Impact of statistical models on the prediction of type 2 diabetes using non-targeted metabolomics profiling

Ten metabolites-based algorithm predicts the future development of type 2 diabetes in Chinese

INTRODUCTION

Type 2 diabetes (T2D) is a heterogeneous metabolic disease with large variations in the relative contributions of insulin resistance and β-cell dysfunction across different glucose tolerance subgroups and ethnicities. A more precise yet feasible approach to categorize risk preceding T2D onset is urgently needed. This study aimed to identify potential metabolic biomarkers that could contribute to the development of T2D and investigate whether their impact on T2D is mediated through insulin resistance and β-cell dysfunction.

METHODS

A non-targeted metabolomic analysis was performed in plasma samples of 196 incident T2D cases and 196 age- and sex-matched non-T2D controls recruited from a long-term prospective Chinese community-based cohort with a follow-up period of ∼ 16 years.

RESULTS

Metabolic profiles revealed profound perturbation of metabolomes before T2D onset. Overall metabolic shifts were strongly associated with insulin resistance rather than β-cell dysfunction. In addition, 188 out of the 578 annotated metabolites were associated with insulin resistance. Bi-directional mediation analysis revealed putative causal relationships among the metabolites, insulin resistance and T2D risk. We built a machine-learning based prediction model, integrating the conventional clinical risk factors (age, BMI, TyG index and 2hG) and 10 metabolites (acetyl-tryptophan, kynurenine, γ-glutamyl-phenylalanine, DG(18:2/22:6), DG(38:7), LPI(18:2), LPC(P-16:0), LPC(P-18:1), LPC(P-20:0) and LPE(P-20:0)) (AUROC = 0.894, 5.6% improvement comparing to the conventional clinical risk model), that successfully predicts the development of T2D.

CONCLUSIONS

Our findings support the notion that the metabolic changes resulting from insulin resistance, rather than β-cell dysfunction, are the primary drivers of T2D in Chinese adults. Metabolomes as a valuable phenotype hold potential clinical utility in the prediction of T2D.

Identification of biomarkers for glycaemic deterioration in type 2 diabetes

We identify biomarkers for disease progression in three type 2 diabetes cohorts encompassing 2,973 individuals across three molecular classes, metabolites, lipids and proteins. Homocitrulline, isoleucine and 2-aminoadipic acid, eight triacylglycerol species, and lowered sphingomyelin 42:2;2 levels are predictive of faster progression towards insulin requirement. Of ~1,300 proteins examined in two cohorts, levels of GDF15/MIC-1, IL-18Ra, CRELD1, NogoR, FAS, and ENPP7 are associated with faster progression, whilst SMAC/DIABLO, SPOCK1 and HEMK2 predict lower progression rates. In an external replication, proteins and lipids are associated with diabetes incidence and prevalence. NogoR/RTN4R injection improved glucose tolerance in high fat-fed male mice but impaired it in male db/db mice. High NogoR levels led to islet cell apoptosis, and IL-18R antagonised inflammatory IL-18 signalling towards nuclear factor kappa-B in vitro. This comprehensive, multi-disciplinary approach thus identifies biomarkers with potential prognostic utility, provides evidence for possible disease mechanisms, and identifies potential therapeutic avenues to slow diabetes progression.

The Potential of Metabolomics in Biomedical Applications

The metabolome offers a dynamic, comprehensive, and precise picture of the phenotype. Current high-throughput technologies have allowed the discovery of relevant metabolites that characterize a wide variety of human phenotypes with respect to health, disease, drug monitoring, and even aging. Metabolomics, parallel to genomics, has led to the discovery of biomarkers and has aided in the understanding of a diversity of molecular mechanisms, highlighting its application in precision medicine. This review focuses on the metabolomics that can be applied to improve human health, as well as its trends and impacts in metabolic and neurodegenerative diseases, cancer, longevity, the exposome, liquid biopsy development, and pharmacometabolomics. The identification of distinct metabolomic profiles will help in the discovery and improvement of clinical strategies to treat human disease. In the years to come, metabolomics will become a tool routinely applied to diagnose and monitor health and disease, aging, or drug development. Biomedical applications of metabolomics can already be foreseen to monitor the progression of metabolic diseases, such as obesity and diabetes, using branched-chain amino acids, acylcarnitines, certain phospholipids, and genomics; these can assess disease severity and predict a potential treatment. Future endeavors should focus on determining the applicability and clinical utility of metabolomic-derived markers and their appropriate implementation in large-scale clinical settings.

