.Laaksonen D.E., Niskanen L., Punnonen K., Nyyssönen K., Tuomainen T.P., Valkonen V.P., Salonen R., Salonen J.T. Testosterone and sex hormone-binding globulin predict the metabolic syndrome and diabetes in middle-aged men. Diabetes Care. 2004;27:1036–1041. doi: 10.2337/diacare.27.5.1036. [DOI] [PubMed] [Google Scholar]

[bib59] 59.Preissner K.T. Structure and biological role of vitronectin. Annu. Rev. Cell Biol. 1991;7:275–310. doi: 10.1146/annurev.cb.07.110191.001423. [DOI] [PubMed] [Google Scholar]

[bib60] 60.Alessi M.C., Nicaud V., Scroyen I., Lange C., Saut N., Fumeron F., Marre M., Lantieri O., Fontaine-Bisson B., Juhan-Vague I., et al. Association of vitronectin and plasminogen activator inhibitor-1 levels with the risk of metabolic syndrome and type 2 diabetes mellitus. Results from the D.E.S.I.R. prospective cohort. Thromb. Haemostasis. 2011;106:416–422. doi: 10.1160/TH11-03-0179. [DOI] [PubMed] [Google Scholar]

[bib61] 61.Yamagishi S.I., Adachi H., Abe A., Yashiro T., Enomoto M., Furuki K., Hino A., Jinnouchi Y., Takenaka K., Matsui T., et al. Elevated serum levels of pigment epithelium-derived factor in the metabolic syndrome. J. Clin. Endocrinol. Metab. 2006;91:2447–2450. doi: 10.1210/jc.2005-2654. [DOI] [PubMed] [Google Scholar]

[bib62] 62.Chen C., Tso A.W.K., Law L.S.C., Cheung B.M.Y., Ong K.L., Wat N.M.S., Janus E.D., Xu A., Lam K.S.L. Plasma level of pigment epithelium-derived factor is independently associated with the development of the metabolic syndrome in Chinese men: a 10-year prospective study. J. Clin. Endocrinol. Metab. 2010;95:5074–5081. doi: 10.1210/jc.2010-0727. [DOI] [PubMed] [Google Scholar]

[bib63] 63.Setoh K., Terao C., Muro S., Kawaguchi T., Tabara Y., Takahashi M., Nakayama T., Kosugi S., Sekine A., Yamada R., et al. Three missense variants of metabolic syndrome-related genes are associated with alpha-1 antitrypsin levels. Nat. Commun. 2015;6:7754. doi: 10.1038/ncomms8754. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib64] 64.Kronenberg F., Kollerits B., Kiechl S., Lamina C., Kedenko L., Meisinger C., Willeit J., Huth C., Wietzorrek G., Altmann M.E., et al. Plasma concentrations of afamin are associated with the prevalence and development of metabolic syndrome. Circ. Cardiovasc. Genet. 2014;7:822–829. doi: 10.1161/CIRCGENETICS.113.000654. [DOI] [PubMed] [Google Scholar]

[bib65] 65.Soggiu A., Piras C., Bonizzi L., Hussein H.A., Pisanu S., Roncada P. A discovery-phase urine proteomics investigation in type 1 diabetes. Acta Diabetol. 2012;49:453–464. doi: 10.1007/s00592-012-0407-0. [DOI] [PubMed] [Google Scholar]

[bib66] 66.Gourgari E., Ma J., Playford M.P., Mehta N.N., Goldman R., Remaley A.T., Gordon S.M. Proteomic alterations of HDL in youth with type 1 diabetes and their associations with glycemic control: a case-control study. Cardiovasc. Diabetol. 2019;18:43. doi: 10.1186/s12933-019-0846-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib67] 67.Laudes M., Oberhauser F., Schulte D.M., Schilbach K., Freude S., Bilkovski R., Schulz O., Faust M., Krone W. Dipeptidyl-peptidase 4 and attractin expression is increased in circulating blood monocytes of obese human subjects. Exp. Clin. Endocrinol. Diabetes. 2010;118:473–477. doi: 10.1055/s-0030-1249014. [DOI] [PubMed] [Google Scholar]