Association of Coronary Artery Disease and Metabolic Syndrome: Usefulness of Serum Metabolomics Approach

Introduction

Individuals with metabolic syndrome (MetS) are at increasing risk of coronary artery disease (CAD). We investigated the common metabolic perturbations of CAD and MetS via serum metabolomics to provide insight into potential associations.

Methods

Non-targeted serum metabolomics analyses were performed using ultra high-performance liquid chromatography coupled with Q Exactive hybrid quadrupole-orbitrap high-resolution accurate mass spectrometry (UHPLC-Q-Orbitrap HRMS) in samples from 492 participants (272 CAD vs. 121 healthy controls (HCs) as cohort 1, 55 MetS vs. 44 HCs as cohort 2). Cross-sectional data were obtained when the participants were recruited from the First Affiliated Hospital of Zhengzhou University. Multivariate statistics and Student's t test were applied to obtain the significant metabolites [with variable importance in the projection (VIP) values >1.0 and p values <0.05] for CAD and MetS. Logistic regression was performed to investigate the association of identified metabolites with clinical cardiac risk factors, and the association of significant metabolic perturbations between CAD and MetS was visualized by Cytoscape software 3.6.1. Finally, the receiver operating characteristic (ROC) analysis was evaluated for the risk prediction values of common changed metabolites.

Results

Thirty metabolites were identified for CAD, mainly including amino acids, lipid, fatty acids, pseudouridine, niacinamide; 26 metabolites were identified for MetS, mainly including amino acids, lipid, fatty acids, steroid hormone, and paraxanthine. The logistic regression results showed that all of the 30 metabolites for CAD, and 15 metabolites for MetS remained significant after adjustments of clinical risk factors. In the common metabolic signature association analysis between CAD and MetS, 11 serum metabolites were significant and common to CAD and MetS outcomes. Out of this, nine followed similar trends while two had differing directionalities. The nine common metabolites exhibiting same change trend improved risk prediction for CAD (86.4%) and MetS (90.9%) using the ROC analysis.

Conclusion

Serum metabolomics analysis might provide a new insight into the potential mechanisms underlying the common metabolic perturbations of CAD and MetS.

Pepper Alkaloids and Processed Meat Intake: Results from a Randomized Trial and the European Prospective Investigation into Cancer and Nutrition (EPIC) Cohort

SCOPE

Processed meat intake has been associated with adverse health outcomes. However, little is known about the type of processed meat more particularly responsible for these effects. This study aims to identify novel biomarkers for processed meat intake.

METHODS AND RESULTS

In a controlled randomized cross-over dietary intervention study, 12 healthy volunteers consume different processed and non-processed meats for 3 consecutive days each. Metabolomics analyses are applied on post-intervention fasting blood and urine samples to identify discriminating molecular features of processed meat intake. Nine and five pepper alkaloid metabolites, including piperine, are identified as major discriminants of salami intake in urine and plasma, respectively. The associations with processed meat intake are tested for replication in a cross-sectional study (n = 418) embedded within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. Three of the serum metabolites including piperine are associated with habitual intake of sausages and to a lesser extent of total processed meat.

CONCLUSION

Pepper alkaloids are major discriminants of intake for sausages that contain high levels of pepper used as ingredient. Further work is needed to assess if pepper alkaloids in combination with other metabolites may serve as biomarkers of processed meat intake.

Triangulating evidence from longitudinal and Mendelian randomization studies of metabolomic biomarkers for type 2 diabetes

The number of people affected by Type 2 diabetes mellitus (T2DM) is close to half a billion and is on a sharp rise, representing a major and growing public health burden. Given its mild initial symptoms, T2DM is often diagnosed several years after its onset, leaving half of diabetic individuals undiagnosed. While several classical clinical and genetic biomarkers have been identified, improving early diagnosis by exploring other kinds of omics data remains crucial. In this study, we have combined longitudinal data from two population-based cohorts CoLaus and DESIR (comprising in total 493 incident cases vs. 1360 controls) to identify new or confirm previously implicated metabolomic biomarkers predicting T2DM incidence more than 5 years ahead of clinical diagnosis. Our longitudinal data have shown robust evidence for valine, leucine, carnitine and glutamic acid being predictive of future conversion to T2DM. We confirmed the causality of such association for leucine by 2-sample Mendelian randomisation (MR) based on independent data. Our MR approach further identified new metabolites potentially playing a causal role on T2D, including betaine, lysine and mannose. Interestingly, for valine and leucine a strong reverse causal effect was detected, indicating that the genetic predisposition to T2DM may trigger early changes of these metabolites, which appear well-before any clinical symptoms. In addition, our study revealed a reverse causal effect of metabolites such as glutamic acid and alanine. Collectively, these findings indicate that molecular traits linked to the genetic basis of T2DM may be particularly promising early biomarkers.

Metabolic Signatures of 10 Processed and Non-processed Meat Products after In Vitro Digestion

The intake of processed meat has been associated with several adverse health outcomes such as type II diabetes and cancer; however, the mechanisms are not fully understood. A better knowledge of the metabolite profiles of different processed and non-processed meat products from this heterogeneous food group could help in elucidating the mechanisms associated with these health effects. Thirty-three different commercial samples of ten processed and non-processed meat products were digested in triplicate with a standardized static in vitro digestion method in order to mimic profiles of small molecules formed in the gut upon digestion. A metabolomics approach based on high-resolution mass spectrometry was used to identify metabolite profiles specific to the various meat products. Processed meat products showed metabolite profiles clearly distinct from those of non-processed meat. Several discriminant features related to either specific ingredients or processing methods were identified. Those were, in particular, syringol compounds deposited in meat during smoking, biogenic amines formed during meat fermentation and piperine and related compounds characteristic of pepper used as an ingredient. These metabolites, characteristic of specific processed meat products, might be used as potential biomarkers of intake for these foods. They may also help in understanding the mechanisms linking processed meat intake and adverse health outcomes such as cancer.

Metabolomic and Lipidomic Signatures of Metabolic Syndrome and its Physiological Components in Adults: A Systematic Review

The aim of this work was to conduct a systematic review of human studies on metabolite/lipid biomarkers of metabolic syndrome (MetS) and its components, and provide recommendations for future studies. The search was performed in MEDLINE, EMBASE, EMB Review, CINHAL Complete, PubMed, and on grey literature, for population studies identifying MetS biomarkers from metabolomics/lipidomics. Extracted data included population, design, number of subjects, sex/gender, clinical characteristics and main outcome. Data were collected regarding biological samples, analytical methods, and statistics. Metabolites were compiled by biochemical families including listings of their significant modulations. Finally, results from the different studies were compared. The search yielded 31 eligible studies (2005–2019). A first category of articles identified prevalent and incident MetS biomarkers using mainly targeted metabolomics. Even though the population characteristics were quite homogeneous, results were difficult to compare in terms of modulated metabolites because of the lack of methodological standardization. A second category, focusing on MetS components, allowed comparing more than 300 metabolites, mainly associated with the glycemic component. Finally, this review included also publications studying type 2 diabetes as a whole set of metabolic risks, raising the interest of reporting metabolomics/lipidomics signatures to reflect the metabolic phenotypic spectrum in systems approaches.

Acyl ethanolamides in Diabetes and Diabetic Nephropathy: Novel targets from untargeted plasma metabolomic profiles of South Asian Indian men

The pathophysiology of diabetic nephropathy (DN) in type 2 diabetes (T2D) patients is minimally understood. We compared untargeted high-resolution accurate mass (HRAM) orbitrap-based plasma metabolomic profiles of 31 T2D-DN (with estimated glomerular filtration rate ≤80 mL/min/1.73 m²), 29 T2D and 30 normal glucose tolerance (NGT) Indian men. Of the 939 plasma metabolites that were differentially abundant amongst the NGT, T2D and T2D-DN (ANOVA, False Discovery Rate – FDR adjusted p-value < 0.05), 48 were associated with T2D irrespective of the renal function of the subjects. Acyl ethanolamides and acetylcholine were decreased while monoacylglycerols (MAGs) and cortisol were elevated in both T2D and T2D-DN. Sixteen metabolites, including amino acid metabolites Imidazolelactate and N-Acetylornithine, changed significantly between NGT, T2D and T2D-DN. 192 metabolites were specifically dysregulated in T2D-DN (ratio ≥2 or ≤0.5 between T2D-DN and T2D, similar abundance in NGT and T2D). These included increased levels of multiple acylcarnitine and amino acid metabolites. We observed a significant dysregulation of amino acid and fatty acid metabolism in South Asian Indian male T2D-DN subjects. Unique to this study, we report a reduction in acyl ethanolamide levels in both T2D and T2D-DN males. Those with dysregulation in acyl ethanolamides, which are endogenous agonists of GPR119, are likely to exhibit improved glycemic control with GPR119 agonists.