[bib68] 68.Winer D.A., Winer S., Shen L., Wadia P.P., Yantha J., Paltser G., Tsui H., Wu P., Davidson M.G., Alonso M.N., et al. B cells promote insulin resistance through modulation of T cells and production of pathogenic IgG antibodies. Nat. Med. 2011;17:610–617. doi: 10.1038/nm.2353. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib69] 69.Harte A.L., McTernan P.G., McTernan C.L., Crocker J., Starcynski J., Barnett A.H., Matyka K., Kumar S. Insulin increases angiotensinogen expression in human abdominal subcutaneous adipocytes. Diabetes Obes. Metabol. 2003;5:462–467. doi: 10.1046/j.1463-1326.2003.00274.x. [DOI] [PubMed] [Google Scholar]

[bib70] 70.Aubert J., Safonova I., Negrel R., Ailhaud G. Insulin down-regulates angiotensinogen gene expression and angiotensinogen secretion in cultured adipose cells. Biochem. Biophys. Res. Commun. 1998;250:77–82. doi: 10.1006/bbrc.1998.9185. [DOI] [PubMed] [Google Scholar]

[bib71] 71.Matsuo Y., Tanaka M., Yamakage H., Sasaki Y., Muranaka K., Hata H., Ikai I., Shimatsu A., Inoue M., Chun T.H., Satoh-Asahara N. Thrombospondin 1 as a novel biological marker of obesity and metabolic syndrome. Metabolism. 2015;64:1490–1499. doi: 10.1016/j.metabol.2015.07.016. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib72] 72.Matsuzawa Y., Funahashi T., Kihara S., Shimomura I. Adiponectin and metabolic syndrome. Arterioscler. Thromb. Vasc. Biol. 2004;24:29–33. doi: 10.1161/01.ATV.0000099786.99623.EF. [DOI] [PubMed] [Google Scholar]

[bib73] 73.Li Q., Xu L.Z., Zou T., Ai P., Huang G.H., Li P., Zheng A.P. Correlation of serum adiponectin and adiponectin gene polymorphism with metabolic syndrome in Chinese adolescents. Eur. J. Clin. Nutr. 2015;10:62–67. doi: 10.1038/ejcn.2014.152. [DOI] [PubMed] [Google Scholar]

[bib74] 74.Wattanapol P., Vichinsartvichai P., Sakoonwatanyoo P. Serum adiponectin is a potential biomarker for metabolic syndrome in peri-and postmenopausal women. Gynecol. Endocrinol. 2020;36:620–625. doi: 10.1080/09513590.2020.1742688. [DOI] [PubMed] [Google Scholar]

[bib75] 75.Kowalska I., Straczkowski M., Nikolajuk A., Adamska A., Karczewska-Kupczewska M., Otziomek E., Kinalska I., Gorska M. Insulin resistance, serum adiponectin, and proinflammatory markers in young subjects with the metabolic syndrome. Metabolism. 2008;57:1539–1544. doi: 10.1016/j.metabol.2008.06.008. [DOI] [PubMed] [Google Scholar]

[bib76] 76.Kim J.Y., Ahn S.V., Yoon J.H., Koh S.B., Yoon J., Yoo B.S., Lee S.H., Park J.K., Choe K.H., Guallar E. Prospective study of serum adiponectin and incident metabolic syndrome: the ARIRANG study. Diabetes Care. 2013;36:1547–1553. doi: 10.2337/dc12-0223. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib77] 77.Bertran L., Portillo-Carrasquer M., Martínez S., Aguilar C., Lopez-Dupla M., Riesco D., Binetti J., Vives M., Sabench F., Del Castillo D., et al. Expression of Jejunal Taste Receptors in Women with Morbid Obesity. Nutrients. 2021;13:2437. doi: 10.3390/nu13072437. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib78] 78.Sharma J.N. Hypertension and the bradykinin system. Curr. Hypertens. Rep. 2009;11:178–181. doi: 10.1007/s11906-009-0032-7. [DOI] [PubMed] [Google Scholar]