Plasma metabolites associated with type 2 diabetes in a Swedish population: a case-control study nested in a prospective cohort

AIMS/HYPOTHESIS

The aims of the present work were to identify plasma metabolites that predict future type 2 diabetes, to investigate the changes in identified metabolites among individuals who later did or did not develop type 2 diabetes over time, and to assess the extent to which inclusion of predictive metabolites could improve risk prediction.

METHODS

We established a nested case-control study within the Swedish prospective population-based Västerbotten Intervention Programme cohort. Using untargeted liquid chromatography-MS metabolomics, we analysed plasma samples from 503 case-control pairs at baseline (a median time of 7 years prior to diagnosis) and samples from a subset of 187 case-control pairs at 10 years of follow-up. Discriminative metabolites between cases and controls at baseline were optimally selected using a multivariate data analysis pipeline adapted for large-scale metabolomics. Conditional logistic regression was used to assess associations between discriminative metabolites and future type 2 diabetes, adjusting for several known risk factors. Reproducibility of identified metabolites was estimated by intra-class correlation over the 10 year period among the subset of healthy participants; their systematic changes over time in relation to diagnosis among those who developed type 2 diabetes were investigated using mixed models. Risk prediction performance of models made from different predictors was evaluated using area under the receiver operating characteristic curve, discrimination improvement index and net reclassification index.

RESULTS

We identified 46 predictive plasma metabolites of type 2 diabetes. Among novel findings, phosphatidylcholines (PCs) containing odd-chain fatty acids (C19:1 and C17:0) and 2-hydroxyethanesulfonate were associated with the likelihood of developing type 2 diabetes; we also confirmed previously identified predictive biomarkers. Identified metabolites strongly correlated with insulin resistance and/or beta cell dysfunction. Of 46 identified metabolites, 26 showed intermediate to high reproducibility among healthy individuals. Moreover, PCs with odd-chain fatty acids, branched-chain amino acids, 3-methyl-2-oxovaleric acid and glutamate changed over time along with disease progression among diabetes cases. Importantly, we found that a combination of five of the most robustly predictive metabolites significantly improved risk prediction if added to models with an a priori defined set of traditional risk factors, but only a marginal improvement was achieved when using models based on optimally selected traditional risk factors.

CONCLUSIONS/INTERPRETATION

Predictive metabolites may improve understanding of the pathophysiology of type 2 diabetes and reflect disease progression, but they provide limited incremental value in risk prediction beyond optimal use of traditional risk factors.

Serum metabolic profiling of type 2 diabetes mellitus in Chinese adults using an untargeted GC/TOFMS

BACKGROUND

Type 2 diabetes mellitus (T2DM) is a huge burden in China. The Chinese patients with T2DM have several special clinical characteristics. Metabolomics studies predominantly have identified several distinguishing metabolites associated with T2DM in Western ancestry population. However, few previous metabolomics studies were conducted in Chinese populations.

METHODS

We performed untargeted serum metabolic profiling between 30 T2DM patients and 30 healthy controls based on GC/TOFMS. Multivariate data analyses were applied to identify the distinguishing metabolites.

RESULTS

Excellent separation was obtained between the two cases. And overall 54 distinguishing metabolites were identified with VIP>1 and P<0.05, which were involved in metabolic pathways of amino acid, carbohydrate, lipids, membrane transport and nucleotides. To further analyze the correlation between the identified metabolites and T2DM, 17 metabolites were selected with FC>2.0, including gentisic acid, citraconic acid, succinic acid, 2-hydroxybutanoic acid and 3-hydroxy-l-proline, the corresponding FC were respectively 5.44, 2.21, 2.10, 2.21 and -2.04.

CONCLUSION

Our results demonstrated that untargeted GC/TOFMS-based metabolic approach processed well performance to identify serum distinguishing metabolites of T2DM in Chinese adults, which may be as potential biomarkers in diagnose and treatment of diabetes. And the results also provided new insight into understand the underlying molecular mechanism.

Early metabolic markers identify potential targets for the prevention of type 2 diabetes

AIMS/HYPOTHESIS

The aims of this study were to evaluate systematically the predictive power of comprehensive metabolomics profiles in predicting the future risk of type 2 diabetes, and to identify a panel of the most predictive metabolic markers.

METHODS

We applied an unbiased systems medicine approach to mine metabolite combinations that provide added value in predicting the future incidence of type 2 diabetes beyond known risk factors. We performed mass spectrometry-based targeted, as well as global untargeted, metabolomics, measuring a total of 568 metabolites, in a Finnish cohort of 543 non-diabetic individuals from the Botnia Prospective Study, which included 146 individuals who progressed to type 2 diabetes by the end of a 10 year follow-up period. Multivariate logistic regression was used to assess statistical associations, and regularised least-squares modelling was used to perform machine learning-based risk classification and marker selection. The predictive performance of the machine learning models and marker panels was evaluated using repeated nested cross-validation, and replicated in an independent French cohort of 1044 individuals including 231 participants who progressed to type 2 diabetes during a 9 year follow-up period in the DESIR (Data from an Epidemiological Study on the Insulin Resistance Syndrome) study.

RESULTS

Nine metabolites were negatively associated (potentially protective) and 25 were positively associated with progression to type 2 diabetes. Machine learning models based on the entire metabolome predicted progression to type 2 diabetes (area under the receiver operating characteristic curve, AUC = 0.77) significantly better than the reference model based on clinical risk factors alone (AUC = 0.68; DeLong's p = 0.0009). The panel of metabolic markers selected by the machine learning-based feature selection also significantly improved the predictive performance over the reference model (AUC = 0.78; p = 0.00019; integrated discrimination improvement, IDI = 66.7%). This approach identified novel predictive biomarkers, such as α-tocopherol, bradykinin hydroxyproline, X-12063 and X-13435, which showed added value in predicting progression to type 2 diabetes when combined with known biomarkers such as glucose, mannose and α-hydroxybutyrate and routinely used clinical risk factors.

CONCLUSIONS/INTERPRETATION

This study provides a panel of novel metabolic markers for future efforts aimed at the prevention of type 2 diabetes.

GC-MS based Gestational Diabetes Mellitus longitudinal study: Identification of 2-and 3-hydroxybutyrate as potential prognostic biomarkers

Gestational Diabetes Mellitus (GDM) causes severe short- and long-term complications for the mother, fetus and neonate, including type 2-diabetes (T2DM) later in life. In this pilot study, GC-Q/MS analysis was applied for plasma metabolomics fingerprinting of 24 healthy and 24 women with GDM at different stages of gestation (second and third trimester) and postpartum (one and three months). Multivariate (unsupervised and supervised) statistical analysis was performed to investigate variance in the data, identify outliers and for unbiased assessment of data quality. Plasma fingerprints allowed for the discrimination of GDM pregnant women from controls both in the 2nd and 3rd trimesters of gestation. However, metabolic profiles tended to be similar after delivery. Follow up of these women revealed that 4 of them developed T2DM within 2 years postpartum. Multivariate PLS-DA models limited to women with GDM showed clear separation 3 months postpartum. In the 2nd trimester of gestation there was also a clear separation between GDM women that were normoglycemic after pregnancy and those with recognized postpartum T2DM. Metabolites that had the strongest discriminative power between these groups in the 2nd trimester of gestation were 2-hydroxybutyrate, 3-hydroxybutyrate, and stearic acid. We have described, that early GDM comprises metabotypes that are associated with the risk of future complications, including postpartum T2DM. In this pilot study, we provide evidence that 2-hydroxybutyrate and 3-hydroxybutyrate may be considered as future prognostic biomarkers to predict the onset of diabetic complications in women with gestational diabetes after delivery.