[bib79] 79.Agarwal R. Bradykinin and inhibition of angiotensin-converting enzyme in hypertension. N. Engl. J. Med. 1999;340:967–968. [PubMed] [Google Scholar]

[bib80] 80.Mandle R.J., Colman R.W., Kaplan A.P. Identification of prekallikrein and high-molecular-weight kininogen as a complex in human plasma. Proc. Natl. Acad. Sci. USA. 1976;73:4179–4183. doi: 10.1073/pnas.73.11.4179. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib81] 81.Ma J., Chen T., Wu S., Yang C., Bai M., Shu K., Li K., Zhang G., Jin Z., He F., et al. iProX: an integrated proteome resource. Nucleic Acids Res. 2019;47:D1211–D1217. doi: 10.1093/nar/gky869. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib82] 82.Chen T., Ma J., Liu Y., Chen Z., Xiao N., Lu Y., Fu Y., Yang C., Li M., Wu S., et al. iProX in 2021: connecting proteomics data sharing with big data. Nucleic Acids Res. 2022;50:D1522–D1527. doi: 10.1093/nar/gkab1081. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib83] 84.Shen B., Yi X., Sun Y., Bi X., Du J., Zhang C., Quan S., Zhang F., Sun R., Qian L., et al. Proteomic and Metabolomic Characterization of COVID-19 Patient Sera. Cell. 2020;182:59–72.e15. doi: 10.1016/j.cell.2020.05.032. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib84] 85.Zhang Y., Cai X., Ge W., Wang D., Zhu G., Qian L., Xiang N., Yue L., Liang S., Zhang F., et al. Potential Use of Serum Proteomics for Monitoring COVID-19 Progression to Complement RT-PCR Detection. J. Proteome Res. 2022;21:90–100. doi: 10.1021/acs.jproteome.1c00525. [DOI] [PubMed] [Google Scholar]

[bib85] 86.Scott M., Lundberg S.-I.L. 2017. A unified approach to interpreting model predictions. (Advances in Neural Information Processing Systems). [Google Scholar]

[bib86] 83.Krämer J., Green J., Pollard J., Jr, Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30:523–530. doi: 10.1093/bioinformatics/btt703. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Population serum proteomics uncovers a prognostic protein classifier for metabolic syndrome

Xue Cai

Zhangzhi Xue

Fang-Fang Zeng

Jun Tang

Liang Yue

Bo Wang

Weigang Ge

Yuting Xie

Zelei Miao

Wanglong Gou

Yuanqing Fu

Sainan Li

Jinlong Gao

Menglei Shuai

Ke Zhang

Fengzhe Xu

Yunyi Tian

Nan Xiang

Yan Zhou

Peng-Fei Shan

Yi Zhu

Yu-ming Chen

Ju-Sheng Zheng

Tiannan Guo

Summary

Graphical abstract

Highlights

Introduction

Results

Serum proteome profiling of a longitudinal cohort

Table 1.

Figure 1.

Figure 2.

A protein-based classifier for predicting MetS insurgence

Figure 3.

Serum proteins regulated in MetS samples

Figure 4.

Molecular characterization of MetS subtypes

Figure 5.

Proteins associated with MetS initiation

Figure 6.

Discussion

Limitations of the study

STAR★Methods

Key resources table

Resource availability

Lead contact

Materials availability

Experimental model and subject details

Ethics approval and consent to participate

Patients and samples

Method details

Protein digestion

Mass spectrometric analysis

Proteome data analysis

Machine learning

Quantification and statistical analysis

Acknowledgments

Author contributions

Declaration of interests

Inclusion and diversity

Declaration of generative AI and AI-assisted technologies in the writing process

Footnotes

Contributor Information

Supplemental information

Date and code availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases